REPRODUCTIVE EFFICIENCY OF 
OSTRICHES (Struthio came/us) 
by 
HELET LAMBRECHTS 
Dissertation submitted to the Faculty of Agriculture, Department of Animal , Wildlife and 
Grassland Sciences, University of the Free State, in partial fulfilment of the requirements 
for the degree 
PHILOSOPHIAE DOCTOR 
Promotor: Dr. D. Swart 
Co-promotor: Prof. J.P.C. Greyling 
Bloemfontein, May 2004 
Dedicated to 
'charlie brown' 
DECLARATION 
I hereby declare that the dissertation, submitted for the degree Philosophiae Doctor at the 
University of the Free State, is my own, unaided work, and has not been previously 
submitted for any degree or examination at any other University. 
Helet Lambrechts Date 
ACKNOWLEDGEMENTS 
I would like to express my sincere thanks to the following people and 
organisations/institutions who rendered support throughout the study: 
DR. DERICK SWART (ARC-Range and Forage Institute) - for his expert and invaluable 
advice and guidance in the design of the study His constant support and encouragement 
is dearly appreciated. 
PROF. JPC GREYLING (Department of Animal, Wildlife and Grassland Sciences, 
University of the Free State)·- for his continuous guidance, interest and support. A special 
word of thanks to Hester Linde, for her assistance during the study period. 
DR. SWP CLOETE (Department of Agriculture, Western Cape) - for his guidance and 
assistance in the statistical analysis of the data. 
Koor VAN SCHALKWYK (Mosstrich Group) - for his guidance and assistance. 
The DEPARTMENT OF AGRICULTURE, WESTERN CAPE, and the ANIMAL PRODUCTION 
RESEARCH TRUST - for the opportunity to conduct the study. A special word of thanks to 
the HET' JANS MARAIS FUND, for their financial contribution to the study. 
The NATIONAL RESEARCH FOUNDATION - for their financial contribution to the study. 
PERSONNEL AT THE KLEIN KAROO AGRICULTURAL DEVELOPMENT CENTRE - a special 
word of thanks to Basie Pfister and Zanell Brand, and all technical personnel for their 
technical support. Without their assistance, the study would not have been a success. 
The KLEIN KAROO GROUP - for the opportunity to conduct research on their breeding 
flock, as well as the use of their research facilities. A special word of thanks to Cecelia 
Mutlow and Dr. Adriaan Olivier, for assistance rendered. 
The ELSENBURG AGRICULTURAL DEVELOPMENT CENTRE LIBRARY - a special word of 
thanks to Wilna Brink and Elizabeth Valentine, for their invaluable assistance. 
The DEPARTMENT OF AGRICULTURE, UNIVERSITY OF WESTERN AUSTRALIA - for the 
opportunity to conduct the RIA studies at their facilities. A special word of thanks to Prof. 
Graeme Martin, John Beesley, Margaret Blackberry, Dr. lrek Malecki, and Sarah Jo 
Smith, for the assistance and guidance rendered during the time abroad. 
ORYX TANNING (PTY) LTD. - for the opportunity to conduct a study on their breeding 
flock. A special word of thanks to Leonie du Toit, Hetta Burger, Weedi and Halihma, 
and all technical personnel, for their invaluable support. 
PHILIPS PIE MEDICAL, SOUTH AFRICA - for their contribution to and assistance rendered 
during the study. 
THEO AND LEONIE DU TOIT - for their friendship and hospitality. 
SOPHIE BooYSENS - for looking after my 'children' during my times away from home. 
MY PARENTS AND FAMILY - for all the opportunities they gave me, and for their continuous 
love and support throughout the study. 
And above all - to Goo, THE ALMIGHTY - all praise and thanks to Him for giving me 
the strength and ability to conduct the study. 
CONTENTS 
Page 
List of tables 
List of figures vi 
List of plates ix 
Chapter 1 : General introduction 
Chapter 2 : Flush feeding, teasing, and a mid-season breeding rest as 
management practices to improve the reproductive performance 
of ostriches (Struthio came/us var. domesticus) in commercial 
farming systems 
Abstract 4 
Introduction 4 
Materials and methods 6 
Results 8 
Discussion 14 
Conclusions 16 
References 17 
Chapter 3: The influence of season and breeding rest periods on the 
reproductive performance of breeding ostriches (Struthio 
came/us var. domesticus) under commercial farming conditions 
Abstract 21 
Introduction 21 
Materials and methods 23 
Results 25 
Discussion 30 
Conclusions 34 
References 34 
Chapter 4: The influence of stocking rate and male:female ratio on the 
reproductive performance of breeding ostriches ( Struthio 
came/us spp.) under commercial farming conditions 
Abstract 38 
Introduction 38 
Materials and methods 39 
Results 42 
Discussion 45 
Conclusions 49 
References 49 
Chapter 5: Genetic and environmental (co)variance estimates for ostrich 
(Struthio came/us var. domesticus) male aggression, shin 
colour, and the egg production performance of companion 
breeding females 
Abstract 53 
Introduction 53 
Materials and methods 54 
Results 57 
Discussion 61 
Conclusions 62 
References 62 
Chapter 6: The relationship between serum testosterone, progesterone, LH 
and prolactin levels, and behavioural and reproduction 
parameters measured in breeding ostriches (Struthio came/us 
var. domesticus) 
Abstract 66 
Introduction 67 
Materials and methods 68 
Results 72 
Discussion 78 
Conclusions 83 
References 83 
Chapter 7: Preliminary results on the use of diagnostic ultrasonography as 
a management tool to quantify egg production potential of 
breeding ostrich (Struthio came/us var. domesticus) females 
Abstract 89 
Introduction 89 
Materials and methods 90 
Results 91 
Discussion 94 
Conclusions 96 
References 96 
Chapter 8: Genetic variability of ultrasound scanning parameters and egg 
production traits of breeding ostrich (Struthio came/us var. 
domesticus) females 
Abstract 99 
Introduction 99 
Materials and methods 101 
Results 105 
Discussion 109 
Conclusions 112 
References ·113 
Chapter 9: (Co)variance estimates for live weight, body measurements and 
reproduction of pair-mated ostrich females 
Abstract 118 
Introduction 118 
Materials and methods 119 
Results 123 
Discussion 131 
Conclusions 136 
References 136 
Chapter 10: Semen quality of ostrich (Struthio came/us var. domesticus) 
males in relation to fertility in a commercial breeding system 
Abstract 139 
Introduction 139 
Materials and methods 140 
Results 143 
Discussion 153 
Conclusions 158 
References 158 
Chapter 11: Management practices and performance indicators of 
reproduction in ostriches - conclusions and recommendations 163 
SUMMARY 170 
OPSOMMING 172 
Appendix 174 
LIST OF TABLES 
Chapter 2 
Table Details Page 
Table 1: The influence of flushing and teasing on the reproduction traits (mean ± SE) of ostrich 
breeding pairs during the 2000/2001 breeding season. 9 
Table 2: The influence of flushing and teasing on the reproduction traits (means ± SE) of ostrich 
breeding pairs during the 2001/2002 breeding seasOll 10 
Table 3: The effect of a mid-season breeding rest on the reproduction traits (means ± SE) of 
ostrich breeding pairs subjected to flushing and teasing during the 2000/2001 breeding 
season. 13 
Table 4: The effect of a mid-season breeding rest on the reproduction traits (means ± SE.) of 
ostrich breeding pairs subjected to flushing and teasing during the 2001/2002 breeding 
season. 13 
Chapter 3 
Table Details Page 
Table 1: The time table of the forced breeding rest periods, as implemented during the 
2002/2003 breeding season. 24 
Table 2: The reproduction and egg traits (means ± SE) of ostrich breeding pairs with a 
continuous breeding season (CS), and breeding pairs joined before (early) or after (late) 
the winter solstice, and subjected to forced breeding rest periods. 26 
Table 3: The influence of timing of breeding and forced breeding rests on the reproduction traits 
(means± SE) of ostrich breeding pairs during the 2002/2003 breeding season. 29 
Chapter 4 
Table Details Page 
Table 1: The stocking rates and male:female (M:F) ratios implemented in the breeding program 
of a commercial ostrich breeding operation maintained near Ladismith in the Western 
Cape, South Africa. 40 
Table 2: The influence of stocking rate on the reproduction traits (means ± SE) of ostriches 
maintained in 1 ha camps during the 2000/2001 breeding season. 42 
Table 3: The influence of stocking rate on the reproduction traits (means ± SE) of breeding 
ostriches maintained in 1 ha camps during the 2001/2002 breeding season. 43 
Table 4: Reproduction traits (means ± SE) for 9-bird breeding flocks maintained in 0.13 ha and 
0.30 ha camps during the 2001/2002 breeding season 43 
Table 5: The influence of stocking rate on the reproduction traits (means ± SE) of breeding 
ostriches maintained in 0.13 ha camps during the 2001/2002 breeding season 44 
Table 6: The influence of male:female ratio on the reproduction traits (mean ± SE) of breeding 
ostriches maintained in 0.06ha breeding camps during the 2000/2001 breeding season. 44 
Table 7: Reproduction traits (means ± SE) recorded for 9-bird and 13-bird breeding flocks 
maintained in 0.30 and 0.35 ha camps, respectively. during the 2000/2001 breeding 
season. 45 
Chapter 5 
Table Details Page 
Table 1: Means and standard deviations, and number of observations for reproduction and 
behavioural traits assessed for ostrich males and females. 58 
Table 2: Log likelihood values for the respective random effect models (model in brackets) 
considered in the analyses, with the best models indicated in bold. 59 
Table 3: Variance components and ratios (±SE) for the egg production performance (EPP) of 
ostrich females, and aggression and shin colour scores assigned to ostrich males. 59 
Table 4: The permanent (PE) and temporary (TE) environment correlations of the egg production 
performance (EPP) of ostrich females with aggression and shin colour of ostrich males, 
as estimated between breeding pairs and on an individual basis. 60 
Chapter 6 
Table Details Page 
Table 1: Means (SD) for live weight, egg production performance (EPP), serum hormone 
concentrations, and male aggression and shin colour measured in breeding ostriches. 72 
Table 2: The log likelihood ratios for the respective random effect models fitted to the data set, 
with the best model indicated in bold. 74 
Table 3: The variance components and ratios for live weight, serum testosterone and LH, male 
aggression and shin colour, and egg production performance (EPP) measured in 
breeding ostriches. 74 
Table 4: The genetic, permanent environmental, environmental, and phenotypic correlations 
between behavioural and reproduction traits measured in breeding ostriches, where at 
least one random effect exceeded 5 % of the overall phenotypic variance. 75 
Table 5: The influence of time of sampling on the live weight, egg production performance 
(EPP), male aggression and shin colour, and the serum hormone levels measured for 
breeding ostriches. The Interaction of gender and time of sampling are presented 
where significant (Ps: 0.05) 78 
Chapter 7 
Table Details Page 
Table 1: Mean (± SE) egg production during the first month, first 2 months, and for the entire 
breeding season, classified according to the number of follicles observed at the 
beginning of the breeding season. 93 
Table 2: Mean (± SE) egg production during the last month, the last two months, and during the 
entire breeding season, when classified according to the number of follicles observed at 
the end of the breeding season. 94 
Chapter 8 
Table Details Page 
Table 1: Follicle classification scores assigned during ultrasonic examination of breeding ostrich 
females. 103 
Table 2: Number of observations, means and standard deviations (SD), and ranges for the 
respective ultrasound parameters and egg production traits for breeding ostrich 
females 105 
Table 3: The correlation of follicle classification score (FCS) with the number of follicles observed 
on the left (L) and right (R) flanks of ostrich females, respectively; and of the number of 
follicles observed on the left flanks (FN-L) with number of follicles observed on the right 
flanks during the 2002/2003 breeding season. 106 
Table 4: Least square means (±SE) depicting the influence of age on the ultrasound parameters 
and egg production traits of breeding ostrich females. 106 
Table 5: Log-likelihood values for models fitting different random effects for ultrasound 
parameters and egg production traits obtained for entire breeding seasons. 107 
Table 6: Log-likelihood values for models fitting different random effects for ultrasound 
parameters and egg production traits obtained at the beginning of the breeding 
seasons. 107 
Table 7: Log-likelihood values for models fitting different random effects for ultrasound 
parameters and egg production traits obtained at the end of the breeding seasons. 107 
iii 
Table 8: Genetic and environmental (co)variance estimates and ratios for follicle classification 
score (FCS), number of follicles (FN), follicle diameter (FD), and egg production 
performance (EPP) over all available records. 108 
Table 9: Genetic and environmental (co)variance estimates and ratios for follicle classification 
score (FCS), number of follicles (FN), follicle diameter (FD), and egg production 
performance (EPP) when recorded at the beginning of the breeding season. 108 
Table 10: Genetic and environmental (co)variance estimates and ratios for follicular classification 
score (FCS), number of follicles (FN), follicle diameter (FD), and egg production 
performance (EPP) recorded at the end of the breeding season 109 
Chapter 9 
Table Details Page 
Table 1: Descriptive statistics for the traits assessed in ostrich breeding females for the 
production years of 1990 to 2001. 123 
Table 2: Log likelihood values for the respective traits under different random effects models, 
with the best model for each trait represented in bold italic figures 126 
Table 3: Variance components and ratios (± SE) for the respective traits assessed in ostrich 
breeding females over the period from 1990 to 2001 127 
Table 4: (Co)variance ratios (± SE) depicting estimates of h2 and c2 as well as genetic, 
permanent environmental, environmental and phenotypic correlations for live weight at 
the beginning (B) or end (E) of the mating season, egg production and chick 
production, as assessed by three-trait analyses in breeding ostrich females. 128 
Table 5: Estimates of genetic, permanent environmental, environmental and phenotyp1c 
correlations (± SE) between traits, as assessed by two-trait analyses in breeding 
ostrich females. Live weight and body traits were recorded at the beginning (B) or end 
(E) of mating. 130 
Chapter 10 
Table Details Page 
Table 1· A scale for the macroscopic assessment of colour and viscosity score of ostrich semen 
samples. 142 
Table 2: Phallus traits and semen quality parameters (mean and SD) measured in ostrich males 
during 4 consecutive breeding seasons. 144 
iv 
Table 3: Means (± SE) of phallus length (Pl_), phallus circumference (PC), and sperm traits 
rneasured in ostriches, as influenced by management practice, breeding season, and 
time of collection. 145 
Table 4: Relationship between age, and colour and viscosity scores (±SE) recorded for ostrich 
semen samples. 147 
Table 5: Variance components and repeatability (t) estimates for phallus traits and semen 
parameters of breeding ostrich males. 148 
Table 6: The correlation of phallus length and circumference with semen parameters of breeding 
ostrich males. 148 
Table 7: The correlations of phallus traits and semen parameters with fertility of eggs produced 
by companion females, determined for ejaculates obtained before the onset of and at 
the end of a breeding for breeding ostrich males. 149 
v 
LIST OF FIGURES 
Chapter 2 
Figure Details Page 
Figure 1: The influence of flushing (F) and teasing (T) on the fortnightly egg production 
performance (EPP; mean ± SE) of ostrich breeding pairs not subjected to a mid-
season breeding rest during the 2000/2001 breeding season. 9 
Figure 2: The influence of flushing (F) and teasing (T) on the fortnightly egg production 
performance (EPP; mean ± SE) of ostrich breeding pairs not subjected to a m1d-
season breeding rest during the 2001/2002 breeding season. 10 
Figure 3: The influence of a mid-season breeding rest on the fortnightly egg production 
performance (EPP; mean ± SE) of ostrich breeding pairs during the 2000/2001 
breeding season. 11 
Figure 4: The influence of a mid-season breeding rest on the fortnightly egg production 
performance (EPP; mean ± SE) of ostrich breeding pairs subjected to flushing (F) and 
teasing (T) management practices during the 2000/200'1 breeding season. 11 
Figure 5: The influence of a mid-season breeding rest on the fortnightly egg production 
performance (EPP; mean ± SE) of ostrich breeding pairs during the 2001/2002 
breeding season. 12 
Figure 6: The influence of a mid-season breeding rest on the fortnightly egg production 
performance (EPP; mean ± SE) of ostrich breeding pairs subjected to flushing (F) and 
teasing (T) management practices during the 2001/2002 breeding season. 12 
Chapter 3 
Figure Details Page 
Figure 1: The influence of timing of breeding on the egg production performance (EPP) of ostrich 
breeding pairs with a continuous breeding season (CS), and breeding pairs joined 
before (ES) and after (LS) the winter solstice. 27 
Figure 2: The influence of forced breeding rest periods on the monthly egg production 
performance (EPP) of ostrich breeding pairs Joined before (ES) and after (LS) the 
winter solstice 28 
Figure 3: The influence of season and forced breeding rest periods on the territorial aggression 
of ostrich males with a continuous breeding season (CS), and ostrich males joined 
before (ES) and after (LS) the winter solstice. 28 
VI 
Chapter 5 
Figure Details Page 
Figure 1: The relationship between male age, aggression and shin colour scores, and female 
egg production performance (EPP). 58 
Figure 2: The average monthly (mean ± SE) egg production performance (EPP) of ostrich 
breeding pairs. 60 
Chapter 6 
Figure Details Page 
Figure 1: The influence of age on the live weight of male and female ostriches. 73 
Figure 2: The influence of age on the serum LH levels (mean ± SE) measured for and egg 
production performance (EPP; mean ±SE) of breeding ostriches. 76 
Figure 3: The relationship between age, serum levels of testosterone and prolactin (mean ±.SE), 
and the egg production performance (EPP; mean ±SE) of breeding ostrich males and 
females. 76 
Figure 4: The relationship between age, serum progesterone concentration (mean ± SE) 
measured in and egg production performance (EPP; mean± SE) of ostrich females. 77 
Chapter 7 
Figure Details Page 
Figure 1: An ultrasonogram of the ovary of breeding ostrich female, with five follicles 
distinguishable in the ultrasonogram. The largest follicle has a diameter of 63mm. 92 
Figure 2: An ultrasonogram of the ovary of a breeding ostrich female, with five follicles 
distinguishable in the ultrasonogram. The largest follicle has a diameter of 58mm and 
the smallest follicle a diameter of 1S mm. 92 
Figure 3: Relationship between female age and number of follicles noted with diagnostic imaging 
at the beginning of the 1999/2000 breeding season. 93 
Chapter 9 
Figure Details Page 
Figure 1: Means (±SE) for live weight at the commencement and cessation of mating in female 
age groups from 2 to 10 years. Trend lines were derived from initial analyses where 
age trends were modeled by cubic splines 124 
vii 
Figure 2: Means (±SE) for egg and chick production in female age groups from 2 to 10 years 125 
Figure 3: Mean hatchability (±SE) in female age groups tram 2 to 10 years. The trend line was 
derived from an initial analysis, where the age trend was modelled by fitting a cubic 
spline 125 
Chapter 10 
Figure Details Page 
Figure 1: The drawing on the left represents an ostrich phallus, with A, B, C and D indicating the 
measurement points for phallus length and circumference (adapted from Jensen et al., 
1992). 141 
Figure 2: The influence of age on phallus length and circumference (±SE) measured in breeding 
ostrich males. 145 
Figure 3: The effect of age on the sperm concentration (±SE) of ejaculates obtained from 
breeding ostrich males. 146 
Figure 4: The influence of male age on the percentage of abnormalities in ejaculates obtained 
from breeding ostrich males. 146 
Figure 5: The effect of male age on the fertility of eggs produced by companion females. 147 
viii 
LIST OF PLATES 
Chapter 5 
Plate Details Page 
Plate 1: Colour chart used to assign monthly shin colour scores to ostrich males. Shin colour 
score indicated in each respective colour box. 55 
Chapter 8 
Plate Details Page 
Plate 1: Sonograms obtained during ultrasound scanning of ostrich females. 103 
Chapter 10 
Plate Details Page 
Plate 1: Plate 1 represents a post-mortem photograph of an ostrich phallus, indicating the 
deviation of the phallus to the left-hand side when everted. 141 
Plate 2: A micrograph depicting morphologically normal ostrich sperm (X200 magnification). 150 
Plate 3: A micrograph depicting morphologically normal ostrich sperm (X400 magnification). 150 
Plate 4: A micrograph depicting acrosomal damage to ostrich sperm (X400 magnification). 151 
Plate 5: A micrograph depicting ostrich sperm with swollen heads (arrows) (X400 
magnification). 151 
Plate 6: A micrograph depicting ostrich sperm with abnormal morphology, i.e. broken and loose 
tails (X 400 magnification) 152 
Plate 7: A micrograph depicting the agglutination of ostrich sperm (X200 magnification). 152 
Plate 8: A micrograph depicting the agglutination of ostrich sperm (X400 magnification). 153 
ix 
Chapter 1 
General introduction 
Ostrich farming in South Africa started as early as 1863, and the commercial population amounted 
to 253 463 ostriches by the end of the 191h century (Van Zyl, 1996). Initially, ostrich feathers were 
the main commercial product and ranked fourth in value, next to gold, diamonds and wool, as a 
South African export product (Smit, 1964; Wagner, 1986). The recognition of the commercia! 
potential of ostrich leather and meat led to the establishment of a commercial ostrich abattoir and 
tannery in 1965 and 1970, respectively, in Oudtshoorn, which was considered as the headquarters 
of the ostrich industry. Today, South Africa is the largest ostrich producer in the world, producing 
approximately 60% of the world's 560 000 birds slaughtered annually. Ostrich leather and meat 
contribute approximately 65% and 27% respectively, to the total GDP of approximately ZAR 5.6 
million generated from slaughter birds (Ostrich Section 7 Committee Report, 2003). 
The emphasis on producing primarily for the export market and high input costs, together with a 
fluctuating exchange rate, necessitates that commercial ostrich tanners produce slaughter birds as 
cost-efficient as possible. The profitability of commercial ostrich production systems is largely 
determined by the number of eggs produced and chicks hatched per female, feed costs, and the 
type of production system used. In contrast to commercial poultry production systems, commercial 
ostrich production systems are characterised by a relatively low reproductive performance, a large 
variation in the number of eggs produced per female, and high chick mortalities (Van Zyl, 2001 ). In 
South Africa, the commercial ostrich industry average of 40 eggs produced per female during an 8-
month breeding season (Van Zyl, 2001 ), a hatch ability of 46.2% (Cloete et al., 1998) and a chick 
survival rate of 95% during the first two days after hatching (Van Zyl, 2001), thus implies that only 
18 viable chicks are produced per female during an 8-month breeding season. 
Almost 80% of the national ostrich breeding population is maintained as breeding flocks, which can 
range from 50 to 100 birds in size, at a ratio of 5 to 6 males for every 1o  females. Identification of 
poor or non-producing individuals is complicated by the fact that eggs are laid in communal nests, 
i.e. more than one ostrich female may lay in a given nest, and an ostrich female may also lay in 
more than one nest (Lambrechts et al., 2002). The reproductive performance of breeding ostriches 
may be influenced by the timing of the breeding season, the long interval between pairing and the 
first oviposition, irregular egg laying sequences, age of breeding birds, nutrition, genetics, breeding 
behaviour of ostrich males and females, breeding environment, and management procedures. 
Feed costs represent almost 83% of the total input costs of a breeding system (Van Zyl, 2001 ), and 
the maintenance of poor or non-producing individuals is thus clearly uneconomical. Methods to 
identify poor or non-producing individuals in commercial ostrich breeding systems, however, are 
scarce and based on historical performance and observations over time, rather than descriptive 
predictions of a breeding ostrich's reproductive potential for future breeding seasons. Such 
techniques need to be developed to assist commercial ostrich farmers in selecting against low or 
non-producing ostrich males and females to improve the production performance of the overall 
farming operation. 
The studies were conducted in the arid Klein Karoo region of South Africa, at longitude 22° 15' E 
and latitude 33° 38' S and at an altitude of 301 m above sea level. The long-term precipitation 
averages 230 mm, with 54 % of the total rainfall occurring during winter (April to September). The 
experimental site is characterized by hot summers, with the average maximum temperature 
exceeding 25°C for the period October to April. 
Management aspects (practices) to synchronise and improve overall reproduction efficiency 
under intensive conditions that were addressed in the study, included the following: 
1. Flush feeding and teasing 
2. Timing of breeding, and forced breeding rest periods 
3. Stocking rate and male:female ratio of breeding ostriches 
Behavioural and physiological aspects that were considered as indirect selection criteria for 
reproduction efficiency of ostriches under commercial conditions were: 
4. Shin colour and aggression of male ostriches and its influence on the reproductive 
performance of companion females 
5. Reproductive hormones and their relationship to behavioural and reproductive parameters in 
breeding ostriches 
6. The extent of follicular development before the onset of reproduction, as determined by 
ultrasound scanning 
7. Live weight and body measurements of breeding ostrich females and the relationship with their 
reproductive performance 
8. Semen quality in relation to fertility under commercial breeding conditions 
Improving the reproductive performance of ostrich males and females through management 
strategies and selection for reproductive performance without increasing input costs will directly 
affect the production cost of day-old chicks. Based on costs calculated during 2000, Van Zyl 
(2001) concluded that a 10% increase or decrease in the number of day-old chicks hatched per 
female will amount to a 9.1 % decrease or increase in the costs of producing a day-old chick. 
Similarly, a 10% increase or decrease in feed costs will amount to a 8.3% increase or decrease in 
the costs of producing a day-old chick (Van Zyl, 2001). The ability to manipulate the reproductive 
cycles of breeding ostriches, i.e. onset and cessation of breeding, as well as the reinitiation of 
reproductive activities, may enable the commercial ostrich farmer to manage a breeding flock 
optimally to ensure the timely production of slaughter birds according to market demand. 
2 
With the implementation of the various management, behaviour and physiological aspects 
addressed in this study, it is hoped that a contribution can be made to increase the overall 
reproduction efficiency of commercial ostrich production systems. 
REFERENCES 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998 Ostrich breeding - progress towards a 
scientifically based strategy. Proc. 2nd Int. Sci. Ratite Congr., 21-25 September, 
Oudtshoorn, South Africa, pp. 55-62. 
LAMBRECHTS H, CLOETE SWP, SWART D & GREYLING, JPC, 2002. Preliminary results on the use of 
diagnostic imaging as a management tool to quantify egg production potential in breeding 
ostrich (Struthio came/us australis) females. J S. Afr. Vet. Assoc. 73(2) 48-52. 
OSTRICH SECTION 7 COMMITIEE REPORT, 2003. Report on the investigation into the effect of 
deregulation on the South African Ostrich Industry, 15 April, South African Ostrich 
Business Chamber, PO Box 952, Oudtshoorn, South Africa. 
SMIT DJ VAN ZYL, 1964. Volstruisboerdery in die Klein Karoo. Department Agricultural Technical 
Services, Pamphlet 358, V&R Printers, Pretoria, South Africa. 
VAN ZYL PL, 2001. 'n Ekonomiese evaluering van volstruisboerdery in die Oudtshoorn omgewing. 
M.Sc. thesis, University of Stellenbosch, Stellenbosch, South Africa. 
VANZYL P, 1996. A global perspective of the ostrich industry. Canadian Ostrich, pp. 8-13. 
WAGNER P, 1986. The ostrich story. Chameleon Press, Cape Town. 
3 
Chapter 2 
Flush feeding, teasing, and a mid-season breeding rest 
as management practices to improve the reproductive 
performance of ostriches (Struthio came/us var. 
domesticus) in commercial farming systems 
Abstract 
Flush feeding, teasing, and a mid-season breeding rest were investigated as management practices to synchronise 
and improve the overall reproductive performance of ostrich breeding pairs (n=136) during breeding seasons. 
Teasing significantly improved (P~0.05) egg production performance of ostrich females. Maintaining breeding 
ostriches as single-sex flocks and visually isolating them during the pre-breeding period, and the re-introduction of 
the ostrich males to neighbouring camps two weeks before being joined, could thus have illicit neuro-endocrine 
responses in ostrich females that resulted in the initiation of ovulation and an earlier onset of egg production. Joining 
breeding ostriches after the winter solstice may be necessary to ensure the use of flushing and teasing as 
management practices to synchronise the reproductive cycles of breeding ostriches for an early onset of 
reproduction. A mid-season breeding rest improved (P~0.05) the egg production performance and consequent chick 
production of the breeding pairs. A mid-season breeding rest was however, only effective in decreasing the time to 
onset of production when the breeding pairs were separated before the photoperiod started to decrease, i.e. when 
birds were still photosensitised enough to react to this management practice. Subjecting breeding ostriches to a 
breeding rest during the middle of spring in the Southern Hemisphere (i.e. mid-October), when ostriches have 
already started to become photorefractory, will most probably limit the effectiveness of this management practice to 
reinitiate reproduction in ostrich males and females. It is thus advisable to separate breeding ostriches for a mid-
season breeding rest before daylight length starts to decrease to ensure the effectiveness of this management 
practice to overcome the seasonal quiescence in reproduction that is characteristic of breeding ostriches. 
INTRODUCTION 
Timing of breeding in temperate-zone birds is determined by the minimum photoperiod or critical 
day length, and the development of photorefractoriness (Leitner et al., 2003). Photoperiod is 
considered as the most reliable and predictive cue that can exert an influence weeks to months 
before the onset of a breeding season. Photoperiodic stimulation has to attain a certain thresh a Id 
before gametogenetic development is initiated and achieved, and additional factors bring the 
gonads to full functional maturity, with subsequent nest-building and egg production. A lack of 
appropriate environmental stimuli may prevent reproduction from occurring. This differs from 
tropical and opportunistic breeders where the reproductive system remains in a 'ready to breed' 
state for a large part of the year, with non-photoperiodic cues determining the onset of 
reproduction. 
4 
Ostriches can be regarded as temperate-zone breeding birds that have developed a rigid form of 
proximate control of their reproductive cycles to ensure the survival of their offspring during the 
time of the year when conditions are favourable (Shanawany, 1995). The timing and the duration 
of ostrich breeding seasons may vary with latitude and altitude (Shanawany, 1994) In birds, 
melatonin is mainly used as a 'circadian clock', and plays a role in the regulation of diurnal activities 
(Siopes, 1983; Lumineau et al., 2002). At the molecular level, the avian circadian clock functions in 
a similar manner to that of the mammalian clocks. In birds, however, photoperiodic responses 
appear not to involve an interaction between the eyes and the pineal gland, but a direct interaction 
between photoreceptors and GnRH neurons (Saldanha et al., 2001). It is unsure whether 
melatonin occurs in ostriches (Skadhauge and Dawson, 1999) 
An emphasis towards the cost-effective production of ostrich leather and meat requires that 
commercial ostrich farmers produce slaughter ostriches as cost-effectively as possible 
(Huchzermeyer, 1998; Van Zyl, 2001 ). The characteristic low egg production performance of, and 
large variation in numbers of eggs produced in ostrich breeding systems necessitate the 
establishment of management practices to assist the ostrich farmer to manage his/her breeding 
flock to improve and optimize reproductive performance under commercial conditions (Cloete et al., 
1998; Bunter, 2002). Several factors may contribute to the low reproductive performance reported 
for ostriches under commercial breeding conditions, and may include a long interval between 
pairing and the first oviposition, irregular egg laying sequences, the age of breeding birds, nutrition, 
genetics, breeding environment, and management. During a breeding season, ostrich males and 
females may experience a period of quiescence that may last as long as 3-4 weeks. This 
quiescent period is characterised by a break in production and a loss of shin colour in the males 
(Stewart, 1989). Separation of ostrich males and females during this quiescent period is said to 
improve overall seasonal fertility and reproductive performance. Contradicting results, however, 
exist on the influence of a mid-season breeding rest period on the production and fertility of 
ostriches. Soley et al. (1991) found a positive relationship between good semen quality and fertility 
of birds not subjected to a breeding rest. 
Flush feeding (flushing) and the use of teaser rams are two management practices used in sheep 
to overcome the seasonal limitation of lamb production (Martin et al., 1986; Signore!, 1990; Nowers 
et al., 1994; Smith et al., 1996; Yildiz et al., 2001 ). Flushing refers to the management practice 
where feed quantity and quality is increased before the onset of a breeding season, with a 
stimulatory effect on the reproductive system of especially the female (McDonald, 1980). In some 
sheep breeds, ewes are preconditioned during spring, with the rams being isolated from the ewes. 
Introduction of the rams will then illicit a neuro-endocrine response that results in estrus, ovulation 
and conception (Lishman and De Lange, 1975; Martin et al., 1986). Ovulation rate can also be 
increased by subjecting ewes to flushing for at least one estrous cycle before being mated .. 
providing that the ewes are in a moderate body condition (Lishman and De Lange, 1975; Haresign, 
1981; Gunn eta/., 1991). 
5 
Management practices that can aid in the synchronization of breeding ostriches in order to identify 
and select early producing females will possibly contribute to the improvement of the overall 
performance of ostrich breeding birds under intensive breeding conditions. Flushing and teasing, 
and a mid-season breeding rest, were consequently investigated as potential management 
practices to synchronize and improve the overall seasonal production of breeding ostriches under 
commercial conditions. 
MATERIALS AND METHODS 
Experimental animals 
An ostrich breeding flock consisting of breeding pairs (n=136) maintained at the Little Karoo 
Agricultural Development Centre outside Oudtshoorn, South Africa, were used for the study. The 
management of the· breeding flock was documented by Van Schalkwyk et al. (1996) and Bunter 
and Graser (2000). The ages of the males and females used in the study varied between 2 and 12 
years. 
Experimental design 
The influence of flushing, teasing, and a mid-season breeding rest, was investigated in a 2X2X2 
factorial design trial, with breeding pairs allocated to their respective treatments groups at the 
beginning of each pre-season rest period. The treatment groups for the flushing and teasing 
management practices were as follows: 
• Control-group: Breeding males and females not subjected to either flushing or teasing 
• FT-group: Breeding males and females subjected to both flushing and teasing 
• F-group: Breeding males and females subjected to flushing only 
• T-group: Breeding males and females subjected to teasing only 
In addition to the above-mentioned treatments, approximately half of the number of breeding pairs 
allocated to each respective treatment group, was also subjected to a mid-season breeding rest. 
Management practices investigated 
Nutritional flushing: 
During each pre-season rest period of 4 months, males and females received a maintenance diet 
(8.5 MJ ME/kg DM and 9.1 % protein) that satisfied the daily maintenance requirements of the 
breeding birds. Two weeks before a breeding season, the diet of the breeding males and females 
was changed to an ad lib flushing diet (9 5MJ ME/kg OM and 16% protein) to prepare both males 
and females for the increased physiological stress experienced during the breeding season. 
6 
Teasing· 
During each pre-season rest period of 4 months, the males and females were pre-conditioned by 
visually isolating the males from the females, i.e. males and females were not allowed any visual 
contact. Two weeks before being joined for the breeding season, the males were introduced to 
camps neighbouring that of the females, and maintained in these camps until being joined. 
During the 2000/2001 breeding season, the ostrich males and females allocated to the FT and T-
groups were maintained as single-sex flocks during the pre-season and mid-season rest periods. 
During the 2001 /2002 breeding season, the ostrich males and females allocated to the Control-
and F-groups, i.e. not subjected to teasing, were maintained as mixed-sex flocks during the pre-
season and mid-season rest periods. 
A mid-season breeding rest 
Breeding pairs that were experiencing a decline in egg production were separated for a mid-season 
breeding rest approximately in the middle of each breeding season, i.e. when ostriches are said to 
experience a natural quiescence in production. Each mid-season rest period lasted 6 weeks. The 
first 4 weeks of the breeding rest simulated a stage where the production cycles of the breeding 
birds were completely arrested by removing the breeding birds from their breeding territories, and 
by placing them on a maintenance diet No visual contact was allowed between the breeding 
males and females during this period. During the remaining 2 weeks of the breeding rest, the 
males and females were again subjected to the flushing and/or teasing practices, as set out above. 
During the 2000/2001 breeding season, the breeding pairs were removed from their breeding 
camps at the end of October 2000, and joined again for breeding at the beginning of December 
2000. During the 2001/2002 breeding season, the breeding pairs were removed at the beginning 
of October 2001, and joined again for breeding in mid-November 2001. 
Reproductive parameters recorded 
Data recorded during the 2000/2001 and 2001/2002 breeding seasons, included the following: 
• Number of days from being paired to the first oviposition - recorded for the pre- and mid-
season rest periods 
• Daily and total egg production 
• Egg production performance (EPP) _ expressed as percentage, and calculated on a fortnightly 
basis by using the following equation (Van Schalkwyk el al., 1996): 
EPP= [total number of eggs produced/(0.5 x number of breeding days)] X 100 
• Number of clutches - a clutch being defined as a set of eggs laid, where each set of eggs are 
laid within 4 days of each other; recorded for pre- and mid-season rest periods 
• Egg weight on day of production {g) 
7 
• Day-old chick weight (g) 
• Number and percentage of chicks hatched The percentage of chicks hatched was calculated 
by using the following equation: 
Percentage of chicks hatched = [Number of chicks hatched I number of eggs set] X 100 
Statistical analysis 
Fixed effects in the analysis on the overall performance of females during the breeding seasons 
included treatment and female age. Least squares procedures were used to account for uneven 
subclasses (Harvey, 1990). The interaction of treatment with female age was included in the 
analysis, but found to be not significant. 
Random animal effects were computed to account for the covariance introduced by the repeated 
sampling of the same animal when the fortnightly egg production performance data were analysed 
ASREML, computer software that is capable of computing random effects in animal breeding, 
whilst also predicting least squares estimates of fixed effect means, was used for this purpose 
(Gilmour et al., 1999). Owing to the fact that typical seasonal patterns were expected for the 
fortnightly production means, these trends were modeled using cubic splines (Verbyla et al., 1999). 
The splines consisted of a fixed linear component, as well as random deviations from linearity 
conforming to a smooth trend. Random deviations from linearity not following a smooth trend were 
also included in initial runs, but were found to be insignificant (P::::0.05), as judged by the likelihood 
ratio test. These trends were interacted with treatment and female age where applicable, and the 
results were presented graphically. 
Various measures of performance were assessed in the 4 treatment groups that were subjected to 
a mid-season breeding rest period during the breeding seasons. As these analyses were also 
based on the same animals being sampled repeatedly, the same basic analyses used for the 
fortnightly egg production data were also applied. The fixed effects in this case included treatment, 
the production period involved (before or after breeding rest) and female age. The interactions 
between these main effects were computed where significant (P:o::0.05). Between female and 
residual variance ratios were obtained from the data sets subjected to the latter two sets of 
analyses. These estimates were used to obtain variance ratios depicting the within season 
repeatability estimates for the respective female traits considered {Turner and Young, 1969). 
RESULTS 
Flushing interacted with teasing during the second year of the study, thus complicating the 
separation of the respective influences of the two management practices. Therefore, the influence 
of flushing and teasing on the reproductive performances of the breeding pairs during both 
breeding seasons are presented in Tables 1 and 2. 
8 
Influence of flushing and teasing 
The reproductive performances of the breeding pairs were not influenced by either flushing or 
teasing during the 2000/2001 breeding season. The reproductive cycles of the breeding males and 
females allocated to the FT-group tended (Ps0.10) to be more synchronised, with time to first 
oviposition being shorter than that of the Control-, F- and T-groups (Table 1). 
Table 1. The influence of flushing and teasing on the reproduction traits (mean ±SE) of ostrich breeding pairs 
during the 2000/2001 breeding season. 
MANAGEMENTPRACT~E TREATMENT 
Flushing (F) Control Control Treatment (F) Treatment (F) 
Teasing (T) Control Treatment (T) Control Treatment (T) 
TRAIT 
Number of breeding pairs 31 33 32 35 
Days to first oviposition 54.6±11.0 58.6 ± 10.8 66.7 ± 10.9 45.8 ± 10.5 
Total EPP (%) 35.6 ± 4.0 36.8 ± 3.9 35.8 ± 3.9 42.8 ± 3.8 
Number of clutches 5.3 ± 0.6 5.2 ± 0.6 5.7 ± 0.6 5.4 ± 0.6 
Egg weight (g) 1426.1 ±20.2 1432.1±204 1450.1±20.1 1464.8±18.4 
Number of chicks hatched 21.8 ± 3.6 24.9 ± 3.6 26.6 ± 3.6 31.1±3.4 
Chicks hatched (%) 19.5 ± 3.2 22.4 ± 3.2 23.5 ± 3.2 27.9±3.1 
Chick weight (g) 877.7 ± 17.3 883.4 ± 17.5 885.7 ± 17.2 916.7 ± 15.8 
When the egg production performance (EPP) of the breeding pairs not subjected to a mid-season 
breeding rest is considered, neither flushing nor teasing significantly influenced the fortnightly EPP 
of the breeding pairs (Figure 1). Although not significant, the T-group had a lower EPP throughout 
the breeding season. The FT- and F-groups tended (Ps0.10) to have higher EPP's throughout the 
breeding season, with the Control-group being intermediate to the above-mentioned three groups. 
80 ----·-·-··-·------
70 - -- ------- ----
60 
50 ·-·------ ·-----
c 40 ±t''~J-
& 30 - -0-FT 
LU E------~~ 
20 
10 
0+-f"!LT~-.----,~-.---:.---.-~.---.--~~~~~~~~-,.-~.,.--, 
-10 ---------·----------------··-·--
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
Production period (fortnightly) 
Figure 1. The influence of flushing (F) and teasing (T) on the fortnightly egg production performance (EPP; 
mean ± SE) of ostrich breeding pairs not subjected to a mid-season breeding rest during the 2000/2001 
breeding season. 
9 
Similar to observations in the 2000/2001 breeding season, almost all reproductive parameters were 
unaffected by either flushing and/or teasing during t~1e 2001/2002 breeding season The only 
exception was the teasing management practice that improved (P:s:0.05) the EPP of the ostrich 
breeding pairs (Table 2). 
Table 2. The influence of flushing and teasing on the reproduction traits (means ± SE) of ostrich breeding 
pairs during the 2001 /2002 breeding season. 
--
MANAGEMENT PRACTICE TREATMENT 
--
Flushing (F) Control Control Treatment (F) Treatment (F) 
Teasing (T) Control Treatment (T) Control Treatment (T) 
-----
TRAIT 
Number of breeding pairs 32 34 33 35 
Days to first oviposition 56.1±8.9 40.2 ± 9.1 54.9 ± 8.7 61.5 ± 8.6 
Total EPP(%) 35.6 ± 3.8b 49.9 ± 3.9' 47.7 ± 3.7" 44.3 ± 3.7h 
Number of clutches 4.1±0.6 4.6 ± 0.6 4.3 ± 0.6 3.5 ± 0.6 
Egg weight (g) 1450.7 ± 20.9 1474.4 ± 20.6 1451.9±20.76 1456.3±18 7 
Chick production (number) 22.2 ± 3.5 314±3.5 26.3 ± 3.4 31.8±3.3 
Chicks hatched (%) 19.4 ± 3.0 28.8 ± 3.1 23.3 ± 2.9 29 8 ± 2.9 
Chick weight (g) 890.7 ± 16.2 903.8 ± ·1s.9 890.6 ± 15.9 886.5 ± 144 
a, D . 
. Columns with different superscripts differ s1g111ficantly (P:>0.05) 
Flushing and/or teasing did not influence the fortnightly EPP of the breeding pairs not subjected to 
a mid-season breeding rest during the 2001/2002 breeding season (Figure 2). There was a 
tendency (P:s:0.10) for the breeding pairs subjected to flushing, teasing, or a combination thereof, to 
have higher EPP's. However, the relative advantage of flushing and teasing decreased as the 
breeding season progressed. 
BO 
70 
60 
50 
c 40 ~rntr.l 
c.. 30 
c.. §
UJ 20 --0-FT ----
10 
0 
-10 
-20 
2 3 4 5 6 7 B 9 10 11 12 13 14 15 16 
Production period (fortnightly) 
Figure 2. The influence of flushing (F) and teasing (T) on the fortnightly egg production performance (EPP; 
mean ± SE) of ostrich breeding pairs not subjected to a mid-season breeding rest during the 2001/2002 
breeding season. 
10 
The mid-season breeding rest also had no detrimental influence on the reproductive performance 
of the rested breeding pairs during the 2001/2002 breeding season The rested breeding birds 
recovered swiftly from the forced break in production, and tended (P50.10) to have improved EPP's 
for the remainder of the production period, when compared to the non-rested breeding pairs 
(Figure 5). 
75 --~--------------------
65 - - - ----~------------.. ~~....,;-~----·-· . . 
55 -~,-:-- ··-·· ------ .. --~·~-- . __:·~,: .- -
45 
. ~--~-·- -- - - -- ----- .. 
[l_ 35 .y ----
[l_ 
UJ 
25 
15 
5 
-5 ---~-----
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
Production period (fortnightly) 
Figure 5. The influence of a mid-season breeding rest on the fortnightly egg production performance (EPP; 
mean± SE) of ostrich breeding pairs during the 2001/2002 breeding season. 
The beneficial influence of flushing and teasing became evident when the reproductive 
performances of the treatment and control groups were considered. The EPP of the breeding pairs 
was quickly restored to levels observed during the pre-rest period, and the FT-, F-, and T-groups 
producing significantly (P50. 01) more eggs than the Control-group during the post-rest period 
(Figure 6). 
75 
65 
55 ... Control 
45 §-o--T  ~ 35 [l_ --¢--F [l_ 
UJ 25 -o--FT 
15 
5 
-5 
.. 15 -~~ - - ~--- - --- --
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
Production period (fortnightly) 
Figure 6. The influence of a mid-season breeding rest on the fortnightly egg production performance (EPP; 
mean ± SE) of ostrich breeding pairs subjected to flushing (F) and teasing (T) management practices during 
the 2001/2002 breeding season. 
12 
The mid-season breeding rest improved (P:S:0.0'1) the post-rest EPP of the FT-group during the 
2000/2001 breeding season (Table 3). The breeding rest also synchronized the reproductive 
cycles of the FT males and females, resulting in an increase (P:S:0.01) in clutch size, i.e. a higher 
EPP together with a lower number of clutches produced, during the post-rest period (Table 3). 
Table 3. The effect of a mid-season breeding rest on the reproduction traits (means± SE) of ostrich breeding 
pairs subjected to flushing and teasing during the 2000/2001 breeding season 
TRAIT Before rest After rest 
EPP(%) 
Control 34.4 ± 7.0 47.7 ± 7.0 
Flushing only 32.2 ± 6.6 36.3 ± 6.9 
Teasing only 41.3 ±7.1 42.1±7.1 
Flushing and teasing 36.5 ± 6.4a 56.2 ± 6.7b 
Chick production (%) 
Control 19.4 ± 6.1 29.8 ± 6.1 
Flushing only 20.7 ± 5.7 21.3 ± 5.9 
Teasing only 25.3 ± 6.2 28.2 ± 6.2 
Flushing and teasing 24.4 ± 5.5 36.7 ± 5.8 
Days to first oviposition 73.3 ± 14.2 69.8 ± 14.5 
Number of clutches 3.3 ± 0.3· 1.9 ± 0.3u 
a, D ..  Rows with different superscripts differ significantly (P:>0.01) 
A mid-season breeding rest significantly (P:S:0.01) improved the post-rest EPP of the F-, T-, and FT-
groups during the 2001/2002 breeding season. A combination of flushing and teasing had to effect 
that fertility and hatchability was improved, resulting in more chicks (P:S:0.01) being hatched for the 
FT-group during the post-rest period. The interval between being joined and the first oviposition 
during the post-rest period was also significantly (P:S:0.01) decreased (Table 4). 
Table 4. The effect of a mid-season breeding rest on the reproduction traits (means ±SE) of ostrich breeding 
pairs subjected to flushing and teasing during the 2001/2002 breeding season. 
TRAIT Before rest After rest 
EPP(%) 
Control 20.9 ± 6.8 32.9 ± 6.8 
Flushing only 42.4 ± 6.6· 57.2 ± 6.6b 
Teasing only 44.8 ± 6.4a 60.8 ± 6.4b 
Flushing and teasing 29.5 ± 6.5· 62.1 ± 6.5b 
Chick production (%) 
Control 11.4±6.2 18.3 ± 6.2 
Flushing only 18.3 ± 6.0 24.9 ± 6.0 
Teasing only 25.3 ± 5.8 32.5 ± 5.8 
Flushing and teasing 23.0 ± 5.9· 41.3 ± 5.9b 
Days to first oviposition 54.9 ± 4.5· 10.9 ± 4.4u 
Number of clutches 2.7 ± 0.3 2.2 ± 0.3 
a, u:  Rows with different superscripts differ s1gmficant1y (P:>0.01) 
13 
DISCUSSION 
The peak production period observed for ostriches in this study confirms the photoperiod 
dependent reproduction pattern of ostriches (Jarvis et al., 1985; Hicks, 1992; Mellett, 1993; 
Horbavczuk and Sales, 1999). Results in this study are in agreement with Hicks-Aldredge (1993), 
i.e. that ostrich breeding seasons are characterised by two production peaks The first production 
peak occurred in late September to late October (spring in the Southern Hemisphere) and the 
second less pronounced production peak during late December (summer in the Southern 
Hemisphere). However, the peak breeding season reported in the study differs from the peak 
production season reported for Kenyan Red ostriches (Struthio came/us massaicus), i.e. from July 
to October (Bertram, 1979), or from September to December, with a peak in November (Hurxthal, 
1979). This supports the contention that the length and timing of ostrich breeding seasons may 
vary with latitude and altitude (Shanawany, 1994). Results also support the studies of Sauer 
(1972), Leuthold (1977), Jarvis et al. (1985), Cloete et al. (1998) and Bunter and Graser (2000), in 
that peak egg production occurred during August to December. A quiescent period in egg 
production that lasted approximately 4 weeks during the 2000/2001 and 6 weeks in the 2001/2002 
breeding season, respectively, was characterised by a decline in egg production. This period of 
quiescence is longer than the 3 to 4 week period of quiescence described by Stewart (1989). 
Average egg production of ostrich females during this study varied between 44 and 61 eggs for an 
8-month breeding season. This is considerably higher than the average egg production reported 
by Jarvis et al. (1985) for ostriches in Zimbabwe (16-17 eggs) under commercial farming conditions 
but falls well within the range of 0 to 120 reported by Van Schalkwyk et al. ( 1996) and Cloete et al. 
(1998). Egg weights reported in the present study fall within the range of 963-1827.89 reported by 
Bunter and Graser (2000). Chick production percentages in this study ranged from 19.37% to 
36.71%, which falls within the range of 0-78.2% reported by Bunter (2002). Another ratite species 
in which reproduction is influenced by photoperiod, the geater rhea (Rhea americana), produce on 
average 24-40 eggs, and 18 eggs per female under captive and natural breeding conditions, 
respectively (Navarro and Martella, 2002), which is lower in both cases than the egg production 
reported for ostrich females in this study. 
Onset of egg production, i.e. days from pairing to first laying, ranged from 40.2 to 66. 7 days. This 
differed from the mean of 35.9 days, and the range of o to 230 days reported by Bunter (2002). 
The number of clutches during a breeding period ranged from 3.5 to 5.7, which is less than the 
average of 6.3 clutches reported by Bunter and Graser (2000). The average clutch size reported in 
this study ranged between 7.7 to 15.3 eggs/clutch, which supports the clutch sizes reported by 
Leuthold (1970), Jarvis et al. (1985) and Horbavczuk and Sales (1999). Jarvis et al. (1985) 
reported clutch sizes of 12-13 eggs /clutch and 16 eggs/clutch for Zimbabwean ostriches in the wild 
and under commercial farming conditions, respectively. Horbavczuk and Sales (1999) reported 
clutches of 12-16 eggs, and Leuthold ( 1970) a clutch size of 22 eggs under natural breeding 
conditions for a single female in Tsavo National Park, Kenya. In emus (Dromaius 
14 
novaehollandiae), artificially inseminated females layed on average 6.7±1.6 eggs per clutch 
(Malecki and Martin, 2002). 
The influence of flushing and teasing 
Flushing, teasing, or a combination of flushing and teasing did not have a significant influence on 
most of the reproductive traits recorded during the study. Teasing was the only management 
practice that significantly influenced the egg production performance of the breeding pairs. Visually 
isolating of ostrich males and females during the pre-breeding rest period, and by allowing visual 
contact two weeks prior to being paired for breeding, elicited a significant physiological response in 
the ostrich females that resulted in an improved egg production performance. The visual presence 
of the male, together with a display of reproductive behaviour, stimulated the females to lay sooner 
after pairing and to produce larger clutches. Total egg and chick production, as well as egg and 
day-old chick weight, were not significantly influenced by flushing and/or teasing. 
Ostrich males and females may differ in terms of their response to proximal stimuli, in this case the 
presence of the opposite sex and the availability of feed with a higher nutritional value. It has been 
found in birds that mere light stimulation may induce complete spermatogenesis in males, whereas 
a female's complete reproductive state may only be achieved under the influence of various 
complementary factors (lmmelman, 1972). In many bird species, it has been found that a female 
will ovulate only after being together with the male for some time, which indicates that the male 
stimulates the female by means of for example courtship and other behavioural interactions. The 
behavioural activities of an ostrich male may thus aid in synchronizing the reproductive cycles of a 
breeding pair. In Houbara bustards (Chlamydotis undulata macqueenii; C. u. undulata) that breed 
in arid to semi-arid areas in Saudi Arabia, males displayed reproductive behaviour well in advance 
of the onset of egg production (Saint Jaime et al., 1996). The stimulatory effect of male courtship 
behaviour and vocalization has been confirmed in several studies for a variety of species 
(Lehrman, 1965; Brockway, 1965; Hinde, 1967). Tactile stimulation as well as auditory signals 
have been proven to be of great significance in stimulating female starlings (Stumus vulgaris), with 
female courtship influencing spermatogenesis in the male starling to a lesser extent than visa versa 
(Burger, 1953). 
Although contradictory results were obtained during the two breeding seasons, it stands to reason 
that the omission of flushing and teasing will have a retarding effect on the development and 
synchronization of ostrich males and females. In sheep, the omission of flushing and teasing 
management practices adversely affected the breeding performance of spring mated ewes 
(Nowers et al., 1994). In chickens, auditory and visual contact with males accelerated the sexual 
maturation of hens reared adjacent to or in mixed-sex flocks. Hens came into lay earlier and had 
larger combs than isolated females (Widowski et al., 1998). Egg production was significantly 
higher in turkey hens under natural mating conditions than in females visually exposed to turkey 
males or completely isolated from males. Females visually exposed to males also laid more eggs 
than those completely isolated from males (Jones and Leighton, 1987). 
15 
The influence of a mid-season breeding rest 
The return of reproductive viability can be regarded as a parameter for determining the successful 
recycling of turkey breeder hens (Moore and Siopes, 2003). In ducks, forced moulting is often 
used to stimulate egg production after a senescence in production (Olver, 1995). Arresting egg 
production in laying chickens by means of a forced moulting had a stimulatory effect on the post-
arrest egg production (Hurwitz et al., 1975). A mid-season breeding rest did not appear to 
negatively influence the EPP of the ostrich breeding pairs, with reproductive activity following the 
rest quickly restored during both years. The extent to which post-rest production was improved, 
however, differed between the two breeding seasons. The breeding rest seemed to have a more 
beneficial effect during the 2001/2002 breeding season, with rested breeding pairs tending to have 
an improved EPP for the remainder of the post-rest period. This tendency was not observed during 
the 2000/2001 mid-season breeding rest; the EPP of the rested breeding pairs appeared improved 
for only a short period and then declined to reach almost the same level as that of the non-rested 
groups. 
This difference in the effect of the breeding rest could be ascribed to the time of the breeding 
season when the breeding rest was enforced. A decreased sensitivity of the pituitary during the 
photorefractory state results from a decreased LHRH output from the hypothalamus (Wingfield et 
al., 1979; Storey and Nicholls, 1983). When the ostrich breeding pairs were separated for a 
breeding rest in 2000/2001, photoperiod was possibly still long enough to support reproduction 
when the breeding pairs were separated. The ostrich males and females, however, had already 
possibly started to enter a state of photorefractoriness due to a gradual decrease in day length. 
During 2001/2002, the breeding rest was enforced one month earlier. The rested breeding pairs 
were possibly still under strong photoperiodic control, and thus still photosensitized. Turkey 
breeder hens showed varying degrees of relative photorefractoriness early in a breeding season, 
with photorefractoriness increasing in severity as the breeding season progressed (Siopes and 
Proudman, 2003). Kumar and Kumar (1991, 1993) concluded that the time to onset of gonada! 
recrudescence and subsequent regression in brahminy myna (Sturnus pagodarum) depended on 
the length of the photophase as well as the time of year when exposed to stimulatory photoperiods. 
Season had a significant effect on the sensitivity of turkey hens to become photorefractory, with 
winter delaying the onset of photorefractoriness (Siopes, 2002). Age did not influence the mean 
time to become photorefractory, with second cycle turkey hens more inclined to become 
photorefractory than younger first cycle hens (Siopes, 2002). 
CONCLUSIONS 
Teasing of breeding ostriches significantly improved the egg production performance of the 
breeding ostrich females. It is possible that the effectiveness of flushing and teasing to 
synchronise the reproductive cycles of breeding ostriches is limited by pairing breeding ostriches 
before the winter solstice. Joining breeding ostriches after the winter solstice, i.e. when 
16 
photoperiod has increased sufficiently to initiate reproduction, may be necessary to ensure the 
successful implementation of flushing and teasing to synchronise the reproduction cycle of 
breeding ostriches. A mid-season breeding rest proved successful in reinitiating and synchronising 
the reproductive cycles of both the ostrich males and females, and also to recycle the ostrich 
females that were experiencing a decline in egg production. The extent to which the reproductive 
cycles of both sexes can be manipulated are most probably influenced by the state of 
photorefractoriness that the birds have reached during that specific stage of the breeding season. 
REFERENCES 
BERTRAM CR, 1979. Ostriches recognise their own eggs and discard others. Nature 710: 233-234. 
BROCKWAY BF, 1965. Stimulation of ovarian development and egg laying by male courtship 
vocalization in Budgerigars (Melopsittacus undulatus). Anim. Behav. 13: 575-578. 
BUNTER KL, 2002. The genetic analysis of reproduction and production traits recorded for farmed 
ostriches (Struthio came/us). Ph.D. dissertation, University of New England, Australia. 
BUNTER K & GRASER, H-U, 2000. Genetic evaluation for Australian ostriches - a report for the Rural 
Industries Research and Development Corporation. Publication no. 00/153, Rural 
Industries Research and Development Corporation, Australia. 
BURGER JW, 1953. The effect of photic and psychic stimuli on the reproductive cycle of the male 
Starling, Sturnus vufgaris. J. Exp. Zoo/. 124: 227-239. 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998. Progress towards a scientifically based 
breeding strategy for ostriches. Proc. of 3fjh Nat. Congr. S. Afr. Soc. Anim. Sci., 5-8 April, 
Stellenbosch, pp. 69-72. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3, NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
GUNN RG, MAXWELL T J, SIM DA, JONES JR & JAMES ME, 1991. The effect of level of nutrition prior 
to mating on the reproductive performance of ewes of two Welsh breeds in different levels 
of body condition. Anim. Prod. 52: 157. 
HARESIGN W, 1981. The influence of nutrition on reproduction in the ewe. 1. Effects of ovulation 
rate, follicle development and luteinizing hormone release. Anim. Prod. 32: 197. 
HARVEY WR, 1990. User's guide for LSMLMW and MIXMDL. PC-2 version, Mimeograph. 
HICKS-ALDREDGE KO, 1993. Ratite reproduction. Mufti-State Big Bird Conf, 24-25 April, 
HICKS KO, 1992. Ratite reproduction. Proc. Assoc. Avian Vet., pp. 318-324. 
HINDE RA, 1967. Aspects of control of avian reproductive development within the breeding season. 
In: Proc. XIV fnternat. Ornitho/. Congr., pp. 135-153, Oxford, Blackwell's, Oxford. 
HORBAvCZUK JO & SALES J, 1999. Reproduction is a crucial problem in ostrich breeding. Warfel 
Poult. 15(5): 28-30. 
HUCHZERMEYER FW, 1998. Ratites in a competitive world. Proc. 2nd Internal. Sci. Ratite Congr., 
21-25 September, Oudtshoorn, South Africa, pp. 1-3. 
17 
HURWITZ S, BORNSTEIN S & LEVY, 1975. Some responses of laying hens to induced arrest of egg 
production. Poult. Sci. 54(2): 415-422. 
HURXTHAL LM, 1979. Breeding behaviour of the ostrich Struthio came/us massaicus Neumann in 
the Nairobi National Park. Ph.D. thesis, University of Nairobi, Kenya. 
IMMELMAN K, 1972. Role of the environment in reproduction as source of "predictive" information. 
In: Breeding biology of birds - Proceedings of a symposium on breeding behaviour and 
reproductive physiology in birds, pp. 121-157 Ed.: OS Farner, National Academy of 
Science, Washington, USA. 
JARVIS MJF, JARVIS C & KEFFEN RH, 1985. Breeding seasons and laying patterns of the South 
African ostrich Struthio came/us. Ibis 127: 442-449. 
JONES MC & LEIGHTON AT JR, 1987. Effect of presence or absence of the opposite sex on egg 
production and semen quality of breeder turkeys. Pou/t. Sci. 66(12): 2056-2059. 
KUMAR BS & KUMAR V, 1993. Photoperiodic control of annual reproductive cycle in subtropical 
brahminy myna, Sturnus pagodarum. Gen. Comp. Endocrinol. 89(1): 149-160. 
KUMAR BS & KUMAR V, 1991. Seasonal reproduction in subtropical brahminy myna, Sturnus 
pagodarum: role of photoperiod Gen. Comp. Endocrinol. 83(3): 354-365. 
LEHRMAN OS, 1965. Interaction between internal and external environments in the regulation of the 
reproductive cycle of the ring dove. In: Sex and behaviour, pp. 355-380. Ed.: F Beach, 
Wiley, New York. 
LEITNER S, VAN'T HOF T J & GAHR M, 2003. Flexible reproduction in canaries is independent of 
photoperiod. Gen. Comp. Endocrino/. 130(2): 102-108. 
LEUTHOLD W, 1970. Some breeding data on Somali ostrich. East Afr. Wild!. J. 8: 206. 
LEUTHOLD W, 1977. Notes on the breeding biology of the ostrich, Struthio came/us, in Tsavo East 
National Park, Kenya. Ibis 119: 541-544. 
LISHMAN AW & OE LANGE GM, 1975. Oestrus activity and lambing rates in ewes following varying 
periods of association with vasoctemized rams prior to the commencement of the mating 
period. Agroanimalia 7: 93. 
LUMINEAU S, GUYOMARC'H C, VIVIEN-ROELS B & HOUDELIER C, 2002. Cyclic melatonin synchronises 
the circadian rhythm of feeding activity in Japanese quail, Coturnix c. japonica. Comptes 
Rendus Biologies 325: 205-212. 
MALECKI IA & MARTIN GB, 2002. Fertile period and clutch size in the emu (Dromaius 
novaehollandiae). Emu: J. ofR. Austr. Ornithol. Umon 102(2): 165-170. 
MARTIN GB, OLDHAM CM, COGNIE y & PEARCE OT, 1986. The physiological responses of 
anovulatory ewes to the introduction of rams - a review. Livestock Prod Sci. 15: 219. 
McDONALD LE, 1980. Veterinary endocrinology and reproduction, 3rd edition, Lea & Feb1ger, 
Philadelphia, USA. 
MELLETT FD, 1993. Ostrich production and products. In: Livestock production systems, principles 
and practice, pp. 187-194. Ed's: C Maree & NH Casey, Agri Development Foundation, 
Pretoria. 
18 
MOORE CB & SIOPES TD, 2003. Immune function in turkey breeder hens during the short day 
prelighting period and renewal of photosensitivity for egg production. Poult. Sci. 82(1 ): 
150-154. 
NAVARRO JL & MARTELLA MB, 2002. Reproductivity and raising of Greater Rhea (Rhea americana) 
and Lesser Rhea (Pterocnemia pennata) - a review. Archiv tor Gef/ugelkunde 66(3): 124-
132. 
NOWERS CB, COETZER WA & MORGENTHAL JC, 1994. Effect of melatonin implants, flushing and 
teasing on the reproductive performance of spring-mated Dohne Merino ewes. S. Afr. J 
Anim. Sci. 24(1): 22-26. 
OLVER MD, 1995. Effect of restricted feeding during the rearing period and a 'forced moult' at 40 
weeks of production on the productivity of Pekin breeder ducks. Brit. Pouft. Sci 36: 737-
746. 
SAINT JALME M, WILLIAMS J, MICKEAELIAN I & PAILLAT P, 1996. Seasonal variation of LH, sex 
steroids, body mass, molt, display, and laying in two subspecies of Houbara bustard, 
Ch/amydotis undulata macqueenii and Ch/amydotis undu/ata undulata, housed in outdoor 
cages under natural conditions. Gen. Comp. Endocrinol. 102(1 ): 102-112. 
SALDANHA CJ, SILVERMAN AJ & SILVER R, 2001 Direct innervation of GnRH neurons by encephalic 
photoreceptors in birds. J Biol. Rhythms 16(1): 39-49. 
SAUER EGF, 1972. Aberrant sexual behaviour in the South African ostrich. Auk 89: 717-737. 
SHANAWANY MM, 1994. The importance of light for ostriches. Ostrich Update 3: 52-54. 
SHANAWANY MM, 1995. Recent development in ostrich farming. WARIRMZ 83: 3-8. 
SIGNORET JP, 1990. The influence of the ram effect on the breeding activity of ewes and its 
underlying physiology. In: Reproductive physiology of Merino sheep: Concepts and 
Consequences, pp. 59-70. Ed's: CM Oldham, GB Martin IW Purvis, School of Agriculture 
(Animal Science), University of Western Australia, Nedlands, Perth, Australia. 
SIOPES TD, 1983. Effect of pinealectomy on broodiness of turkey hens. Poult. Sci. 62(11): 2245-
2248. 
SIOPES TD, 2002. Photorefractoriness in turkey breeder hens is affected by age and season. 
Poult. Sci. 81(5): 689-694. 
SIOPES TD & PROUDMAN JA, 2003. Photoresponsiveness of turkey breeder hens changes during 
the egg-laying season: relative and absolute photorefractoriness. Poult. Sci. 82(6): 1042-
1048. 
SKADHAUGE E & DAWSON A, 1999. Physiology. In: The ostrich - biology, production and health, pp. 
5181. Ed.: DC Deeming, CABI Publishing, Wallingford, Oxon, United Kingdom. 
SMITH KC, MORGAN KL & PARKINSON TJ, 1996. Mating patterns and reproductive wastage in 
commercial lowland ewes in Somerset. Vet. Rec. 139(23): 563-566. 
SOLEY JT, BERTSCHINGER HJ, ROBERTS DC, DE LANGE DH & BURGER WP, 1991. C,orrelation 
between semen quality and egg fertility in ostriches. Seminar on Ostrich Research, July, 
Oudtshoorn, South Africa, pp. 45-47. 
STEWART JS, 1989. Husbandry and medical management of ostriches. Proc. Assoc. Avian Vet., 
pp. 208-212. 
19 
STOREY CR & NICHOLLS T J, 1983. Responses of photosensitive and photorefractory intact and 
castrated male canaries Serinus canarius to treatment with synthetic mammalian 
luteinizing hormone-releasing hormone. Ibis 125: 228-234. 
TURNER HN & YOUNG SSY, 1969. Quantitative genetics in sheep breeding. Ithaca, New York. 
VAN SCHALKWYK SJ' CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations 
for body weight and reproduction in commercial ostrich breeding pairs. Brit. Poult. Sci. 37: 
953-962. 
VANZYL PL, 2001. 'n Ekonomiese evaluering van volstruisboerdery in die Oudtshoorn omgewing 
M.Sc. thesis, University of Stellenbosch, Stellenbosch, South Africa. 
VERBYLA AP, KENWARD BR & WELHAM SJ, 1999. The analysis of designed experiments and 
longitudinal data by using smoothing splines. Appl. Statist. 48(3): 269-311. 
WIDOWSKI TM, Lo FO, WONG OM & DUNCAN IJ, 1998. Rearing with males accelerates onset of 
sexual maturity in female domestic fowl. Poult. Sci. 77(1): 150-155. 
WINGFIELD JC, CRIM JW, MATTOCKS PW & FARNER OS, 1979. Responses of photosensitive and 
photorefractory male white-crowned sparrows (Zonotrichia leucophrys gambelit) to 
synthetic mammalian LHRH. Biol. Reprod. 21: 801-806. 
YILDIZ S, GUVEN B, COLAK A & UZUN M, 2001. Progesterone profiles of fat-tailed tuj ewes following 
tactile and visual separation of rams at the beginning of the breeding season. J. Vet. Med. 
Physiof. Pathof. Cfin. Med. 49(3): 132-136. 
20 
Chapter 3 
The influence of season and breeding rest periods on the 
reproductive performance of breeding ostriches ( Struthio 
came/us var. domesticus) under commercial farming 
conditions 
Abstract 
Breeding ostriches were paired for breeding at different times of the year to investigate the influence of onset of 
breeding on the reproductive performance under commercial farming conditions. Breeding was also interrupted at 
different times during the breeding season to determine the potential of forced rest periods to reinitiate reproduction 
and thus improve overall reproductive performance and fertility of breeding ostriches. Peak egg production was 
recorded during July and January, with lower production recorded during February to June. The time of the year 
when the ostriches were paired for breeding had a significant influence on the reproductive traits, with pairing off 
after the winter solstice positively influencing almost all of the reproduction parameters. A pre-season rest period 
was beneficial in terms of all reproductive traits, with breeding pairs subjected to a continuous breeding season 
performing worse throughout the entire breeding period. Forced rest periods during breeding significantly improved 
egg production but not fertility in the treatment groups. This implies that reproduction could be successfully 
reinitiated in ostrich females, and maintained in ostrich males. Although fertility was not improved by the forced rest 
periods, the ostrich males had to successfully fertilize the higher number of eggs produced by their companion 
females and accomplished this without a compromise in fertility. The time of year when the forced rest period was 
applied determined the extent to which reproductive activities could be reinitiated in ostriches. The fact that the 
reproductive activities of the breeding pairs that were paired before the winter solstice were synchronised to a higher 
degree than breeding pairs joined after the solstice, implies that the inhibitory effect of a short photoperiod, i.e. 
insufficient to stimulate reproduction, may be overcome by subjecting breeding ostriches to enforced breeding rest 
during a time of the year when the birds are still photosensitive enough to respond to this management practice. 
This may enable commercial ostrich farmers to manipulate the reproductive cycles of their breeding ostriches in 
order to ensure the timely production of slaughter birds according to market demand. 
INTRODUCTION 
Reproductive cycles of birds breeding in temperate regions are regulated by the seasonal change 
in photoperiod, and these species have become almost completely dependent on a change in 
photoperiod to initiate gonad development and reproductive activities (lmmelman, 1972). 
Photoperiod is by far the most reliable and predictive factor that can exert its influence weeks to 
months before the onset of breeding, with timing of breeding determined by the minimum 
photoperiod or "critical day length" and the development of photorefractoriness (Leitner et al., 
21 
2003). Critical day length varies between seasons and are a dynamic characteristic of 
photoperiodic processes (Siopes, 1994 ). 
An increase in photoperiod is thought to initiate two processes, firstly the secretion of the 
gonadotrophic hormones and subsequent gonadal maturation and reproduction, and secondly the 
development of photorefractoriness (Dawson et al., 1986; Foster et al., 1987). Photorefractoriness 
represents a period where the reproductive system becomes insensitive to photoperiod, with 
subsequent gonadal regression to a pre-breeding state (Nicholls et al., 1988; Rani et al., 2001; 
Leitner et al., 2003; Siopes and Proudman, 2003). Photorefractoriness represents almost the most 
highly developed termination scheme, and is a special adaptation to prevent birds from maintaining 
production when environmental conditions become unfavourable to the rearing of young, which 
usually takes place during the months after the summer solstice (Lofts and Murton, 1968). Once in 
a state of photorefractoriness, exposure to short day lengths is necessary to render birds sensitive 
to photoperiod again, thus initiating gonadal recrudescence and steroid synthesis and secretion 
(Wingfield and Farner, 1980; Farner, 1986; Kumar and Kumar, 1991 ). 
Photoperiodic responses in birds appear to involve direct interaction between photoreceptors and 
GnRH neurons, with the photic cues that regulate the timing of seasonal reproductive cyclicity 
detected by non-retinal, non-pineal deep brain photoreceptors. These photoreceptors 
communicate directly with GnRH-neurons, and represents a means by which photoperiodic 
information reaches the reproductive axis (Saldanha et al., 2001 ). In mammals, photoperiod also 
serves as an important proximate cue to initiate the onset of breeding, with the pineal gland and 
melatonin being an integral part in the regulation of the breeding cycle of e.g .. ewes (Bittman, 1984; 
Kennaway et al., 1982; Nowers et al., 1994; Bentley et al., 2000). In birds, however, melatonin 
does not fulfill the same role and is mainly used as a circadian 'clock' to regulate diurnal or daily 
activities (Siopes, 1983; Lumineau et al., 2002). It is unsure whether melatonin occurs in ostriches 
(Skadhauge and Dawson, 1999). 
Ostriches are generally regarded as seasonal breeders, with breeding seasons coinciding with an 
increase in photoperiod (Hicks, 1992; Mellett, 1993; Horbavczuk and Sales, 1999). Timing and 
duration of ostrich breeding seasons can also vary with latitude and altitude (Shanawany, 1994; 
Bertram, 1979; Sauer and Sauer, 1966a, 1969b). In southern Africa, ostrich breeding seasons are 
normally from June to January, with some seasons extending to the end of February (Jarvis et al., 
1985; Mellett, 1993). Deeming and Ar (1999), however, postulated that because ostriches occur in 
arid areas on both sides of the equator that their laying season is opportunistic rather than 
determined by daylength, they are good candidates for domestication and all-year round laying. 
The cost-effectiveness of a commercial ostrich enterprise is hampered by the low number of eggs 
produced and chicks hatched per female, and chick mortalities (Van Zyl, 2001). When compared 
to commercial poultry production systems, ostrich production systems are characterised by a low 
reproductive performance and large variation in egg and chick production, and consequently 
22 
slaughter bird production. Commercial ostrich producers need to manage their commercial 
breeding flocks optimally to ensure the cost-effective production of slaughter birds. The present 
study had a dual purpose, firstly to determine the influence of onset of breeding on the reproduction 
performance of ostriches over a breeding season, and secondly to determine the potential of rest 
periods during the breeding season to reinitiate reproduction in ostrich males and females. The 
ability to manipulate the reproductive cycles of breeding ostriches, i.e. onset and cessation of 
breeding, as well as the reinitiation of reproductive activities, may enable the commercial ostrich 
farmer to manage his or her breeding flock to ensure the timely production of slaughter birds 
according to market demand. 
MATERIALS AND METHODS 
Experimental animals 
An ostrich breeding flock maintained at the Little Karoo Agricultural Development Centre outside 
Oudtshoorn, South Africa, were used for the study. The management of the breeding flock is 
documented by Van Schalkwyk et al. (1996) and Bunter and Graser (2000). The ages of the males 
and females used in the study varied between 2 and 12 years. 
Experimental procedure 
Influence of season 
Continuous breeding (CS): At the end of the 2001/2002 breeding season, 24 randomly selected 
breeding pairs were treated for external and internal parasites, vaccinated for the Newcastle 
Disease virus, and returned to their breeding camps. These breeding pairs were not subjected to 
either a pre-season rest period, or flushing and teasing management practices (Materials and 
Methods, Chapter 2). The breeding pairs were paired on 1 March 2002 and separated on 8 April 
2003. 
Early season (ES): Fifty-one randomly selected breeding pairs were subjected to a 3-month rest 
period, as well as flushing and teasing management practices, before being paired for a 10-month 
breeding season The breeding pairs were joined on 2 May 2002 and separated on 19 February 
2003. 
Late breeding season (LS): Fifty-three randomly selected breeding pairs were subjected to a 5-
month rest period, as well as flushing and teasing management practices, before being paired for a 
10-month breeding season. The breeding pairs were joined on 1 July 2002 and separated on 8 
April 2003. 
23 
Influence of forced rest periods 
Twenty five breeding pairs of the ES group, and 26 breeding pairs of the LS group, respectively, 
were subjected to a series of forced rest periods throughout their respective breeding seasons. 
The time table for the separation and re-joining of the breeding pairs is set out in Table 1. 
Table 1. Time table of the forced breeding rest periods implemented during the 2002/2003 breeding season. 
Early breeding season Late breeding season 
Number of breeding pairs 25 26 
First forced rest period 
Start 10 July 2002 16 September 2002 
End 21 August 2002 21 October 2002 
Second forced rest period 
Start 30 October 2002 20 December 2002 
End 11 December 2002 28 January 2003 
The management of the breeding pairs during the rest periods consisted of the ostrich males and 
females being visually isolated during the first 4 weeks, and the males and females being fed a 
standard maintenance diet. Two weeks before being joined again with their respective breeding 
companions, the breeding pairs were subjected to standard flushing and teasing management 
practices (Materials and Methods, Chapter 2). 
Data recorded 
Data recorded for all breeding pairs included: 
• Time to first lay - defined as the number of days between the pairing of a breeding pair and the 
production of the first egg; recorded for each production period. 
• Number of clutches - defined as groups of eggs laid, where each set of eggs comprises of eggs 
laid within four days of each other; calculated for each production period 
• Daily egg production 
• Number of chicks - defined as number of chicks hatched from eggs set 
• Egg production performance (EPP) - to standardize for the difference in the number of 
production days, the total number of eggs produced was expressed as a percentage by using 
the following equation (Van Schalkwyk et al., 1996): 
EPP= [total number of eggs produced/(0.5 X number of production days)] X 100 
• Fertility- fertility of eggs produced by the companion female was regarded as the fertility of the 
breeding pair; expressed as a percentage and calculated by using the following equation: 
Fertility = [(number of eggs set - number of infertile eggs) I number of eggs set] X 100 
• Male territorial aggression - territorial aggression exhibited by the ostrich males was evaluated 
on a monthly basis (Materials and Methods, Chapter 5). 
24 
Statistical analysis 
The fixed effects in the analyses on the egg production performance (EPP) of the ostrich females 
over the entire breeding season included treatment and female age. Least squares procedures 
were used to account for uneven subclasses (Harvey, 1990). The interaction of treatment with 
female age was included in the analyses, but was found to be insignificant. The results pertaining 
to female age were consistent with literature reports (Bunter, 2002; Cloete et al., 1998), and are 
thus not reported. Selected linear contrasts were estimated between treatments to assist in the 
interpretation of the results. 
The analyses of monthly EPP, as well as male aggression and shin colour were complicated by the 
fact that the same animals were sampled repeatedly. These observations could thus not be 
considered as independent, as required by analysis of variance. Random animal effects were 
computed to account for the covariance introduced by the repeated sampling of the same animal. 
ASREML, computer software that is capable of the computation of various random effects in 
animals breeding, while also predicting least squares estimates of selected treatment means, was 
used for this purpose (Gilmour et al., 1999). The fixed effects included in the analyses were 
treatment, female age and month of production. Owing to the fact that typical seasonal patterns 
were expected in monthly production means, these trends were modeled using cubic splines 
(Verbyla et al., 1999). The splines consisted of a fixed linear component, as well as random 
deviations from linearity conforming to a smooth trend. Random deviations from linearity not 
following a smooth trend were also included in initial runs, but were found to be insignificant, as 
judged by the likelihood ratio test. These trends were interacted with treatment and female age 
where applicable. 
Various measures of performance were assessed in the two treatment groups that were subjected 
to three forced rest periods during the breeding season. Since these analyses were also based on 
the same animals being sampled repeatedly, the same basic analyses used for the monthly data 
were applied. Fixed effects in this case included the timing of the commencement of breeding 
(early or late), the production period involved (first, second or third) and female age. Interaction 
between these main effects were computed where significant (Ps0.05). Between female and 
residual variance ratios were obtained from the data sets subjected to the latter two sets of 
analyses. These estimates were used to obtain variance ratios depicting the within season 
repeatability of the respective female traits considered (Turner and Young, 1969). 
RESULTS 
The time when the breeding pairs were joined for breeding had a significant influence on almost all 
of the reproduction traits. The time from being joined to the first laying was significantly (Ps0.01) 
shorter for the breeding pairs paired for a late breeding season (LS), compared to the breeding 
pairs joined for an early breeding season (ES), and with a continuous breeding season (CS) (Table 
25 
2). When the ES females were compared to the CS females, the onset of production was not 
stimulated (P=0.26) by the time when the breeding pairs were paired for breeding (Table 2). The 
number of clutches was also not influenced by the time when the breeding pairs were paired for 
breeding. However, the LS females produced more eggs per clutch (P:s;0.05) than either the ES or 
CS females. The ES and CS females, on the other hand, produced clutches of similar sizes (Table 
2). 
When egg and consequent chick production are considered, the CS breeding pairs performed 
worse (P:s;0.01) than the ES and LS breeding pairs that were subjected to a pre-season rest period 
(Table 2). The CS breeding pairs also had lower (P:s;0.01) egg production performances (EPP's) 
than the breeding pairs subjected to, and not subjected to forced breeding rest periods, 
respectively (Table 2). Chick production was similarly affected. Average egg weight and average 
day-old chick weight was not influenced (P::::0.15) by the onset of breeding or the forced breeding 
rest periods (Table 2). The fertility of eggs produced by the CS breeding pairs was significantly 
(P:s;0.05) lower than that of breeding pairs subjected to a pre-season rest (Table 2). The fertility of 
the ES breeding pairs were improved (P:s;0.01) to a higher degree, when compared to the CS 
breeding pairs. The fertility of the LS breeding pairs tended (P:s;O. 1O ) to be higher than that of the 
CS breeding pairs. The fertility of the breeding pairs not subjected to forced breeding rest periods 
was significantly higher than that of the CS (P:s;0.01) and rested (P:s;0.05) breeding pairs. 
Table 2. Reproduction and egg traits (means ± SE) of ostrich breeding pairs with a continuous breeding 
season (CS), and breeding pairs paired before (early) or after (late) the winter solstice, and 
subjected to forced breeding rest periods. 
TREATMENT 
Continuous Early breeding season Late breeding season 
TRAIT season 
-
Not rested Rested Not rested Rested 
No. of breeding pairs 24 26 25 27 26 
Days to first egg 85 ± 11 60 ± 12 79 ± 12 24±13 20 ± 13 
Number of clutches 5.7 ± 0.8 5.9 ± 0.8 5.2 ± 0.8 5.5 ± 0.9 5.7 ± 0.9 
Eggs/clutch 6.6 ± 1.3a 7.32 ± 1.4a 7.8±1.3" 9.8±1.5b 9.9 ± 1.5b 
EPP(%) 16.9 ± 3.6 30.2 ± 4.0 33.5 ± 3 8 34.2 ± 4.4 45.7 ± 4.3 
Chick production 6.6 ± 3.0 17.9 ± 3.3 18.8±3.1 20.0 ± 3.6 23.0 ± 3.6 
-
No. of observations 20 25 22 27 26 
Egg weight (g) 1392 ± 39 1430 ± 27 1428 ± 32 1372 ± 29 1446 ± 29 
Water loss(%) 12.0 ± 0.9 11 7 ± 0.6 12.0 ± 0.8 12.1±0.7 11.4 ± 0. 7 
Chick weight (g) 852 ± 27 881 ± 19 869 ± 22 843 ± 20 893 ± 20 
Fertility(%) 66.3 ±4.4a 86.9 ± 5.0b, c, d 78.0 ± 4.6b' c,. 84. 1 ± 5.3b, d 69.2 ± 5.3b, e 
a, band d, e: Rows with different superscripts differ significantly (P,,0.05) 
a, c: Rows with different superscripts differ significantly (P,,0.01) 
When the average monthly EPP's of the treatment groups are considered, it is evident that 
treatment interacted with the linear and non-linear components of the spline for month (Figure 1). 
The breeding pairs subjected to a continuous breeding season performed on average worse than 
26 
the ES and LS breeding pairs The EPP of CS breeding pairs oscillated between 16.4% and 
31.8% for the first 8 months of the breeding season, before declining to below 10%. Joining 
breeding pairs two months later had to effect that the EPP of ES females not rested, increased 
from 16.8% to 39.0% in the month after being joined. The EPP of the ES, not rested group, 
continued to be higher than that of the CS group, and ranged between 42.4% and 48.9% in the 
following 4 months, declining to 2.4% at the end of their 10-month breeding season. The LS 
breeding pairs had an average higher EPP of 45.9% in the month of joining, which increased to 
between 52.6% and 60.5% for the following 3 months of their breeding season. Although their EPP 
declined in the following month, it continued to be higher than that of the CS and ES females during 
the 11th month of the study (Figure 1) . 
winter solstice 
70.0 I 13.0 
I 
60.0 12.0 'i:" _.. r. .. L..  Cll 50.0 L..  r. .. .I 11.0 ~ 10.0 E. 
c 40.0 _ I 9.0 ~ 
n. ::J fu  30.0 - 8.0 _g 
7.0 :c 
20.0 .~ 
6.0 >. 
10.0 Ill 
5.0 0 
0.0 '---+-----t--1-----J--1---------t----t---- V--+--U--t 4.0 
N N N N N N N N N N M M 
~ '[ ~ ~ ~ '5, 9 ~ 0 0 0 0 > c: .c 
~ < ~ ~ ~ ~ ~ 0 0 "Cl) z  Ill Cl) 0 ~ LL. 
Month 
r=o='Es~L8-=o=.cs-:-:--i..=-:~Num'ber<X~y1ightli;~1 
Figure 1. The influence of timing of breeding on the egg production performance (EPP) of ostrich breeding 
pairs with a continuous breeding season (CS), and breeding pairs joined before (ES) and after (LS) the winter 
solstice. 
The response curves for the ES and LS females subjected to forced rest periods were initially 
similar to ES and LS females not rested (Figure 2). The forced rest periods had a stimulatory 
influence on the EPP of the ES and LS females, for their EPP was elevated to levels generally 
exceeding those of the corresponding groups of females not subjected to resting. For example, the 
average EPP of rested ES females 1 month after the first rest period was 61.4%, compared to 
42.4% of the non-rested ES females. However, the relative advantage of the time when the 
breeding pairs were joined for breeding, and forced breeding rest periods, disappeared almost at 
the end of the study period (Figure 2). The EPP of the rested ES females was on average 33.8%, 
compared to 26.4% of the rested LS females. 
27 
t wmter solstice I 80.0 
700 ·-
60.0 
50.0 
~ 
a. 40.0 
a. 
w 30.0 
20.0 
10.0 
0.0 
N N N N N N N N N N M M M 
.0c  '2 0 0 c 0 0 >- :..;,  Cl sc . s 
0 0 0 0 0 
c. > (.) .c.,  .c .c (.) "' .:.:,i  ::i " 0 G> <I: G> 0 z "' G> u. i: :iii :!E <I: (/) 0 :!E 
Month 
Figure 2. The influence of forced breeding rest periods on the monthly egg production performance (EPP) of 
ostrich breeding pairs joined before (ES) and after (LS) the winter solstice. 
The benefit of a pre-season rest on the reproductive behaviour of ostrich males became evident 
during the later half of the breeding season, when the CS males exhibited territorial aggression to a 
lesser extent than the males with a pre-season rest period. The males subjected to forced 
breeding rest periods tended (P:s:0.10) to exhibit lower levels of territorial aggression than their non-
rested contemporaries (Figure 3). 
10 winter s~lstice - 12.0 
L:' 
11.0 Cl> 
.Q.. 9 .c ) 
0 10.0 E :l 
0 
rn .s 
c: 8 
0 D 
"ii) 
rn 
i:! 7 
Cl 
Cl 
< 6 
5 
N N N N N N N N N M M 
0 0 0 0 0 c :; a s 0 0 0 0 .c Cl 0 c .c .. (.) > (..)  ll. .:.:,i  .., ::i G> G> <I: <I: G> 0 0 :!E (/) z 0 ..".',  u. 
Month 
--¢--ES-no rest - - + - -ES-rested -0--- LS-no rest 
--&--cs - - O - - Daylight hours 
-----~-------------------------__, 
Figure 3. The influence of season and forced breeding rest periods on the territorial aggression of ostrich 
males with a continuous breeding season (CS), and ostrich males joined before (ES) and after (LS) the winter 
solstice. 
28 
The time when the breeding pairs were joined for breeding generally interacted with the first, 
second and third productive periods following the forced break in production of the breeding pairs 
(Table 3) The interval between pairing and the first laying of the breeding pairs paired after the 
winter solstice were shorter (P~0.01) than that of the breeding pairs paired before the winter 
solstice. The same effect was observed after the second forced rest period. The reproductive 
behaviour of the ES males and females, however, tended to be more synchronised after the 
second forced rest, with the number of days from being joined till the first laying being shorter than 
that reported for the LS males and females. The number of clutches produced, as well as eggs 
produced per clutch, followed roughly the same trend (Table 3). 
The EPP and consequent chick production of the LS females was, respectively 4.7 and 5. 1 higher 
(P~0.01) than that of the ES females during the first production period (Table 3). The beneficial 
influence of joining the breeding pairs after the winter solstice seemed to disappear as the season 
progressed, for no significant differences were found in terms of the egg and chick production of 
the ES and LS breeding pairs during the second production period (Table 3). A second forced 
breeding rest had a significant (P:s;0.01) influence on the reproductive performance of the ES 
females, for the EPP of the ES females was almost double that of the LS females (Table 3). Chick 
production of the ES breeding pairs was similarly affected, and was 3.5 times higher than that 
reported for the LS breeding pairs. 
Table 3. The influence of timing of breeding and forced breeding rests on the reproduction traits (means ± 
SE) of ostrich breeding pairs during the 2002/2003 breeding season. 
PRODUCTION PERIOD 
TRAIT AND TREATMENT 
First Second Third 
Days to laying 
Early breeding season 73.7±4.1a, 2 15.5 ± 4.51 18.0 ± 4.?1 
2 
Late breeding season 16.4 ± 4.3b, 1' 2 12.5 ± 4.31 25.3 ± 5.0
Number of clutches 
1 2 
Early breeding season 1.2±0.3°' 1 2.2 ± 0.32 1.9±0.3 '
1 
Late breeding season 2.5 ± 0.3b, 2 1.9 ± 0.41 2 ' 1.4 ± 0.3
Eggs/clutch 
2 2 
Early breeding season 3.2±1.6"' 1 9.7±1.6 7.7±1.6°'
1 
Late breeding season 12.2 ± 1.6b' 2 8.8 ± 1.?2 3.6±1.7b'
EPP(%) 
Early breeding season 13.7 ± 5.0•· 1 
3 
51.6 ± 5.0 39.3 ± 5.0 .. 
2 
2 1 
Late breeding season 64. 1 ± 5.0b, 3 49.8 ± 5.2 20. 1 ± 5.2b, 
Fertility (%) 
Early breeding season 85.2 ± 5.0 85.9 ± 4.6 89.4 ± 4.8 
Late breeding season 72.3 ± 4.6 76.5 ± 4.6 77.2 ± 5.0 
Chick production 
2 2 
Early breeding season 5.8±4.3°' 1 30.2 ± 4.3 26.4 ± 4.3°' 
2 1 
Late breeding season 29.3 ± 4.4b, 2 24.8 ± 4.4 7.6±4.4b,
a, b ..  values in columns drffer srgnr'f'r cantly (P:s0.01) 
'"": values in rows differ significantly (P:s0.01) 
29 
The repeatabilities of intra-season EPP and chick production were low to moderate, ie. 0.20 and 
0.41. The repeatability of the number of days from joining to the first oviposition, number of 
clutches and clutch size was weakly repeatable at 0. 11, 0.24, and 0. 10, respectively. Fertility and 
male aggression were highly repeatable, 0.41 for fertility and 0.56 for male aggression, 
respectively. 
DISCUSSION 
The influence of time of season when joined for breeding 
A delicate balance exists between a variety of environmental factors that protect and regulate the 
gametogenetic activity in birds (Lofts, 1975). Only after the reproductive system has been 
activated by photoperiod can it be sensitive to additional stimuli such as food availability and rainfall 
that then exert an influence on the hypothalamo-pituitary-gonad axis (Wingfield, 1983; Wingfield et 
al., 1992). An increase in photoperiod leads to a so-called neuro-endocrine cascade that is 
initiated by the secretion of the gonadotrophic releasing hormones from the hypothalamus (Murton, 
1975; Hau, 2001). This in turn stimulates the secretion of the pituitary hormones such as FSH and 
LH, which in turn stimulates gonadal growth and steroid hormone production (Ball, 1993). 
The extent to which the production traits of ostriches responded clearly illustrates their photoperiod-
dependent breeding strategy. This differs from the general breeding strategy of birds in arid or 
desert environments, where breeding tends be dissociated from photoperiodic control, and 
breeding may be more opportunistic (lmmelman, 1972; Lofts, 1975; Vleck, 1993) in the present 
study. Peak production of ostriches in the study occurred between the winter and summer solstice, 
with lower production observed between April and June. This was supported by the findings of 
Jarvis et al. (1985), but not that of Leuthold (1977). Jarvis et al. (1985) reported that for 
Zimbabwean ostriches, peak production occurred between July and end December-mid January. 
Leuthold (1977) speculated that ostriches in the Tsavo National Park in Kenya were more 
opportunistic in their breeding strategy, with rainfall playing a role in triggering reproductive 
activities. According to Vleck (1993), breeding schedules of birds in arid areas do not operate 
uniformly over all regions, and may differ in breeding periodicity. 
In the present study, onset of reproductive activity was significantly influenced by the time of year 
when the birds were paired for breeding. Pairing the birds after the winter solstice, i.e. after 
daylength starts to increase, had a beneficial influence on almost all the reproductive traits 
considered. The reproductive cycles of the ostrich males and females joined after the winter 
solstice were synchronized to a greater degree, with time to first laying being significantly shorter. 
Clutch size was also larger for birds paired after the winter solstice. The number of clutches 
produced, however, were unaffected by the time when the birds were paired for breeding. 
30 
The EPP and consequent chick production were significantly higher for breeding pairs joined after 
the winter solstice. This supports various studies that reported on the photoperiodic nature of 
ostrich reproduction (Smith et al., 1995; Bertram, 1979; Sauer and Sauer, 1996a, 1966b). The 
earlier onset to egg production and the higher EPP of the birds paired after the winter solstice must 
have resulted from the gonads that were fully matured because the breeding ostriches were in a 
fully photosensitized state when they were paired for breeding. This is in contrast to the breeding 
pairs joined before the winter solstice that were still presumably in a photorefractory state, and 
daylight length was too short to allow birds to react to the stimulating influence of the long 
daylengths, and thus allow the gonads to undergo recrudescence. The incubation parameters, i.e. 
egg weight and day-old chick weight, were not influenced by the timing of pairing. 
Short days, i.e. less than twelve hours of daylight can terminate the photorefractory state in birds 
and lead to the restoration of the photosensitive state (Wilson, 1992). This was clearly observed in 
this study for the birds paired before the winter solstice. The gradual increase in photoperiod had 
the effect that birds pairs paired off before the winter solstice gradually recovered from being 
photorefractory, with egg production gradually increasing after being paired for breeding. The 
decrease in egg production before the summer solstice is a result of the breeding pairs becoming 
less photosensitive and entering a photorefractory state, even though photoperiod was still 
sufficiently long enough to support production. In brahminy myna (Stumus pagodarum), it appears 
that birds do not enter a refractory state unless full gonadal maturation is achieved and that 
photorefractoriness develops in advance of gonadal regression. However, it was unsure whether 
photorefractoriness was fixed by daylength before the gonads have developed fully or whether it is 
fixed by the available daylight at the end of the gonadal growth period, as in starlings (Kumar and 
Kumar, 1991). 
The fertility of the breeding ostriches joined before the winter solstice, was significantly higher than 
breeding pairs with a continuous breeding season. The fertility of breeding pairs joined after the 
winter solstice, however, only tended to be higher than breeding pairs with a continuous breeding 
season. 
The influence of a pre-season rest period 
The beneficial influence of a pre-season rest period became evident when the reproductive 
performance of the breeding flock are considered. The cs males and females performed worse in 
terms of almost all reproductive traits. The CS breeding pairs had significantly lower EPP's than 
their ES and LS contemporaries. The physiological stress of the previous breeding season 
presumably depleted the reserves of the CS breeding pairs. The absence of a pre-season rest 
period, in which the ostrich males and females could restore their body reserves, possibly 
contributed to the lower EPP's observed. The EPP of the CS females oscillated between 
approximately 16 and 32% for the first two thirds of the breeding season, and declined and 
remained below 20% for the remainder of the season. The ES females performed better, with EPP 
31 
increasing from 16.4% to 39.0% in the first month after being joined. The LS females, however, 
outperformed the CS and ES females, with egg production 17.5% higher than that observed for the 
ES females during the first month after being paired for breeding. Although the EPP of the LS 
females declined as the season progressed, it was found to be higher than that of the CS and ES 
females. 
The influence of forced breeding rest periods 
The production curves of the ES and LS females subjected to forced breeding rest periods initially 
followed the same pattern as that of their non-rested contemporaries. When the production curves 
after each respective rest period are considered, however, it was found that the EPP of the rested 
breeding pairs generally exceeded that of their non-rested contemporaries. The advantage of the 
forced rest periods seemed to be almost absent at the end of the breeding period. 
The breeding males not subjected to forced breeding rest periods had significantly higher fertilities 
than males without a pre-season rest and males subjected to breeding rests. Fertility of the males 
subjected to forced breeding rests however, were maintained or even improved, for the breeding 
males joined before the winter solstice had to fertilize the higher number of eggs produced by their 
breeding companions during the post-rest periods. This was not the case for the ostrich males 
paired after the winter solstice, where fertility was unaffected by, and the EPP of their companion 
females not improved by the forced breeding rests. The maintenance and even potential 
improvement in the fertility of the breeding birds paired for an early breeding season, is 
contradictory to findings of Van Schalkwyk (1991 ). In the latter study, egg production was 
significantly improved for breeding birds paired before the winter solstice. The fertility of the 
breeding pairs, however, was compromised in the latter study, because more infertile eggs were 
reported for the post-rest period. 
The effectiveness of forced breeding rest periods and degrees to which the respective reproductive 
traits were affected, were influenced by the time of the breeding season when the forced breeding 
rest was implemented. A forced breeding rest period before the traditional quiescent period, i.e. 
from middle August to September, had a more pronounced effect on the reproductive traits than a 
forced breeding rest implemented later during the breeding season. This can be explained by the 
fact that daylight length was still sufficiently long to ensure that the breeding birds remained in a 
photosensitive state, and that the gonads have not yet started to regress. Forced rest periods had 
the effect that almost all of the reproduction traits of ES breeding pairs were improved in the 
second and third production periods. Reproductive traits that were further improved after the first 
forced rest, included days to nest, days to first egg, EPP, and number of clutches. Forced rest 
periods did not have a beneficial influence on the reproductive traits of the LS breeding pairs. The 
onset of a decrease in daylength had the effect that the LS males and females presumably entered 
a photorefractory state, and were becoming less sensitive to photoperiod that was still long enough 
to ensure and support reproduction. 
32 
The development of photorefractoriness (PR) is not an 'all-or-nothing' response. It can be 
considered as a complex interaction of age and previous experience, and may possibly be 
influenced by genotype (Siopes and Wilson, 1981 ). The incidence and onset of PR were found to 
be highly variable in first-year turkey hens Turkey hens also differed in terms of becoming 
absolutely or relatively photorefractory (S1opes, 2001) The incidence of PR was influenced by age 
of female turkey breeders, with 100% of recycled hens expressing PR. and 89% of first year 
breeders expressing PR (Siopes, 2002). First and second year hens did not differ in terms to the 
time to onset of PR (Siopes, 2002). The onset of PR is influenced by season, with hens stimulated 
in winter, taking longer to becoming photorefractory, when compared to hens stimulated in spring 
(Siopes, 2002). Varying degrees of PR may be experienced throughout a breeding season, with 
incidence and severity increasing as a breeding season progresses (Siopes and Proudman, 2003). 
Relative PR is experienced early in a breeding season, and increases in incidence as well as 
severity as a breeding season progresses. At the end of a breeding season when the photoperiod 
is still long enough to support the reproductive system, birds become refractory, with a subsequent 
decrease in gonadal steroid secretion and an decreased sensitivity of the reproductive system to 
long day lengths (Nicholls et al., 1988). It is thus possible the gonads of the LS males and females 
were less likely to respond to any changes in the steroid hormones that might have resulted from 
the potential stimulatory influence of the flushing and teasing management practices. The second 
forced rest, applied when photoperiod was progressively decreasing, had no significant influence 
on production, for the LS birds had already entered a relative to absolute refractory state, rendering 
them insensitive to any efforts to reinitiate reproduction. In turkey hens, longer intervals between 
and greater variation in LH surges were found late in the breeding season when compared to 
earlier in the season. This increase in intervals and variation in LH surges were also associated 
with lower egg production (Liu et al., 2002). Baseline LH levels were found to be low at the end of 
a long laying period, and exposure to long day photoperiod (14h:10h) had the effect that overall 
and baseline LH concentrations were increased (Bacon and Long, 1996). 
Light management is used in poultry species to stimulate reproduction and can be used to identify 
females of superior laying ability during the first cycle of lay (Lewis and Perry, 1995). The 
application of such lighting programs however, is impractical in the case of ostriches given their 
physical size as well as their specific behavioural requirements (Lambrechts et al., 1998). 
Selecting for responsiveness in broiler males and females resulted in a positive influence on egg 
production in females, and selection for early sexual maturity has potential for positively affecting 
reproductive performance and photosensitivity of female broilers (Eitan and Soller, 2000). 
Selecting for ostrich males and females that comes into production earlier may result in improving 
the overall performance of a commercial breeding flock. 
33 
CONCLUSIONS 
The study confirms the photoperiod-dependent breeding strategy of ostriches. Peak production 
occurred between July (winter in the southern hemisphere) and January (late summer in the 
southern hemisphere). Given the absence of reproductive performance-related selection in the 
ostrich industry, it is advisable that ostrich breeding flocks are joined after the winter solstice, when 
breeding birds have achieved full gonadal maturation. Subjecting breeding birds paired before the 
winter solstice to forced breeding rest periods, however, provided a means to overcome the 
influence of photoperiod on the reproductive activities of breeding ostriches. This may enable 
commercial ostrich farmers to manipulate the reproductive cycles of their breeding birds in order to 
ensure the timely production of slaughter ostriches. Forced breeding rest periods, however, were 
only effective in reinitiating reproductive activities during the time of season when ostriches are still 
photosensitive and have not yet entered a relative to absolute state of photorefractonness. 
REFERENCES 
BACON WL & LONG OW, 1996. Secretion of luteinizing hormone during a forced molt in turkey 
hens. Poult. Sci. 75(12): 1579-1586. 
BALL GF, 1993. The neural integration of environmental information by seasonally breeding birds. 
Am. Zoo/. 33: 185-199. 
BENTLEY GE, SPAR BO, MACDOUGALL-SHACKLETON SA, HAHN TP & BALL GF, 2000. Photoperiodic 
regulation of the reproductive axis in male zebra finches, Taeniopygia guttatta. Gen. 
Comp. Endocrinol. 117: 449-455. 
BERTRAM BCR, 1979. Ostriches recognise their own eggs and discard others. Nature 279: 233-
234. 
BITIMAN EL, 1984. Melatonin and photoperiodic time measurement: Evidence from rodents and 
ruminants. In: The pineal gland, pp. 155-192. Ed.: RJ Reiter, Raven Press, New York. 
BUNTER KL, 2002. The genetic analysis of reproduction and production traits recorded for farmed 
ostriches (Struthio came/us). Ph.D. dissertation, University of New England, Australia. 
BUNTER K & GRASER H-U, 2000. Genetic evaluation for Australian ostriches - a report for the Rural 
Industries Research and Development Corporation. Publication no. 00/153, Rural 
Industries Research and Development Corporation, Australia. 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998. Progress towards a scientifically based 
breeding strategy for ostriches. Proc. of the 361h Nat. Congr. S. Afr. Soc. Anim. Sci., 5-8 
April, Stellenbosch, pp. 69-72. 
DAWSON A, GOLDSMITH AR, NICHOLLS T J & FOLLETI BK, 1986. Endocrine changes associated with 
the termination of photorefractoriness by short daylengths and thyroidectomy in starlings 
(Sturn us vu/garis). J Endocrino/. 110: 73-79. 
34 
DEEMING DC & AR A, 1999. Factors affecting the success of commercial incubation. In: The 
ostrich - biology, production and health, pp. 159-190 Ed. DC Deeming, CASI Publishing, 
CAB International, Wallingford, Oxon, United Kingdom. 
EITAN Y & SOLLER M, 2000 Two-way selection for high and low responsiveness to 
photostimulation in broiler strain males. Poult. Sci. 79(9): 1227-1232. 
FARNER OS, 1986. Generation and regulation of annual cycles in migratory passerine birds. Am. 
Zoo/. 26: 493-501. 
FOSTER RG, PLOWMAN G, GOLDSMITH AR & FOLLETT BK, 1987. Imm unohistochem ical 
demonstration of marked changes in the LHRH system of photosensitive and 
photorefractory European starlings (Stum us vu/garis). J. Endocrinol. 115: 211-220. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3, NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
HARVEY WR, 1990. User's guide for LSMLMW and MIXMDL. PC-2 version, Mimeograph. 
HAU M, 2001. Timing of breeding in variable environments: tropical birds as model systems. 
Harm. Behav. 40: 281-290. 
HICKS KO, 1992. Ratite reproduction. Proc. Assoc. Avian Vet., pp. 318-324. 
HORBAvczuK JO & SALES J, 1999. Reproduction is a crucial problem in ostrich breeding. World 
Pou/t. 15(5): 28-30. 
IMMELMAN K, 1972. Role of the environment in reproduction as source of "predictive" information. 
In: Breeding biology of birds - Proceedings of a symposium on breeding behaviour and 
reproductive physiology in birds, pp. 121-157. Ed.: OS Farner, National Academy of 
Science, Washington, USA 
JARVIS MJF, JARVIS C & KEFFEN RH, 1985. Breeding seasons and laying patterns of the South 
African ostrich Strut/1io came/us. Ibis 127: 442-449. 
KENNAWAY DJ, GILMORE TA & SEAMARK RE, 1982. Effect of melatonin feeding on serum prolactin 
and gonadotropin and the onset of estrous cyclicity in sheep. Endocrinol. 110: 1766. 
KUMAR V & KUMAR BS, 1991. The development of photorefractoriness in termination of the 
breeding season in the tropical brahminy myna: role of photoperiod. Reprod. Nutr. Dev. 
31:27-36. 
LAMBRECHTS H, HUCHZERMEYER FW & SWART D, 1998. Integrating ostrich behaviour in commercial 
farming systems. Proc. of the 2nd !ntemat. Sci. Ratite Congr., 21-25 September 1998, 
Oudtshoorn, South Africa, pp. 167-169. 
LEITNER S, VAN'T HOF T J & GAHR M, 2003. Flexible reproduction in canaries is independent of 
photoperiod. Gen. Comp. Endocrinol. 130(2): 102-108. 
LEUTHOLD W, 1977. Notes on the breeding biology of the ostrich, Struthio came/us, in Tsavo East 
National Park, Kenya. Ibis 119: 541-544. 
LEWIS PD & PERRY GC, 1995. Effects of lighting on reproduction in poultry. In: Poultry Production, 
World Animal Science, volume 9, pp. 359-388. Ed.: P Jensen, Elsevier Publishers, 
Amsterdam, Holland. 
35 
Liu H-K, LONG OW & BACON WL, 2002. Interval between preovulatory surges of lutein1zing 
hormone increases late in the reproductive period in turkey hens. Biol. Reprod. 66(4): 
1068-1075. 
LOFTS B, 1975. Environmental control of reproduction. In: Avian Physiology, pp. 177-197, Ed. M 
Peaker, Academic Press, London, United Kingdom. 
LOFTS B & MURTON RK, 1968. Photoperiodic and physiological adaptations regulating avian 
breeding cycles and their ecological significance. J. Zoo/. Land. 155: 327-394. 
LUMINEAU S, GUYOMARC'H C, VIVIEN-ROELS B & HOUDELIER C, 2002. Cyclic melatonin synchronises 
the circadian rhythm of feeding activity in Japanese quail, Coturnix c. japonica. Comptes 
Rendus Biologies 325: 205-212. 
MELLETT FD, 1993. Ostrich production and products. In: Livestock production systems, principles 
and practice, pp. 187-194. Ed's: C Maree & NH Casey, Agri Development Foundation, 
Pretoria. 
MURTON RK, 1975. Ecological adaptation in avian reproductive physiology. In: Avian physiology, 
pp. 149-175, Ed.: M Peaker, Academic Press, London, United Kingdom. 
NICHOLLS T J, GOLDSMITH AR & DAWSON A, 1988. Photorefractoriness in birds in comparison with 
mammals. Physiol. Rev. 68: 133-176. 
NOWERS CB, COETZER WA & MORGENTHAL JC, 1994. Effect of melatonin implants, flushing and 
teasing on the reproductive performance of spring-mated Dohne Merino ewes. S. Afr. J. 
Anim. Sci. 24(1 ): 22-26. 
RANI S, SINGH S, MISRA M & KUMAR V, 2001. The influence of light wavelength on reproductive 
photorefractoriness in migratory blackheaded bunting (Emberiza melanocephala). Reprod. 
Nutr. Dev. 41(4): 277-284. 
SALDANHA CJ, SILVERMAN AJ & SILVER R, 2001. Direct innervation of GnRH neurons by encephalic 
photoreceptors in birds. J. Biol. Rhythms 16(1 ): 39-49. 
SAUER EGF & SAUER EM, 1966a. Social behaviour of the South African ostrich Struthio came/us 
australis. Ostrich (suppl.) 6: 183-191. 
SAUER EG & SAUER EM, 1966b. The behaviour and ecology of the South African ostrich. Living 
Bird 5: 45-75. 
SHANAWANY MM, 1994. The importance of light for ostriches. Ostrich Update 3: 52-54. 
SIOPES TD, 1983. Effect of pinealectomy on broodiness of turkey hens. Poult. Sci. 62(11): 2245-
2248. 
SIOPES TD, 1994. Critical day lengths for egg production and photorefractoriness in the domestic 
turkey. Poult. Sci. 73(12): 1906-1913. 
SIOPES TD, 2001. Temporal characteristics and incidence of photorefractoriness in turkey hens. 
Poult. Sci. 80(1): 95-100. 
SIOPES TD, 2002. Photorefractoriness in turkey breeder hens is affected by age and season. 
Poult Sci. 81(5): 689-694 
SIOPES TD & PROUDMAN JA, 2003. Photoresponsiveness of turkey breeder hens changes during 
the egg-laying season: relative and absolute photorefractoriness. Pou/t Sci. 82(6): 1042-
1048 
36 
SIOPES TD & WILSON WO, 1981. Termination of photorefractoriness in the chukar partridge 
(Alectoris graeca chukar) by low light intensity. J. Reprod. Fert. 63(1): 125-128. 
SKADHAUGE E & DAWSON A, 1999. Physiology. In: The ostrich - biology, production and health, pp. 
5181. Ed.: DC Deeming, CABI Publishing, Wallingford, Oxon, United Kingdom. 
SMITH WA, CILLIERS SC, MELLETI FD & VAN SCHALKWYK SJ, 1995. Ostrich Production - a South 
African perspective. In: Biotechnology in the Feed Industry - Proc. of Alltech's 11th Annual 
Symp., pp. 175-197. Ed.: T Lyons & KA Jacques. Nottingham University Press 
TURNER HN & YOUNG SSY, 1969. Quantitative genetics in sheep breeding. Ithaca, New York. 
VAN SCHALKWYK SJ, 1991. Die invloed van seisoene en rusperiodes op die reproduksieprestasie 
van volstruise. Ostrich Research Seminar, 17-18 July, Oudtshoorn, pp. 34-41. 
VAN SCHALKWYK SJ' CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations 
for body weight and reproduction in commercial ostrich breeding pairs. Brit. Poul!. Sci. 37: 
953-962. 
VANZYL PL, 2001. 'n Ekonomiese evaluering van volstruisboerdery in die Oudtshoorn omgewing. 
M. Sc. thesis, University of Stellenbosch, Stellenbosch, South Africa. 
VERBYLA AP, KENWARD BR & WELHAM SJ, 1999. The analysis of designed experiments and 
longitudinal data by using smoothing splines. Appl. Statist. 48(3): 269-311. 
VLECK CM, 1993. Hormones, reproduction, and behaviour in birds of the Sonoran desert. In: Avian 
Endocrinology, pp.73-86 Ed.: PJ Sharp, Journal of Endocrinology Ltd, Burgess Scientific 
I 
Press, Bristol, United Kingdom. 
WILSON FE, 1992. Photorefractory Harris' sparrows (Zonotrichia querula) exposed to a winter-like 
daylength gradually regain photosensitivity after a lag. Gen. Comp. Endocrinol. 87: 402-
409. 
WINGFIELD JC, 1983. Environmental and endocrine control of avian reproduction: an ecological 
approach. In: Avian endocrinology, pp. 265-288, Ed.'s: S Mikami, K Homma & M Wada, 
Springer-Verlag, Berlin. 
WINGFIELD JC & FARNER DS, 1980. Control of seasonal reproduction in temperate zone birds. 
Prag. Reprod. Biol. 5: 62-101. 
WINGFIELD JC, HAHN TP, LEVIN R & HONEY P, 1992. Environmental predictability and control of 
gonadal cycles in birds. J. Exp. Zoo/. 261: 214-231. 
37 
Chapter 4 
The influence of stocking rate and male:female 
ratio on the reproductive performance of breeding 
ostriches (Struthio came/us spp.) under commercial 
farming conditions1 
Abstract 
The reproductive performance of breeding ostriches, as influenced by stocking rate and male:female (M:F) ratio, 
were investigated under intensive breeding conditions. The stocking rates for large flocks ranged from 114 to 210 
birds per hectare, and the stocking rates for smaller flocks ranged between g and 13 birds in 0.13ha (1250 m2) and 
2
0.30ha (3000 m ) camps, respectively. The different M:F ratios investigated were 1M  1F  (pairs), 1M :2F (trios) and 
1M:3F (quads), for breeding systems maintained in 0.06ha (625 m2) camps. Total and average egg production, 
fertility, and hatchability were compromised when the stocking rate was increased in almost all the breeding 
systems. High stocking rates were detrimental to the reproductive performance and reproductive behaviour of the 
flocks. Increasing the number of females per male had no negative influence on the reproductive traits, with a 
significantly higher production observed for breeding quads. These findings indicated that ostrich breeding flocks 
can be maintained at stocking rates higher than those presently used on commercial ostrich farms. Ostrich 
breeding pairs, trios and quads can also be maintained on smaller areas, with acceptable production levels. 
Increasing the stocking rate of breeding flocks will potentially have a inhibitory effect on the establishment of 
territories and use of space by breeding ostriches, thereby impacting negatively on the reproductive behaviour of 
ostrich females and males in large flocks, respectively. 
INTRODUCTION 
Ostriches are the main farming interest in the Little Karoo region of South Africa, and used to 
occur in areas of the Little Karoo that are used mainly for agricultural purposes today. At present, 
almost 65% of the South African ostrich breeding population is found in the Little Karoo region 
(Van Zyl, 2001 ). On commercial ostrich farms where the entire or part of the production cycle, i.e. 
breeding birds, artificial incubation and raising of chicks and slaughter birds are represented, 
ostriches are concentrated on relatively small areas (South African Ostrich Business Chamber, 
2003). The natural vegetation in the Little Karoo is of low nutritional quality and natural veld areas 
are most often used only as holding areas for breeding flocks, which are then fed commercial 
breeder diets. This practice differs from those employed for other livestock farming systems, e.g. 
sheep and beef cattle, where natural vegetation serve as the main food source. 
1 Published in the S. Afr. J. Anim. Sci 34(2): 87-96 
38 
The maintenance of breeding ostriches in large flocks is a well-known practice in South Africa and 
Israel. Breeding flocks can range in size from 50 to 200 birds, respectively (Deeming and Bubier, 
1999). Flocks are usually maintained in camps several hectares in size, with no clear guidelines 
on the optimal stocking rate for ostriches under free-range conditions. Male:female ratios (M:F) in 
South African breeding flocks are 5-6 males for every ten females, which is similar to the 1M  2F 
ratio in Israel (Deeming and Bubier, 1999). Guidelines of the South African Ostrich Business 
Chamber (2003) require a maximum camp size of 0.25 ha for breeding pairs or trios when birds 
are fed a breeding diet. In the case of larger breeding flocks, the minimum stocking rate is one 
breeding bird/1 Oha for an 8-month breeding season. Breeding pairs, trios and quads are used in 
breeding systems in Australia, with the breeding quads consisting of 2M:2F (More, 1997). In 
Poland, ostriches are bred as pairs, trios or small flocks and most farmers pay special attention to 
the space requirements of breeders to create optimum exercise conditions (Horbanczuk, 2002). 
According to the draft recommendations of the Standing Committee of the European Convention 
for the Protection of Animals Kept for Farming Purposes, the minimum space requirement for the 
maintenance of ratite breeding birds is 1200 m2 per bird, with larger areas recommended to 
ensure well-being and proper exercise conditions (Horbanczuk, 2002). According to the Code of 
Conduct for animal welfare in New Zealand (Animal Welfare Advisory Committee, 1998), the 
stocking rate for ostrich breeding pairs in open conditions should not exceed 20 birds per ha for 
breeding flocks. The minimum camp size required for breeding pairs and trios is 0.06 ha and 0.20 
ha, respectively. No status in terms of percentage vegetation cover was mentioned in this 
document. 
The intensive nature of ostrich farming presents both the commercial farmer and the 
conservationist with a challenge to ensure that conservation of natural resources can co-exist 
together with commercial ostrich farming. Scientific research on the production of ostriches in 
large commercial breeding systems are scarce, and this complicates the development of best 
practice techniques to enable producers to farm as cost-effectively as possible (/\dams and 
Revell, 2003). Most of the available literature on the egg production performance of ostriches 
under intensive farming and wild conditions are compiled for breeding pairs, breeding trios, and 
small flocks (Hurxthal, 1979; Jarvis et al., 1985; Foggin and Van Niekerk, 1995; Hicks-Aldredge, 
1996; More, 1997; Deeming and Bubier, 1999; Bunter and Graser, 2000; Lambrechts et al., 
2002). This study presents results on the reproductive performance of a large commercial 
breeding population, and the influence of different stocking rates and male:female ratio on 
reproductive traits. 
MATERIALS AND METHODS 
Experimental animals 
Production data was obtained from an ostrich breeding flock of 4500 birds during two consecutive 
breeding seasons. The breeding flock were maintained as a single breeding operation near 
39 
Lad1sm1th. in the Western Cape Province of South Africa. The breeding flock consisted of birds 
maintained continuously on the farm. and breeding birds translocated from another commercial 
breeding farm in the North-Western Province of South Africa. The North-Western birds were 
translocated to the Ladismith breeding farm approximately two months before the onset of the 
2000/2001 breeding season. 
Experimental design 
The different stocking rates and M:F ratios studied during the two consecutive breeding seasons 
are presented in Table 1. 
Table 1. Stocking rates and male:female (M:F) ratios implemented in the breeding program of a commercial 
ostrich breeding operation maintained near Ladismith in the Western Cape, South Africa. 
Stocking rate M:F ratio Breeding season studied 
Camp size (ha) Camp size (m2) 
(breeding system) 
--- -· 
Breeding pair 1: 1 0.06 and 0.13 625 and 1250 2000/2001 & 2001/2002 
Breeding trio 1: 2 006and0.13 625 and 1250 2000/2001 & 2001/2002 
Breeding quad 1 :3 0.06 625 2000/2001 
9-bird flock 1: 2 0.13 and 0.30 1250 and 3000 2000/2001 & 2001/2002 
13-bird flock 1 :2.25 0.35 3500 2000/2001 
114 bird flock 1 :2 1 10 000 2001/2002 
120-bird flock 1: 2 1 10 ODO 2000/2001 
130-bird flock 1: 2 1 10 ODO 2001/2002 
141-bird flock 1 :2 1 10 ODO 2001/2002 
150-bird flock 1:2 1 10 000 2000/2001 
210-bird flock 1: 2 1 10 000 2000/2001 
During the 2000/2001 breeding season, the breeding birds maintained continuously on the 
Ladismith farm were allocated to only large breeding units, and the imported breeding birds were 
allocated to both large and small breeding units. The breeding birds were joined in July 2000 for 
a 7-month breeding season. At the end of August 2000, 40 females and 20 males were removed 
from each of four 210-bird flocks to establish two 120-bird and four 150-bird flocks. As a result of 
the shorter breeding season for the 120-bird and 150-bird flocks, only reproductive data for the 
remaining 5 months of the 2000/2001 breeding season are presented. 
Based on the results of the 210-bird flocks during the 2000/2001 breeding season, three breeding 
systems, i.e. 114-bird, 130-bird, and 141-bird flocks, were established. The breeding quads and 
210-bird flocks did not form part of the experimental groups in 2001/2002 breeding season. Due 
to the allocation of a larger number of birds to large breeding flocks (results not presented here) 
that were maintained in approximately 40-70ha camps, only 114 and 141 birds were allocated to 
the 120-bird and 150-bird breeding systems, respectively. The breeding birds were joined in July 
2001 and results presented here represent data recorded for the entire 7-month breeding period. 
40 
During both breeding seasons, the breeding birds received a complete balanced breeder diet at 
2.5kg/bird.day and had free access to fresh, clean water. Eggs were collected daily and identified 
by means of camp number and date of production. Production was recorded individually for each 
breeding system. After collection, the eggs were cleaned by dry wiping and disinfected by 
fumigation (80g potassium permanganate in 130ml 40% formaldehyde solution per m3 for 20m1n; 
Smith et al., 1995). All eggs not suitable for incubation were noted for each breeding system and 
removed from storage Reasons for rejection included broken/cracked eggshells, chalky 
eggshells and loose air cells (i.e. where the air cell was damaged during transport to the 
incubation facilities). 
All eggs were stored in an upright position with the air cell uppermost for a maximum of 10 days at 
17-18 °C and 75% relative humidity, and were turned once daily through an angle of 45°. The 
eggs were artificially incubated in Buckeye® electronic incubators at 36°C and 28% relative 
humidity for a period of 38 days before being transferred to the hatchers for the remainder of the 
incubation period. Eggs were candled on day 14 to establish early embryonic deaths and fertility, 
and on day 38 to establish late embryonic deaths. After hatch, chicks were kept in the hatcher for 
a period of 24h to allow sufficient time for the navel to close, before being dispatched to contract 
growers. All eggs and reasons for not hatching during incubation were noted throughout the 
study. 
Definitions for reproduction traits are as follows: 
• Total egg production = total number of eggs produced per breeding system during the 
production period 
• Chicks (number) = number of day-old chicks hatched 
• Average EP!female = average number of eggs produced per female during the production 
period 
• Fertility(%)= ((Total eggs set- number of infertile eggs)/total eggs set] x 100 
• Hatchability (%) =( Number of chicks hatched/Total number of eggs incubated) x 100 
Statistical analysis 
The influence of stocking rate and male:female ratio were assessed by one-way analysis of 
variance procedures. In cases where trends were expected and where degrees of freedom for 
treatments exceeded one, the degrees of freedom for the number of birds were partitioned in 
orthogonal polynomials depicting linear tendencies. The LSMLMW computer program (Harvey, 
1990) was used for this purpose. When more than two means were compared, least significant 
differences were computed, provided that the comparisons were protected by a significant value 
in the ANOVA table (Snedecor and Cochran, 1967)). When only two means were compared, 
differences between means were tested for significance by using the F-test derived from the 
ANOVA table (Snedecor and Cochran, 1967). 
41 
RESULTS 
The 210-bird flocks had on average a higher total egg production during the 2000/2001 breeding 
season, when compared to the 150-bird and 120-bird flocks (Table 2). Although the F-value far 
treatment (flock size) in the analysis of variance table did not reach statistical significance 
(P=0.13), there was a suggestion of a linear increase (P:S0.05) in total egg production per camp 
with an increase in colony size. This increase in egg production amounted to 11.9±5.4 eggs per 
female, and accounted for 99.6% of the variation associated with treatments. The same tendency 
was observed for the number of chicks hatched, with an increase of 3.11 ±1.86 chicks hatched per 
breeding bird (P;::0.12). However. fertility and hatchability declined significantly P:S0.05) with an 
increase in stocking rate. The linear regression coefficient for fertility amounted to -0.06±0.02% 
2
per bird (R =0.99), and for hatchability to -0.11±0.03% per bird (R 2=0.96). Average egg 
production was independent of stocking rate. 
Table 2. Influence of stocking rate on the reproductive traits (means ± SE) of ostriches maintained in 1 ha 
camps during the 2000/2001 breeding season. 
SIZE OF BREEDING FLOCK (1M:2F) 
PARAMETER 120 birds 150 birds 210 birds 
Number of breeding systems 2 4 8 
Total EP 1748±515 2181±364 2859 ± 257 
Number of chicks hatched 515±176 707 ± 124 844 ± 88 
Average EP/female 20.0 ±4.1 22.9 ± 2.9 20.9 ± 2.0 
Fertility(%) 85.8 ± 1 9• 83.2 ± 1.48 79 9 ± 0 9b 
Hatchability (%) 70.1 ± 2.6" 64.6 ± 1.9b 59.4 ± 1.3c 
a, 0 1 C: Columns with different superscripts differ s1gn1f1cantly (Ps:0.05) 
During the 2001/2002 breeding season, egg production was compromised with an increase in 
stocking rate in the 1 ha camps. Total egg production for the 141-bird flock declined significantly 
(P:S0.01) by 21.3% and 20.2%, when compared to the 130-bird and 114-bird flocks, respectively. 
Chick production was similarly affected, and declined by 21.2% and 22.0%, when compared to 
the 130-bird and 114-bird flocks, respectively (Table 3). The average egg production declined 
2
significantly (P:s::0.01) by 0.44 ± 0.01 eggs per bird as flock size increased (R =0.94). Fertility and 
hatchability did not differ between the respective treatments (P=0.13 and P=0.29, respectively). 
However, when the degrees of freedom were partitioned in orthogonal polynomials, a linear 
regression coefficient amounting to -0.15 ± 0.07% per breeding bird were observed for fertility 
2
(P:S0.05; R2=0.99). Hatchability similarly tended to decline (-0.09 ± 0.05%, P=0.1 O; R =0.99) with 
an increase in stocking rate 
42 
Table 3. Influence of stocking rate on the reproduction traits (means ± SE) of breeding ostriches maintained 
in 1 ha camps during the 2001/2002 breeding season. 
SIZE OF BREEDING FLOCK 
PARAMETER 114 birds (38M:76F) 130 birds (43M:87F) 141 birds (47M:94FJ 
Number of breeding systems 8 4 8 
Total EP 2793 ± 121• 2834±171 3 2230 ± 121b 
3 
Number of chicks hatched 1938 ± 873 1917±123 1510 ± 87b 
Average EP/female 3 34.3±1.43 30.5±1.9 22.1±1.4b 
Fertility(%) 83.6±1.3 81.2±1.9 79.6±1.3 
Hatchability (%) 784±1.0 77.3 ± 1.5 76.0 ± 1.0 
a, u, 
. Columns with different superscripts differ s1gnif1cantly (Ps0.05) 
Under a constant male to female ratio of 1: 2, fertility was improved by 9.5% for the 9-bird flocks 
maintained in the 0.13 ha camps, when compared to the 9-bird flocks maintained in 0.30ha camps 
(Table 4). A similar tendency was observed for hatchability, which was 6.0% higher for the 9-bird 
flocks maintained in the smaller 0.13ha camps. Nine-bird flocks maintained in 0.13ha camps 
produced on average 23.5 more chicks (P:s:0.05) than the 9-bird flocks maintained in the larger 
0.30 ha camps. 
Table 4. Reproduction traits (means ±SE) for 9-bird breeding flocks (1 M:1 F) maintained in 0.13 ha and 0.30 
ha camps during the 2001/2002 breeding season. 
CAMP SIZE 
PARAMETER 0.13 ha 0.30 ha 
Number of breeding systems 50 50 
Total EP 256 ± 11• 233 ± 11 b 
3 
Number of chicks hatched 182.6 ± 84 147.8 ± 84b 
Average EP/female 34.1±14
8 29.7±14b 
Fertility(%) 81.6±14
8 74.5 ± 1.4b 
Hatchability (%) 75.1 ± 1.0• 70.6 ± 1.0b 
I ' 
a, D : Columns with different superscripts differ significantly (P,;0.05) 
Average egg production per female was 68% higher (P:s:0.01) for breeding trios, when compared 
to the 9-bird flocks (Table 5). However, fertility and hatchability of eggs produced by 9-birds 
flocks was significantly improved by 19.6% and 28.9%, respectively (Table 5). When the average 
egg production per female are compared on a per trio basis, i.e. the 9-bird flock can be 
considered equivalent to two breeding trios, the 9-bird flocks produced on average 15.8% more 
eggs on a per trio basis than did the breeding trios. 
43 
Table 5. Influence of stocking rate on the reproductive traits (means ±SE) of breeding ostriches maintained 
in 0.13 ha camps during the 2001 /2002 breeding season. 
FLOCK SIZE (M:F) 
PARAMETER Trios (1 M:2F) 9-bird colony (3M 6F) 
Number of breeding systems 83 50 
Total EP 114.7 ± 5.8° 255.6 ± 74b 
Number of chicks hatched 67.7 ± 4.6° 182.6 ± 5.9b 
Average EP/female 57.4 ± 1.8° 34.1 ± 2.4b 
Fertility(%) 68.2 ± 1.3° 81.6±1.7b 
Hatchability (%) 58.2 ± 1.3° 75.1 ± 1.6b 
a, D: 
Columns with different superscripts denote significance (Ps0.05) 
As can be expected, total egg and chick production increased linearly (P:'.'.:0.01) with an increase in 
the number of females per breeding system (Table 6). Average egg production per female, and 
fertility of eggs produced, however, did not differ between the different treatments. Hatchability 
was significantly higher (Ps0.05) for eggs produced by the breeding quads (Table 6). 
Table 6. Influence of male:female ratio on the reproduction traits (mean ± SE) of breeding ostriches 
maintained in 0.06ha breeding camps during the 2000/2001 breeding season. 
BREEDING SYSTEM (M:F) 
PARAMETER Pairs (1M:1F) Trios (1M:2F) Quads (1 M:3F) 
Number of breeding systems 50 125 8 
Total EP 25.9 ± 2.3° 52.1 ± 1.4b 92.6 ± 5.7° 
Number of chicks hatched 21.9 ± 2.0• 42.6 ± 1.3b 80.5 ± 5.0c 
Average EP/female 21.6±1.1 21.7±0.7 25.8 ± 2.7 
Fertility(%) 74.3±2.1 77.6 ± 1.3 83.8 ± 5.2 
Hatchability (%) 70.0 ± 2.2• 73.5 ± 1.4b 81.1±5.4c 
a, 0 1 C: Columns with different superscripts differ s1gn1f1cantly (Ps0.05) 
Although maintained in slightly larger camps (0.30 vs. 0.35 ha) and at a higher stocking rate, the 
13-bird flocks performed on average better in terms of all reproduction traits, when compared to 
the 9-bird flocks (Table 7). Total egg production was significantly (Ps0.01) increased for the 13-
bird flocks. A similar tendency (P:'.'.:0.007) was observed for average egg production per female. 
Thirteen-bird flocks produced on average 233.7 ± 8.6 eggs in total and 21.7 ± 0.9 eggs/female, 
compared to 129.7 ± 8.6 eggs in total and 18.0 ± 0.9 eggs/female for the 9-bird flocks. Fertility 
and hatchability was also improved (Ps0.01) in the 13-bird flocks. 
44 
Table 7. Reproduction traits (means ± SE) recorded for 9-bird and 13-bird breeding flocks maintained in 
0.30 and 0.35 ha camps, respectively, during the 2000/2001 breeding season. 
FLOCK SIZE (M:F) 
PARAMETER 9-birds ( 1M  2F) 13-birds (1M2.25F) 
Number of breeding systems 25 25 
Total EP 129.7 ± 8.6· 233.7 ± 8.6b 
Number of chicks hatched 107.9 ± 7.4. 198.3±7.4b 
Average EP/female 180 ± 0 9• 21.6 ± 0 9b 
Fertility(%) 84.9 ± 1.3" 90.3 ± 1.3b 
Hatchability (%) 79.4±1.4" 86.1±1.4b 
a, D: 
Columns with different superscripts differ significantly (Ps0.01) 
DISCUSSION 
The influence of stocking rate 
During the 2000/2001 breeding season, egg production was higher for the 210-bird per hectare 
stocking rate, with egg production increasing linearly as the stocking rate was increased. There 
also was a tendency for the number of chicks hatched per bird to be increased as stocking rate 
was increased. Fertility and hatchability, however, decreased as stocking rate increased. The 
respective stocking rates did not influence average egg production. During the 2001/2002 
breeding season, however, egg production was compromised when stocking rate was increased. 
Total egg production decreased in the 130-bird and 141-bird flocks, when compared to the 114-
bird flocks maintained in the 1 ha camps. Average egg production decreased by 0.44±0.007 eggs 
per bird as stocking rate was increased. Fertility was lower and hatchability tended to decrease 
for the higher stocking rates. 
The contradictory results obtained during the two breeding seasons can be explained by the 
translocation of a proportion of the breeding population to a new breeding environment two 
months before the onset of breeding during 2000/2001 breeding season. These newly imported 
birds were allocated to a number of the 210-bird flocks. These breeding birds may not have had 
sufficient time to adapt to their new environment. Translocation to a new and unfamiliar 
environment results in disorientation and the consequent manifestation of abnormal behavioural 
patterns (Lambrechts et al., 1998b). Results from this study suggest that the respective flocks 
were still acclimatizing to the new breeding environment that differed in terms of vegetation, 
climate and management. With the onset of breeding during the 2001/2002 breeding season, the 
imported birds had sufficient time to acclimatize and this was reflected in the production figures of 
the flock at the lower 114 birds per hectare stocking rates. 
The tendency for fertility and hatchability to be negatively influenced in the larger flocks during 
both breeding years, support the reasoning that too high a stocking rate adversely affects the 
45 
reproductive performance of ostriches. It 1s also possible that cort1costerone levels were higher in 
females housed at the higher stocking rates and this may have had a negative influence on 
ovulation and subsequent oviposition. Corticosterone levels were found to be higher in laying 
hens housed in higher rates (Mashaly et al., 1984). Abnormal doses of corticosteroids resulted in 
premature LH release and an irregular laying patterns in laying hens (Gilbert, 1971 ). 
The negative influence of the higher stocking rates on normal reproductive behaviour also 
possibly contributed to the lower fertility observed for the large flocks during both years. 
Courtship and mating behaviour of ostriches have been well-documented, with reproductive 
behaviour in captive environments closely resembling that observed in natural environments 
(Bolwig, 1973; Stewart, 1994; Berendsen, 1995; Hicks-Aldredge, 1996; Deeming, 1997; Bubier et 
al., 1998). Any change in the immediate environment may disrupt normal reproductive behaviour 
(Lambrechts et al., 1998b). The establishment of a male's territory is usually initiated by the 
making of a nest, and borders of males' territories seldom overlap (Sauer and Sauer, 1966b; 
McKeegan and Deeming, 1997). During breeding seasons, males spend most of their daily 
activity budget on territorial defense, and this is more pronounced in small breeding groups. The 
high stocking rates prevented males from establishing clear territories, which may resulted in 
them spending relatively more time on aggressive interactions with other males and Jess time on 
territorial defense and reproductive interaction with females. Although no behavioural study was 
done to substantiate this, instances were observed where more dominant males interrupted 
mating sessions by chasing or running a copulating male off a female. Aggressive displays 
between males and towards onlookers were also observed. The aggressive encounters that were 
more prevalent in the 210-bird flocks during the study period may be indicative that the defense 
and maintenance of territories were more difficult in the larger flocks. The increase in fighting 
bouts between males also possibly disturbed the laying activity of females in these flocks. 
Wechsler and Schmid (1998) found that aggressive interactions were significantly increased in 
multiple-male groups of Japanese quail (Coturnix japonica). Aggressive encounters between 
females, as observed by Sauer and Sauer (1966a, 1966b), were seldom observed during the 
present study. 
The high stocking rates of 150 birds, 141 birds and 210 birds per hectare represented stressful 
conditions that had a negative influence on the normal behaviour of the breeding ostriches. 
Cases were observed where females and males displayed abnormal homosexual behaviour and 
this was more prevalent at the highest stocking rate of 21 O birds per hectare. Homosexual 
behaviour was also observed during the non-breeding season, when males and females were 
separated and maintained in separate-sex flocks at a rate of approximately 150 birds per hectare 
Feather pecking was observed for both genders maintained at the higher stocking rates. with 
females being pecked to a greater extent than males. McKeegan and Deeming (1997), 
Lambrechts and Cloete (1998) and Lambrechts et al. (1998a) indicated that ostriches tend to 
have rather conservative behavioural patterns and maintaining breeding ostriches in large flocks, 
may give rise to certain behavioural problems (Stewart, 1994; Hicks-Aldredge, 1996; Deeming, 
46 
1997). Abnormal homosexual behaviour in female ostriches in captive environments (Sambraus, 
1994) and imprinting on humans (Bubier et al., 1998) have also been reported. Our observations 
in this study lend support to suggestions of Stewart (1994 ), Hicks-Aldredge (1996), Deeming 
(1997) and Sambraus (1995 cite Deeming and Bubier, 1999) that abnormal behaviour patterns 
may develop as a result of too high stocking rates. The abnormal behaviour patterns observed at 
the higher stocking rates agree with findings in poultry where high stocking rates had a negative 
influence on the reproduction performance and well-being of laying hens (Hughes and Wood-
Gush, 1977; Sherwin and Kelland, 1998; Blokhuis and Wiepkema, 1998; Martrenchar et al., 1999; 
Savory et al., 1999; Bilcik and Keeling, 2000; Klein et al., 2000). Feather pecking was found to be 
transmitted socially in groups of laying hen chicks (Zeltner et al., 2000). Oden et al. (1999) found 
that the presence of males had a reducing effect of agonistic behaviour in laying hens, but not on 
the degree of feather pecking among laying hens housed at high stocking rates. 
In contrast to the influence of too high stocking rate of the production of the large flocks, the 9-bird 
flocks maintained in the smaller 0.13ha camps outperformed the 9-bird flocks in the larger 0.30ha 
camps in terms of almost all production parameters. Total eggs produced tended to be 9.5% 
higher (P?:0.15) for 9-bird flocks maintained in the smaller 0.13ha camps. Average egg 
production was improved by 14.5% in the 9-bird flocks maintained in the 0.13 ha camps. The 
fertility and hatchability of 9-bird flocks in the 0.13ha camps were improved by 9.5% and 6.4%, 
respectively. Chick production was 23.5% higher for the 9-bird flocks maintained in the smaller 
camps. A possible explanation for this improvement in the respective production traits is that the 
increased frequency of encounters between males and females in the smaller camps were more 
frequent than in the larger 0.3ha camps. The smaller camp size of 0.13ha did not appear to 
influence reproductive behaviour of the birds. It would appear that the establishment of clear 
territories were not inhibited during the study period. The smaller 0.13ha camps actually might 
have contributed to females visiting different territories more frequently than in the larger 0.30 ha 
camps and this possibly resulted in a higher frequency in sexual encounters, which in turn 
possibly contributed to the higher production performance of the 9-bird flocks in the smaller 0.13 
ha camps. 
When the production of breeding trios and 9-bird flocks in o.13ha camps are considered, the 9-
bird flocks performed on average better for all reproduction traits, with the exception of average 
egg production per female. The improvement in almost all the reproduction traits in the 9-bird 
flocks can be ascribed to the higher number of females and males present in this breeding 
system, when compared to the trios. Average egg production however, was compromised for the 
higher stocking rate of 9 birds per 0.13 hectare. The improved fertility and hatchability for the 9-
bird flocks maintained in 0.13ha camps can be ascribed to the stimulating effect of more than one 
male in this breeding system. Females had the opportunity to visit more than one male's territory 
and this possibly contributed to more frequent sexual encounters and the higher fertility and 
hatchability recorded for the 9-birds flocks. 
47 
When production of the smaller 9-bird and 13-bird flocks are compared, the 13-bird flocks 
performed significantly better than the 9-bird flocks in terms of all reproduction traits. ie. total egg 
production, average egg production per female, number of chicks hatched, fertility and 
hatchability. When the two breeding systems are compared on an area/bird basis, it became 
evident that reproductive performance is not only stocking density dependent. The higher density 
in the 0.35 ha camps (269.2m2/bird vs. 333 3m2/bird). possibly had to effect that sexual 
encounters between the males and females were more frequent, which in turn contributed to the 
improved fertility and hatchability. It is difficult; however, to reach a conclusion with regard to the 
improved production of the 13-bird flocks, for the slight increase in camp size might also have had 
a positive influence on the production performance of the 13-bird flocks. 
The influence of male.female ratio 
Increasing the number of females per male in the 0.06ha camps had to effect that egg production 
increased linearly with 29.1 ± 2.3 eggs per camp, with a corresponding linear increase of 24.2±2.0 
chicks per camp. Average egg production per female, however, was not similarly affected. 
Fertility and hatchability were significantly improved for the breeding quads. An important 
observation was that neither fertility nor hatchability was compromised when the number of 
females was increased. A possible explanation for this is that the breeding quads are similar to a 
breeding harem that is usually established by males during a breeding season. Ostriches are 
gregarious by nature and the stimulus of an increase in the number of individuals in the respective 
breeding systems is reflected in the improvement in almost all the reproduction traits of the 
breeding quads in the study. Deeming (1996) found that breeding pairs and trios were more 
productive than larger flocks. Results obtained in this study support the findings of Deeming 
(1996). His study however, did not include a comparison of breeding quads with the larger flocks. 
Based on the results in the present study it can be assumed that breeding quads will be more 
productive on a per female basis than larger flocks. However, it has to be conceded that the 
number of replicates studied for the breeding quads was relatively small when compared to other 
studies (Craig et al., 1977; Bates et al., 1987; Deeming and Wadland, 2001; Campo and Davilla, 
2002). It is therefore possible that the breeding quads could have performed exceptionally well 
under these circumstances. 
The present results are contradictory to results reported by More (1997), who found that the laying 
performance of breeding pairs tended to be better than that of both trios and a breeding flock. 
However, although female age significantly influenced egg production, no mention was made on 
the composition of the respective breeding systems. The improvement in the fertility and 
hatchability for the breeding quads are contradictory to the decline in fertility reported when the 
number of females were increased in single-male groups of Japanese quail (Coturnix japonica) 
and commercial pheasants (Phasianus colchinus) (Bates et al., 1987; Wechsler and Schmid, 
1998; Deeming and Wadland, 2002). Single-male groups containing 8, 12, 16 and 20 hens had 
fertility percentages of 92%, 84%, 77% and 69%, respectively (Wechsler and Schmid. 1998) 
48 
Increasing the number of females from 1M:12F to 1M:18F in pheasants had to effect that egg 
production was increased. with a subsequent decrease in fertility. Campo and Davilla (2002) 
investigated the influence of four different mating ratios on stress and fearfulness indicators in 
chickens. Their results suggested that too high mating ratios increased physiological and 
psychological stress responses. Deeming and Wadland (2002) investigated the effect of two 
mating ratios, i.e. 1M:8F and 1M:12F, in commercial pheasant flocks. Flocks with a mating ratio 
of 1M :8F produced significantly more eggs, and fertility and hatchability were improved 
significantly for the 1M :8F flocks. The higher M:F ratio of the breeding quads therefore indicates 
that ostrich males can successfully keep and service a harem consisting of at least three females. 
CONCLUSIONS 
The results indicate that breeding ostriches could be maintained at stocking rates higher than 
what is currently used on commercial ostrich farms. This has important implications in terms of 
the intensification of ostrich farming, especially in areas that are characterised by vegetation that 
are exposed to the trampling effect of ostriches. However, too high stocking rates may have a 
negative influence on normal reproductive behaviour, and thereby may not be beneficial for 
reproduction as well as well-being of ostriches under intensive farming conditions. 
Results indicated that a "trade-off" in performance was observed when either the stocking rate 
was increased, or when the number of females per male was increased. For example, when 
average egg production per female was improved, fertility and hatchability was compromised. The 
2
114-breeding flocks performed the best in 1 ha (10 000m ) breeding camps, in terms of all the 
2
reproductive traits. Increasing the stocking rate in smaller 0.13 ha (625m ) breeding camps had 
the effect that, although average egg production/female was lower, fertility and hatchability were 
improved, possibly as a result of more frequent sexual interaction of males and females. 
Breeding pairs, trios and quads could be maintained on smaller areas than what is normally 
prescribed, with acceptable production levels. Although the number of quads in the study was 
relatively small, it indicates that harem size under intensive conditions may be as high as three 
females per male, which supports the polygamous nature of ostrich males. 
REFERENCES 
ADAMS J & REVELL BJ, 2003. Ostrich farming - a review and feasibility study of opportunities in 
the EU. Available at: http:\\www.mluri.sari.ac.uk\ 
ANIMAL WELFARE ADVISORY COMMITTEE, 1998. Code of Recommendations and minimum 
standards for the welfare of ostrich and emu. Code of Animal Welfare No. 21, ISBN 0-478-
07476-X, Ministry of Agriculture and Forestry, PO Box 2526, Wellington, New Zealand. 
49 
BATES DP, HANSON LE, COOK ME, WENTWORTH BC, SUNDE ML & BITGOOD JJ, 1987 Lighting and 
sex ratio for breeding ringnecked pheasants in confined housing. Poult Sci. 66(4): 605-
612. 
BERENDSEN KD, 1995. Behaviour. In: Ostrich Farm Management, pp. 25-32. Ed.'s: A Kreibich & 
M Sommer, Landwirtschaftsverlag GmbH, Munster-Hiltrup, Germany. 
BILCIK B & KEELING LJ, 2000. Relationship between feather pecking and ground pecking in laying 
hens and the effect of group size. Appl. Anim. Behav. Sci. 68(1): 55-66. 
BLOKHUIS HJ & WIEPKEMA PR, 1998. Studies of feather pecking in poultry. Vet. Quat. 20(1 ): 6-9. 
BOLWIG N, 1973. Agonistic and sexual behaviour of the African ostrich (Struthio came/us). 
Condor 7 5 100-105. 
BUBIER NE, PAXTON CGM, BOWERS p & DEEMING DC, 1998. Courtship behaviour of farmed 
ostriches in relation to humans. Brit. Pou/t. Sci. 39: 477-481. 
BUNTER KL & GRASER H-U, 2000. Genetic evaluation for Australian ostriches - a report for the 
Rural Industries Research and Development Corporation. Publication no. 00/153, Rural 
Industries Research and Development Corporation, Australia. 
CAMPO JL & DAVILLA SG, 2002. Influence of mating ratio and group size on indicators of 
fearfulness and stress of hens and cocks. Poult. Sci. 81 (8): 1099-1103. 
CRAIG JV, AL-RAWI B & KRATZER DD, 1977. Social status and sex ratio effects on mating 
frequency of cockerels. Poult. Sci. 56(3): 767-772. 
DEEMING DC, 1996. Production, fertility and hatchability of ostrich (Struthio came/us) eggs on a 
farm in the United Kingdom. Anim. Sci. 67: 329-336. 
DEEMING DC, 1997. Ratite egg incubation - a practical guide. Ratite Conf, Buckinghamshire, 
United Kingdom. 
DEEMING DC & BUBIER NE, 1999. Behaviour in natural and captive environments. In: The Ostrich: 
biology, production and health, pp. 83-104. Ed.: DC Deeming, CABI Publishing, CAB 
International, Wallingford, Oxon, United Kingdom. 
DEEMING DC & WADLAND D, 2001. Observations on the patterns of embryonic mortality over the 
laying season of pheasants. Brit. Pou/tr. Sci. 42(5): 569-573. 
DEEMING DC & WADLAND 0, 2002. Influence of mating sex ration in commercial pheasant colonies 
on bird health and the production, fertility and hatchability of eggs. Brit. Poult. Sci. 43(1): 
16-23. 
FOGGIN CM & VAN NIEKERK A, 1995. Ostriches in the wild, colony breeding and foster rearing. 
Proc. 5th Austr. Ostrich Assoc. Conf, pp. 111-116. 
GILBERT AB, 1971. control of ovulation. In: Physiology and Biochemistry of the Fowl, volume 3, 
pp. 1225-1235. Ed.'s: O.J. Bell & B.M. Freeman, Academic Press, London. 
HARVEY WR, 1990. User's guide to for LSMLMW and MIXMDL. PC-2 version, Mimeograph. 
HICKS-ALDREDGE KO, 1996. Reproduction. In: Ratite management, medicine and health, pp. 47-
57. Ed.'s: TN Tully & SM Shane, Krieger Publishing Co., Malabar, Florida. 
HORBANCZUK JO, 2002. The history and current status of ostrich farming in Poland. Proc. World 
Ostrich Congr., 26-29 September, Warsaw, Poland, pp. 7-13. 
50 
HUGHES BO & Wooo-GusH DG, 1977. Agonistic behaviour in domestic hens: the influence of 
housing method and group size. Anim. Behav. 25(4) 1056-1062. 
HURXTHAL LM, 1979. Breeding behaviour of the ostrich Struthio came/us massaicus Neumann on 
the Nairobi National Park. Ph. D thesis, University of Nairobi, Kenya. 
JARVIS MJF, JARVIS C & KEFFEN RH, 1985. Breeding seasons and laying patterns of the southern 
African ostrich Struthio came/us. Ibis 127: 442-449. 
KLEIN T, ZELTNER E & HUBER-EICHER B, 2000. Are genetic differences in foraging behaviour of 
laying hens chicks paralleled by hybrid-specific differences in feather pecking? Appl. Anim. 
Behav. Sci. 70(2): 143-155. 
LAMBRECHTS H & CLOETE SWP, 1998. Activity budgets of adult breeding ostriches classified 
according to annual egg production of the previous breeding season. Proc. 3f/h Annual 
Congr. S. Afr. Soc. Anim. Sci., 5-8 April, Stellenbosch, South Africa. 
LAMBRECHTS H, CLOETE SWP, SWART D, VAN SCHALKWYK SJ & GREYLING JPC, 2002. Egg 
production and fertility of ostriches as influenced by stocking rate and male:female ratio. 
Proc. 151 Joint Congr. of the Grassland Soc. S. Afr. and the S. Afr. Soc. Anim. Sci., 13-16 
May 2002, Christiana Aventura, South Africa, p. 103. 
LAMBRECHTS H, CLOETE SWP & DAVIS HJ, 1998a. Influence of a L-carnitine-magnesium 
supplement on the behaviour of adult breeding ostriches. Proc. 2nd lnternat. Ratite Scient. 
Congr., pp. 170-172, 21-25 September, Oudtshoorn, South Africa. 
LAMBRECHTS H, HUCHZERMEYER FW & SWART D, 1998b. Integrating ostrich behaviour in 
commercial farming systems Proc. 2nd lntemat. Ratite Scient. Congr. . pp. 167-169, 21-25 
September, Oudtshoorn, South Africa. 
MARTRENCHAR A, HUONNIC D, COTIE JP, BOILLETOT E & MORISSE JP, 1999. Influence of stocking 
rate on behavioural, health and productivity traits of turkeys in large colonies. Brit. Pauli. 
Sci. 40(3): 323-331. 
MASHALY MM, WEBB ML, YOUTZ SL, ROUSH WB & GRAVES HB, 1984. Changes in serum 
corticosterone concentration of laying hens as a response to increased population rate. 
Poult. Sci. 63(11): 2271-2274. 
MCKEEGAN DEF & DEEMING DC, 1997. Effects of gender and group size on the time-activity 
budgets of adult breeding ostriches (Struthio came/us) in a farming environment. Appl. 
Anim. Behav. Sci. 51: 159-177. 
MORE SJ, 1997. Monitoring the health and productivity of farmed ostrich colonies. Austr. Vet. J. 
75(8): 583-587. 
ODEN K, VESTERGAARD KS & ALGERS B, 1999. Agonistic behaviour and feather pecking in single-
sexed and mixed groups of laying hens. Appl. Anim. Behav. Sci. 62: 219-231. 
SAMBRAUS HH, 1994. The circadian rhythm in the behaviour of ostriches (Struthio came/us) kept 
in pens. Deutsche Tierartzliche Wochenschrift 107: 339-341. 
SAUER EGF & SAUER EM, 1966a. Social behaviour of the South African ostrich Struthio came/us 
australis. Ostrich (suppl) 6: 183-191. 
SAUER EGF & SAUER EM, 1966b. The behaviour and ecology of the South African ostrich. Living 
Bird 5: 45-75. 
51 
.....,.,. 
SAVORY CJ, MANN JS & MACLEOD MG, 1999. Incidence of pecking damage in growing bantams 1n 
relation to food form, group size, stocking rate, dietary tryptophan concentration and dietary 
protein source. Brit. Poult Sci. 40(5): 579-584. 
SHERWIN CM & KELLAND A, 1998. Time-budgets, comfort behaviours and injurious pecking of 
turkeys housed in pairs. Brit. Poult. Sci. 39(3): 325-332. 
SMITH WA, CILLIERS SC, MELLETT FD & VAN SCHALKWYK SJ, 1995. Ostrich Production - a South 
African perspective. In: Biotechnology in the Feed Industry - Proc of Alltech's 11th Annual 
Symp, pp. 175-197. Ed.: T Lyons & KA Jacques, Nottingham University Press 
1
SNEDECOR GW & COCHRAN WG, 1967. Statistical methods, 6 h edition. Iowa State University, 
U.S.A. 
SOUTH AFRICAN OSTRICH BUSINESS CHAMBER, 2003. Riglyne vir die benutting van natuurlike veld 
deur volstruisboerdery: 'n gedragskode vir volstruisprodusente. South African Ostrich 
Business Chamber, PO Box 952, Oudtshoorn, South Africa. 
STEWART JS, 1994. Ostrich behaviour and behavioural problems. Assoc. Avian Vet, Proc., pp. 
103-109. 
VANZYL PL, 2001. 'n Ekonomiese evaluering van Volstruisboerdery in die Oudtshoorn omgewing. 
M.Sc. thesis, University of Stellenbosch, South Africa. 
WECHSLER B & SCHMID I, 1998. Aggressive pecking by males in breeding groups of Japanese 
quail (Coturnix japonica). Brit. Poult. Sci. 39(3): 333-339. 
ZELTNER E, KLEIN T & HUBER-EICHER B, 2000. Is there social transmission of feather pecking in 
groups of laying hen chicks? Anim. Behav. 60(2): 211-216. 
52 
Chapter 5 
Genetic and environmental (co)variance estimates for 
ostrich ( Struthio came/us var. domesticus) male 
aggression, shin colour and egg production performance 
of companion breeding females 
Abstract 
Monthly records on male aggression and shin colour scores, as well as for female egg production performance 
(EPP), were obtained for ostrich breeding pairs (n=136) over the period 1999 to 2002. All traits were affected by 
animal age, generally increasing up to intermediate ages of 4-7 years, and decreasing subsequently. Direct 
heritability estimates (±SE) were 0.30±0.04 for male aggression score, 0.03±0.02 for shin colour score and 
0.08±0.04 for EPP. The animal permanent environmental variance (PE - association of records obtained across 
years) was significant for EPP only, at 0.07±0.03. The animal temporary environmental variances (TE - associations 
between records obtained within a breeding season) amounted to 0.23±0.02, 0.29±0.02, and 0.13±0.02, 
respectively. When assessed on a per breeding pair level, phenotypic correlations of male aggression score and 
male shin colour score with female EPP were estimated at 0.29±0.02 and 0.19±0.02 respectively. The partitioning of 
these correlations indicated that the TE components contributed most to the observed phenotypic relationships, at 
0.42±0.07 and 0.51±0.07, respectively. On the PE and genetic levels, the observed correlations were low, mostly 
negative in sign, and not significant. The derived phenotypic relationships would thus not be valuable as indirect 
selection criteria for the improvement of egg production in ostriches in the current flock or future generations. The 
significant genetic variation for male aggression score, however, suggests that it could be possible to change the 
temperament of ostriches, to enable them to adapt better to routine husbandry operations, without impacting 
negatively on the egg production performance of companion breeding females. 
INTRODUCTION 
In Africa, ostriches have been hunted for centuries for their meat and feathers (Holtzhausen and 
Kotze, 1990; Drenowatz et al., 1995; Deeming, 1999). The first record of commercial ostrich 
farming in South Africa dates back as far as the 1860's, with feathers being the main source of 
income (Smit, 1964). Since then, the emphasis has shifted to leather and meat as the major 
sources of income, with feathers contributing the least to the total income generated from a 
slaughter bird (Wagner, 1986; Van Zyl, 2001 ). 
The intensification of commercial ostrich farming and an increase in the emphasis on the welfare of 
animals under intensive farming conditions during the last few decades, had the effect that 
commercial ostrich farmers need to integrate the specific behavioural requirements of ostriches into 
their management programs to ensure an optimum breeding environment (Stewart, 1994; Reiner et 
53 
al., 1996; Lambrechts et al., 1998; Mohamed et al., 2003). Identification and integration of the 
behavioural patterns of ostriches in commercial farming systems may reduce potential stress-
causing environments that may have an adverse effect on production performance. Overt 
aggressive behaviour in ostrich males is considered as an abnormal behaviour, and may contribute 
to lowered reproductive performance in ostrich breeding systems. There are, however, few studies 
to support these contentions. 
Reports on ostrich behaviour have mostly been concerned with natural environments (Sauer, 1972; 
Bolwig, 1973; Hurxthal, 1975; Bertram, 1992; Sambraus, 1995). Ostriches are gregarious in nature 
during non-breeding seasons (Sauer and Sauer, 1966a, 1966b; Bertram, 1980, 1992; Burger and 
Gochfeld, 1988). During breeding seasons, male and female behaviour changes considerably, 
with both genders exhibiting characteristic behavioural repertoires (Sauer and Sauer, 1966b; 
Sambraus, 1994; Deeming and Bubier, 1999). Behaviour exhibited under commercial farming 
conditions closely resembles that observed in the wild (Bolwig, 1973; Berendsen, 1995; Hicks-
Aldredge, 1996; Deeming, 1997, 1999; Bubier et al., 1998). Display of territorial aggression, 
together with the change in colour of the shins, are commonly used by ostrich farmers as an 
indication of the readiness of a male to enter a breeding season. Both traits were previously shown 
to be positively related to egg production of female companions on a phenotypic level (Lambrechts 
et al. (2000). Ostrich males appear to develop courtship behaviour later than females, with 
colouration of the beak, thighs and shins only after the onset of breeding (Bolwig, 1973; Hicks, 
1990; Samson, 1996; Deeming and Bubier, 1999). During breeding seasons the shins of ostrich 
males change in colour from light pink to crimson red. 
Territorial behaviour in adult ostrich males, which is most often combined with aggression and 
vocalisation, is observed to a higher degree at the onset of and during breeding seasons (Sauer, 
1972; Bolwig, 1973; McKeegan and Deeming, 1997; Lambrechts, 1998). This increased level of 
aggression during a breeding season is associated with elevated levels of testosterone (Degen et 
al., 1994 ). Male aggression is mainly associated with territorial behaviour, and is characterised by 
males ramming each other with their chests and by forward kicking (Huchzermeyer, 1997), and by 
kantling displayed to other males and intruders in or near their territory (Bertram, 1992). The aim 
of this study was to determine the genetic and environmental (co)variance components for ostrich 
male aggression, male shin colour, and the correlation of these traits with the reproductive 
performance of their companion breeding females under commercial conditions. 
MATERIALS AND METHODS 
Experimental animals 
An ostrich breeding flock consisting of 136 breeding pairs, maintained at the Little Karoo 
Agricultural Development Centre outside Oudtshoorn, South Africa, was used for the study. The 
54 
management of the breeding flock is documented by Van Schalkwyk et al (1996) and Bunter and 
Graser (2000). The ages of the males used in the study varied between 2 and 12 years 
Experimental procedure 
Each male was evaluated on a monthly basis throughout the 1999/2000, 2000/2001, 200112002 
and 2002/2003 breeding seasons, by using a subjective scale to assign aggression and shin colour 
scores. 
Assignment of aggression score: 
The study focussed on territorial behaviour exhibited by ostrich males, i.e. aggressive behaviour 
displayed towards intruders in its breeding territory. Aggression scores were assigned by 
evaluating males on a monthly basis in terms of the extent of aggressive behaviour displayed 
throughout the breeding season. Scores were assigned on a scale, according to the following 
criteria: 
• Low aggression score (5-6) assigned to males exhibiting no territorial aggression in a specific 
production month, and eggs could be collected without additional assistance to divert a male's 
attention away from the nest Males appeared subdued and also disinterested in the activities 
in and around their breeding territories. 
• Medium aggression score (7-8): assigned to males exhibiting a more extensive degree of 
territorial aggression, i.e. additional assistance were required to collect eggs on certain, but not 
all, occasions during a specific production month. Males frequently mock-charged any 
intruders, but on confrontation retired to another part of his breeding territory. 
• High aggression score (9-10): assigned to males that exhibited extreme territorial aggression 
towards any intruders into their breeding territories in a specific month. In all cases, additional 
assistance was required to collect eggs 
Assignment of shin colour score 
A shin colour evaluation chart was compiled by using commercial paint colour charts. The colour 
chart used to assign shin colour scores is represented by Plate1. A score of 0 represented a white 
shin colour, with 5 being light pink and 10 a bright crimson red. 
1 2 3 4 5 6 
Plate 1. Colour chart used to assign monthly shin colour scores to ostrich males. Shin colour score indicated 
in each respective colour box. 
55 
Egg production 
Daily egg production was noted for each pair throughout the breeding season. During the 
2000/2001, 2001/2002 and 2002/2003 breeding seasons, the breeding pairs were subjected to 
different management practices, resulting in the breeding pairs differing in the number of 
production days. Breeding pairs were also subjected to mid-season breeding rests during the 
respective breeding seasons, resulting in their overall breeding shorter than their non-rested 
contemporaries. To standardize for the number of production days per breeding pair and also to 
accommodate months that differed in terms of production days, the egg production performance 
(EPP) for each breeding pair was calculated by using the following equation (Van Schalkwyk et al., 
1996): 
EPP= [(total number of eggs produced/(0.5 X number of production days)] X 100 
Statistical analysis 
The ASREML program (Gilmour et al., 1999) was used for the estimation of the fixed effects, and 
also subsequently to derive variance components for female egg production, and male aggression 
and shin colour scores in univariate analyses. Fixed effects that were considered included 
production year (2000-2002) and treatment (rested or not). Trends with regard to animal age and 
month of the breeding season were modelled, using cubic splines (Verbyla et al., 1999). The cubic 
splines consisted of a fixed linear component and a random non-linear component depicting 
random deviations conforming to a smooth trend. Initially, random deviations from linearity not 
conforming to a smooth trend were also fitted, but this component was not found to be significant 
(P?:0.05) for any of the variables. The first analyses involved fitting various combinations of fixed 
effects, random spline components and interactions between them to obtain an operational model 
(termed as Model 1). Effects found to be significant (P..s:0.05) in these preliminary analyses were 
retained in subsequent analyses. Random terms were then added to the operational model, 
resulting in the following genetic models for analyses (in matrix notation): 
y = Xb + Z1a + e (2) 
y = Xb + Z2PE + e (3) 
y = Xb + z3TE + e (4) 
y = Xb + Z1a + Z2PE + e (5) 
y = Xb + Z1a + z3TE + e (6) 
y = Xb + Z2PE + z3TE + e (7) 
y = Xb + Z1 a + Z2PE + z3TE + e (8) 
In these models, y was a vector of observations for female production or reproduction traits; b, a, 
permanent environment (PE) and temporar; environment (TE) were vectors of fixed effects, direct 
additive genetic effects, animal permanent environmental effects and animal temporary 
environmental effects, respectively. PE represented the factors that could contribute to variation 
observed between breeding seasons, whereas TE represented factors that could contribute to the 
56 
variation observed within a breeding season. X, Z1, z,, and Z3 were the corresponding incidence 
matrices relating the respective effects to y, and e the vector of residuals. 
It was assumed that: 
2
V(a) =h :/a; V(PE) = l02pE; V(rE) =l 02rE; V(e) =l 0 e, 
2 2 2 2 
with A being the numerator relationship matrix, I being an identity matrix; and 0 a. 0 PE, 0 rE and 0 e 
being the direct genetic variance, animal permanent environmental variance, animal temporary 
environmental variance and environmental (residual) variance respectively. All analyses included 
the full pedigree file, consisting of 601 individuals, the progeny of 135 males and 132 females. 
Likelihood ratio tests (LRT) were conducted to determine the most suitable model for each trait in 
univariate analyses. The LRT was based on testing twice the increase in Log likelihood resulting 
from adding an additional random term to the model of analysis as a Chi2 statistic. Alternatively, for 
two models with the same number of random effects, and assuming identical fixed effects models, 
the one with the higher likelihood fits the data better. 
Subsequently, two-trait animal models were fitted to the data in an attempt to estimate genetic, 
environmental and phenotypic correlations between the respective traits. Direct additive variance 
components in the study were relatively small, particularly in the case of shin colour score. It was 
attempted to use the numerator relationship matrix to estimate genetic correlations between sex-
limited traits measured in either males or females, viz. between EPP and aggression score, as well 
as between EPP and shin colour score. Since these traits were not recorded on the same gender, 
these correlations were not stable, with relatively large standard errors. To add to these results, 
within and between breeding pair (co)variance components were also obtained to estimate 
permanent and temporary environmental correlations between these traits on a breeding pair level. 
Temporary environmental correlations were seen as associations between measures of breeding 
pair performance within breeding seasons, while permanent environmental correlations were 
defined as associations across breeding seasons. 
RESULTS 
The monthly egg production performance (EPP) of the ostrich females was exceedingly variable, 
as indicated by a standard deviation exceeding the mean (Table 1 ). A relatively large proportion of 
females had monthly EPP records of zero in some months, contributing to the observed variation. 
The corresponding coefficient of variation for male aggression score was 30%. Male shin colour 
score was less variable, with a coefficient of variation below 10%. 
57 
Table 1. Means and standard deviations, and number of observations for reproductive and behavioural traits 
assessed for ostrich males and females. 
TRAIT Mean (SD) Number of observations 
Monthly EPP(%) (C() 32.1 (34.2) 4025 
Aggression score (o) 6.62 (2.0) 4406 
Shin colour score (o) 7.75 (0.68) 4406 
Male aggression increased with age, with young males (i.e. 2 years of age) exhibiting less 
aggressive behaviour than their older contemporaries. Aggression reached a peak in males aged 
6 to 7 years, and declined with an increase in age. When shin colour scores are considered, males 
aged 5 to 7 years received the highest scores for shin colour (8.02±0.04). Two-year old males 
received lower shin colour scores of on average 7.36±0.09 (Figure 1) . 
8.5 60.0 
8.0 
50.0 
7.5 
40.0 
Q) 7.0 
0 c 
(.J 30.0 D.. 
en 6.5 D.. w 
20.0 
6.0 
5.5 10.0 
5.0 0.0 
2 3 4 5 6 7 8 9 10 11 12 
Age (years) 
[--o-- Aggression -¢----Shin colour --0-::~] 
Figure 1. The relationship between male age, aggression and shin colour scores, and female egg production 
performance (EPP). 
The likelihood ratio tests revealed that Model 8 of the statistical analysis, including the direct 
additive genetic effects, female permanent environmental effects and female temporary 
environmental effects as random factors, fitted the data best (P:0:0.01) for monthly EPP (Table 2). 
Model 7, i.e. the model including only the direct additive genetic and male temporary environmental 
effects, fitted the data best (P:0:0.01) for aggression and shin colour scores of males (Table 2) 
58 
Table 2. Log likelihood values for the respective random effect models (model in brackets) considered in the 
analyses, with the best models indicated in bold. 
·-
EFFECT EPP Aggression score Shin colour score 
------
Fixed only (1) -18857.3 -5454.9 -525.83 
+ h2 (2) -18525.9 -5464.22 -349.34 
+PE (3) -18525.3 -5457.9 -339.02 
+TE (4) -18465.8 -5351.1 -154.18 
+ h2 +PE (5) -18519.8 -5451.7 -349.34 
+ h2 +TE (6) -18429.9 -5302.5 -152.10 
+PE+ TE (7) -18431.8 -5408.0 -153.22 
+h2 + PE +TE (8) -18426.8 -5301 1 -152.10 
Heritability (h2 ) of the male traits ranged from 0.03 to 0.30 for shin colour and aggression, 
respectively (Table 3). The monthly EPP of ostrich females were lowly heritable at 0.08. The 
female permanent environmental variance ratio (PE) was only significant (P:s:0.05) for EPP. The 
animal temporary environmental variance ratios were significant (P:00.05) for all traits, and 
exceeded 0.10 throughout. These results suggested that the animal temporary environmental 
variances (i.e. the correspondence between monthly records on the same animal within a season) 
contributed markedly to all the traits measured on male and female ostriches in the present study. 
Table 3. Variance components and ratios (±SE) for the egg production performance (EPP) of ostrich females, 
and aggression and shin colour scores assigned to ostrich males. 
VARIANCE EPP Aggression score Shin colour score 
Component 
CT2A 80.623 1.2919 0.0155 
2
CT PE 74.282 
cr"TE 138.030 0.9829 0.1347 
CT2E 752.026 1.9657 0.3217 
cr2p 1044.961 4.2405 0.4719 
Ratios 
h2 0.08 ± 0.04 0.30 ± 0.04 0.03 ± 0.02 
PE 0.07 ± 0.03* 
TE 0.13 ± 0.02* 0.23 ± 0.02* 0.29 ± 0.02* 
*Ps0.05 
The significant phenotypic correlation between monthly EPP and aggression score (Table 4), could 
mostly be attributed to a highly significant (P:s:0.01) TE component. This result suggests that short-
term, seasonal differences mostly accounted for the positive phenotypic and environmental 
correlations. The PE correlation (i.e. the long-term correlation between breeding pair relationship 
across breeding seasons) was positive, low, and smaller in magnitude than the corresponding 
standard error. The genetic correlation between male aggression and shin colour was low, at 0.08 
(Table 4). The positive and significant phenotypic correlation once again depended mostly on a 
significant (P:00.01) TE correlation. 
59 
Table 4. The permanent (PE) and temporary (TE) environment correlations of the egg production 
performance (EPP) of ostrich females with aggression and shin colour of ostrich males, as 
estimated between breeding pairs and on an individual basis 
-
TRAITS AND BETWEEN BREEDING PAIR INDIVIDUAL 
---
CORRELATIONS PE TE +-Environmental -- -- Phenotypic 
Number of combinations 299 546 4406 4406 
EPP x 
Aggression score 0.15 ± 0.12 0.42 ± 0.07*' 0.31±002 0.29 ± 0.02 
Shin colour score -0.12 ± 0.35 0.51 ± 0 07*' 0.12±0.02 0.19±0.02 
-
Genetic TE Environmental Phenotypic 
--~---------- --~---
Aggression score x 
Shin colour score 0.08 ± 0.24 0.32 ± 0.07** 0.15 ± 0.02 0.18±0.02 
---
**P:S0.01 
When the numerator relationship matrix was used for the estimation of a genetic correlation, the 
derived figure was negative, correspondingly low and not different from zero at -0_ 10 ± 0.31. 
Results pertaining to the between breeding pair correlation of monthly EPP with shin colour score 
yielded estimates that were broadly in agreement. The estimated genetic correlation was larger in 
magnitude and negative, but still not significantly (P-:::0.05) different from zero at -0.55 ± 0.42. 
The average monthly EPP of the breeding pairs showed a typical seasonal pattern during the 
study, with the lowest EPP of 22.2±2.7% reported during June (i.e. the first month of a breeding 
season). The highest EPP of 49.1±2.7% was reported for September, and egg production declined 
progressively to reach an average of 30.0±2.7% during January (Figure 2). 
60.0 
50.0 
40.0 
~ 
[l. 30.0 
[l. 
UJ 
20.0 
10.0 
0.0 
Jun Jul Aug Sept Oct Nov Dec Jan 
Month 
Figure 2. The average monthly (mean ± SE) egg production performance (EPP) of ostrich breeding pairs 
60 
DISCUSSION 
Annual egg production records of ostrich females were previously found to be highly variable, with 
coefficients of variation exceeding 50 % (Cloete et al., 1998; Bunter et al., 2001). In the present 
study, monthly figures showed even more variation and the derived standard deviation exceeded 
the mean. No corresponding literature estimates were found for male aggression and shin colour 
scores, but these traits generally showed less variation. 
The increase in territorial aggression with age may be ascribed to previous breeding experience of 
the older males. Kim and Zuk (2000) stated that previous breeding experience and aggression 
may determine the position of a male in the dominance hierarchy in a breeding environment of 
junglefowl. Broiler females showed distinct preferences for certain males, using male behaviour 
and not male morphology as a basis for their choices (Millman and Duncan, 2000). Rintamaki et 
al. (1998) found that female black grouse (Tetrao tetrix) mating with the males with the highest 
social ranking produced the biggest clutches, with females in good conditions mating with the 
higher ranking males. The tendency towards a decline in aggression in ostrich males older than 7 
years may simply reflect that such males are past their prime. It coincides with a finding that 
fertility of ostrich eggs depends on male age, with maximum fertility at intermediate ages of 4 - 6 
years (Bunter, 2002). The mates of younger and older males produced eggs with a lower level of 
fertility. 
The heritability of monthly egg production figures of females was below 0.10. Estimates based on 
annual female performance ranged from 0.12 - 0.23 (Bunter et al., 2001; Cloete et al., 2004). The 
suggestion of a lower estimate for monthly EPP in the present may be related to the inherently 
greater variability of the trait. The significant genetic variation for EPP, however, opens up 
avenues for the use of part-records obtained over a fraction of the normal breeding season for the 
selection of ostrich females for egg production. In a previous study, Lambrechts et al. (2004) 
correspondingly found that EPP recorded over as short as 14 days were highly related to egg 
production over the entire season on a genetic level. Genetic correlations with annual egg 
production were unity for part-records derived over periods as short as one month of an 8-month 
breeding season. Female PE effects in the literature were generally higher than in the present 
study, at 0.18 - 0.32 (Bunter et al., 2001; Cloete et al., 2004). 
No comparable results for ostriches were found in the literature as far as the significant genetic 
variation for male aggression score was concerned. Other domestic livestock species were 
subjected to extensive selection for temperament, resulting in modern genotypes that are well 
adapted to general husbandry routines (Boissy et al., 2002). It is reassuring that the present 
results suggest that an aspect of temperament in ostriches can also be modified by selection. The 
implications of such a program should, however, be understood properly before it could be 
recommended unconditionally. 
61 
In a previous investigation, Lambrechts et al. (2000) reported that male aggression and shin colour 
scores were positively related to the EPP of their companion females on a phenotypic level. 
Respective phenotypic correlations amounted to 0.29 and to 0.21 for aggression and shin colour 
scores, values that were in close agreement with those obtained in the present study. The 
previous findings suggested that these traits could play a part in the phenotypic selection of 
breeding pairs for egg production. In the present study, it was possible to partition the covariance 
components to show that the correlations reported earlier depended mostly on the temporary 
environment common to both traits. Short term climatic and other environmental stimuli thus 
probably played the major role in these relationships, rendering it of no value as far as selection is 
concerned. Permanent environment (for progress in the current flock) and genetic correlations (for 
future generations) with EPP, were low and not significant, both for male aggression and shin 
colour scores. On the other hand, these findings indicate that continued selection for egg 
production is unlikely to result in excessive levels of aggression in breeder males as a correlated 
response. The significant phenotypic correlation between male aggression score and male shin 
colour score, also partitioned mostly towards the TE component. On a genetic level, these traits 
will unlikely depend on the same sets of genes. 
CONCLUSIONS 
Male behavioural traits do not seem to have value as indirect selection criteria for egg production in 
females, either in the current flock or in future generations. It is nevertheless important to 
recognize the fact that a measure of temperament exhibited genetic variation in ostrich males. This 
opens up the possibility that the temperament of ostriches may be changed by genetic means, to 
allow them to adapt better to routine husbandry routines, without impacting negatively on the egg 
production performance of companion breeding females. This is, however, an entirely new concept 
in the keeping of ostriches. Extensive research on the genetic components of ostrich behaviour is 
needed for the further development of this train of thought. Monthly egg production records from 
the ostrich females in the study also exhibited significant genetic variation. The selection of 
breeding ostrich females on partial egg production records, i.e. obtained over part of a breeding 
season, in ostrich breeding programs should be also considered in future research. 
REFERENCES 
BERENDSEN K-0, 1995. Behaviour. In: Ostrich farm management, pp. 25-32. Ed.'s A Kreibich & 
M Sommer, Landwirtschaftsverlag, GmbH, Munster-Hiltrup, Germany. 
BERTRAM BCR, 1992. The ostrich communal nesting system. Princeton University Press, 
Princeton, New Jersey. 
BERTRAM BCR, 1980. Vigilance and group size in ostriches. Anim. Behav. 28 278-286. 
62 
BOISSY A, FISHER A, BOUIX J, BOIVIN x & LE NEINDRE p' 2002. Genetics of fear and fearfulness in 
domestic herbivores. Proc. th World Congr. Gen. Appl. Livestock Prod, 19-23 August, 
Montpellier, France, pp. 3-10. 
BOLWIG N, 1973. Agonistic and sexual behaviour of the African ostrich (Struthio came/us). Condor 
75: 100-105. 
BUBIER NE, PAXTON CGM, BOWERS P & DEEMING DC, 1998. Courtship behaviour of ostriches in 
captivity. Proc. 151 /nternat. Scient. Congr, 27-29 March, Manchester, United Kingdom, pp. 
19-20. 
BUNTER KL, 2002. The genetic analysis of reproduction and production traits recorded for farmed 
ostriches (Struthio came/us). Ph.D. dissertation, University of New England, Armidale, 
Australia. 
BUNTER KL, CLOETE SWP, VAN SCHALKWYK SJ & GRASER H-U, 2001. Factors affecting 
reproduction in farmed ostriches. Proc. Assoc. Advmnt Anim. Breed Genet. 14: 43-46. 
BUNTER K & GRASER H-U, 2000. Genetic evaluation for Australian ostriches - a report for the Rural 
Industries Research and Development Corporation. Publication no. 00/153, Rural 
Industries Research and Development Corporation, Australia. 
BURGER J & GOCHFELD M, 1988. Effect of group size and sex on vigilance in ostriches (Struthio 
came/us): antipredator strategy or mate competition? Ostrich 59: 14-20. 
CLOETE SWP, VAN SCHALKWYK, SJ & BRAND Z, 1998. Ostrich breeding - progress towards a 
scientifically based strategy. Proc. 2nd Intern. Ratite Cong., 21-25 September 1998, 
Oudsthoorn. pp 55-62. 
CLOETE SWP, BUNTER KL, LAMBRECHTS H, BRAND Z, SWART D & GREYLING JPC, 2004. 
(Co)variance estimates for live weight, body measurements and reproductive traits of pair-
mated ostrich females. Brit. Poultry Sci., submitted. 
DEEMING DC, 1997. Ratite egg incubation - a practical guide. Ratite Conference, 
Buckinghamshire, United Kingdom. 
DEEMING DC, 1999. Introduction. In: In: The ostrich - biology, production and health, pp. 1-11. 
Ed.: DC Deeming, CABI Publishing, CAB International, Wallingford, Oxon, United 
Kingdom. 
DEEMING DC & BUBIER NE, 1999. Natural and Captive Behaviour. In: The ostrich - biology, 
production and health, pp. 83-104. Ed.: DC Deeming, CABI Publishing, CAB International, 
Wallingford, Oxon, United Kingdom. 
DEGEN AA, WEIL S, ROSENSTRAUCH A, KAM M & DAWSON A, 1994. Seasonal plasma levels of 
luteinizing and steroid hormones in male and female domestic ostriches (Strut/Jio 
came/us). Gen. Comp. Endocrinol. 93: 21-27. 
DRENOWATZ C, SALES JD, SARASQUETA DV & WEILBRENNER A, 1995 History and geography. In: 
The Ratite Encyclopedia, pp. 3-29, Ed.: C Drenowatz, Ratite Records, San Antonio, Texas. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3. NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
HICKS KO, 1990. Male reproduction. Proc.1990 Ostrich Workshop, Texas A&M University, pp. 3-5. 
63 
HICKS-ALDREDGE KO, 1996. Ratite reproduction. In: Ratite management, medicine and health, pp. 
47-57. Ed.'s: TN Tully & SM Shane, Krieger Publishing Co., Florida, USA. 
HOLTZHAUSEN A & KOTZE M, 1990. The Ostrich. CP Nel Museum, Oudtshoorn, South Africa 
HUCHZERMEYER FW, 1997. Behavioural problems in farmed birds. Ostrich News 4(1): 17-18, 20-
21. 
HURXTHAL LM, 1979. Breeding behaviour of the ostrich in Nairobi National Park. Ph.D. thesis, 
University of Nairobi, Kenya. 
KIM T & ZUK M, 2000. The effects of age and previous experience on social rank in female red 
junglefowl, Gal/us gal/us spadiceus. Anim. Behav. 60(2): 239-244. 
LAMBRECHTS H, 1998. Activity budgets of adult breeding ostriches classified according to annual 
egg production of the previous breeding season. Proc. 3rjh Nat. Congr. S. Afr. Soc. Anim. 
Sci., pp. 93-94. 
LAMBRECHTS H, CLOETE SWP, SWART D & PFISTER AP, 2000. Influence of territorial aggressiveness 
of ostrich males on egg production of companion females. S. Afr. J. Anim. Sci. 30 (suppl 
1): 68-69. 
LAMBRECHTS H, HUCHZERMEYER FW & SWART D, 1998. Integrating ostrich behaviour in commercial 
farming systems. Proc. 2nd lntemat. Scient. Ratite Congr., 21-25 September 1998, 
Oudtshoorn, South Africa, pp. 167-169. 
LAMBRECHTS H, SWART D, CLOETE SWP, GREYLING JPC & PFISTER AP, 2004. Genetic and 
environmental (co)variance estimates for ultrasound scanning and egg production 
parameters in breeding ostrich females. Livest. Prod. Sci., submitted. 
MCKEEGAN DEF & DEEMING DC, 1997. Effects of gender and group size on the time-activity 
budgets of adult breeding ostriches (Struthio came/us) in a farming environment. Appl. 
Anim. Behav. Sci. 51: 159. 
MILLMAN ST & DUNCAN IJH, 2000. Do female broiler breeder fowl display a preference for broiler 
breeder or laying strain males in a Y-maze test? Appl. Anim. Behav. Sci. 69(4): 275-290. 
MOHAMED MA, FARES IM, KHALIL AM & AKMAL HH, 2003. A behavioural vision in analysis of ostrich 
performance under various systems of breeding. Assiut Vet. Med. J. 49(97): 78-87. 
REINER G, SEITZ K & DZAPO V, 1996. A survey of farming environments and ostrich behaviour in 
51 
Germany. Proc 1 lntemat. Scient. Congr, 27-29 March, Manchester, United Kingdom, 
pp. 23-24. 
RINTAMAKI PT, LUNDBERG A, ALATALO RV & HOGLUND J, 1998. Assortative mating and female 
clutch investment in black grouse. Anim. Behav. 56(6): 1399-1403 
SAMBRAUS HH, 1995. Social ranking in groups of African ostriches ( Struthio came/us) in Namibia. 
J. Anim. Breeding Gen. 112: 234-238. 
SAMBRAUS HH, 1994. The circadian rhythm in the behaviour of ostriches (Struthio came/us) kept in 
pens. Deutsche Tierartzliche Wochenschrift 107: 339-341. 
SAMSON J, 1996. Behavioural problems of farmed ostriches in Canada. Can. Vet. J. 37: 412-414. 
SAUER EGF, 1972. Aberrant sexual behaviour in the South African ostrich. Auk 89(4): 717-737. 
SAUER EG & SAUER EM, 1966a. Social behaviour of the South African ostrich (Struthio came/us 
australis). Ostrich suppl. 6: 183-191. 
64 
SAUER EG & SAUER EM, 1966b. The behaviour and ecology of the South African ostrich. Living 
Bird 5: 45-75. 
SMIT DJ VAN ZYL, 1964. Volstruisboerdery in die Klein Karoo. Department Agricultural Technical 
Services, Pamphlet 358, V&R Printers, Pretoria, South Africa. 
STEWART JS, 1994. Ostrich behaviour and behavioural problems. Assoc. Avian Vet.: 103-109. 
VAN SCHALKWYK SJ, CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations 
for body weight and reproduction in commercial ostrich breeding pairs. Brit. Poult. Sci. 37: 
953-962. 
VANZYL PL, 2001. 'n Ekonomiese evaluering van volstruisboerdery in die Oudtshoorn omgewing. 
M.Sc. thesis, University of Stellenbosch, Stellenbosch, South Africa. 
VERBYLA AP, KENWARD BR & WELHAM SJ, 1999. The analysis of designed experiments and 
longitudinal data by using smoothing splines. Appl. Statist. 48(3): 269-311. 
WAGNER P, 1986. The ostrich story. Chameleon Press, Cape Town. 
65 
Chapter 6 
The relationship between serum testosterone, 
progesterone, LH and prolactin levels, and behavioural and 
reproduction parameters measured in breeding ostriches 
( Struthio came/us var. domesticus) 
Abstract 
A total of 669 ostrich (320 female and 349 male) serum samples was assayed to determine the relationship between 
serum levels of testosterone, LH, progesterone, and prolactin, male aggression and shin colour, and female egg 
production performance (EPP). The serum levels of almost all the hormones, with serum prolactin being the exception, 
were exceedingly variable, with the coefficients of variation exceeding 60% in almost all cases. Serum testosterone and 
LH levels varied on a permanent environmental level, in contrast to progesterone and prolactin that were influenced by 
the time of sampling and/or the gender of the bird. The animal permanent environment contributed significantly to the 
variation observed for testosterone, LH, aggression, live weight, and EPP. The only significant (P:s:0.01) correlation 
reported in this study was that of serum LH with the EPP of ostrich females (0.40). Serum LH levels increased with an 
increase in male age, in contrast to serum LH levels of females, which declined with an increase in age. Serum 
testosterone levels tended to decline with an increase in male age, and the highest levels were reported for 6 year old 
ostrich males (0. 7±0.1 ng.mL-1). The highest female testosterone levels were reported for 3- (0.6±0 1 ng.mL-1) and 1O -
year old (0. 7±0.2 ng.ml-1) females, respectively. Ostrich males and females did not differ in terms of serum prolactin 
levels. Serum progesterone concentration tended to increase with an increase in female age, with highest levels 
measured in 11-year old females (0.7±0.1 ng.mL-1). Gender and time of sampling interacted for serum LH and 
testosterone concentrations, implying that the extent of gonad differentiation and the degree of photorefractoriness the 
breeding ostriches were experiencing influenced the serum testosterone and LH levels measured at the beginning and 
end of breeding. The maintenance of territorial aggression in males at the end of a breeding season, when lower levels 
of serum testosterone were measured, imply that factors other than LH may contribute to the maintenance of 
reproductive behaviour at the end of a breeding season. The study also reports on the sexual dimorphism observed in 
the breeding ostriches. The larger size of ostrich males reflects the increased parental investment in the care of their 
young. Heritabilities reported for live weight and EPP amounted to 0.30 and 0.31, respectively. The potential of, and the 
extent to which the reproductive cycles can be manipulated by means of exogenous gonadotropins and steroid hormones 
to improve the reproductive performance of breeding ostriches under commercial farming conditions warrants studies to 
determine baseline values of the gonadotropins and steroid hormones for breeding ostriches. 
66 
INTRODUCTION 
Birds may use a variety of breeding strategies to enable them to breed under diverse conditions and 
environments. Synchronisation of the endocrine and behavioural factors involved in reproductive 
activities is necessary to ensure successful reproduction. Several studies in birds have been 
performed to illustrate this, but literature on endocrine factors that may influence behaviour and 
reproduction in ostriches, are however, scarce Previous studies on ostriches focussed mainly on 
determining haematological parameters (Van Heerden el al, 1985; Dawson et al., 1996; Mushi et al., 
1998; Spinu et al, 1999; Romdhane et al., 2000), and biochemical properties of the hypothalamic and 
pituitary hormones (Papkoff et al., 1982; Bonna-Gallo et al., 1983; Licht et al., 1983; Powell et al, 
1987; Korde et al., 1996; Maney et al, 1997; Skadhauge and Dawson, 1999). In another ratite 
species, Williams et al ( 1998) investigated the potential of surgical and hormonal gonadectomy to 
overcome the seasonal effect on productivity of emus. Degen et al. ( 1994) presented the first 
information on the level of LH, testosterone and estradiol in breeding ostriches. Prolactin levels were 
measured in a few cases in ostriches, with the prolactin levels being higher in an incubating male 
(Skadhauge and Dawson, 1999). Marked seasonal changes in circulating levels of prolactin occur in 
most birds species studied, but have not been established for ostriches. 
Two forms of gonadotropin-releasing hormone (GnRH) are found in birds, ie. GnRH-1 that elicits 
gonadotropin release, and GnRH-11 that plays a role in reproductive behaviour, apparently independent 
from GnRH-1 (Maney et al., 1997). An increase in photoperiod stimulates the release of hypothalamic 
GnRH, with a subsequent secretion in pituitary LH and FSH (Deviche et al. 2000). These three 
hormones induce gonadal development and the subsequent secretion of gonadal hormones that play a 
role in behavioural changes during a breeding period, and also the development of the secondary 
sexual characteristics (Paster, 1991; Deviche et al., 2000). FSH is responsible for the maturation of 
follicles in the female and the Sertoli cells in the testes of the male, while LH stimulates the interstitial 
cells of the ovary and the Leydig cells of the testes to produce the steroid hormones. 
Estrogens, androgens and progesterone are the most important steroid hormones involved in avian 
reproduction, and have a variety of effects that may include sexual differentiation, the development of 
secondary sexual characteristics, behaviour, metabolism, gamete production, moult and feedback 
inhibition of LH secretion (Howarth, 1995). In the avian female, progesterone is produced by the 
granulosa layers of the largest preovulatory follicles in reaction to LH, and has a positive feedback on 
the pituitary that results in an LH surge that results in subsequent ovulation. The estrogens are 
synthesised by the theca layer, with maximum levels during ovarian development and egg laying The 
estrogens stimulate the liver to produce vitellogenin that is transported via the blood to the ovary where 
it is deposited as yolk in the developing follicles, and is also involved in oviduct development and 
induction of female reproductive behaviour (Farner and Wingfield, 1980; Etches, 1995). Prolactin is a 
67 
pituitary peptide hormone, with secretion stimulated by a decrease in photoperiod. High circulating 
levels of prolactin coincide with the development of photorefractoriness and subsequent gonadal 
regression (Deviche et al., 2000). Vasa-active intestinal polypeptide (VIP) primarily regulates the 
synthesis and release of prolactin in birds, with increased VIP levels associated with decreased GnRH 
expression (El Halawani et al., 1995; Deviche et al., 2000). 
Testosterone is the most important androgen produced by the testes, and secretion shows a typical 
seasonal pattern (Donham, 1979; Dittami, 1981; Fraissinet et al., 1987; Degen et al., 1994) 
Testosterone, in the presence of FSH, plays a role in the regulation and maintenance of 
spermatogenesis, with increased levels associated with the onset and the later stages of 
spermatogenesis. The Sertoli cells are also under the control of FSH, and may play a role in the 
conversion of androgens to estrogens via a process of aromatization (Howarth, 1995). The testes 
become functional earlier than the ovary in birds (Farner and Wingfield, 1980). Increased male 
testosterone levels are associated with aggressive, territorial displays at the onset of courtship 
behaviour, and is responsible for libido and mating behaviour (Farner and Wingfield, 1980; Howarth, 
1995). 
Behavioural patterns exhibited during the reproduction period can provide a good indication of an 
individual's reproductive ability (Ottinger and Mench, 1989). The present study investigated the 
relationship between male aggression and shin colour, female egg production performances recorded 
for, and serum levels of testosterone, LH, progesterone, and prolactin sampled in ostriches at the 
beginning and end of breeding seasons. 
MATERIALS AND METHODS 
Experimental animals 
An ostrich breeding flock consisting of breeding pairs (n=136) maintained at the Little Karoo 
Agricultural Development Centre outside Oudtshoorn, South Africa, were used for the study. The 
management of the breeding flock is documented by Van Schalkwyk et al. (1996) and Bunter and 
Graser (2000). The ages of the breeding birds varied between 2 and 12 years. 
Experimental procedures 
Collection and processing of blood samples 
Blood sampling was performed at the beginning and end of each breeding season. Blood samples 
were collected by venipuncture of the wing vein (Van Heerden et al., 1985) and processed as soon as 
68 
possible after collection. The blood samples were not collected at a set time and collections were 
performed throughout the day. In cases where samples were not processed on the day of collection, 
samples were stored at 4°C for a maximum of one day. Blood samples were centrifuged at 3000 rpm 
for 10 minutes and serum was transferred to sterile vials, and stored at -20°C for later analysis. 
Freeze drying of samples 
One millilitre of each sample was aliquoted to sterile glass serum vials, labelled and transferred to a 
deep freeze at -80°C, where the samples were kept until subjected to freeze-drying. After completion 
of the freeze drying process, the samples were stored at 4 °c until analysis. 
Hormone assays 
All radio-immuno assays were performed at the Faculty of Agriculture, University of Western Australia, 
Perth, Australia. After being subjected to gamma irradiation, serum samples were reconstituted by 
adding 1m l of sterile, distilled water (Millipore) to each sample. Before the reconstituted serum 
samples were subjected to radio-immuno assay (RIA) procedures, samples were clarified by 
polyethyleneglycol (PEG) precipitation, according to a method described by Van Cleeff et al. (1998). 
Emu serum samples were clarified for RIA with PEG, with minimal hormonal loss (Van Cleeff et al., 
1998). Serum samples destined for testosterone and progesterone assays were not subjected to PEG 
precipitation. The respective media preparation guides and assay procedures are set out in Appendix 
A. All assays were performed in duplicate. 
LH assay 
Serum concentrations of LH were measured by the RIA method described by Sharp et al. ( 1987), using 
chicken-LH (PRC-AE1-1 ), prepared and used for antisera production, as described by Talbot et al. 
(1988). The antiserum was raised in a rabbit and designated anti-ch-LH 3/3 (anti 3/3). The PRC-AE1-
1 was purified by gel filtration to remove contaminating proteins of high and low molecular weight, and 
the resulting PRC-AE1-1-s-1 was labelled with 1251 using chloramine-T. On day one, 50µ1 of the first 
antibody was added to all the tubes and incubated at 4 °C for 24h. On day two, 50µ1 of tracer (1 25 1-
PRC-AEl-s-1) was added to each tube and incubated for 48h. On day four of the assay, 50µ1 of normal 
rabbit serum (NRS) and 50µ1 of the second antibody were added to all tubes, except the total counts. 
On day five, 1m l of basic LH buffer with 6% PEG was added to all tubes, except the total counts, and 
centrifuged at 3000rpm for 30 minutes. The supernatant was then aspirated with a Pasteur pipette 
attached to water pump. Tubes were then counted for 60 seconds. 
69 
Prolactin assay 
Serum prolactin concentrations were assayed as described by Talbot and Sharp (1994). 
Recombinant-derived chicken prolactin was prepared as described by Hanks et al. (1989 cite Talbot 
and Sharp, 1994) and was used for tracer and antibody production. On day one, 50µL of the first 
antibody was added to all tubes except the total count (TC) and non-specific binding (NSB) tubes, and 
incubated to 4°C for 24h. On day two, 50µL of tracer (radioactive prolactin) was added to all tubes and 
incubated at 4°C for 24h. On day three, 50µL of NRS and 50µl of the second antibody was added to 
all tubes, vortexed and incubated at 4°C for 24h. On day four, 1ml of PRL buffer containing 6%PEG 
was added to all samples and centrifuged at 4°C and 3000rpm for 25 minutes. The supernatant was 
then aspirated off and all tubes were counted. 
Testosterone assay 
Serum testosterone concentrations were measured in unextracted samples, as described by Malecki et 
al. (1998). An antiserum raised against testosterone-3-carboxymethyloxime-human serum albumin, 4-
androsten-17~-ol-3-one as standard and 1, 2, 6, 7-3H-testosterone as tracer was used in the assay. 
On day one, 100µl of the first antibody was added to all tubes except the NSB tubes, to which 100µl 
GPB was added. On day two, 1O Oµl of the tracer was added to all tubes and samples were vortexed 
and incubated at 4 °C. On day three, 1O Oµl NRS and 100µ1 of the second antibody was added to all 
tubes. On day four, 1m l of PBS with 2%PEG was added to all tubes except the total counts. All 
samples, with the exception of the total counts, were centrifuged at 4°C and 3000rpm for 25 minutes. 
After centrifugation, the supernatant was aspirated and 0.5ml 0. 05M HCI added to tubes. The 
samples were then vortexed and dispensed into the counting vials that contained 2ml of scintillant 
each. Each sample was gently shaken and left in the dark for at least 1h  before being counted. 
Progesterone assay 
Serum progesterone concentrations were measured in unextracted samples by using a method 
developed by the Animal Science RIA laboratory of the University of Western Australia. Samples were 
extracted by adding 2ml of hexane to 100µl of serum and vortexed for 3 minutes. The aqueous 
phase was frozen by using an acetone-dry ice bath. The solvent phase was then transferred to assay 
tubes (10X75 mm) and was then dried under nitrogen gas. After extraction on day one, 200µL of the 
first antibody were added to samples, vortexed and incubated at 4 °C overnight. On day two, 1O Oµl 
tracer was added and samples were again vortexed and incubated overnight at 4 °C. On day three, 
100µl of the second antibody (15µl donkey serum 5 batch 1 + 90µl PBS) was added, with samples 
again vortexed and incubated overnight at 4 "C. On day four, 1m l of 2% PEG was added and 
samples were spinned for 25 minutes at 4°C and 3000rpm. The supernatant was aspirated and 0.5ml 
70 
0.05M HCI was added to all tubes, except the total count tubes. Samples were then vortexed and 
dispensed into counting vials with 2ml of scintillant. Samples were shaken and left 1n dark for a 
minimum of one hour before being counted. 
Data recorded 
Each male was evaluated on a monthly basis in terms of territorial aggression and shin colour, as 
described in Materials and Methods, Chapter 5. Daily egg production was recorded for each breeding 
pair throughout the breeding season. Post-collection treatment and subsequent artificial incubation of 
eggs was performed as described by Van Schalkwyk (1998). To account for any diffe1·ences in the 
number of production days, the number of eggs produced per breeding pair was calculated as egg 
production performance (EPP) by using the following equation, and expressed as a percentage (Van 
Schalkwyk et al., 1996): 
EPP= [(total number of eggs produced/(0.5 X number of production days)] X 100 
Statistical analysis 
The ASREML program (Gilmour et al., 1999) was used for the estimation of the fixed effects, and also 
subsequently to derive the variance components for live weight, hormone concentrations, EPP and 
hormone concentrations in univariate analyses. The fixed effects that were considered in all the 
analyses included sampling time (at the beginning and end of breeding), and gender where applicable 
(progesterone concentrations and EPP were measured in females only, and aggression and shin 
colour scores in males only). Age was not included as a fixed effect and could be described as 
longitudinal data. Correlations were estimated only for traits where one or more of the random effects 
exceeded 5 % of the overall phenotypic variance, since spurious results may be obtained with very 
small variance components. 
On the assumption that a specific trend would be discernible, a smoothing spline was fitted to the data 
where applicable (Verbyla et al., 1999). The spline consisted of three components, i.e. a fixed linear 
component, random deviations from linearity following a smooth trend, and random deviations from 
linearity not conforming to a smooth trend. The latter effects were found to be not significant, and were 
excluded from the final analyses. Random terms were then added to the operational model, resulting 
in the following models for analyses (in matrix notation): 
y=Xb+Z1a+e ( 1) 
y = Xb + Z2c + e (2) 
y = Xb + Z1 a + Z2c + e (3) 
In these models, y was a vector of observations for production traits or hormone concentrations; b, a 
and c were vectors of fixed effects, direct genetic effects and permanent environmental effects; x, z, 
71 
and Z2 were the corresponding incidence matrices relating the respective effects to y, and e the vector 
of residuals. 
It was assumed that: 
V 2 2 (a)= ACT a. V(CpE) = l02 c; V(e) =la e. 
2 2 2 
with A being the numerator relationship matrix, I being an identity matrix; and 0 a. 0 c, and 0 e being 
the direct genetic variance, permanent environmental variance, and environmental (residual) variance 
respectively. All analyses included the full pedigree file, consisting of 601 individuals, the progeny of 
135 males and 132 females. Random terms were added to analytical models sequentially, as noted 
above, with Likelihood Ratio Tests (LRT) performed to assess the significance of their contribution to 
improvements in the model for analysis. The LRT is based on testing twice the increase in Log-
2 
likelihood resulting from adding n random terms to the model of analysis as a Xn statistic. Alternatively, 
for two models with the same number of parameters, and assuming identical fixed effects models, the 
one with the higher likelihood fits the data better. Subsequently, 2-trait animal models were fitted. 
These analyses allowed the calculation of all relevant correlations between traits, together with their 
appropriate standard errors. 
RESULTS 
The coefficient of variation for live weight of the breeding birds was 12.4% (Table 1) . Serum hormone 
levels were exceedingly variable, with the respective coefficients of variation exceeding 60% in almost 
all cases. The exception was prolactin, with a coefficient of variation of below 30%. The coefficient of 
variation for shin colour score was approximately 10%, while that of aggression score was slightly 
below 30%. The coefficient of variation of EPP across the two breeding seasons approached 50%. 
Table 1. Means (SD) for live weight, egg production performance (EPP), serum hormone concentrations, and 
male aggression and shin colour measured in breeding ostriches. 
TRAIT Number of observations Mean (SD) CV(%) Range 
Live weight (kg) 692 121 (15) 12.4 76 -160 
EPP(%) 220 41.6(19.4) 46.6 0- 89.9 
Testosterone (ng.ml-1) 628 0.52 (0.57) 109.6 0 - 6.4 
Prolactin (ng.mL 1) 690 0.84 (0.23) 27.4 0.01 - 1.75 
LH (ng.ml-1) 683 3.25 (2.97) 91.4 0.16 -17.B 
Progesterone (ng.mL-1) 320 0.58 (0.40) 68.9 0.14-2.63 
Shin colour score 349 7.42 (0.77) 0.10 5 - 10 
Aggression score 349 6.20 (1.82) 29.4 5-·10 
-----
72 
The ostrich males were slightly heavier (P~0.05) than females overall, i.e. 125 ± 2 kg vs. 119 ± 2 kg, 
respectively (Figure 1 ). Live weigt1t and EPP of ostriches generally increased with animal age, to 
reach a maximum at 5-6 years of age for females and 6 years for males As the trends at older ages 
(9-11 years) were based on fewer observations, they were more variable as indicated by the larger 
standard errors (Figure 1) , and did not differ from live weight at 6 years of age (P~0.05). 
140.0 
130.0 ------ ----
.... --i---f-.~ 
..c: 
·~ 120.0 ·- --- --- ------- --- r=+- M~\e-l 
=Q=)  
> '1' -----n--0--1 l--0---F~ 
:i 
110.0 -
100.0 -~--,~--.-·~---·----,--------1~--- -----,--· ------,-- --~---------1-- -------1 
2 3 4 5 6 7 8 9 10 11 
Age (years) 
Figure 1. The influence of age on the live weight of male and female ostriches. 
The observed likelihood ratios indicated that the models fitting both direct genetic and animal 
permanent environmental variances fitted the data best for the production traits of animal live weight 
and EPP (Table 2). The animal permanent environment was the only significant (P~0.05) random 
effect in the cases of serum testosterone and LH concentrations. Serum prolactin and progesterone 
concentrations were independent of the random sources of variation considered, and a fixed effect 
model fitted the data best. The best model for shin colour score included the direct additive effect as a 
significant (P~0.05) additional random effect. The male permanent environment contributed 
significantly to aggression score. 
73 
Table 2. The log likelihood ratios for the respective random effect models fitted to the data set, with the best 
model indicated in bold. 
-
EFFECT(-S) IN MODEL 
TRAIT 
Fixed only + h2 (1) + c2 (2) + h2 + c2 (3) 
Live weight -2151.3 -2001.7 -2002.5 -1994.3* 
Egg production performance -806.3 -789.6 -790.0 -787.5* 
Shin colour score -64.97 -62.58* -62.83 -62.58 
Aggression score -393.2 -363.3 -363.0* -361.4 
Serum LH -791.7 -755.1 -771.8* -770.8 
Serum testosterone 46.84 47.46 49.01* 49.01 
Serum progesterone 127.5* 128.1 127.5 128.1 
Serum prolactin 642.6* 642.8 643.6 643.6 
·-
*Ps:0.05 
The heritability (h 2 ) for live weight amounted to 0.30, whereas the animal permanent environment (c2 ) 
contributed 42 % towards the overall phenotypic variance (Table 3). The h2 for EPP was 0.31, with a 
corresponding c2 estimate of 0.29. The animal permanent environment contributed 9%, 29% and 43% 
respectively, to the variation observed for serum testosterone, LH, and male aggression score, 
respectively. Although the log likelihood ratios in Table 2 indicated a significant h2 estimate for shin 
colour score, this estimate was below 0.05 (Table 3) and smaller than the corresponding standard 
error. 
Table 3. The variance components and ratios for live weight, serum testosterone and LH, male aggression and 
shin colour, and egg production performance (EPP) measured in breeding ostriches. 
COMPONENTS TRAIT 
AND RATIOS Live weight Testosterone LH Shin Aggression EPP 
colour 
Component 
2 
()' a 57.37 0.0164 129.3 
2 
cr PE 80.82 0.0265 1.027 1.405 118.8 
2 
cr e 54.26 0.2814 2.509 0.4918 1.897 167.7 
2 
()' p 192.44 0.3079 3.536 0.5082 3.302 415.9 
Ratio 
h2 0.30 ± 0.11* 0.03 ± 0.04 0.31± 0.16 
c• 0.42 ± 0.10** 0.09 ± 0.04* 0.29 ± 0.05** 0.43 ± 0.06** 0.29 ± 0.16 
--~ 
*Ps0.05; **Ps0.01 
The genetic correlation between live weight and EPP were positive, and relatively high (Table 4). Live 
weight was negatively correlated with EPP on a permanent environmental level and in the medium 
range, compared to most of the other correlations, which were below 0.30. Environmental correlations 
were in most cases below 0.15, with the only exception the correlation between serum LH 
74 
concentration and EPP (0 40; Table 4). The phenotypic correlations between live weight and EPP 
(0.17) were significant (P:e:0.05) and the phenotypic correlation between LH concentration and EPP 
(0.16) approached significance. 
Table 4. The genetic, permanent environmental, environmental, and phenotypic correlations between behavioural 
and reproduction traits measured in breeding ostriches, where at least one random effect exceeded 5 % 
of the overall phenotypic variance. 
----·---- -
CORRELATION 
TRAITS INVOLVED Genetic Permanent environmental Environmental Phenotypic 
Live weight x 
Serum LH -0.01±0.12 0.05 ± 0.05 0.02 ±0.04 
Serum testosterone -0.17 ± 0.20 0.09 ± 0.05 0.01 ±0.04 
Aggression score -0.07 ± 0.15 0.11 ± 0.07 0.02 ±0.06 
EPP 0.67 ± 0.22** -0.34 ± 0.42 0.03 ± 0.10 0.17 ± 0.08* 
Serum testosterone x 
Serum LH 0.20 ± 0.23 0.05 ± 0.05 0.07 ± 0.04 
Aggression score -0.04 ± 0.25 -0.02 ± 0.07 -0.02 ± 0.05 
EPP 0.30 ± 0.44 -0.14±0.11 -0.04 ± 0.07 
Serum LH x 
Aggression score -0.07 ± 0.14 -0.04 ± 0.06 -0.05 ± 0.05 
EPP -0.22 ± 0.33 0.40±0.11** 0.16±0.09 
*Ps0.05; **Ps0.01 
The influence of age 
Serum LH levels increased almost linearly with age in males, and ranged between 4.9±0.2 ng.ml-1 (2-
1 
year old) to 6.8±0.4 ng.mL· (11-year old). Female serum LH levels, however, did not follow the same 
pattern (Figure 2). The highest serum LH concentrations were reported for young 2-year old females, 
and the lowest serum LH concentrations reported for older 11-year old females. The serum LH 
1 
concentrations reported for females ranged between 0.9±0.2 ng.ml- and 0.8±0.5 ng.ml-1, for 2-year 
and 11-year old females, respectively. 
75 
8.0 60.0 
- 7.0. 50.0 
E 
Ci 6.0 
c: 40.0 
; 5.0 ~ 
0 
:.;::; 
... 4.0 
30.0 ~ 
<ti a.. 
.....  w 
c: 3.0 20.0 
Q) 
(.) 
c: 2.0. 
0 
(.) 10.0 
1.0 
0.0 0.0 
2 3 4 5 6 7 8 9 10 11 
Age (years) 
1;-·--~-----.-···---··1 
Lc:::::J Male c::::J Female --- EPP 
Figure 2. The influence of age on the serum LH levels (mean ± SE) measured for and egg production 
performance (EPP; mean ±SE) of breeding ostriches. 
The serum testosterone concentrations measured in the ostrich males was higher in birds 6 and 8 
1 
years, respectively, and varied between 0.7±0.1 ng.mL·1 and 0.8±0.1 ng ml (Figure 3). Serum 
1 
testosterone levels tended (P:s:0.10) to decline with an increase in age, and measured 0.3±0.3 ng.mL·
in 11-year old males. There was a tendency (P:s:0.10) for the serum testosterone levels of females 
aged 9 years and younger to be lower, when compared to that measured in the males. Exceptions 
were in the 3-year old and 9 years and older age groups, where the ostrich females had higher levels 
of serum testosterone than the males (Figure 3). 
Serum prolactin levels ranged between 0.8±0.1 ng.mL·1 and 0.9±0.1 ng.mL·1 The serum prolactin 
levels measured in males ranged between 0.7±0.1 ng.mL·1 and 0.9±0.1 ng.mL·1 (Figure 3). 
1.2 60.0 
1.0 50.0 
c: -
tl "E :i: ;.s ;r ~ 
~ gi ·. ~ :V . -,,
i  ~ .I :i:---t 0.8 ~ l ~ 40.0 
n: ~ ::
a; 0 :.:. 
 :3 
 ~~ ·~ § 
~.  :: ~-=s  
c :;; 0.6 :: ; .. 30.0 a.. 
.o. ...n.s  . . I . ~ a.. 
.s ~ 11 :..:  ·~ I~ ~ ~ . ~ ',.. ·  . 
~ w 
gi ~ 0.4 ·. § ' 20.0 u; § ·. I·,. :3 ~ ;: ,. :- ~3 . .·. ~..  . ' ~ 0 ~ : ~ 
0.2 :; 10.0 ~: ~ ~ ,. :~:  
'. ~ ~ : 
~ .~  ., . ~ . ' ..-.·~   •' 
0.0 ~= -: .. · ' " 0.0 
2 3 4 5 6 7 8 9 10 11 
Age (years) 
[C:=:nest (M) c::::nest (F) - Prol (M) ~Pro! (Fl_-0- EPP-~ 
Figure 3. The relationship between age, serum levels of testosterone and prolactin (mean ± SE), and the egg 
production performance (EPP; mean ±SE) of breeding ostrich males and females. 
76 
Serum progesterone levels tended (P:o::0.10) to increase with age, and ranged between O 5±0 04 
ng mL·1 in 2-year old females, to 0.7±0.1 ng.ml-1 in 11-year old females (Figure 4). 
0.9 60.0 
0.8 
c: 
0 50.0 
·.;:; 0.7 
ce:  
Q) 0.6 40.0 
<.> 
8c:  -E -o.5 ~ 
30.0 ~ 
Q) Ci 
ec: -c: 0.4 a.. w 
2 0.3 20.0 
If) 
Q) 
Cel  0.2 
c.. 10.0 
0.1 
0 +- -+ +- 0.0 
2 3 4 5 6 7 B 9 10 11 
Age (years) 
Figure 4. The relationship between age, serum progesterone concentration (mean ± SE) measured in and egg 
production performance (EPP; mean ± SE) of ostrich females. 
The influence of gender and time of sampling 
Gender and the time of sampling interacted (P:o::0.05) for serum testosterone and LH. Serum 
testosterone concentrations measured in the breeding males and females were similar at the beginning 
of breeding in 2000 (Table 5). Female serum testosterone concentrations tended to be higher at the 
end of breeding in January 2001, whereas it was approximately 30 % higher (Ps0.05) in the males at 
the beginning of the 2001 breeding season. Male serum LH concentrations tended to increase during 
the 2000 breeding season, being higher at the end of the season. The opposite trend was observed 
for the breeding females (Table 5). Male serum LH concentrations measured at the beginning of the 
2001 breeding season stayed at almost the same level, while female serum LH concentrations tended 
(P?0.05) to be higher at the beginning of the 2001 breeding season (Table 5) The serum 
progesterone concentrations measured in the ostrich females at the end of the 2000 breeding season 
were higher (P:o::0.05) than that measured at the beginning of the 2001 breeding season (Table 5). The 
serum prolactin concentrations were not affected by the stage of sampling, but the ostrich females had 
slightly higher concentrations than the males overall (0.9 ± 0.01 vs. 0.8 ± 0.01 ng.ml-1 respectively). 
The total EPP during the 2001 breeding season was 26 % higher (Ps0.01) than in the 2000 breeding 
season. Male shin colour scores assessed at the end of the 2000 breeding season were higher 
(Ps0.01) than scores obtained at the beginning of breeding (Table 5). No significant tendency was 
found for male aggression scores 
77 
Table 5. The influence of time of sampling on the live weight, egg production performance (EPP), male 
aggression and shin colour, and the serum hormone levels measured for breeding ostriches. The 
interaction of gender and time of sampling are presented where significant (Ps 0.05). 
TRAIT AND GENDER FOR -h---·---<------·-·--SAMPLING PER-10__0 _________________  
SIGNIFICANT INTERACTIONS Beginning 2000 End 2000 Beginning 200f--
---
Live weight (kg) 122.0 ± 2.0 123.0 ± 1.0 121.0±1.0 
Total EPP(%) 35.5 ± 3.0 44.6 ± 3.0 
Shin colour score 7.11±0.08 7.80 ± 0.06 7.21 ± 0.07 
Aggression score 5.92±0.19 6.23±0.16 6.20 ± 0.16 
Serum testosterone (ng.ml· ) 
Male 0 53 ± 0.06 0.51 ± 0.05 0.46 ± 0.05 
Female 0.50 ± 0.07 0.57 ± 0.05 0.31 ± 0.05 
Serum LH (ng.ml-1) 
Male 4.84 ± 0.21 5.49 ± 0.17 5.44 ± 0.18 
Female 0.91±0.24 0.66±0.18 0.79±0.17 
Serum progesterone (ng.ml-1) 0 59 .± 0.05 0.67 ± 0.04 0.50 ± 0.03 
Serum prolactin (ng.mL-1) 0.82 ± 0.02 0.84 ± 0.01 0.85 ± 0.01 
------------------------~---~----~----~ 
DISCUSSION 
Serum levels of testosterone, LH, and progesterone were extremely variable, with the coefficient of 
variation exceeding 60% in all cases. Serum prolactin was the only exception, showing the least 
variation. The animal permanent environment contributed significantly (P:s;0.05) to the observed 
variation in the serum testosterone and LH concentrations in both males and females. Variation in the 
serum levels of prolactin in both the breeding males and females, and serum progesterone in the 
breeding females could be attributed to the time of sampling and/or the gender of the animal. The 
observed variance in male aggression and shin colour (both traits measured only in the ostrich males) 
could be attributed to the direct additive variation and the permanent environment, respectively. When 
the correlations of the behavioural and reproduction traits with the various hormones measured are 
considered, no significant correlations were evident for the various traits considered. The only 
exception was the correlation between serum LH concentration and the EPP of the breeding females, 
which was found to be significant and positive on a genetic, environmental, and phenotypic level. 
An interesting observation was the sexual dimorphism in terms of live weight, with males being the 
larger gender. This presents the first evidence of sexual dimorphism in ostriches, as sexual 
dimorphism is generally not found for live weight in ostriches (Cloete et al., 1998a, b; Bunter et al., 
1999; Cloete et al., 1999a, b; Bunter and Graser, 2000; Cloete et al., 2002). The latter studies 
focussed on slaughter ostriches of up to 14 months of age. Variation in sexual dimorphism in birds is 
generally attributed to differences in social mating systems, ie. polygamy leads to the competitive sex 
78 
being larger and more ornate (Owens and Hartley, 1998). This is, however, not the case in all bird 
species, for it can also be associated with gender differences in parental care (Owens and Hartley, 
1998). Males and females can also differ in less overt ways such as levels of activity, regulation of 
body weight and aggression level, with some gender-specific actions associated with systematic 
dissimilarities in certain parts of the brain (Crews, 1994). The observed dimorphism in size in the 
breeding males and females in the present study is supported by the fact that the ostrich male invests 
more time and energy in the raising of chicks than the female does. Characteristic of all ratite species 
is an increased parental investment by the male, with ostriches being the only group where the female 
will assist in incubating the eggs, as well as participating in raising of the chicks. 
The heritability reported for live weight was in the medium range, while the permanent environment 
contributed 42% to the overall phenotypic variance. Live weight was positively and correlated (P~0.05) 
with EPP on a genetic level, although a negative correlation was reported with EPP on a permanent 
environmental level. This negative correlation support previous studies by Van Schalkwyk et al. 
(1996), where a negative correlation was reported for live weight and EPP. 
The influence of age 
Serum LH concentration measured in the breeding males tended to increase almost linearly with age, 
with highest levels reported for 11-year old males. This tendency was not observed for the breeding 
females, for highest concentrations were reported for the 2-year old females and lowest concentrations 
for the 11-year old females. Serum LH levels measured in the males were considerably higher than 
those measured in the females, which indicates the extent of involvement of LH in the various 
reproductive processes. In males, LH is involved in steroid synthesis, as well spermatogenesis, and in 
females LH is involved in steroid synthesis and ovulation. Serum testosterone levels measured in the 
breeding ostriches varied between 0 and 6.4 ng.ml 1, with levels being higher in males in most cases. 
Male serum testosterone levels tended to increase with in age, with highest levels reported for males 
aged 5-6 years. Serum prolactin levels did not differ between the males and females, with the females 
having slightly higher concentrations than the males. Serum progesterone levels measured in the 
breeding females increased with age, and ranged between 0.1 ng.mL· 1 and 2.7 ng ml-1. Age did not 
influence the level of circulating progesterone and estradiol in domestic chicken hens, with the older 
hens producing larger but fewer eggs than younger females (Joyner et al, 1987). 
The influence of gender and time of sampling 
The ostrich males tended to have higher serum testosterone concentrations at the beginning of a 
breeding season, with the females having higher concentrations at the end of a breeding season. The 
establishment of territories, attracting breeding partners, and high serum LH levels are possible factors 
79 
contributing to the higher serum testosterone levels measured. Several studies have highlighted the 
relationship between circulating levels of testosterone, reproduction, and territorial behaviour of males 
during breeding periods in birds [song sparrows (Wingfield, 1984 ). Northern pintail ducks (Penfold et 
al., 2000); Korean ring--necked pheasants (Kim and Yang, 2001 )]. In male Houbara bustards, the 
highest testosterone and LH levels were associated with territorial display before onset of lay and after 
cessation of laying (Saint Jaime et al., 1996). In the current study, serum testosterone levels were 
slightly lower at the end of breeding, possibly as a result of decreased LH activity due to the ostrich 
males becoming photorefractory. In male emus, plasma LH and testosterone concentrations were 
highest during the breeding season, with a gradual decrease in LH and a rapid decrease in 
testosterone concentrations at the end of a breeding season (Malecki et al., 1998). Males of another 
ratite species, the North Island brown kiwi (Apteryx australis mantel/!}, had low levels of testosterone 
during the non-breeding season (<0.18ng.ml·\ which rose to a maximum of 1.9±0.8 ng.ml-1 during 
the peak breeding season. In contrast, testosterone levels in the kiwi females remained lower than 
0.10 ng.ml-1 (Potter and Cockrem, 1992). Values reported for the ostrich males were intermediate to 
that reported for the kiwi males, and values reported for ostrich females in this study were higher than 
that reported for kiwi females. The relative higher testosterone values reported for the ostrich females, 
when compared to the kiwi females, can possibly be ascribed to the fact that the ostrich female 
participate in nest-defence as well as incubation behaviour, whereas the female kiwi does not. 
Studies of Gomez et al. (1998) suggested that complex mechanisms regulate steroid synthesis in the 
avian ovary. In the avian female, androgens of ovarian origin is also thought to be involved in oviduct 
maturation (Gilbert, 1971 ). The extent of follicular development of a female's ovary at the time of 
sampling could possibly have contributed to the differences observed for beginning and end of season 
measurements. Follicular development is less extensive at the onset of breeding, and becomes more 
extensive as a breeding season progresses. Testosterone is synthesised by the theca cells of the 
avian ovary, and a less extensive follicular development would possibly have the effect that levels were 
lower than when measured at the end of a breeding season when follicular development is more 
extensive. At the end of the breeding period, it is possible that the ovary was still sensitive enough to 
the given daylength and continued to synthesise testosterone to support the oviduct. Testosterone 
stimulated progesterone synthesis in cultured granulosa cells of Japanese quail (Sasanami and Mori, 
1999). Higher female testosterone levels observed at the end of the breeding season might also be 
the result of a decrease in aromatase activity and estradiol production. Female zebra finches had 
higher aromatase activity, but lower testosterone levels than males (Vockel et al., 1990). It is thought 
that the steroid modulation of aromatase activity observed in the latter study may be related to the 
activation of sexual, aggressive, and nest-building behaviours (Vockel et al., 1990). Tanabe et al. 
(1983) found that plasma levels of testosterone and 17~ estradiol were higher in female than in male 
embryos. Unfortunately, aromatase activity and estradiol levels were not measured in this study. 
80 
The lower serum- LH levels measured in the ostrich females at the end of a breeding season can be 
ascribed to the ostrich females becoming photorefractory, with a subsequent decrease in serum LH 
concentration. LH levels show a characteristic seasonal cycle in domestic Hungarian geese, with a 
definite decrease at the onset of photorefractoriness (Hargitai et al., 1993). In rose-ringed parakeets 
(Psittacula kramen), plasma LH decreased at the end of a breeding season (Sailaja et al, 1988). A 
sharp decline in plasma LH levels was also observed in wild Svalbard ptarmigans (Lagopus mutus 
hyperboreus), a non-photo periodic breeder (Stokkan et al, 1986). The higher LH levels measured in 
males at the end of the breeding season can be ascribed to the fact that ostrich males do not react as 
strongly as ostrich females to a decrease in photoperiod, and that LH synthesis was maintained in 
order to support the synthesis of testosterone and reproductive behaviour Males tend to defend their 
breeding territory throughout the breeding season, which is usually accompanied by territorial 
aggression. Testosterone levels in brown dipper (Cine/us pa!lasi1) males were associated with 
intensive territory-defence behaviour (Kofuji et al., 1993), and free-living white ibises (Eudocimus 
a/bus) had higher testosterone levels before a breeding season, and lower levels throughout the 
remaining stages of the reproductive season, i.e. copulation, incubation and chick rearing. 
Degen et al. (1994) found that peak LH levels in ostrich males and females coincided with the initiation 
of an egg production cycle. However, the amplitude for the reported annual cycle was small, when 
compared to that of other avian species. Malecki et al. (1998) reported a similar subdued cycle in LH 
for emus. Degen et al. ( 1994) measured both testosterone and estradiol levels in ostriches and found 
peak levels to coincide with egg laying. The values however, were lower than values reported for other 
seasonally breeding birds (Degen et al., 1994) 
Lower serum levels of progesterone were measured before the onset of breeding than at the end of 
breeding. This can possibly be ascribed to the degree of development on the ovary which is not as 
extensive as during the later stages of the breeding period. Progesterone is synthesised by the larger 
yellow or pre-ovulatory follicles in the follicular hierarchy, and a lower number of the pre-ovulatory 
follicles may contribute to the lower levels of progesterone measured in the study. The presence of a 
larger number of large follicles towards the end of breeding, partly as a result of failure to be ovulated 
as a result of lower LH levels may contribute to the higher progesterone levels measured at the end of 
breeding. In two species of Houbara bustard, seasonal variation in LH and progesterone in females 
was synchronised with egg production (Saint Jaime et al., 1996). In female white ibises, progesterone 
levels increased during display and remained high throughout the breeding season, whereas male 
progesterone levels did not show the same seasonal pattern, and were more variable (Heath et al., 
2003). 
Prolactin is a pituitary peptide hormone that is involved in the development of broodiness and 
regression of gonads, with marked seasonal changes in circulating levels of prolactin occurring in most 
bird species studied. Prolactin levels were measured in a few cases in ostriches, with prolactin level 
81 
being higher in an incubating male (Skadhauge and Dawson, 1999). In this study, eggs were collected 
on a daily basis that suppressed the development of broody behaviour in both males and females. It 
can thus be accepted that the mnge reported for breeding ostriches in this study (0.80 - 0.87 ng mL-1) 
represents low values for breeding, non-incubating ostriches, and that levels measured in incubating 
ostriches could probably be higher Malecki et al. (1998) found prolactin levels to increase in non-
breeding emu males, with the highest levels coinciding with a rapid decrease in testosterone level, 
which marked the end of the breeding season. Serum prolactin concentration peaked several weeks 
after the peak of the breeding season (Malecki et al., 1998). Prolactin levels measured in male emus 
1 
varied according to the number of eggs present in the nest, ie. between 1.2±0.5 ng.mL· (6 eggs in the 
nest) and 6.6±1.8 ng.mL· 1 (8 and more eggs in nest), respectively (Van Cleeff et al., 1999). In 
mallards, prolactin levels were higher during incubation, and LH levels remained high for longer than 
that observed in females (Goldsmith and Williams, 1980). 
Laying and non-laying canvasback duck (Aythya valisineria) females were compared in terms of levels 
of serum LH, prolactin, estradiol, and progesterone during the pre-laying, laying and incubation periods 
(Bluhm et al., 1983). Lower levels of LH, prolactin, estradiol, and progesterone were recorded for non-
laying females during the respective three periods. Bluhm et al. (1983) suggested that the lack of 
reproduction could be ascribed to the failure of the hypothalamus to release GnRH hormones and not 
the inhibitory effect of prolactin, for injection with LHRH had the effect that serum LH levels were 
significantly elevated. In domestic hens, plasma LH levels were found to be significantly higher in hens 
with a high rate of egg production, when compared to hens with a poor rate of egg production (Wilson, 
1978). Serum LH levels varied between 1.7±0.3 ng.mL-1 and 2.6±0.7 ng.mL-1 for arrested and laying 
turkey hens, respectively, and progesterone levels between 1.8±0.2 ng.mL-1 and 4. 7±1.3 ng mL-1 for 
normal and arrested laying turkey hens, respectively (Liu et al., 2001 ). 
Domestication may have an influence on the hormonal cycles of certain avian species, which may lead 
to certain reproductive problems (Paster, 1991 ). Seasonal changes in the production of the steroid 
hormones seem to control the seasonal changes in sexual behaviour of birds, with influences on e.g. 
courtship, copulation and territorial displays (Gahr, 2001 ). The sex steroids may influence the 
development and expression of sexual behaviour by direct actions on the central nervous system 
(Schultz and Schlinger, 1999). Selecting ostriches for production performance should not have a 
negative impact on the hormonal profiles of females. Turkey hens selected for high growth had 
multiple ovulations, and low plasma concentrations of oestrogen, with a correlated response in the 
steroidogenic capacity of the ovarian tissue (Buchanan et al., 2002). 
82 
.,.. 
CONCLUSIONS 
The only significant correlation reported was between serum LH concentration and the egg production 
performance of the ostrich females, on an environmental level. The fact that the blood samples were 
collected throughout the day in this study and not at a specific time may have affected the serum LH 
concentrations measured in this study. It is known that LH secretion follow a diurnal pattern in female 
birds, and it is possible that peak levels were measured in some females and low levels on other may 
have contributed to the observed variation. 
Gender and time of sampling interacted for serum LH and testosterone concentrations, implying that 
the extent of gonad differentiation and the degree of photorefractoriness the breeding ostriches were 
experiencing influenced the serum testosterone and LH levels measured at the beginning and end of 
breeding. The environment thus played a major role in the regulation of reproduction and reproductive 
performance of the breeding birds. The maintenance of territorial aggression in males at the end of a 
breeding season, when lower levels of serum testosterone were measured, imply that factors other 
than LH may contribute to the maintenance of reproductive behaviour at the end of a breeding season. 
The potential of, and the extent to which the reproductive cycles can be manipulated by means of 
exogenous gonadotropins and steroid hormones to improve the reproductive performance of breeding 
ostriches under commercial farming conditions warrants studies to determine baseline values of the 
gonadotropins and steroid hormones for breeding ostriches. 
Future studies need to focus on establishing the circulating levels of LH, testosterone, estradiol, 
progesterone, as well as prolactin, in ostriches during non-breeding and breeding seasons, and also on 
a diurnal basis. This may provide better insight into the endocrine control of reproduction in both males 
and females, and how this is linked to reproductive behaviour exhibited by both genders. This will also 
enable the potential application of reproductive technologies, as in the case of breeding emu male and 
females, to improve overall reproduction efficiency of ostriches in commercial breeding systems 
(Williams et al., 1998). 
REFERENCES 
BLUHM CK, PHILLIPS RE & BURKE WH, 1983. Serum levels of luteinizing hormone (LH), prolactin, 
estradiol, and progesterone in laying and nonlaying canvasback ducks (Aythya va/isineria). 
Gen. Comp. Endocrinol. 352(1): 1-16. 
BON NA-GALLO A, LICHT P & PAPKOFF H, 1983. Biological and binding activities of pituitary hormones 
from the ostrich, Struthio came/us. Gen. Comp. Endocrinol. 51: 50-60. 
83 
BUCHANAN S, ROBERTSON GW & HOCKING PM, 2002. Comparative changes in plasma concentrations 
of progesterone, oestradiol and LH during the ovulatory cycle in a multiple ovulating male line 
and a single ovulating traditional line of turkeys. Reprod. 123(1) 127-133. 
BUNTER K & GRASER, H-U, 2000. Genetic evaluation for Australian ostriches - a report for the Rural 
Industries Research and Development Corporation. Publication no. 00/153, Rural Industries 
Research and Development Corporation, Australia. 
BUNTER KL, CLOETE SWP & VAN SCHALKWYK SJ, 1999. Significant genetic parameters for egg, chick 
and juvenile weight traits in ostriches. Proc. Assoc. Advancem. Anim. Breeding Gen. 13: 476-
479. 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998a. Ostrich breeding - progress towards a 
scientifically based strategy. Proc. 2nd lnternat. Scient. Ratite Cont., 21-25 September, 
Oudtshoorn, South Africa, pp. 55-62. 
CLOETE SWP, VAN SCHALKWYK SJ & PHISTER AP, 1998b. Interrelationships between production traits 
of commercial slaughter ostriches. Proc. 2nd lnternat. Scient. Ratite Cont., 21-25 September, 
Oudtshoorn, South Africa, pp. 133-136. 
CLOETE S, BUNTER K, VAN SCHALKWYK SJ & BRAND Z, 1999a. Seleksiemoontlikhede by volstruise -
onlangse vordering. Animal Production Progress Report 1999/2000, pp. 338-341, Animal 
Production Division, Department of Agriculture, Western Cape. 
CLOETE SWP, BUNTER KL, VAN SCHALKWYK SJ & BRAND Z, 1999b. Selekteer vir swaarder volstruise. 
Animal Production Progress Report 1999/2000, pp. 342-343, Animal Production Division, 
Department of Agriculture, Western Cape. 
CLOETE SWP, VAN SCHALKWYK SJ & BUNTER KL, 2002. Progress in ostrich breeding research. Proc. 2nd 
World Ostrich Congr., September 2002, Warsaw, Poland, pp. 23-37. 
CREWS D, 1994. Animal sexuality. Scient. Amer.: 108-114. 
DAWSON A, DEEMING DC, DICK ACK & SHARP PJ, 1996. Plasma thyroxine concentrations in farmed 
ostriches in relation to age, body weight, and growth hormone. Gen. Comp. Endocrinol. 
103(3): 308-315. 
DEGEN AA, WEIL S, ROSENSTRAUCH A, KAM M & DAWSON A, 1994. Seasonal plasma levels of 
luteinizing and steroid hormones in male and female domestic ostriches ( Struthio came/us). 
Gen. Comp. Endocrinol. 93:21-27. 
DEVICHE P, SALDANHA CJ & SILVER R, 2000. Changes in brain gonadotropin-releasing hormone- and 
vasoactive intestinal polypeptide-like immunoreactivity accompanying reestablishment of 
photosensitivity in male dark-eyed juncos (Junco hyemalis). Gen. Comp. Endocrino/. 117: 8-
19. 
DITIAMI JP, 1981. Seasonal changes in the behaviour and plasma titers of various hormones in 
barheaded geese, Anser indicus. Z. Tierpsychol. 55: 289-324. 
DONHAM RS, 1979. Annual cycles of plasma LH and sex hormones in male and female mallards (Anas 
plathyrynchos). Biol. Reprod. 21: 1273-1285. 
84 
EL HALAWANI ME, YOUNGREN OM, ROZENBOIM I, Pins GR, SILSBY JL & PHILLIPS RE, 1995. 
Serotonergic stimulation of prolactin secretion is inhibited by vasoactive intestinal peptide 
immunoneutralization in the turkey. Gen. Comp. Endocrinol. 99 69-74. 
ETCHES RJ, 1995. Physiology of reproduction: the female. In: Poultry production, World Animal Series 
-volume 9, Ed.: P Hunton, Elsevier, Amsterdam, pp.221-241. 
FARNER DS & WINGFIELD JC, 1980. Reproductive endocrinology of birds. Ann. Rev. Physiol. 42 457-
472. 
FRAISSINET M, VARRIALE B, PIERANTONI R, CALIENDO MF, DI MATTEO L, BonONI L & MILONE M, 1987. 
Annual testicular activity in the gray partridge (Perdix perdix L.). Gen. Comp. Endocrinol. 
68:28-32. 
GAHR M, 2001. Distribution of sex steroid hormone receptors in the avian brain: functional implications 
for neural sex differences and sexual behaviours Microsc. Res. Tech. 55(1 ): 1-11. 
GILBERT AB, 1971. The endocrine ovary in reproduction. In Physiology and biochemistry of the 
domestic fowl-volume 3, pp. 1449-1468. Ed.: DJ Bell & BM Freeman, Academic Press, New 
York. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3, NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
GOLDSMITH AR & WILLIAMS DM, 1980. Incubation in mallards (Anas platyrhynchos): changes in plasma 
levels of prolactin and luteinizing hormone . .J  Endocrinol. 86(2): 371-379. 
GOMEZ Y, VELAZQUEZ PN, JUAREZ-OROPEZA MA & PEDERNERA E, 1998. Steroid metabolism in the 
granulosa and theca interna cells from preovulatory follicles of domestic hen (Gallus 
domesticus). Anim. Reprod. Sci 52(1): 81-91. 
HARGITAI C, FORGO V, MEZES M, Do THI DONG XUAN & PECZELY P, 1993. Seasonal and circadian 
fluctuation of plasma LH level and its change in the domestic goose as an effect of GnRH 
treatment. Acta Biol. Hung. 44(2-3): 255-268. 
HEATH JA, FREDERICK PC, EDWARDS TM & GUILLETIE LJ JR, 2003. Reproductive physiology of free-
living white ibises (Eudocimus a/bus) in the Florida Everglades. Gen. Comp. Endocrinol. 
133(1): 118-131. 
HOWARTH B, 1995. Physiology of reproduction: the male In: Poultry production, World Animal Series 
-volume 9, Ed.: P Hunton, Elsevier, Amsterdam, pp.243-270. 
JOYNER CJ, PEDDIE MJ & TAYLOR TG, 1987. The effect of age on egg production in the domestic hen. 
Gen. Comp. Endocrinol. 65(3): 331-336. 
KIM IS & YANG HH, 2001. Seasonal changes of testicular weight, sperm production, serum 
testosterone, and in vitro testosterone release in Korean ring-necked pheasants (Phasianus 
colchicus karpowt). J. Vet. Med. Sci. 63(2): 151-156. 
KOFUJI H, KANDA M & OISHI T, 1993 Breeding cycles and fecal gonadal steroids in the brown dipper 
Cine/us pallasii. Gen. Comp. Endocrino/. 91(2): 216-223. 
85 
KOIDE Y, PAPKOFF H & KAWAUCHI H, 1996 .. Complete amino acid sequences of follitropin and lutropin in 
the ostrich. Struthio came/us. Eur J. Biochem. 240: 262-267. 
LICHT P, PAPKOFF H, BONA-GALLO A & AGGARWAL B, 1983. Subunits of an avian (ostrich) follicle-
stimulating hormone and their hybridization with subunits of mammalian gonadotropins. Gen. 
Comp. Endocrinol. 51: 414-422. 
Liu H-K, LONG DW & BACON WL, 2001. Concentration change patterns of luteinizing hormone and 
progesterone and distribution of hierarchical follicles in normal and arrested laying turkey hens. 
Poult. Sci. 80: 1509-1518. 
MALECKI IA, MARTIN GB, O' MALLEY PJ, MEYER GT, TALBOT RT & SHARP PJ, 1998. Endocrine and 
testicular changes in a short-day seasonally breeding bird, the emu (Dromaius 
novaehol/andiae), in southwestern Australia. Anim. Reprod. Sci. 53(1-4): 143-155. 
MANEY DL, RICHARDSON RD & WINGFIELD JC, 1997. Central administration of chicken gonadotropin-
releasing hormone-II enhances courtship behaviour in a female sparrow. Harm. Behav. 32(1): 
11-18. 
MUSHI EZ, BINTA MG, CHABO RG, ISA JF & MODISA L, 1998. Serum biochemical values of farmed 
ostriches ( Struthio came/us) in Botswana. Onderstepoort J. Vet. Res. 65(3): 189-193. 
OTTINGER MA & MENCH JA, 1989. Reproductive behaviour in poultry: implications for artificial 
insemination technology. Brit. Pou/t. Sci. 30(2): 431-442. 
OWENS !PF & HARTLEY IR, 1998. Sexual dimorphism in birds: why are there so many different forms of 
dimorphism? Proc. R. Soc. Land. B 265: 397-407. 
PAPKOFF H, LICHT P, BONA-GALLO A, MACKENZIE OS, OELOFSEN W & OOSTHUIZEN MMJ, 1982. 
Biochemical and immunological characterization of pituitary hormones from the ostrich 
(Struthio came/us). Gen. Comp. Endocrino/. 48: 181-195. 
PASTER MB, 1991. Avian reproductive endocrinology Vet. Clin. North Am. Small Anim. Pract. 21 (6): 
1343 -1359. 
PENFOLD LM, WILDT DE, HERZOG TL, LYNCH W, WARE L, DERRICKSON SE & MONFORT SL, 2000. 
Seasonal patterns of LH, testosterone and semen quality in the Northern pintail duck (Anas 
acuta). Reprod. Fert. Dev. 12(3-4): 229-235. 
POTTER MA & CocKREM JF, 1992. Plasma levels of sex steroids in the North Island brown kiwi 
(Apteryx australis mantel!!) in relation to time of year and stages of breeding. Gen. Comp. 
Endocrinol. 87(3): 416-424. 
POWELL RC, JACH H, MILLAR R & KING JA, 1987. Identification of Gln8GnRH and His5Trp7Tyr8GnRH in 
the hypothalamus and extrahypothalamic brain of the ostrich Struthio came/us. Peptides 8: 
185-190. 
ROMDHANE BS, ROMDANE MN, MHIRI S, MILED BMA & KORTAS M, 2000. Biochemical and 
haematological blood parameters of the ostrich ( Struthio came/us) in Tunisian breeding. Rev. 
Med. Vet. 151 (3): 231-238. 
86 
SAILAJA R, KOTAK VC, SHARP PJ, SCHMEDEMANN R & HAASE E, 1988. Environmental, dietary, and 
hormonal factors in the regulation of seasonal breeding in free-living female Indian rose-ringed 
parakeets (Psittacu/a kramen). Harm. Behav. 22(4): 518-527. 
SAINT JALME M, WILLIAMS J, MICKAELIAN I & PAILLAT P, 1996. Seasonal variation of LH, sex steroids, 
body mass, molt, display, and laying in two subspecies of Houbara bustard, Chlamydotis 
undulata macqueenii and Chlamydotis undu/ata undulata, housed in outdoor cages under 
natural conditions. Gen. Comp. Endocrino!. 102(1): 102-112. 
SASANAMI T & MORI M, 1999. Effects of oestradiol-17beta and testosterone on progesterone 
production in the cultured granulosa cells of Japanese quail. Brit. Pou/t. Sci. 40(4): 536-540. 
SCHULTZ JD & SCHLINGER BA, 1999. Widespread accumulation of [3H]testosterone in the spinal cord of 
a wild bird with an elaborate courtship display. Neurobiol. 96( 18): 10428-10432. 
SHARP PJ, DUNN IC & TALBOT RT, 1987. Sex differences in the LH responses to chicken-LHRH-1 and -
II in the domestic fowl. J. Endocrinol. 115: 323-331. 
SKADHAUGE E & DAWSON A, 1999. Physiology. In: The ostrich - biology, production and health, pp. 
51-81. Ed.: DC Deeming, GABI Publishing, Wallingford, Oxon, United Kingdom. 
SPINU M, SPINU 0 & DEGEN AA, 1999. Haematological and immunological variables in a domesticated 
and wild subspecies of ostrich (Struthio came/us). Brit. Poult. Sci. 40(5): 613-618. 
STOKKAN KA, SHARP PJ & UNANDER S, 1986. THE annual breeding cycle of the high-arctic Svalbard 
ptarmigan (Lagopus mutus hyperboreus). Gen. Comp. Endocrinol. 61 (3): 446-451. 
TALBOT RT & SHARP PJ, 1994. A radioimmunoassay for recombinant-derived chicken prolactin suitable 
for the measurement in other avian species. Gen. Comp. Endocrino/. 96: 361-369. 
TALBOT RT, SHARP PJ, HARVEYS, WILLIAMS JB, DUNN IC, STERLING RJ & BAHR JM, 1988. Comparison 
of the fractionation and assay of domestic duck and fowl pituitary gonadotrophins. Brit. Poult. 
Sci. 21: 81-92. 
TANABE Y, YANO T & NAKAMURA T, 1983. Steroid hormone synthesis and secretion by testes, ovary, 
and adrenals of embryonic and post-embryonic ducks. Gen. Comp. Endocrino!. 49(1 ): 144-
153. 
VAN CLEEFF J, BLACKBERRY M & MARTIN GB, 1998. Hyperlipaemic emu serum is clarified for 
radioimmunoassay by polyethyleneglycol (PEG) precipitation. Austr. & NZ Soc. Comp. 
Physiol. Biochem., Perth, Australia. 
VAN CLEEFF J, PRENDERGAST E, BLACHE D & MARTIN GB, 1999. A critical number of eggs stimulates 
prolactin secretion in male emus during the breeding season. Austr. Soc. Reprod. Biof., 1999 
meeting, Melbourne, Australia. 
VAN HEERDEN J, DAUTH J, JARVIS MJF, KEFFEN RH, DENNY JEFM, DREYER MJ & KRIEK NPJ, 1985. 
Blood chemical and electrolyte concentrations in the ostrich Struthio came/us. J. S. Afr. Vet. 
Assoc. 56(2): 75-79. 
VAN SCHALKWYK SJ, 1998. Improvement of fertility and hatchability of artificially incubated ostrich eggs 
in the Little Karoo. M.Sc. thesis, Rhodes University, Grahamstown 
87 
VAN SCHALKWYK SJ, CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations for 
body weight and reproduction in commercial ostrich breeding pairs. Brit. Poult. Sci. 37: 953. 
962. 
VERBYLA AP, KENWARD BR & WELHAM SJ, 1999. The analysis of designed experiments and 
longitudinal data by using smoothing splines. Appl. Statist. 48(3): 269-311. 
VocKEL A, PROVE E & BAL THAZART J, 1990. Effects of castration and testosterone treatment on the 
activity of testosterone-metabolizing enzymes in the brain of male and female zebra finches. 
J. Neurobiol. 21(5); 808-825. 
WILLIAMS KM, BALCHE D, MALECKI IA, SHARP PJ, TRIGG TE, RIGBY ROG & MARTIN GB, 1998. Growth, 
sexual development and carcass composition in intact and surgically or hormonally 
gonadectomised male and female emus. Proc 2nd /nternat. Scient. Ratite Congr., 21-25 
September, Oudtshoorn, South Africa, pp. 75-80 
WILSON SC, 1978. Relationship between plasma concentration of luteinising hormone and intensity of 
lay in the domestic hen. Brit. Pou/t. Sci. 19(5): 643··650 
WINGFIELD JC, 1984. Environmental and endocrine control of reproduction in the song sparrow, 
Me/ospiza melodia. I. Temporal organization of the breeding cycle. Gen. Comp. Endocrinol. 
56(3): 406-416. 
88 
Chapter 7 
Preliminary results on the use of diagnostic 
ultrasonography as a management tool to quantify egg 
production potential in breeding ostrich ( Struthio 
came/us australis) females 1 
Abstract 
An ostrich breeding flock, joined as individual breeding pairs (n=136 pairs), was used to investigate the possibility of 
diagnostic ultrasonography as a method to predict the reproductive performance of ostrich females during a breeding 
season. Follicular activity was easily detected and quantified using diagnostic ultrasonography. One to eight 
follicles were recorded in 25% of the females when scanned at the beginning of the nine-month breeding season. 
At the end of the breeding season, one to three follicles were observed in 28.7% of the females. Females in which 
follicular activity was observed before paired came into production earlier than their contemporaries in which no 
follicles were observed, with the mean(± SE) number of days to the production of the first egg being 22.3 ± 12.5 and 
87.4 ± 7.2 days, respectively. Females in which follicular activity was observed with scanning at the beginning of the 
breeding season, produced on average 181% more eggs during the first month of the breeding season (P<0.01) 
when compared to females in which no follicular activity was observed (6.67 ± 0. 70 vs. 2.37 ± 0.41 eggs). Egg 
production over the first two months of breeding and over the entire breeding season were similarly affected (P < 
0.01), with the mean number of eggs produced over the first two months of the breeding season being 14. 7 ± 1.5 for 
females with observed follicular activity and 7.4 ± 0.9 eggs for females with no observed follicular activity Females 
in which follicular activity was observed at the end of the breeding season produced on average 108% more eggs 
(P<0.01) during the last month of the breeding season than females in which no follicular activity was observed (2. 77 
± 0.43 vs. 1.33 ± 0.27 eggs). There was a tendency (P=0.06) for egg production over the last two months to be 
similarly affected (6.10 ± 0.85 vs. 4.19 ± 0.54 eggs). No relationship with egg production over the entire breeding 
season was found for the end-of-the-breeding-season observations. Diagnostic ultrasonography can thus be used 
as a management tool to identify reproductively healthy ostrich females and also females with a higher egg 
production potential over a period of two months after or prior to assessment. Future studies should focus on the 
development of the technique to predict reproductive performance over entire breeding seasons for selection 
purposes. 
INTRODUCTION 
Egg production in commercially farmed ostriches were shown to be extremely variable (Van 
Schalkwyk et al, 1996; Cloete et al., 1998). Egg production over an 8-month breeding season was 
shown to vary from zero to more that 100 eggs per female. It is estimated that 70 - 80 % of ostrich 
eggs produced in South Africa originate from breeding flocks in free-ranging systems. Male and 
1 Published in J. S. Afr. Vet. Assoc.73(2): 48-52, April 2002 
89 
female ostriches are commonly kept in flocks with a size of 50-100 birds at a sex ratio of 6 males to 
10 females (Swart, 1986). Eggs are laid in communal nests and several females may visit and lay 
in a single nest. Commercially viable methods for the identification of eggs according to the female 
that produced them are presently not in place. It is therefore difficult to determine either egg 
production on a per female basis or the parentage of the eggs produced. It is, however, most likely 
that the same large variation in egg production observed under pair-breeding conditions will prevail 
under flock mating conditions. 
Breeder birds are fed a complete breeder diet during the breeding season, with the average daily 
feed consumption of ostrich females estimated at 2.5 - 3.0 kg feed (ca. 3% of live mass) per bird 
per day. It is evident that the maintenance of breeding birds is a major source of expenditure for 
commercial breeders and the maintenance of females with a poor egg production would be 
uneconomical. The repeatability of egg production in ostriches were shown to be fairly high (t "' 
0.45) (Van Schalkwyk et al., 1996; Cloete et al., 1998), meaning that female ostriches would tend 
to achieve similar rankings for egg production both within and across production seasons. It is 
likely that feed costs may be reduced substantially without a marked reduction in egg production if 
poor producers in breeding flocks can be identified and culled. No practical system for this purpose 
is in place at present. The success of diagnostic ultrasonography in the determination of 
reproduction success in for example domesticated farm species (Wani, 1981; Fowler and Wilkins, 
1982, 1984; Owens and Armstrong, 1985; Fukui et al., 1986; Grace et al., 1989; Kilgour, 1992; 
Knopf et al., 1989; Lambrechts and Pfister, 1999), llamas (Bourke et al., 1992) and camels 
(Skidmore et a/.1996) triggered the idea of using diagnostic ultrasonography as a tool to assess the 
reproductive fitness of the ostrich under commercial farming conditions. 
Against this background we investigated the feasibility of diagnostic ultrasonography for the 
quantification of egg production potential of adult breeding ostrich females. This paper presents 
preliminary findings with regard to the method as well as the relationship of the ultrasonograms 
with female egg production. 
MATERIALS AND METHODS 
Animals, management and recordings 
The ostrich breeding flock (n=136 breeding pairs) at the Little Karoo Agricultural Development 
Centre, Oudtshoorn, was used in the study. The flock, its origin and management are well 
documented by Van Schalkwyk (1998) and Van Schalkwyk et al. (1996). The entire breeding flock 
was subjected to a 3-month rest period, before paired in 0.25 ha camps. The breeding birds 
received a balanced breeder diet at 2.5kg/bird.day and had free access to clean drinking water. 
The breeding season commenced on 24 May 1999 (day 1) and ended on 29 February 2000 (day 
281 ). The ages of the females in this study ranged between 2 and 13 years. 
90 
All females were subjected to diagnostic ultrasonography at the beginning and end of the 
1999/2000 breeding season. A Pie Medical 100LC Vet scanner fitted with a 3.5/5.0 MHz dual 
frequency curved array transducer (Philips South Africa (Pty) Ltd.) was used for this purpose. 
Females were scanned at a frequency of 3.5 MHz and ultrasound coupling gel was used during 
each scanning procedure to ensure sufficient lubrication between the surface of the transducer and 
the body of a female. The transducer was placed on the ventral, non-feathered area, caudal to the 
right thigh of each female. During each scanning procedure, as many as possible follicles in the 
scanning field (i.e. a single plane image) were captured and stored on disc for later analysis. Each 
ultrasonogram was printed out at 100% of the size of the on-screen version and was used to count 
the number of follicles and to measure follicle diameter. The presence and number of discernable 
follicles were noted for each individual ultrasonogram, and the average diameter (in millimeters) of 
the observed follicles was measured in each case. 
Eggs produced by each female were collected daily during the afternoons and were recorded on an 
individual basis for each fE;male. 
Statistical analysis 
Monthly egg production was averaged for all breeding females. Average egg production of females 
with different numbers of follicles discernable at the beginning of the breeding season were 
calculated for the first month of production, the first two months of production, and the entire nine 
month breeding period. One-way analysis of variance procedures were used for this purpose 
(Snedecor and Cochran, 1967). The average subsequent egg production of females where ;:: 1 
follicle was observed was also compared to that of females where no follicles could be discerned. 
The same basic procedure was followed for relating ultrasonograms at the end of the breeding 
season to egg production during the last month of breeding, the last two months of breeding and 
during the entire breeding season. 
RESULTS 
Follicles (or ova) of different sizes and stages of maturity were detected in the ostrich females 
during scanning. The different views obtained during the imaging are depicted in Figures 1 and 2 
that represent ultrasonograms of ovaries with an extensive degree of follicular development. A 
developing follicle appears as a round anechoic image with a slightly more hyperechoic central 
area in the middle of the follicle. This hyperechoic area appears to be the developing follicle's 
attachment to the ovary. A normal ostrich ovary will contain a number of follicles of different sizes, 
ranging from 12.6mm to 120 mm in diameter. The extent of follicular development of an ovary 
determined the number of follicles that could be captured in each field. 
91 
Figure 1. An ultrasonogram of the ovary of a breeding ostrich female, with five follicles distinguishable in the 
ultrasonogram. The largest follicle has a diameter of 63mm. 
Figure 2. An ultrasonogram of the ovary of a breeding ostrich female, with five follicles distinguishable in the 
ultrasonogram. The largest follicle has a diameter of 58mm and the smallest follicle a diameter of 18mm. 
One or more follicles could be discerned in 34 females (25%) at the beginning of the breeding 
season and in 39 females (28.7%) at the end of breeding season. At the beginning of the breeding 
season, females of three years and younger generally showed considerably less follicular 
development than females aged 4 years and older, with no follicles observed in 2-year old females 
(Figure 3). The number of follicles observed was unrelated to female age when assessed at the 
end of the breeding season. 
92 
9 
...  • y =0  2649x - 04361 8 R2 =0 .1356 7 . ... • 
~6 
.s • 
I/) 5 ... 
Q) • • • 
:§ 
4 
0 • • 
u. 
3 • • • • • • 
2 •· 
• 
0 
0 2 4 6 8 10 12 
Female age (years) 
Figure 3. Relationship between female age and number of follicles noted with diagnostic imaging at the 
beginning of the 1999/2000 breeding season. 
Egg production over the first month and the first two months generally was greater (P<O. 05) in 
females where two or more follicles were observed during scanning at the beginning of the 
breeding season. When females where at least one follicle was observed (Group A, n=34) were 
pooled, this group deviated from those where no follicles were observed (Group B, n=102) (Table 
1) . Group A came into production earlier than Group B, with the mean (± SE) period in days from 
the commencement of the breeding season to the production of the first egg being 22.3 ± 12.5 and 
87.4 ± 7.2 days, respectively. Group A produced on average 181 % more eggs during the first 
month of breeding (mean ± SE) than Group B (P ::; 0.01; 6.7 ± 0.7 vs. 2.4 ± 0.4 eggs). Egg 
production over the first two months of breeding and over the entire breeding season were similarly 
affected (P<0.01 ). The mean number of eggs produced over the first two months of the breeding 
season were 14.7 ± 1.5 for Group A, compared to 74 ± 0.9 eggs produced by Group B. 
Corresponding mean egg production was, respectively, 48.5 ± 4.4 and 35.3 ± 2.5 eggs over the 
breeding period of 281 days. 
Table 1. Mean (± SE) egg production during the first month, first 2 months, and for the entire breeding 
season, classified according to the number of follicles observed at the beginning of the breeding 
season. 
NUMBER OF NUMBER OF EGG PRODUCTION 
,_____ 
FOLLICLES OBSERVATIONS Firs t month First 2 months Entire season 
OBSERVED 
0 102 2. 4 ±04 7.4 ± 0.9 35.1±2.5 
1 4 2. 8 ± 2.0 7.5 ± 4.3 41.3 ± 12.7 
2 4 7. 5 ± 2.0 16.5 ± 4.3 52.0±12.7 
3 12 6. 3 ± 1 2 14.3 ± 2.5 43.4 ± 7.4 
4 3 10. 0±24 23.3 ± 5.0 66.0 ± 14.7 
5 5 9. 0 ± 1.8 16.6 ± 3.9 42.8±114 
6+ 6 6._ 0_ ±_ _1_.7 ---·--·-- ____ 13.2 ± 3.5 
53.3 ± 10.4 
,___ _______________ ----------
93 
The number of eggs produced during the last month of the breeding season, during the last two 
months of the breeding season and during the entire breeding season in relation to the number of 
follicles observed at the end of the breeding season are presented in Table 2. Fewer follicles, but 
with a larger diameter, were observed at the end of the breeding season. Females where follicular 
activity was observed at the end of the breeding season produced on average 108% (P<0.01) 
more eggs during the last month of the breeding season than females where no follicular activity 
was observed (2.8 ± 0.4 vs. 1.3 J: 0.3 eggs). There was a tendency (P=0.06) for egg production 
over the last two months to be similarly affected (6.1 ± 0.9 vs 4.2 ± 0.5 eggs). Egg production over 
the entire season was unrelated (P 2': 0.05) to the number of follicles observed at the end of the 
breeding season. 
Table 2. Mean (± SE) egg production during the last month, the last two months, and during the entire 
breeding season, when classified according to the number of follicles observed at the end of the 
breeding season 
NUMBER OF FOLLICLES NUMBER OF EGG PRODUCTION 
OBSERVED OSTRICHES Last month 1-Lost month• Entire season 2 
0 97 1.3±0.3 4.2 ± 0.5 37.4 ± 2.6 
1 18 3.3 ± 0 6 7 8 ± 1.3 40.7 ± 6.1 
2 8 2.4 ± 1.0 4.8 ± 1.9 44.8 ± 9.2 
3+ 13 2.2 ± 0.7 ___ J_ __ 4.5±1.5 38.0 ± 7.2 
~ -·-~· 
The number of follicles observed was negatively correlated (P<0.01) with mean follicle diameter, 
with the correlation between number of follicles observed and mean follicle diameter amounting to 
-0.69 at the beginning of the breeding season and -0.72 at the end of the breeding season. The 
phenotypic correlation between the number of follicles observed at the beginning of the breeding 
season and that observed at the end of the breeding season was relatively low (r=0.20; P<0.05). 
DISCUSSION 
Follicle activity in ostrich females was clearly discernable with diagnostic ultrasonography. During 
imaging, it is impossible to capture all possible follicles in one image field and in each case, as 
many follicles as possible were captured and stored on disc for later analysis. Because only a 
limited number of follicles (i.e. less than what actually could be observed in each female) could be 
captured in each ultrasonogram, each female was given a score (see Chapter 8) in terms of the 
extent of ovary development. These data, however, are not presented in this paper but will be 
discussed in a subsequent paper. In this study, females were scanned on the right side only. 
Although the ovary is located in the left half of the female's body, better ultrasonograms have been 
obtained in most cases from the right hand side. In a follow-up phase of this study, females will be 
scanned on both sides to determine the correlation between ultrasonograms obtained from the left 
and right sides of a female. 
94 
The number of follicles observed was related to the egg production of individual females over the 
short term, i.e. within one or two months of assessment at the beginning of the breeding season. 
Selection of females based on the visibility of one or more follicles at the onset of the breeding 
season would consequently have resulted in marked improvements in egg production over the first 
two months following assessment. A similar conclusion was reached with regard to scanning at the 
end of the breeding season. Although lower levels of egg production were attained during the last 
month of the breeding season, it was still possible to identify females with higher levels of egg 
production over the last month prior to scanning. The ability to accurately identify high producing 
individuals over a longer term (> two months), however, was compromised to an extent This can 
be ascribed to the fact that egg production of ostrich females start to decline at the specific time of 
year, i.e. end of January, when the end of season scanning was performed. 
This study is in agreement with previous reports that egg production in ostriches is highly variable 
(Deeming, 1996; Van Schalkwyk et al., 1996; Cloete el al., 1998). The wide variation in egg 
production within female age groups is clearly illustrated in Figure 3. The fact that no follicles were 
observed in the 2-year old females that were in their first breeding season, forced the regression of 
number of follicles scanned on female age in a positive direction. The absence of follicular activity 
in two-year old females at the beginning of the breeding season and their consistently lower egg 
production can probably be ascribed to the age at which females become reproductively mature. 
Culling of ostrich females solely on egg production and I or ultrasonograms is therefore likely to 
discriminate against young females under circumstances where flock breeding is practiced, 
especially when their age is unknown. Previous results have shown that egg production of ostrich 
females at a young age predicted subsequent egg production fairly well (Van Schalkwyk et al., 
1996) and that egg production generally increased with age to reach a maximum at approximately 
nine years of age (Cloete et al., 1998). It is advisable to take this information into account under 
conditions where ostrich females of unknown age are scanned for egg production. 
The correlation between ultrasonograms obtained at the beginning and end of the breeding season 
was relatively low (r=0.20). Factors contributing to this observation is that older, higher producing 
birds entered a natural rest period at the end of the breeding season, and the 2-year old females 
only came into production towards the end of the breeding season. 
Two aspects remain as cause for concern with regard to the practical application of the 
ultrasonography technique in ostriches. First, in the present study only a small percentage of 
females that were scanned actually showed follicular activity at assessment. It is clearly not viable 
to retain such a minority of the total number of females available when breeding under commercial 
conditions. The only possible exception may be a breeding operation with an almost unlimited 
access to resources in terms of female breeding birds. Given the high cost of growing and finishing 
breeder birds, such conditions are unlikely. Second, the ability to predict egg production over the 
entire breeding season was limited, especially when the females were scanned at the end of the 
breedi;ig season. Based on the results in the present study, diagnostic ultrasonography can be 
95 
used to identify reproductively healthy females but cannot be used as a selection tool to cull 
females at the end of a breeding season. 
In order for the practice of diagnostic ultrasonography to become a viable proposition in the 
management of commercial ostrich breeding flocks, these issues need to be addressed. It is also 
necessary to ensure that a greater proportion of the breeding females is in a stage of ovulation 
when scanning is performed. The objective should thus be to find ways of strategically stimulating 
ostrich females to higher levels of sexual activity at the beginning of a breeding season. This could 
be done by for example, delaying the onset of the breeding season, when natural sexual activity 
ought to be higher. Alternatively, other practices like teasing and flushing that were found to 
enhance sexual activity in other livestock species (Pearce and Oldham, 1984; Scaramuzzi and 
Campbell, 1990; Signore!, 1990) and ostriches (Chapter 3) can also be considered. These 
practices are also widely advocated in the ostrich industry (Swart, 1986; Lambrechts et al, 1998) 
CONCLUSIONS 
Diagnostic ultrasonography can be used to identify reproductively healthy females and can be used 
as a diagnostic tool to identify potentially high producing females at the beginning of a breedr ng 
season. Certain issues, however, needs to be addressed before it can be applied on a large scale 
in commercial ostrich enterprises. In the present study, only a small proportion of females 
displayed follicular activity when scanned, making these results less suitable for selection purposes 
under most conditions where only small numbers of breeding females are available. It was also 
complicated by the fact that younger females were less likely to show follicle development. 
Farming practices to maximize and synchronize ovulation and egg laying cycles at tile 
commencement of the breeding season should also be investigated. The investigation of teasing 
and flushing effects on ostrich egg production could not only provide clarity with regard to tile 
importance of these practices to efficient production systems, but could also prove to be 
indispensable for the practical application of diagnostic ultrasonography in the broader industry It 
is also necessary to address the need to improve long term predictions in terms of egg production 
for ostrich females, based on a single scanning measurement. These aspects should be 
addressed in the further development and adaptation of diagnostic ultrasonography as a 
management tool to improve reproduction efficiency in commercial ostrich breeding flocks. 
REFERENCES 
BOURKE DA, ADAM CL & KYLE CE, 1992. Ultrasonography as an aid to controlled breeding in the 
llama (Lama glama). Vet. Rec. 130(19): 424-428 
96 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998. Ostrich breeding - progress towards a 
scientifically based strategy. Proc. 2nd Internal. Ratite Congr, Oudtshoorn, South Africa, 
21-25 September 1998, pp. 55-62. 
DEEMING DC, 1996. Production, fertility and hatchability of ostrich (Struthio came/us) eggs on a 
farm in the United Kingdom. Anim. Sci. 63: 329-336. 
FOWLER DG & WILKINS JF, 1982. The identification of twin bearing ewes with real time ultrasound 
and its potential use in sheep and wool production. Wool Technology and Sheep Breeding 
(September-October): 111-117. 
FOWLER DG & WILKINS JF, 1984. Diagnosis of pregnancy and number of foetuses in sheep by real-
time ultrasonic imaging. I. Effects of number of foetuses, stage of gestation, operator and 
breed of ewe on accuracy of diagnosis. Livestock Prod. Sci. 11:437-450. 
FUKUI Y, KOBAYASHI M, TSUBAKI M, SHIMODA K & ONO H, 1986. Comparisons of two ultrasonic 
methods for multiple pregnancy diagnosis in sheep and indicators of multiple pregnancy 
ewes in the blood. Anim. Prod. Sci. 11: 25-33. 
GRACE ND, BEACH AD, QUINLIVAN TD & WARD B, 1989. Multiple pregnancy diagnosis of ewes 
using real time ultrasonic body scanner and video-fluoroscopy systems. Proc. NZ Soc. 
Anim. Prod. 49: 107-111. 
KILGOUR RJ, 1992. Lambing potential and mortality in Merino sheep as ascertained by 
ultrasonography. Austr. J Exp. Agric. 32: 311-313. 
KNOPF L, KASTELIC JP, SCHALLENBERGER E & GINTHER OJ, 1989. Ovarian follicular dynamics in 
heifers: test of two-wave hypothesis by ultrasonically monitoring individual females. 
Domest. Anim. Endocrinol. 6(2): 111-119. 
LAMBRECHTS H, HUCHZERMEYER FW & SWART D, 1998. Integrating ostrich behaviour in commercial 
farming systems. Proc. 2nd lnternat. Ratite Congr., Oudtshoorn, South Africa, 21-25 
September, pp. 167-169. 
LAMBRECHTS H & PFISTER AP, 1999. The use of real-time linear ultrasonography to predict egg 
production in female ostriches (Struthio came/us). Proc. 3fh Nat. Congr. S. Afr. Soc. 
Anim. Sci., Pretoria, 27-29 July 1999, p.31. 
OWENS JL & ARMSTRONG JR, 1985. Diagnosis of foetal number in prolific sheep. Proc. NZ Soc. 
Anim. Prod. 45: 155-157. 
PEARCE OT & OLDHAM CM, 1984. The ram effect, its mechanism and application in the 
management of sheep. In: Reproduction in sheep, pp. 26-34. Ed.'s: DR Lindsay and OT 
Pearce, Australian Academy of Science, in conjunction with the Australian Wool 
Corporation. Canberra, Australia. 
SCARAMUZZI RJ & CAMPBELL BK, 1990. Physiological regulation of ovulation rate in the ewe: A new 
look at an old problem. In: Reproductive Physiology of Merino sheep - Concepts and 
Consequences, pp. 71-84. Ed.'s: CM Oldham, GB Martin and IW Purvis IW, School of 
Agriculture (Animal Science), The University of Western Australia, Nedlands 6009, Perth. 
SIGNORET JP, 1990. The influence of the ram effect on the breeding activity of ewes and its 
underlying physiology. In: Reproductive Physiology of Merino sheep - Concepts and 
97 
Consequences, pp. 59-70. Ed.'s: CM Oldham, GB Martin and IW Purvis IW, School of 
Agriculture (Animal Science), The University of Western Australia, Nedi ands 6009, Perth. 
SKIDMORE JA, BILLAH M & ALLEN WR, 1996. The ovarian follicular wave pattern and induction of 
ovulation in the mated and non-mated one-humped camel (Came/us dromedarius). 
Reprod. Fert. Dev. 106(2): 185-192. 
SNEDECOR GW & COCHRAN WG, 1967 Statistical methods - 6th edition. Iowa State University, 
USA. 
SWART D, 1986. Nuutste bevindinge ten opsigte van broeitegnieke by volstruise. Pamflet, 
Volstruisboeredag, Oudtshoorn. 
VAN SCHALKWYK SJ, 1998. Improvement of fertility and hatchability of artificially incubated ostrich 
eggs in the Little Karoo. M.Sc. dissertation, Rhodes University, Grahamstown, South 
Africa. 
VAN SCHALKWYK SJ' BRAND Z, CLOETE SWP & BROWN CR, 1999. The effects of time of egg 
collection and pre-incubation treatment on blastoderm development and embryonic 
mortality in ostrich embryos. S. Afr. J. Anim. Sci. 29: 154-163. 
VAN SCHALKWYK SJ, CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations 
for body weight and reproduction in commercial ostrich breeding pairs. Brit. Pou/t. Sci. 37: 
953-962. 
WANI GM, 1981. Ultrasonic pregnancy diagnosis in sheep and goats - A review. World Rev. Anim. 
Prod. Vol. XVIII, no. 4 (October-December): 43-48. 
98 
Chapter 8 
Genetic variability of ultrasound scanning parameters 
and egg production traits in breeding ostrich ( Struthio 
came/us var. australis) females1 
Abstract 
The use of ultrasound scanning as a management tool to identify reproductively healthy ostrich females and to 
determine the genetic variability of ultrasound parameters was investigated during four consecutive breeding 
seasons. A total of 1334 records was analysed, with follicle diameter only measured in 900 cases where follicular 
activity was observed. Total egg production records were analysed in 491 cases where scanning was performed 
before the onset and at the end of the breeding seasons. Heritabilities were low, but significant (P:o;0.05) in the case 
of follicular classification score (0.07) and number of follicles observed (0.06), and increased in magnitude as a 
breeding season progressed. Overall egg production and ultrasound recordings at the beginning of a breeding 
season were highly correlated (0.41±0.06), suggesting that females may be selected on partial production records 
for use in ostrich breeding programmes. Ultrasound parameters were closely related to short-term egg production 
on a genetic level, indicating the potential of the technique to identify females that start producing earlier in a 
breeding season. Duration of lay after commencement of a breeding season was also strongly associated with 
reproductive success. Results also indicate that age of ostrich females needs to be considered when performing 
ultrasound scanning when breeding selection is based on phenotypic data alone to prevent discrimination against 
young breeding female ostriches. 
INTRODUCTION 
Commercial breeding systems for ostriches can range from extensive free-range flock mating to 
intensive pair or trio breeding systems Ostrich farming in South Africa is mostly based on open 
range flock mating systems, as well as breeding pairs and trios, with almost 70% of ostrich eggs 
produced in flock breeding systems. In flock breeding systems, more than one female contributes 
to a single nest and females may also lay in more than one nest (Jarvis et al., 1985; Lambrechts et 
al., 2002a). This behaviour of ostrich females complicates the identification of females with low or 
no egg production. 
Commercial ostrich breeding systems are characterised by low reproductive performance, with egg 
production varying between 0 and 100 eggs produced per female in an 8-month breeding season 
(Van Schalkwyk et al., 1996; Van Schalkwyk, 1998; Deeming and ,a,r, 1999; Cloete et al., 2002a). 
1 Submitted for publication in Livestock Production Science, with co--authors D Swart, SWP Cloete, JPC Greyling and AP 
Pfister 
99 
Low average production performance of ostrich females and the extreme variation in egg 
production is major constraints in the reproduction efficiency of ostrich breeding systems. Egg 
production in ostriches has been found to be highly repeatable (t ~ 0.45), implying that females with 
a certain egg production ranking would tend to achieve similar rankings in subsequent breeding 
seasons (Cloete et al., 1998). Feeding costs of breeding ostriches contribute up to almost 80% of 
input costs, and the identification and culling of poor or low producers has the potential to reduce 
feeding costs substantially in a commercial ostrich breeding system (Brand, 2002). 
The ostrich female's reproductive tract consists of the single left ovary and accompanying left 
oviduct, and lies ventral to the cranial part of the left kidney and dorsal to the abdominal air sacs 
(Duerden, 1912; Bezuidenhout, 1986; Fowler, 1991; Hicks, 1993; Soley and Groenewald, 1999). 
As in poultry, the ostrich female's reproductive tract is completely internalised, making visual 
evaluations of the reproductive status of the female difficult (Juli, 1952; Gilbert, 1971; Gilbert, 1979; 
Solomon, 1983). The reproductive status of the female or the extent of follicular development on 
the ovary can be determined if a female is killed, and thus only lifelong egg production can be 
inferred. This, together with the cost of sacrificing breeding females, necessitated the development 
of a commercially viable technique for the identification of poor or non-producing ostrich females 
and the determination of the reproductive potential of ostrich females for selection purposes. 
The success of ultrasound scanning as a tool to assess reproductive fitness in humans and 
animals is well documented (Allan and Meredith, 1981 (dogs); Fowler and Wilkins, 1982 (sheep); 
Sengoku et al., 1985 (humans); Bourke et al., 1992 (llamas); Boyd, 1995 (cattle); Skidmore et al., 
1996 (camels); Casares et al., 1997 (tortoises); Krautwald-Junghans et al., 1998; Shaffer et al., 
1998 (rhinoceros); Hildebrandt et al., 2000). In domestic chickens, ultrasound has been used 
successfully to identify individual males and females with acceptable testicular development and 
well-formed, normal follicular hierarchies (Kirby et al., 2001; Melnychuk et al., 2002). In ostriches, 
Lambrechts and Pfister (1999) and Lambrechts et al. (2002a) investigated the potential of 
ultrasound scanning as a management tool to predict the reproductive performance of ostrich 
females. Ultrasound scanning could be used successfully to identify reproductively healthy 
females, as well as females with a higher egg production during the first two months after or prior to 
scanning. Bronneberg and Taverne's (2003) findings supported the potential of the technique as a 
management tool to identify actively producing ostrich females. 
The application of genetic selection as a tool to improve reproduction and production performance 
has been well illustrated in livestock industries, with the poultry industry mostly used as a model to 
demonstrate the possible benefits of selection in the ostrich industry. The ostrich industry, 
however, has a short history in terms of selection. Literature reports on genetic parameters for 
ostriches are scarce, and are in general agreement with genetic correlations and selection 
responses reported for poultry (Petitte and Davis, 1999; Van Schalkwyk et al., 1995; Van 
Schalkwyk et al., 1996; Horbaficzuk and Sales, 1999; Cloete et al., 1998; 2002a; 2002b; Bunter, 
2002). The use of poultry genetic parameters is, however, not ideal, given the basic physiological 
100 
differences between poultry and ostriches and the different management conditions under which 
these species are maintained. The long generation interval in ostriches and the age at which 
sexual maturity is attained suggest that genetic progress will be slower than that reported for 
commercial poultry species. Large numbers of breeding birds with unknown pedigree and lack of 
production records infer the hazard of inbreeding depressions (Hicks-Aldredge, 1993; Badley, 
1997; Deeming and Ar, 1999). The determination of genetic effects in ostriches is complicated by 
the fact that paternal and maternal effects are confounded. The fact that the same male 1s often 
mated repeatedly with the same female under the same management conditions also hampers the 
partitioning of the direct and maternal genetic effects (Cloete et al., 1998; Bunter, 2002). 
When the successes that have been achieved in poultry are considered, it is evident that there is 
considerable scope for improvement of egg production in commercial ostrich breeding programs 
(Nestor et al., 1969; Gowe and Fairfull, 1985; Poggenpoel et al., 1996). Cloete et al. (2002a) 
recently reviewed the potential role of genetic improvement in the ostrich industry. The review 
reported estimates for several reproduction-related parameters, but no studies have yet been 
conducted on the genetic variability of ultrasound scanning parameters. The aim of this study was 
therefore to establish genetic and environmental estimates for ultrasound scanning parameters for 
ostrich females that may assist the commercial ostrich farmer in making sound selection decisions 
MATERIAL AND METHODS 
Experimental animals and management 
An ostrich breeding flock (n=136 breeding pairs) maintained at the Little Karoo Agricultural 
Development Centre outside Oudtshoorn, South Africa, were used in the study. The management 
of the breeding flock is well documented (Van Schalkvvyk et al., 1996; Bunter and Graser, 2000). 
The ages of females used in the study ranged between 2 and 13 years. Data were recorded during 
four consecutive (1999/2000, 2000/2001, 2001 /2002 and 2002/2003) breeding seasons. 
The breeding birds were maintained in single-sex flocks during the 4-month rest period preceding 
breeding (approximately February to May). Two weeks prior to being paired as breeding pairs in 
0.25ha camps, males and females were placed in adjoining camps (i.e. teasing) and were fed a 
flushing diet (Lambrechts et al., 2002b). Females were introduced to their breeding camps before 
the males. 
During the 1999/2000 and 2000/2001 breeding seasons, 90 of the 136 breeding pairs were 
subjected to a feeding trial that investigated the influence of different energy and protein levels on 
the production of the males and females. During the 1999/2000 breeding season, the three energy 
levels were 8.5 MJ ME/kg DM, 9 5MJ ME/kg OM and 10.5MJ ME/kg OM. The three protein levels 
wens s13.5%, 15.0% and 16.5%. During the 2000/2001 breeding season, the energy levels were 
101 
7.5MJ ME/kg OM, 8.5MJ ME/kg OM and 9.5MJ ME/kg OM The tt1ree protein levels were 10.5%, 
12.0% and 13.5%. 
During the 2000/2001 and 2001 /2002 breeding seasons, 63 of the 136 breeding pairs were 
subjected to a mid-season rest period of 6 weeks During this rest period , the males and females 
were visually separated and placed onto a maintenance diet for a period of four weeks. Two 
weeks prior to being joined again, the males and females were subjected to flushing and teasing, 
as described above. 
During the 2002/2003 breeding season the breeding flock were divided into five treatment groups. 
Group 1 (n=25 breeding pairs) was not subjected to a rest period . Breeding pairs in this group 
were paired immediately after being treated for external and internal parasites, and vaccinated for 
Newcastle Disease. They were separated on 8 April 2003. Groups 2 (n=25 breeding pairs) and 3 
(n=25 breeding pairs), after being subjected to flushing and teasing , were joined on 1 May 2002 
and separated on 19 February 2003 . Group 3 was subjected to a series of simulated production 
and rest periods. The duration of each production and rest period was approximately 10 weeks 
and six weeks, respectively. During each rest period , males and females were separated visually 
and placed on a maintenance diet. During the last 2 weeks before being joined again, males and 
females were subjected to flushing and teasing, as described above. Groups 4 (n=25 breeding 
pairs) and 5 (n=25 breeding pairs) , after being subjected to flushing and teasing, were joined on 1 
July 2002 and separated on 8 April 2003. Group 5 was subjected to a series of production and 
rests cycles, with treatment during each rest period the same as for Group 3. 
Diagnostic imaging 
A Philips Pie Medical Scanner 100LC (Pie Medical Equipment BV, Maastricht, The Netherlands), 
fitted with a 3.5/5.0 MHz dual frequency curved array transducer, was used to obtain the diagnostic 
images. Females were scanned at a frequency of 3.5 MHz and a hypoallergenic water-based 
ultrasonic coupling gel (Ultra/Phonic Conductivity Gel, Pharmaceutical Innovations Inc., Newark, 
New Jersey, 07114, USA) was used to ensure sufficient contact between the transducer surface 
and the body of the female . The transducer was placed on the ventral non-feathered area, caudal 
to the right thigh of a female (Lambrechts and Pfister, 1999). During each evaluation , as many as 
possible follicles were captured in each scanning field, i.e single plane image (Figure 1), and stored 
on computer for later analysis . Each sonogram obtained was used to calculate the number of 
follicles and to measure follicle diameter (Lambrechts et al., 2002a). 
102 
Plate 1. Sonograms obtained during ultrasound scanning of ostrich females 
As it was impossible to capture all follicles in a given sonogram, each female was assigned a 
follicular classification score (FCS), depending on the number of follicles observed during each 
examination. The follicle classification scores assigned to each female are set out in Table 1. 
Table 1. Follicle classification scores assigned during ultrasonic examination of breeding ostrich females 
Number of follicles observed during examination Follicle classification score 
1-----------------------1------------------~--
No  follicles observed 0 
1 to 3 follicles observed 
4 to 6 follicles observed 2 
7+ follicles observed 3 
------------·---~------ ------------~--
Ultrasonograms were obtained at the beginning and end of each respective breeding season 
(Lambrechts et al, 2002a). In addition to the pre- and post-season scanning, approximately half of 
the breeding flock's females were subjected to scanning before and after a forced mid-season rest 
during the 2000/2001, 2001/2002 and 2002/2003 breeding seasons During the 2002/2003 
breeding season, the females used in the study were subjected to ultrasound scanning on both the 
left and right sides to determine the correlation between the FCS and the number of follicles 
observed in each sonogram. 
Data recorded 
Eggs laid by the breeding females were collected daily during the afternoons and recorded for each 
female throughout each breeding season. Egg collection, handling, disinfecting and incubation 
procedures were performed as documented by Van Schalkwyk (1998). The number of eggs 
produced were expressed as egg production percentage (EPP, Van Schalkwyk et al., 1996) which 
were calculated using the following equation: 
EPP= [(number of eggs produced)/(O 5Xnumber of production days)) X 100 
103 
Statistical analysis 
The ASREML program (Gilmour et al., 1999) was used for the estimation of the fixed effects, and 
to derive variance components for the respective follicle and egg production parameters in 
univariate analyses. Although fixed effects of treatments within production years (as defined 
previously) were included in the model of analysis, these effects are not presented or discussed in 
detail because they formed part of separate studies, to be reported elsewhere. Initially, all 
available scanning and egg production records were incorporated (numbering 1334 observations in 
total). Recordings made at the onset and end of each mating season were analysed separately, 
with these data sets consisting of 491 records each. Total egg production during the breeding 
season was only considered in analyses involving scanning at the beginning or end of the mating 
season. 
The analysis of discrete data such as follicle classification score and the number of follicles 
observed with parametric methods used in the present study are not optimal, as was outlined by 
Purvis and Hillard (1997). The close approximation of outcomes from linear models to that derived 
from non-linear methods resulted in recommendations that the former methods could be employed 
until alternative software packages become available (Jorgensen, 1994; Brien et al., 2002). Fixed 
effects that were considered in all the analyses included year (1999-2002) and age of the female. 
When scanning results at the beginning of the season were incorporated, teasing and flushing 
treatments were also considered during the 2000/2001 and 2001/2002 breeding seasons. Random 
terms were then added to the operational model, resulting in the following models for analyses (in 
matrix notation): 
y = Xb + Z1a + e (1) 
y = Xb + Z2Ctema1e + e (2) 
y = Xb + Z1a + Z2Ctema1e + e (3) 
In these models, y was a vector of observations for female follicle or egg production parameters; b, 
a and Ctemaie were vectors of fixed effects, direct genetic effects and female permanent 
environmental effects; X, Z1 and Z2 were the corresponding incidence matrices relating the 
respective effects to y, and e the vector of residuals. 
It was assumed that: 
V(a) =Aria; V(CpE) = l021ema1e; V(e) = 1028 , 
where A represents the numerator relationship matrix, I an identity matrix, 02a . cr21emaie. and 0
2 
e the 
direct genetic variance, female permanent environmental variance, and environmental (residual) 
variance, respectively. All analyses included the full pedigree file, consisting of 578 individuals, the 
progeny of 129 males and 127 females. Log likelihood tests were conducted to determine the most 
suitable model for each parameter in univariate analyses, and subsequently, two-trait animal 
models were fitted (Snyman et al., 1996). These analyses allowed the calculation of all relevant 
104 
correlations between parameters, together with their appropriate standard errors. The Analysis 
Tool pack of the Microsoft Excel rM package was used to calculate the correlation between the 
number of follicles observed on both flanks and the assigned FCS. 
RESULTS 
A total of 1334 records were analyzed, with follicle diameter (FD) measured in only 900 cases 
where at least one follicle was observed. Total egg production records were analyzed in only 491 
evaluations where scanning was performed at the beginning and end of a breeding season. The 
fixed effect of year was significant (Ps0.05) in some instances. It is, however, not tabulated for this 
report as differences between year means resulted from various interacting management and 
husbandry practices, with the prevailing climate also possibly exerting an influence. 
The number of follicles (FN) observed ranged between O and 9, with an average of 1.92±1 l7 
(Table 2). Follicle diameter measured ranged between 2.0 and 120mm, with an average of 
45.3±19.2mm All egg production (EP) parameters were highly variable, as suggested by the high 
coefficients of variation and extreme ranges. Standard deviations exceeding the mean and 
coefficients of variation exceeding 50%, were commonly found. 
Table 2. Number of observations, means and standard deviations (SD), and ranges for the respective 
ultrasound parameters and egg production traits for breeding ostrich females 
Parameter-·-------- Numb-w -- Mean (SD) ----·---CV(%) Range 
roliic18 c1assificatTor;--·-~----- -~-- 1334 --- - --To7 (O 91) __ _ --·--··----- -------85 1 0-3 
Number of follicles 1334 1.92 (1 l7) 92.2 0-9 
Follicle diameter (mm) 900 45.3 (19.2) 424 2.0 - 120.0 
EP % over 14 days (EPP14d) 1334 23.9 (32.8) 137.2 0-133.3 
EP % over 1 month (EPP1m) 1334 31.6 (32.8) 103.8 0-115.2 
EP % over 2 months (EPP2m) 1334 334 (29.6) 88.6 0 - 109.7 
EP% for entire breeding season (EPP101a1) 491 43.1 (24.0) 55.7 0-109 
Sonograms obtained on the right flanks of ostrich females provided a better indication of the 
follicular development in the ovary, with correlation of the follicle classification score (FCS) with the 
sonograms obtained on the left and right flanks being 0.43 and 0.90, respectively (Table 3). 
105 
Table 3. The correlation of follicle classification score (FCS) with the number of follicles observed on the left 
(L) and right (R) flanks of ostrich females, respectively; and of the number of follicles observed on 
the left flanks (FN-L) with number of follicles observed on the right flanks during the 2002/2003 
breeding season. 
- -
FCS L R L+R FN-L 
- - --
Number of follicles observed on 
Left flank only (L) 0.01 
Right flank only (R) 0.01 -0.19 
Left and right flanks (L +R) 0.37 -0.37 -0.42 
Number of follicles observed on: 
Left flank (FN-L) 0.43 0.25 -0.44 0 59 
Right flank 0.90 -0.25 0.10 0.39 0.25 
------ ------ -------
The production and reproduction traits of females in the older age groups were generally higher 
than the performance levels of 2-year old females. Two-year old females performed worse 
(Ps:0.05) than their older contemporaries in terms of FCS and egg production performance (EPP) 
for the first 14 days after or prior to scanning, one month and 2 months after or prior to scanning 
and also for the entire breeding season (Table 4). Mean FN and FD did not differ significantly 
between the three age groups (Table 4, P~0.05). 
Table 4. Least square means (±SE) depicting the influence of age on the ultrasound parameters and egg 
production traits of breeding ostrich females. 
- ------------·--
Parameter A ge group 
2 Years 3-6 Years ~-7+ Years 
Follicle classification score 0.81 ± 0.07• 1 02 ± 0 04°- 1 10 ± 0 06° 
Number of follicles 1.57 ± 0.14 1.82±0.08 1.92±0.11 
Follicle diameter (mm) 44.1±1.8 45.8 ± 0.9 47.2 ± 1.2 
EPP14ci 11.1±2.7. 23.3±1.6b 24.0±2.1b 
EPP1m 14.4 ± 2.7· 30.5±1.7b 34.2 ±2.2b 
EPP2m 14.0 ± 2.7· 32.8±1.9b 35.3 ±2.2b 
EPP total 21.7 ± 3.2• 44.8 ± 2.4b 45.4 ± 2.8b 
a, o. .. . Different superscripts denote s1gnif1cance (P:s;0.05) 
Models fitting the data best in the overall analysis mostly included the direct additive genetic 
variation as the only significant random effect (Table 5). The only exception was FD, where none 
of the additional random effects fitted to the data improved the log likelihood. The same was found 
when the production traits were assessed at the beginning (Table 6) and end of the breeding 
season (Table 7). EPP14d was the exception when end of season results were analysed. The 
model fitting additional random effects did not improve the log likelihood. 
106 
Table 5. Log-likelihood values for models fitting different random effects for ultrasound parameters and egg 
production traits obtained for entire breeding seasons. 
Parameter ----~--Fixed -effects 
- --
h" +PE h' PE 
-- ----------!---·------- --
Follicle classification score -435.6 -420.9 -421.2* -424.4 
--f--------------~-------- ---
Number of follicles -1340.5 -13294 -1329.4* -1333.6 
---
Follicle diameter (mm) ------=z-990_8*-
--
-2989.2 -29894 -2989 4 
--f-- --
EPP14d -5279.5 -5270.2 -5269.9* -5274 2 
--- --------- --
EPP1m -5266.7 -5249.3 -5249.0* -5255.1 
EPP2m -5125.1 -5086.9 -5087.3* -5091.9 
-----
h_ 2 -- hentab1-l1- ty, PE -- permanent environment, * -- P:S0.05 
Table 6. Log-likelihood values for models fitting different random effects for ultrasound parameters and egg 
production traits obtained at the beginning of the breeding seasons. 
Parameter Fixed effects h 2 +PE h" PE 
--
Follicle classification -151.673 --145.153 -145.156* -148432 
Number of follicles -501.873 -499.193 -498.756* -501.84 
--
Follicle diameter (mm) -1028.51* -1027.85 -102848 -1027.85 
--
EPP14d -1733.20 -1729.28 -1729.48* -1730.77 
-- ------
EPP1m -1814.16 -1806.00 -1806.00* -1809.96 
---
EPP2m --1827.75 -1805.54 -1805.60* -1811.34 
EPP1otal -1757.48 -1721.76 -1722.83* -1726.87 
h_ 2 -- hentab1-l-ity, PE = permanent environment, *= P:>0.05 --
Table 7. Log-likelihood values for models fitting different random effects for ultrasound parameters and egg 
production traits obtained at the end of the breeding seasons. 
Parameter Fixed effects ---i?+PE- h' --PE 
-- f--- --
Follicle classification -166.852 -162 089 -161.990* -165.36 
---~- ----- --
Number of follicles -467.732 -462.403 -462.404* -465.463 
--
Follicle diameter -979.182* --977.679 -978.262 -977.689 
--- f--- --
EPP14d -1904.28* -1903.24 -1903.26 -1903.52 
--
EPP1m -1830.82 -1827.34 -1827.35* -1828.42 
--
EPP2m -1763.06 -1760.20 -1760.19* -1761.25 
-- ---
EPP total -1764.88 -1728.32 -1729.66* -1732.58 
2 ----h =heritabi-l-ity; PE= permanent environment, . - P:>0.05 --
Heritability estimates (in bold and on the diagonal) in the overall analysis were generally low, in 
most cases < 10% (Table 8). In the case of EPP, estimates tended to increase as data were 
recorded over a longer period. The genetic correlations (above the diagonal) of FCS and number 
of follicles with EPP were generally positive, but failed to reach statistical significance (ie. a level of 
double the corresponding standard error). Genetic correlations among EPP parameters were very 
high and outside the parameter space in some instances. Environmental correlations (below the 
diagonal) between the various measures of egg production were generally high. The 
environmental correlation between FCS and number of follicles observed approached 80%. The 
107 
latter parameters were generally positively related to the various measures of EPP (P:o;0.05) on an 
environmental level, although the estimates failed to reach 0.20 
Table 8. Genetic and environmental (co)variance estimates and ratios for follicle classification score (FCS), 
number of follicles (FN), follicle diameter (FD), and egg production performance (EPP) over all 
available records. 
Variances and-- FOLLICLE PARAMETERS ·----1 EGG PRODUCTION TRAITS 
parameters I traits FCS FN FD 14 da ys 1 month 2 months 
+----
Variances 
Phenotype 0.6926 2.700 268.4 1 017 .5 997.4 815.6 
Genotype 0.0517 0.158 7 9 53.7 77.4 121.4 
Environment 0.6409 2.542 278.5 9 63. 8 919.0 694.2 
·-- --- --t--------+----------
Parameter 
FCS 0.07±0.02 0.96±0.04 0.00±0.41 0.2 3±0 .23 0.27±0.21 0.19±0.18 
FN 0 78±0.01 0.06±0.02 -0 64±0.36 0.2 6±0 24 0.30±0.22 0.27±0.19 
FD 0 16±0.03 -0.30±0.03 0.03±0.02 0.2 5±0 .45 0.14±0.39 0.25±0.31 
EPP14d 0.09±0.03 0.11±0.03 0 08±0 04 0.0 5±0 .02 0.96±0.05 1 00±0 04 
EPP1M 0.12±0.03 0.12±0 03 0.09±0.04 0.7 0±0 .02 0.08±0.02 1.05±0.01 
EPP2M 0.14±0.03 0 15±0.03 0.05±0.04 0.6 9±0 .02 0.87±0.01 0.15±0.03 
Both measures of follicular activity were highly correlated with EPP close to the assessment, and 
declined and became insignificant (P~0.05) as data were recorded over a longer period. When 
related to overall egg production, genetic correlations were negative in sign, but small in magnitude 
(Table 9). Environmental correlations resembled those reported in Table 8, but were generally 
higher when the measures of follicular activity were related to egg production. Heritability of overall 
egg production was fairly high at 41 %. 
Table 9. Genetic and environmental (co)variance estimates and ratios for follicle classification score (FCS), 
number of follicles (FN), follicle diameter (FD), and egg production performance (EPP) when 
recorded at the beginning of the breeding season. 
Variances and FOLLICLE PARAMETERS -- EGG PRODUCTION TRAITS 
parameters/traits FCS FN FD 14 days 1 month 2 months Overall 
Variances 
Phenotype 0.6402 2.722 190.2 446.7 623.5 665.0 521.2 
Genotype 0.0809 0.193 2.34 39.2 85.9 182.0 215.3 
Environment 0.5596 2.529 187.8 507.5 537.6 483.0 305.9 
Parameter 
FCS 0.13±0.05 1.11±0.13 n.e. 1.04±0.26 0.56±0.26 0.25±0.22 -0.14±0.21 
FN 0.69±0.03 0.07±0.04 n.e. 0 99±0.34 0.48±0.33 0.19±0.28 -0 13±0.25 
FD n.e. n.e. 0.01±0.05 n.e. n.e. n.e. n.e. 
EPP1<d 0.17±0.05 0.14±0.05 n.e. 0.09±0.05 1.00±0.05 0.97±0.09 0.74±0.17 
EPP1M 0.23±0.05 0.19±0.05 n.e. 0 81±0.01 0.15±0.05 0.98±0.03 0 68±0.14 
EPP2M 0.21±0.05 0.20±0.05 n.e. 0 60±0 03 0 84±0.02 0.27±0.06 0 78±0.08 
EPP total 0.12±0.06 0.15±0.05 n.e. 0.25±0.05 0.37±0.05 0.50±0.04 0.41±0.06 
= L--------~------ -- --n.e. not estimable 
108 
2 
Results pertaining to h estimates and environmental correlations at the end of the breeding 
season corresponded with those presented in Table 8 and Table 9 in most instances (Table 1O ) 
Genetic correlations between follicle parameters and egg production however, were different from 
those obtained for the entire season and at the beginning of the breeding season All end of 
season correlations were negative and substantial in magnitude for EPP14d· 
Table 10. Genetic and environmental (co)variance estimates and ratios for follicular classification score 
(FCS), number of follicles (FN), follicle diameter (FD), and egg production performance (EPP) 
recorded at the end of the breeding season. 
Variances and --FOLLICLE PARAMETERS EGG PRODUCTION TRAITS 
--
parameters/traits FCS FN FD 14 days 1 month 2 months Overall 
-- ---~-~ ~ -~ 
Variances 
Phenotype 0.6890 2.383 252.8 895.9 665 g 502.4 521.2 
Genotype 0.0634 0.235 29.9 46.8 72.2 48.5 215.3 
Environment 0.6256 2.148 222.9 849.1 593.0 445.9 305.9 
-
Parameter 
FCS 0.09±0.04 0.97±0.05 -0.85±0.35 -0.80±0.56 -0.20±0.36 -0.30±0.37 -0.16±0.24 
FN 0.86±0.01 0.10±0.04 -0.76±0.27 -0.64±0.54 -0.06±0.34 -0.15±0.36 0.05±0.23 
FD -0.22±0.07 -0.39±0.06 0.12±0.07 0.74±0.55 -0.26±0.46 0.09±0.08 -0.22±0.30 
EPP1,d 0.11±0.05 0.13±0.05 -0.04±0.07 0.05±0.04 1.04±0.20 0.97±0.10 0.78±0.18 
EPP1M 0.13±0.05 0.14±0.05 0.09±0.07 0.46±0 04 0.11±0.05 1.01±0 04 0.97±0.09 
EPP2M 0.15±0.05 0.15±0.05 0.03±0.07 0.75±0.02 0 86±0 01 0.10±0.05 0.94±0.09 
EPPtotol 0.16±0.06 0.18±0.05 -0 01±0.07 0.38±0.05 0.45±0.04 0.47±0 04 0.41±0.06 
DISCUSSION 
Ultrasound scanning proved to be an effective and stress-free method to assess the extent of 
follicular development on the ostrich ovary. The scanning procedure is non-invasive and can be 
performed easily and time-efficiently, subjecting the bird to minimal stress if the proper handling 
techniques are employed. Observations on ovarian appearance and different sized follicles during 
the study are in agreement with observations on the female reproductive tract in poultry and ratites 
(Duerden, 1912; Waddington et al., 1985; Bezuidenhout, 1986; Hocking et al., 1987; Hicks, 1992; 
Soley and Groenewald, 1999; Williams, 1998). 
A prominent feature of the data is the variability of all ultrasound parameters and egg production 
traits considered, as reflected by the high coefficients of variation and extreme ranges in Table 2. 
Despite the large variation, no evidence of skewness in the distributions of any of the parameters 
or traits was detected by standard descriptive statistics methods. Egg production percentages 
calculated for the various production periods after or prior to scanning showed extreme variability 
and agree with the marked variation in EPP recorded in previous studies (Van Schalkwyk, 1998; 
Bunter and Graser, 2000; Lambrechts et al., 2002a, b; Cloete et al., 2002a) The range of follicles 
observed was larger than that observed by Lambrechts et al. (2002a) and Bronneberg and Taverne 
(20U3). Follicles with a minimum size of 30mm were considered in the latter study, with no cases 
109 
where no follicles were observed. In this study, a larger range of different sized follicles was 
observed than reported by Bronneberg and Taverne (2003). The number and diameter of follicles 
observed in this study also differs from that observed in domestic chicken and turkey hens (Gilbert 
et al., 1983; Hocking et al., 1987; Melnychuk et al., 2002). Although the ostrich ovary is located on 
the left side of the body, sonograms obtained on the right flank of the ostrich females provided a 
better indication of the extent of follicular development on the ostrich ovary. 
Females younger than 3 years in this study performed worse in terms of all ultrasound parameters 
and egg production traits than their older contemporaries. This supports findings that ostrich 
females younger than 3 years produced significantly less eggs than older females and 
contemporaries and suggested that immaturity was the most important factor affecting the laying 
performance of the females (More, 1997; Cloete et al., 1998; Bunter, 2002). It also agrees with 
findings of Lambrechts et al. (2002a) that ostrich females of three years and younger showed 
considerably less follicular development than females four years of age and older In the latter 
study, two-year old ostrich females showed no follicular activity at the beginning of the breeding 
season. Ovary size is in part determined by the age of an ostrich female, with the ovary of a young 
immature female being thin, flattened and almost rectangular in appearance. The ovary of a 
mature female consists of a stroma with a series of embedded, developing follicles, which is absent 
in reproductively immature females (Cho et al., 1984). 
Jn the absence of reliable data on ostriches, it is appropriate to consider information regarding the 
laying period in other avian species. The laying period of domestic hens can be divided into three 
phases, i.e. the onset of lay, the main period of lay and the end of laying. During the onset of Jay, 
the ovary might be very sensitive to gonadotropic levels and action. The ovary and the oviduct 
appear to be asynchronized, as about 40% of ovulations did not result in oviposition during this 
phase in domestic hens (Gilbert and Wood-Gush, 1971). The hormones responsible for oviductal 
growth and function, i.e. androgens, progestagens and oestrogens, are dependent on the 
hormones responsible for follicular maturation and ovulation. Lack of synchronization between or 
co-ordination of the ovary and oviduct may be a result of bringing birds into lay precociously. 
During the main production period, the egg production pattern becomes regular, with each hen 
having a characteristic production pattern, and at the end of the laying period production ceases as 
a result of the cessation of ovulation and the production of ova. The infindibulum also fails to 
engulf almost 40% of the ova that are shed (Gilbert and Wood-Gush, 1971). The reason for this is 
unclear since the oviduct was shown to remain active longer than the ovary (Gilbert and Wood-
Gush, 1971 ). The duration of the processes involved from ovulation to oviposition in the ostrich is 
unknown, but it can be assumed that it takes 48 h to travel through the oviduct (Irons, 1995). 
It is possible that the reproductive system of the young ostrich females in this study were not yet 
sufficiently developed to ensure the production of the reproductive hormones at levels high enough 
to support follicular recruitment, development, and the subsequent ovulation processes (Card and 
Nesheim, 1972; Hafez, 1987). This is reflected in the significantly lower egg production observed in 
110 
this study. The presence of follicles in young (in this study 2-year old) females does not 
necessarily indicate that a female will produce eggs. In cases where the ovaries of 2-year old 
females might have been sufficiently developed to support the synthesis and secretion of the 
reproductive hormones, the ovary and oviduct might not have been synchronised at the beginning 
of and throughout the breeding season to ensure successful egg production. Very small follicles 
that occur in young female chickens are indicative of the absence of gonadotropin support by the 
follicles. In laying hens, it appears that the maturation and the initiation of egg-laying is dependant 
on an age-related increase in ovarian sensitivity to gonadotropin stimulation (Etches, 1995). The 
circulating levels of FSH and LH may be sufficient to ensure follicular recruitment and development, 
but not ovulation and subsequent egg production in the young females. The first oviposition is not 
necessarily the first indication of sexual maturity since certain sexual organs may be functional 
before others (Gilbert, 1971 ). No correlation of significance has been found between the number of 
visible ova and the egg-laying ability of the hen (Card and Nesheim, 1972). Laying hens (Gallus 
domesticus) of all ages were able to produce mature follicles. Birds with poor laying performance 
had reduced ovulation rates due to the loss of large follicles by atresia, which was rarely observed 
in birds with good laying performances (Waddington et al., 1985). 
The fact that not all females come into lay at the beginning of a breeding season implies that 
2 
genetic factors are involved. The h estimates for the ultrasound parameters and EP traits reported 
in this study were generally lower than that reported for poultry species, but significant for almost all 
ultrasound parameters and EP traits considered. The exception was follicle diameter. Heritability 
estimates of EP reported for domestic chickens were approximately 0.1 O and for common ducks 
(Anas platyrhynchos) between 0.15 and 0.47 (Gowe and Fairfull, 1995; Brun and Larzul, 2003). 
Heritability reported for the total number of eggs produced by three Catalan poultry breeds, were 
estimated to be 0.20, 0.31 and 0.33 respectively (Francesh et al., 1997) Heritability estimates for 
egg number for purebred and crossbred chicken lines were 0.54 to 0.74 and 0.04 to 0.51, 
respectively (Wei and Van der Werf, 1995). A long-term goose breeding program for egg 
production and crammed liver weight (Shalev et al., 1991) found that heritability estimates were 
found to be high for most of the parameters, with phenotypic and genetic correlation coefficients 
being rather low. 
2 
There was a suggestion for h estimates to be higher when assessed at the beginning of a 
breeding season. No literature could be found to support or refute these trends. Overall EP had a 
2 
higher h of 0.41. Previous studies on the same flock yielded h2 estimates of approximately 0.15, 
with permanent environmental variance ratios of 0.25-0.30 (Bunter, 2002; Cloete et al., 2002b). 
The best model in the present study however, did not partition the genetic and permanent 
environmental variances as the previous studies did. The size of the present data set was not 
large enough to provide the necessary degrees of freedom to allow for the partitioning of genetic 
and permanent environmental effects. Alternatively, a lack of pedigree depth could also have 
contributed to the Jack of partitioning genetic effects from the overall female variances in the earlier 
studies. Bunter (2002) stressed this aspect since relatively few of tr1e hens that were assessed 
111 
were descendants from females with production records. Total between-female variance ratios 
(i.e. the sum of the direct additive and female permanent environmental effects), however, 
remained relatively stable at 0.40-0.45 across all studies. 
All genetic correlations between FCS and number of follicles were effectively unity, illustrating that 
it is effectively the same parameter on a genetic level. The same reasoning is applied to part-
season records for egg production. The very high genetic correlation of parameters for follicular 
activity with egg production during the first 14 days or month into a breeding season indicate that 
ultrasound can be used to predict egg production accurately over the short term. The relationship 
however, grew weaker as the interval between scanning and the recording of egg production 
increased. Lambrechts et al. (2002a) reported similar results on a phenotypic level. 
The discrepancy between parameters for follicular activity and egg production at the beginning and 
end of the breeding season is possibly the most striking feature of the data set Although the 
general trends observed in the present study agree with each other, the derived genetic correlation 
differed in sign. It would appear that follicular activity at the end of a breeding season was 
associated with a low egg production just prior to scanning at the end of the breeding season on a 
genetic level. It is possible that the decrease in day length had the effect that synthesis of the 
gonadotropin decreased to such a level that follicle recruitment and growth could be supported, but 
that the circulating levels were not high enough to support ovulation and oviposition at the end of 
the season. Therefore the presence of follicles on the ovary would not necessarily result in 
successful oviposition at the end of the season. Further research is required for the proper 
understanding of this phenomenon. 
Price et al. (1988) also stated that pairs of birds that start to breed earlier in the season have higher 
reproductive success than those who start breeding later. When the high genetic correlations 
between part-season records for egg production and overall egg production is considered, our 
results would support this contention. The results from this study did not, however, hold true for 
follicular activity reflected in the sonograms, with activity being poorly related to overall egg 
production. Environmental correlations between follicular activity and egg production were positive 
in all instances, suggesting that an environment favourable for follicular activity would also benefit 
egg production. The derived correlations however, were too low for accurate predictions of egg 
production. The potential of ultrasound scanning to predict the egg production for an entire 
breeding season is therefore limited. 
CONCLUSIONS 
The study presents novel information on genetic estimates of ultrasound parameters and their 
relationship with egg production traits in breeding ostrich females. Ultrasound scanning can be 
used to identify females that come into lay early in a breeding season, and the duration of lay after 
112 
the commencement of the breeding season was strongly associated with reproductive success. 
The correlation reported for early season scanning and overall egg production suggest that females 
can be selected on partial production records for use in breeding programs 
Progress can be made in the current flock through the selection of ostrich females with well-
developed follicular hierarchies at the beginning of a breeding season. Purvis and Hillard (1997) 
reported acceptable rates of genetic progress under similar conditions in sheep. Partial records for 
a given season may, however, not be a reliable indicator of the ability of a female if the 
commencement of lay were delayed for the specific female or if the age of the female was 
unknown. The age of females needs to be considered when selection decisions are based on 
phenotypic data. Failure to do so may discriminate against young females. 
REFERENCES 
ALLEN WE & MEREDITH MJ, 1981. Detection of pregnancy in the bitch; a study of abdominal 
palpation, A-mode ultrasound and Doppler ultrasound techniques. J. Small Anim. Pract. 
22: 609-622. 
BADLEY AR, 1997. Fertility, hatchability and incubation of ostrich. Poult. Avian Biol. Rev. 8: 53-76. 
BEZUIDENHOUT AJ, 1986. The topography of the thoraco-abdominal viscera in the ostrich (Struthio 
came/us). Onderstepoort J. Vet. Res. 53: 111-117. 
BOYD JS, 1995. Real-time diagnostic ultrasound in bovine reproduction. In: Goddard, P.J. (Ed.), 
Veterinary Ultrasonography, CAB International, United Kingdom, pp. 233-255. 
BOURKE DA, ADAM CL & KYLE CE, 1992. Ultrasonography as an aid to controlled breeding in the 
llama (Lama glama). Vet. Rec. 130(19): 424-428. 
BRAND Z, 2002. The effect of energy and protein nutritional levels on production of breeding 
ostriches. M.Sc thesis, Rhodes University, Grahamstown, South Africa. 
BRIEN FD, KONSTANTINOV KV & GREEFF JC, 2002. Comparison of linear and threshold models for 
predicting direct and maternal genetic effects on number of lambs weaned in Western 
Australian Merinos. Proc. fh World Congr. Gen. Appl. Livestock Prod. 32, 18-23 August, 
Montpellier, France, pp. 307-310. 
BRONNEBERG RGG & TAVERNE MAM, 2003. Ultrasonography of the female reproductive organs in 
farmed ostriches (Struthio came/us spp.). Theriogenof. 60(4): 617-633. 
BRUN JM & LARZUL C, 2003. Inheritance of reproductive parameters of female common ducks 
(Anas platyrhynchos) in pure breeding and in inter-generic crossbreeding with Muscovy 
ducks ( Cairina moschata). Brit. Pou/t. Sci. 44(1 ): 40-45. 
BUNTER KL, 2002. The genetic analysis of reproduction and production parameters recorded for 
farmed ostriches (Struthio camefus). Ph.D. dissertation, University of New England, 
Australia. 
113 
BUNTER K & GRASER H-U, 2000. Genetic evaluation for Australian ostriches - a report for the Rural 
Industries Research and Development Corporation. Publication no. 00/153, Rural 
Industries Research and Development Corporation, Australia. 
CARD LE & NESHEIM MC, 1972. Poultry Production, 11th edition. Lea & Febiger, Philadelphia. 
CASARES M, ROBEL A & HONEGGER RE, 1997. Observations on the female reproductive cycle of 
captive giant tortoises (Geoche/one spp) using ultrasound scanning. J Zoo Wild!. Med. 
28(3); 267-273. 
CHO P, BROWN R & ANDERSON M, 1984. Comparative gross anatomy of ratites. Zoo Biol. 3: 133-
144. 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998. Progress towards a scientifically based 
breeding strategy for ostriches. Proc. 3ffh Nat. Congr. S. Afr. Soc. Anim. Sci., 5-8 April, 
Stellenbosch, pp. 69-72. 
CLOETE SWP, BUNTER KL & VAN SCHALKWYK SJ, 2002a. Progress towards a scientific breeding 
strategy for ostriches. Proc. fh World Congr. Gen. Appl. Livestock Prod., 18-23 August, 
Montpellier, France, volume 30: 561-568. 
CLOETE SWP, BUNTER KL, VAN SCHALKWYK SJ & BRAND Z, 2002b. Genetic parameters for live 
weight and reproduction in ostrich breeding females. Proc. 151 Joint Congr. Grassland Soc. 
S. Afr. and S. Afr. Soc. Anim. Sci., 13-16 May, Christiana Aventura, p. 81. 
DEEMING DC & AR A, 1999. Factors affecting the success of commercial incubation. Jn: Deeming, 
D.C. (Ed.), The ostrich - biology, production and health, CAB/ Publishing, CAB 
International, Wallingford, Oxon, United Kingdom, pp. 159-190. 
DUERDEN JE, 1912. Experiments with ostriches XX. The anatomy and physiology of the ostrich. C. 
The internal organs. S. Afr. J. (April/May): 1-27. 
ETCHES RJ, 1995. Physiology of reproduction: the female. In: Hunton, P. (Ed.), World Animal 
Science, C9 - Poultry Production, Elsevier, New York, pp. 221-241. 
FOWLER ME, 1991. Comparative clinical anatomy of ratites. J Zoo Wild!. Med. 22: 204-227. 
FOWLER DG & WILKINS JF, 1982. The identification of twin bearing ewes with real time ultrasound 
and its potential use in sheep and wool production. Wool Technol and Sheep Breeding 
(Sep-Oct): 111-117. 
FRANCESH A, ESTANY J, ALFONSO L & IGLESIAS M, 1997. Genetic parameters for egg number, egg 
weight, and egg shell colour in three Catalan poultry breeds. Pou/t. Sci. 76(12): 1627-
1631. 
GILBERT AB, 1971. The Ovary. In: Bell D.J., Freeman, B.M. (Ed.'s), Physiology and biochemistry 
of the domestic fowl (volume 3), pp. 1163-1202. 
GILBERT AB, 1979. Female genital organs. In: King, A.S., Mclelland, J. (Ed.'s), Form and function 
in birds, Academic Press, New York. 
GILBERT AG & Wooo-GusH DGM, 1971. Ovulatory and ovipository cycles. In: Bell, D.G., Freeman, 
B.M. (ed.'s), Physiology and biochemistry of the domestic fowl, Academic Press, London, 
pp. 1353-1378. 
114 
GILBERT AB, PERRY MM, WADDINGTON D & HARDIE MA, 1983. Role of atresia in establishing the 
follicular hierarchy in the ovary of the domestic hen (Gallus domesticus). J. Reprod. Fert. 
69(1): 221-227. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3. NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
GOWE RS & FAIRFULL RW, 1985. The direct response to long-term selection for multiple 
parameters 1n egg stocks and changes in genetic parameters with selection. In Hill, W.G., 
Manson, j.M., Hewitt, D. (Ed's), Poultry Breeding and Genetics, Longman Group, Harlow, 
pp. 125-145. 
GowE RS & FAIRFULL RW, 1995. Breeding and genetics of egg laying chickens. In: Hunton, P 
(Ed.), World Animal Science, C9 - Poultry Production, Elsevier, New York, pp. 435-455. 
HAFEZ ESE, 1987. Reproduction in Farm Animals, 5th ed. Lea & Febiger, Philadelphia. 
HICKS KO, 1993. Ostrich reproduction. In: Fowler, M.E. (Ed.), Zoo and Wildlife Medicine. Current 
therapy 3, W.B. Saunders, Philadelphia, pp. 203-206. 
HICKS KO, 1992. Ratite reproduction. Proc. Assoc. Avian Vet., pp. 318-324. 
HICKS-ALDREDGE KO, 1993. Ratite Reproduction. Multi-State Big Bird Conf, pp. 1-20 
HILDEBRANDT TB, HERMES R, JEWGENOW K & GORITZ F, 2000. Ultrasonography as an important 
tool for the development and application of reproductive technologies in non-domestic 
species. Theriogeno/. 53(1 ): 73-84. 
HOCKING PM, WALKER MA, WADDINGTON D & GILBERT AB, 1987. Observations on the size of the 
follicular hierarchy in the turkey hen and a case of arrested follicular development. Brit. 
Poult. Sci. 28(4): 755-757. 
HoRBAvCZUK JO & SALES J, 1999. Reproduction is a crucial problem in ostrich breeding. World 
Poultry 15(5): 28-30. 
IRONS P, 1995. Ostrich reproductive research: quantum physics or back to the drawing board?., 
Proc.rfh Ann. Congr. Livestock Health Prod. Group S. Afr. Vet. Assoc., June, Warmbaths, 
pp. 140-148. 
JARVIS MJF, JARVIS C & KEFFEN RH, 1985. Breeding seasons and laying patterns of the South 
African ostrich Struthio came/us. Ibis 127, 442-449. 
JORGENSEN JN, 1994. Estimation of genetic parameters for litter size in sheep using a non-linear 
method. Acta Agriculturae Scandinavica. A. Anim. Sci. 44, 8-11. 
JULL MA, 1952. Poultry breeding, 3rd edition. John Wiley & Sons, Inc., New York. 
KIRBY JD, RORIE RW, MELNYCHUK v & ANTHONY NB, 2001. Ultrasound as a tool to assess 
reproductive status in poultry. J. Dairy Sci. 84 (suppl. 1), abstract no.734. 
KRAUTWALD-JUNGHANNS M-E, KOSTKA VM & HOFBAUER H, 1998. Observations on the significance 
of diagnostic findings in egg-binding of Psittaciformes. Vet. Rec. 143: 498-502. 
LAMBRECHTS H, CLOETE SWP, SWART D & GREYLING JPC, 2002a. Preliminary results on the use of 
diagnostic ultrasonography as a management tool to quantify egg production potential in 
breeding ostrich (Struthio came/us australis) females. J. S. Afr. Vet. Assoc. 73(2): 48-52 
115 
LAMBRECHTS H, CLOETE SWP, SWART D & GREYLING JPC, 2002b. Preliminary results on teasing 
51 
and flushing practices for breeding ostriches. Proc. 1 Joint Congr: Grassland Soc. S.Afr. 
and S. Afr. Soc. Anim. Sci., 13-16 May, Christiana Aventura, South Africa, p. 103 
LAMBRECHTS H & PFISTER AP, 1999. The use of real-time linear ultrasonography to predict egg 
production in fem ales ostriches ( Struthio came/us) Proc. 3fh Nat. Congr. S Afr. Soc. 
Anim. Sci., 27-29 July, Pretoria, p. 31. 
MELNYCHUK VL, COOPER MW, KIRBY JD, RORIE RW & ANTHONY NB, 2002. Use of ultrasonography 
to characterise ovarian status in chicken Poult Sci. (June) 81 (6): 892-895. 
MORE SJ, 1997. Monitoring the health and productivity of farmed ostrich flocks. Austr. Vet. J. 8: 
583-587. 
NESTOR KE, McCARTNEY MG & BACHEV N' 1969. Relative contributions of genetics and 
environment to turkey improvements. Poult. Sci. 48: 1944-1949 
PETITIE JN & DAVIS G, 1999. Breeding and Genetics. In: Deeming, D.C. (Ed.), The Ostrich -
Biology, Production and Health, CAB! Publishing, CAB International, Wallingford, United 
Kingdom, pp. 275-292. 
POGGENPOEL DG, FERREIRA GF, HAYES JP & DU PREEZ JJ, 1996. Response to long-term selection 
foregg production in laying hens. Brit. Poult. Sci. 37(4): 743-756. 
PRICE T, KIRKPATRICK M & ARNOLD SJ, 1988. Directional selection and the evolution of breeding 
date in birds. Science 240(4853): 798-799. 
PURVIS IW & HILLARD M, 1997. Biology and genetics of reproduction. In: Piper, L., Ruvinsky, A. 
(Ed.'s), The genetics of sheep, CAB International, Wallingford, Oxon, United Kingdom, pp. 
275-394. 
SENGOKU K, ISHIKAWA M, YAMASHITA K, MIZOGUCHI H, KASAMO M, SHIMIZU T & TANAKA K, 1985. 
Assessment of follicular development and ovulation by ultrasound. Nippon Sanka Fujinka 
Gakkai Zasshi (Dec) 34(12): 2205-2211. 
SHAFFER N, BRYANT W, AGNEW D, MEEHAN T & BEEHLER B, 1998. Ultrasonographic monitoring of 
artificially stimulated ejaculation in three rhinoceros species (Ceratotherium simum, 
Diceros bicornis, Rhinoceros unicornus). J. Zoo Wild!. Med. 29(4 ): 386-393. 
SHALEV BA, DVORIN A, HERMAN R, KATZ z & BORNSTEIN S, 1991. Long-term goose breeding for egg 
production and crammed liver weight. Brit. Pou/t. Sci. 32(4): 703-709. 
SKIDMORE JA, BILLAH M & ALLEN WR, 1996. The ovarian follicular wave pattern and induction of 
ovulation in the mated and non-mated one-humped camel (Came/us dromedarius). J. 
Reprod. Fert. Dev. 106(2): 185-192. 
SNYMAN MA, OLIVIER JJ & OLIVIER WJ, 1996. Variance components and genetic parameters for 
body weight and fleece parameters of Merino sheep in an arid environment. S. Afr. J. 
Anim. Sci. 26: 11-14. 
SOLEY JT & GROENEWALD HB, 1999. Reproduction. In: Deeming, D.C. (Ed.), The Ostrich -
Biology, Production and Health, CAB! Publishing, CAB International, Wallingford, United 
Kingdom, pp. 129-158. 
SOLOMON SE, 1983. Oviduct. In: Freeman, B.B. (Ed.), Physiology and Biochemistry of the 
Domestic Fowl -volume 4, Academic Press, London, pp. 379-419. 
116 
VAN SCHALKWYK SJ, 1998. Improvement of fertility and hatchability of artificially incubated ostrich 
eggs in the Little Karoo. M.Sc. thesis, Rhodes University, Grahamstown 
VAN SCHALKWYK SJ, CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations 
for body weight and reproduction in commercial ostrich breeding pairs. Brit. Poult Sci. 37: 
953-962. 
VAN SCHALKWYK SJ, CLOETE SWP, DE KOCK JA, BRAND z & PFISTER B, 1995. The repeatability of 
egg production in the ostrich: preliminary results. Proc. 341h Annual Congr S. Afr. Soc. 
Anim. Sci., 10-13 April, Bloemfontein, p. 68. 
VANZYL SMIT DJ, 1964. Volstruisboerdery in die Klein Karoo Pamphlet no. 358, Department of 
Agricultural Technical Services, Republic of South Africa, Pretoria. 
WADDINGTON D, PERRY MM, GILBERT AB & HARDIE MA, 1985. Follicular growth and atresia in the 
ovaries of hens (Gallus domesticus) with diminished egg production rates. J Reprod. Fert. 
74(2): 399-405. 
WEI M & VANDERWERF JH, 1995. Genetic correlation and heritabilities for purebred and crossbred 
performance in poultry egg production parameters. J Anim Sci 73(8): 2220-2226. 
WILLIAMS J, 1998. Ratites: diagnostic imaging. Vet. Clinics North Am Food Anim. Pract. (Nov) 
14(3): 421-435. 
ZAKARIA AH, MIYAKI T & !MAI K, 1983. The effect of ageing on the ovarian follicular growth in laying 
hens. Poult. Sci. 62(4): 670-674. 
117 
Chapter 9 
(Co)variance estimates for live weight, body 
measurements and reproductive traits of pair-mated 
ostrich females1 
Abstract 
Estimates of genetic parameters for live weight, body measurements and reproductive traits were obtained using data 
from a pair-mated ostrich flock located at Oudtshoorn in South Africa. Live weight, chest circumference and tail 
circumference were recorded at the commencement and end of each breeding season. Reproductive traits included total 
egg and chick production, along with the percentage of infertile eggs. The data set consisted of approximately 1000 
ostrich female-year records, from 283 hens, obtained for the period 1990-2001 Estimates of heritabilities (h 2 ) were 0.20 
to 0.34 for live weight, 0.12 for chest circumference and 0.30 to 0.38 for tail circumference. In the case of the 
reproductive traits, h2 was 0.23 for egg production, 0.20 for chick production and 0.10 for hatchability. Estimates of 
female permanent environmental effects (c2 ) were 0.32 to 0.36 for live weight and 0.23 to 0.32 for chest circumference; 
0.18 for egg production, 0.18 for chick production and 0.21 for hatchability. Service sire exerted significant effects only 
on hatchability (0.22) and consequently chick production (0.09). Correlations between live weight recorded at the 
beginning and end of the breeding season were unity for additive genetic and permanent environmental effects, but the 
residual correlation was substantially lower. Genetic correlations of live weight or the body measurements with 
reproductive traits were generally low to moderate and variable and failed to reach a significant difference from zero. 
Egg and chick production were highly correlated genetically and phenotypically, with the genetic correlation exceeding 
the theoretical limit. Hatchability was positively related to chick production among others at the male (service sire) level. 
Selection gains in the current flock and future generations are likely in ostriches. No significant adverse relationships 
were found between live weight, body measurements and reproductive traits. This will enable the development of a dual-
purpose ostrich strain, where emphasis may be placed on both growth traits and reproduction. 
INTRODUCTION 
Marked advances in the genetic evaluation and sustained improvement of production traits in livestock 
have been achieved over the past few decades, particularly in the more intensive industries. The 
ostrich industry is an exception in this regard, and limited knowledge of genetic parameters of growth 
and reproduction is characteristic of the industry (Pettite and Davis, 1999). Genetic parameter 
estimates for live weight and reproductive traits are limited to a few studies, generally on the same 
resource population (Bunter et al., 1999; 2001 a; Cloete et al., 2002a). This is in stark contrast to 
1 Submitted for publication in British Poultry Science, with co-authors SWP Cloete, KL Bunter, D Swart, JPC Greyling and z 
Brand 
118 
experience with small domestic poultry species, where extensive knowledge is readily available for as 
growth and development (Marks, 1995) and egg production (Gowe and Fairfull, 1995) of commercial 
stocks. 
There is general consensus that growth and development is unfavourably correlated genetically with 
several reproductive traits in domestic poultry species (Marks, 1995; Gowe and Fairfull, 1995; Bunter, 
2002). At this stage, there is almost a total lack of knowledge in this regard as far as ostriches are 
concerned. Phenotypic correlations between live weight or body measurements and reproductive traits 
have been reported (Van Schalkwyk and Cloete, 1996; Van Schalkwyk et al., 1996; Lambrechts et al., 
1998; 1999). These studies were inconclusive as far as the sign and magnitude of correlations of 
growth traits with reproduction were concerned, possibly because they were estimated from relatively 
small databases. The nature of this correlation is, however, very important for developing an 
appropriate breeding policy for the broader industry (Bunter, 2002). The formation of a single, dual-
purpose line, where emphasis is placed on both growth and reproductive performance would become 
feasible if this correlation is more favourable. Alternatively, the formation of specialist lines, one for 
growth and one for reproduction, could be contemplated, with crossbreeding of these lines at the 
commercial level (Cloete et al., 2002a). 
The present study provides estimates of genetic and environmental parameters for mature female live 
weight and body measurements, as well as reproductive traits for ostrich females, including estimates 
of genetic correlations between these traits. 
MATERIALS AND METHODS 
Experimental animals and location 
Data were obtained from the commercial ostrich breeding flock at the Klein Karoo Agricultural 
Development Centre, recorded during the breeding seasons from 1990 to 2001. The experimental site 
is situated in the arid Klein Karoo region of South Africa, at longitude 22° 15' E and latitude 33° 38' S 
and at an altitude of 301 m above sea level. The long-term precipitation averages 230 mm, with 54 % 
of the total rainfall occurring during winter (April to September). The site is characterized by hot 
summers, with the average maximum temperature exceeding 25°C for the period October to April. 
The origin and management of the ostrich population is well-documented (Van Schalkwyk et al., 1996; 
Bunter, 2002). The data used for this paper were obtained from between 33 and 136 pair-breeding 
paddocks over the period 1990 to 2001. In total, 1177 female-year records of 283 females were 
represented in the reproductive data. The animals were paired off (one male to one female) for a 
period of approximately 8 months in most years included in the study, but the breeding season was 
119 
extended to a 9th month in 1999. The actual mean (SD) number of production days of individual 
females in the study was 235 (36) days per annum. These data included part-records in cases where 
females were replaced due to death or injury during the course of the breeding season. The females in 
the study were between 2 and 22 years of age. The age structure of the breeding flock differed 
considerably during the experimental period. Initially, the breeding stock was relatively old, with few 
age groups represented. From 1992 onwards, a breeding flock with a balanced age structure and 
balanced age groups was strived for. During the most recent breeding season (2001 ), female age 
groups ranging from 2 to 11 years were represented in the data. 
The tendency of ostrich producers to continuously mate the same male-female combinations in the 
same breeding paddock completely confounded data obtained from earlier analyses (e.g. Van 
Schalkwyk et al., 1996; Cloete et al., 1998; Lambrechts et al., 1999). In later years breeding pairs 
were intentionally reallocated to different paddocks, while natural attrition of both males and females 
resulted in different male-female combinations becoming available (Bunter, 2002). These changes 
facilitated the partitioning of variances for animal (or female), breeding paddock and service sire effects 
in later years. The breeding stock were mostly commercial birds obtained from a variety of sources, 
including birds derived from South African feather strains. The classification of birds as commercial or 
feather strains was somewhat arbitrary, and no significant differences between these groups or 
indications of heterosis in crossbred individuals has been observed previously for reproduction, egg 
traits and live weight to slaughter (Bunter, 2002). Against this background, strain effects were not 
considered in the present study. 
Breeding pairs were housed in paddocks of 250 m2 when paired off for breeding. Although some 
vegetation in the paddocks was available for browsing, all birds received a breeding diet at a level of 
2.5 kg per bird per day throughout the entire breeding season. During the course of the data 
recording period, a number of experiments were carried out on the experimental flock. During 1998 
and 1999, the flock were subjected to a feeding trial involving a series of dietary energy and protein 
concentrations (Brand et al., 2000; 2003). Studies on the flushing and teasing, as well as a mid-
season rest, of breeding birds were conducted during 2000 and 2001 (Lambrechts et al., 2002; 
Chapter 2). These treatments were considered as fixed effects in this study, solely to account for the 
variation in performance of breeding females it controlled. 
Data recorded 
Live weight and body measurements were recorded on individual females at the beginning and end of 
each breeding season for the entire experimental period. Live weight at the commencement of mating 
was not available for females introduced in mid-season to replace birds incapacitated due to death or 
injury. Birds leaving the breeding flock during the mating period were likewise not weighed at the 
cessation of breeding. Body measurements followed previous definitions provided by Van Schalkwyk 
120 
and Cloete (1996), and included the thoracic circumference measured cranial to the hind limb (chest 
circumference) and the largest circumference of the most caudal aspect of the abdomen in the vicinity 
of the caudal callus (tail circumference). These data were recorded selectively on predominantly 
younger birds during the initial years of the experiment, but were recorded routinely for all potential 
breeders from the 1996-breeding season onwards. Eggs produced during the breeding season were 
collected daily, and incubated artificially, as described by Van Schalkwyk (1998) and Bunter (2002). 
Daily egg production was known for individual breeding paddocks and was used to derive production 
records of individual females within a specific season. During the artificial incubation process, the 
outcome of individual eggs (hatched, infertile, dead in shell) was known (Van Schalkwyk, 1998; Van 
Schalkwyk et al, 2000; Bunter, 2002) This information was used to derive the hatchability of eggs set, 
as well as the number of chicks produced per breeding female for a breeding season. Hatchability 
was only considered in females where more than five eggs were set during a specific breeding season. 
Statistical analyses 
The ASREML program (Gilmour et al., 1999) was used for the estimation of the fixed effects, and 
subsequently to derive (co)variance components for each trait in univariate analyses. ASREML 
estimates variance components for mixed models by restricted maximum likelihood, employing an 
average information algorithm that concurrently provides estimates of standard errors for parameter 
estimates (Gilmour et al., 1995). A series of two-trait analyses were then performed to estimate 
correlations between traits. From the outcomes of these analyses, it was decided to perform three-trait 
analyses, involving live weight at the commencement and cessation of breeding, as well as egg or 
chick production. 
Systematic effects 
Fixed effects that were considered for all traits included the production year (1990-2001 ), age of the 
female at the start of each breeding season, dietary treatments (1998 and 1999), as well as flushing, 
teasing and resting (reproductive management) treatments (2000 and 2001). Two-factor interactions 
between main effects were considered initially, where applicable. Interactions were, however, found to 
be not significant, and were excluded from the final, operational models. Length of the breeding season 
(in days) was fitted as a linear covariate for egg and chick production traits. 
The means for females belonging to the respective age groups could be described as longitudinal data. 
On the assumption that a specific trend would be discernable, a smoothing cubic spline was initially 
fitted to the data (Verbyla et al., 1999) using ASREML (Gilmour et al, ·1999). The spline consisted of 
three components, namely: a fixed linear component, random deviations from linearity following a 
smooth trend, and random deviations from linearity not conforming to a smooth trend. In some cases 
this approach did not result in marked improvements in the interpretation of the results. The final 
121 
operational models thus included female age as a fixed effect, although outcomes of the analyses 
involving splines will be presented where applicable. In view of the change in the age structure of the 
breeding flock, it was decided to obtain age trends for the period from 1996 to 2001, when all age 
groups from 2 to 1O  years were represented annually. The data file used for this purpose consisted of 
708 female-year records for reproduction and between 665 and 692 records for live weight and body 
measurements. The approach to use data on a limited number of age groups is motivated further by 
recommendations that there is little to be gained from the maintenance of ostrich female age groups 
older than 1O to 11 years of age (Cloete et al., 1998; Bunter 2002). 
Random effects 
Parameter estimates for each trait were obtained under an animal model, attributing each record to an 
individual female. Repeated records were accommodated by fitting a random permanent 
environmental effect for each trait. In initial analyses, random paddock effects were also fitted to the 
data. This term was, however, not significant and dropped from the final operational models. The 
following models were considered for analyses of female live weight and body measurements (in 
matrix notation): 
y = Xb + Z1a + e (1) 
y = Xb + Z2C1ema1e + e (2) 
y =X b + Z1a + Z2C1ema1e + e (3) 
The additional random effect of service sire was assessed for all the reproductive traits, resulting in the 
following model: 
(4) 
where y is a vector of observations; b, a, Ciemale and Cservice sire are vectors of fixed effects, additive 
genetic effects and permanent environmental effects of hens or their mates; X, Z1, Z2, and Z3 are the 
corresponding incidence matrices relating the records to effects; and e is the vector of residuals 
It was assumed that: 
2 2
V(a) =Ar/a; V(Ciemale) = lcr 1ema1e; V(Cservice sire)= lcr2 service sire V(e) = lcr e, 
where A is a matrix describing relationships between animals (i.e., the numerator relationship matrix), I 
2 2
is an identity matrix; and cr a, 2 2 CY temale> 0' service sire and cr e are the variation attributable to additive genetic 
effects, permanent environmental effects of the female and her mate and environmental (residual) 
effects respectively. 
122 
Random terms were added to analytical models sequentially. Likelihood Ratio Tests (LRT) were 
performed to assess the significance of the contribution of each random term to improvements in the 
model for analysis. The LRT is based on testing twice the increase in Log-likelihood resulting from 
2 
adding a random term to the model of analysis as a Chi statistic. Alternatively, for two models with the 
same number of random terms, and assuming identical fixed effects modeling, the model with the 
higher value for the Log-likelihood fits the data better. Variance components were expressed relative to 
the overall phenotypic variance for each trait to derive estimates of the appropriate variance ratios. All 
analyses included the full pedigree file, consisting of 601 individuals, the progeny of 135 males and 
132 females. 
RESULTS 
Data description 
Female live weight had coefficients of variation ranging from 12.2 % at the beginning of the breeding 
season and 15.3 % at the end of breeding (Table 1) Corresponding coefficients of variation for body 
measurements were below 10 %. Reproductive traits were exceedingly variable, with coefficients of 
variation ranging from 51.6 to 82.5 %. 
Table 1. Descriptive statistics for the traits assessed in ostrich breeding females for the productive years 1990 to 
2001 
TRAIT Number of observations MEAN (SD) CV(%) 
At the beginning of mating: 
Live weight (kg) 1131 115(14) 12.2 
Chest circumference (cm) 846 122 (6) 4.9 
Tail circumference (cm) 849 102 (10) 9.8 
At the end of mating: 
Live weight (kg) 1112 111(17) 15.3 
Chest circumference (cm) 836 128 (12) 9.4 
Tail circumference (cm) 815 100 (8) 0.1 
Reproduction traits: 
Eggs produced (n) 1177 46.3 (26.5) 57.2 
Chicks produced (n) 1177 22.9 (18.9) 82.5 
Hatchability (%) 1089 51.0 (26.3) 51.6 
------- ------·- ---·------
123 
Systematic effects 
The effects of diet, flushing, teasing and a mid-season rest are not considered, since it formed part of 
other studies (Brand et al, 2000; 2003; Chapter 2). Effects due to production year are mediated 
through a number of interacting climatic and husbandry factors. Since year effects tend not to be 
repeatable, these effects are not reported. The length of breeding season affected egg and chick 
production. The production of eggs increased (P:s;0.01) by 0.166 ± 0.016 per day in the period paired 
off. Chick production correspondingly increased (P:s;0.01) by 0.088 ± 0.013 per additional day paired 
off. 
Live weight at the commencement of breeding was relatively stable from 2 years of age to 7 years of 
age, and no significant differences between age groups were evident (Figure 1 ). In subsequent years, 
live weight tended to decline with a further increase in age When live weight at the cessation of 
breeding was considered, there was a suggestion of an increase up to an age of five years, followed by 
a general decline in later years. Differences between age groups were, however, relatively small and 
generally not significant (P~0.05). Linear body measurements showed less variation with female age 
than live weight, and no conclusive age trends were noted. These results are thus not presented. 
125 ------ -
:0::1.  115 - . 
:E 
OJ 
«ii 
:: 
Q) 
> 
:J 
105 
95 ---r-----··- -,-------·------,-- -· ------.,------------i------- --,---------- --T--------r-----·-, 
2 3 4 5 6 7 8 9 10 
Age (year) 
~egi~-~ing D End ] 
Figure 1. Mean (±SE) live weight at the commencement and cessation of mating in female age groups from 2 to 
1o  years. Trend lines were derived from initial analyses where age trends were modeled by cubic splines 
When age effect on egg and chick production were considered, the main contribution in preliminary 
analyses came from the random non-linear component of the spline not conforming to a smooth trend. 
On this basis, it was decided to include age as a fixed effect in the appropriate operational models Egg 
124 
and chick production increased (P:S0.05) from 2-year-old females to 3-year-old females (Figure 2). At 
later ages, egg production tended to remain stable with no significant differences between age groups 
In the case of chick production, on the other hand, there was a suggestion towards a slight decline at 
later ages. 
50 ... --------·-·-·-·--
40 ·---
'2 
c:: 30 
0 
:;::; 
(,) 
::::s 
"'C 20 
.0... . 
a. 
10 
0 r--------r-~----,-~--------------,-------T------r-----------1 
2 3 4 5 6 7 8 9 10 
Age (year) 
[Ii E_ggs -0-Chicks] 
Figure 2. Means (±SE) for egg and chick production in female age groups from 2 to 1O  years. 
Hatchability increased (Ps:0.05) from 2-year-old females to 3-year-old females (Figure 3). A gradual 
decline was observed at subsequent ages. 
60 
50 
20 
10 -1-----.-------.-~----,---~--,-----,-------·-r-----,-------
2 3 4 5 6 7 8 9 10 
Age (year) 
Figure 3. Mean hatchability (±SE) in female age groups from 2 to 1O  years. The trend line was derived from an 
initial analysis, where the age trend was modelled by fitting a cubic spline. 
125 
Random effects 
The likelihood ratio tests indicated that models including additive genetic and permanent environmental 
effects due to the female, fitted the data best for live weight and chest circumference at the beginning 
and end of breeding (Table 2). Only additive genetic effects were required for analyses of tail 
circumference. Models for chick production and hatchability included additive genetic, permanent 
environmental (female) effects along with the additional random effect for service sire. Not 
surprisingly, service sire effects were not significant (P:c::0.05) for egg production. 
Table 2. Log likelihood values for the respective traits under different random effects models, with the best model 
for each trait represented in bold italic figures 
Trait Fixed only + h2 + c2 + h2 + c2 + h2 + c2 + ss2 
(Model 2) (Model 1) (Model 3) (Model 4) 
--~ - ----
At the beginning of mating: 
Live weight (kg) -3457.6 -3100.3 -3104.5 -3094.2 -
Chest circumference (cm) -1914.6 -1827.3 -1821.8 -1819.4 -
Tail circumference (cm) -2138.9 -2088.7 -2097.6 -2088.4 -
At the end of mating: 
Live weight (kg) -3489.9 -3297.4 -3297.0 -3290.2 -
Chest circumference (cm) -1968.4 -1912.3 -1910.6 -1907.7 -
Tail circumference (cm) -2010.4 -1940.5 -1952.7 -1940.3 -
Reproduction traits: 
Eggs produced (n) -4187.1 -4078.6 -4082.5 -4075.9 -4074.8 
Chicks produced (n) -3915.4 -·3788.7 -3789.3 -3783.5 -3779.3 
Hatchability (%) -6349.1 -6260.8 -6256.6 -6255.2 -6236.9 
h2 - Direct heritability c, - Female permanent environment 552 - service sire 
Ratios of variance components 
Estimates of heritability (h2) for female live weight were o 34 at the beginning of mating and 0.20 at the 
end of mating (Table 3). Estimates derived for chest circumference were lower at 0.12, while the h2 for 
tail circumference ranged from 0.30 to 0.38 at the beginning and end of breeding. Estimates of h1 were 
slightly above 0.20 for egg and chick production, and 0.1 O for hatchability. All estimates were 
significant (P:s;0.05), with the exception of h2 values for chest circumference and hatchability. 
126 
Table 3. Variance components and ratios (± SE) for the respective traits assessed in ostrich breeding females 
over the period from 1990 to 2001. 
--
TRAIT VARIANCE COMPONENTS VARIANCE RATIOS 
2 cr2 2 2 h2 
·---------------- ----
c cr ss c2 ss2 
--
CT A CT E 
----
Beginning of mating: 
Live weight (kg) 62.2 66.3 ·- 55.5 0.34±0 11 * 0.36±0.10 -
Chest circumference (cm) 4.25 11.7 - 20.1 0.12±0.07 0.32±0 07 -
Tail circumference (cm) 19.1 - -- 44.6 0.30±0.04* - -
End of mating: 
Live weight (kg) 43.5 69.9 - 102 020±0.09* 0.32±0.08 -
Chest circumference (cm) 5.39 10.1 - 28.6 0.12±0.07 0.23±0.07 -
Tail circumference (cm) 22.0 - -- 35.9 0.38±0.05* - -
Reproduction traits: 
Eggs produced (n) 121. 97.9 - 314 0.23±0.09* 0.18±0.08 -
Chicks produced (n) 64.1 57.1 28.28 176 0.20±0.08* 0.18±0.08 0.09±0.04 
Hatchability (%) 74.0 153 162.39 336 0.10±0.07 0.21±0.08 0.22±0.05 
: Effect not considered or not significant cr2A: Direct additive variance component 
cr 2c: Female permanent environmental variance component cr2 88: Service sire variance component 
cr2E: Environmental variance h2 : Direct heritability; * = P:o;0.05 
c2 : Female permanent environment ss 2 Service sire 
Permanent environmental variation (c2 ) accounted for between 0.23 and 0.36 of the phenotypic 
variation for chest circumference and live weight Estimates of c2 effects for the reproduction traits 
were relatively stable, at approximately 0.20 for egg and chick production. Service sire (ss), as 
expected, only exerted a significant influence (P:::::0.05) on hatchability and subsequently chick 
production, accounting for between 0.22 and 0.09 of the observed phenotypic variation m these traits. 
The estimates of h2 (upper value) and c2 (lower value) are represented by bold figures on the diagonal 
in Table 4. The genetic (upper value) and permanent environmental (lower value) correlations are 
given above the diagonal. The environmental (upper value) and phenotypic (lower value) correlations 
are given below the diagonal. The estimates of h2 and c2 for live weight at the beginning of mating in 
the three-trait analyses with final live weight and egg or chick production closely resembled estimates 
derived from the two-trait analyses (Table 4). Variances for live weight at the end of breeding were 
partitioned somewhat more towards h2 and less towards c2 in the three-trait analyses, although 
changes were not significant (P2'.0.05) according to the appropriate standard errors. Estimates of h2 
and c2 for egg and chick production were within 0.02 of those derived from the two-trait analyses. The 
ss2 estimate for chick production was slightly higher (0.11 ± 0.04) in the three-trait analysis, compared 
to the corresponding estimate obtained from the two-trait analysis. 
127 
Table 4. (Co)variance ratios(± SE) depicting estimates of h2 and c2 as well as genetic, permanent environmental, 
environmental and phenotypic correlations for live weight at the beginning (B) or end (E) of the mating 
season, egg production and chick production, as assessed by three-trait analyses in breeding ostrich 
females. 
TRAIT LIVE WEIGHT (8) LIVEWE IG HT (E) EGG PRODUCTION 
-----
Live weight (B) 0.33 ± 0.11 0.97 ± 0.04 010±0.27 
0.36 ± 0.10 0.90 ± 0.05 -0 07 ± 0.23 
Live weight (E) 0.17 ± 0.03 0.25 ± 0.09 -0.14 ± 0.27 
0.62 ± 0.03 0.26 ± 0.08 -0.04 ± 0.24 
Egg production 0.18 ± 0.03 -0.0 6± 0.04 0.22 ± 0.08 
0.09 ± 0.05 -0 0 8± 0.05 0.19 ± 0.07 
TRAIT LIVE WEIGHT (B) LIVEW EIG HT (E) CHICK PRODUCTION 
Live weight (B) 0.33 ± 0.11 0.97 ± 0.04 0.07 ± 0.28 
0.36 ± 0.10 0.90 ± 0.05 -0.08 ± 0.24 
Live weight (E) 0.17 ± 0.03 0.25 ± 0.09 -0.16 ± 0.28 
0.62 ± 0.03 0.26 ± 0.08 -0.12 ± 0.25 
Chick production 0.08 ± 0.04 -0.11 ± 0.04 0.22 ± 0.09 
0.03 ± 0.05 -0.1 2± 0.04 0.19±0.09 
Correlations 
Correlations between additive genetic and permanent environmental effects for live weight and the 
linear body measurements were high to very high (Table 5). When the same trait was assessed at the 
beginning and end of the breeding season, these correlations did not differ from unity. The genetic 
correlation between chest circumference at the beginning and end of the breeding season went outside 
the boundary of the parameter space (i.e. exceeded one). A similar result was obtained for the 
permanent environmental correlation between live weight and chest circumference at the end of the 
breeding season. Environmental correlations among the live weight and body measurements were 
generally lower (below 0.30). Overall, moderate to strong estimates of phenotypic correlations 
between live weights and body measurements were evident, with the strongest correlations generally 
between the same trait recorded at different points in time. 
The genetic and phenotypic correlations between female live weight or body measurements at the 
commencement of breeding with reproductive traits were generally favourable but low when compared 
to the corresponding standard errors (Table 5). The only genetic correlations between body 
measurements with reproductive traits that actually exceeded their standard errors were those for tail 
circumference measured at the beginning of breeding with egg and chick production. Corresponding 
permanent environmental and residual correlations were low and variable in sign. The environmental 
correlations between live weight and egg production, and the phenotypic correlations between tail 
128 
,, --
circumference with egg and chick production, were positive and significantly (P:s:0.05) different from 
zero. 
The genetic and permanent environmental correlations of live weight or body measurements recorded 
at the end of the breeding season with reproductive traits were low and generally negative (Table 5). 
Corresponding environmental and phenotypic correlations were consistently negative. Environmental 
correlations of live weight with chick production and hatchability were significant (P:s;0.05), but relatively 
low (-0.1 o to -0.12). Phenotypic correlations between live weight and egg production or hatchability 
also reached significance (P < 0.05), at -0.08 and -0.12 respectively. 
Genetic correlations among reproductive traits were very high, and outside the boundary of parameter 
space in the case of egg and chick production (Table 5). The high genetic correlation between egg 
production and hatchability is notable, suggesting that females with a genetically high egg production 
are also likely to produce eggs with a high hatchability. However, correlations between these two traits 
for other random effects were only moderate, contributing to a moderate phenotypic correlation 
between egg production and hatchability overall. Egg and chick production were highly correlated at all 
levels of random effects, as were chick production and hatchability. The ss correlation between 
hatchability and chick production was high, as might be expected, at o. 77 ± 0.09. 
When compared to the two-trait analyses, the three-trait analyses yielded identical correlations 
between initial and final live weight of breeding hens at all levels. Genetic correlations between live 
weight at the commencement of breeding and reproduction were similar in sign, but the magnitude was 
approximately half that of corresponding correlations derived from two-trait analyses in absolute terms. 
Correlations of live weight at the cessation of breeding with reproduction traits were similar in sign and 
within 0.02 in magnitude from those derived from two-trait analyses. Female permanent environmental 
correlations obtained from the three-trait analyses were similar in sign, but consistently smaller in 
magnitude than those obtained from the two-trait analyses. All environmental and phenotypic 
correlations between live weight and reproduction closely resembled those derived from two-trait 
analyses. 
129 
Table 5. Estimates of genetic, permanent environmental, environmental and phenotypic correlations (± SE) 
between traits, as assessed by two-trait analyses in breeding ostrich females. Live weight and body 
traits were recorded at the beginning (8) or end (E) of mating. 
-------
TRAITS CORRELATION 
Genetic Female PE Environment Phenotyp~ 
·---- --
Live weight (B) x 
Chest circumference (B) 0.89 ± 0.09 0.96 ± 0.05 0.30 ± 0.04 0.63 ± 0.03 
fail circumference (B) 0.93 ± 0 06 - 0.13 ± 0.04 0.42 ± 0 04 
Live weight (E) 0.97 ± 0 04 0.90 ± 0.05 017±0.03 0.62 ± 0.03 
Chest circumference (E) 0.80 ± 0.15 0.87 ± 0.11 0.18 ± 0.04 0.49 ± 0.03 
Tail circumference (E) 0.92±005 - 0.05 ± 0.04 0.43 ± 0.04 
Egg production 0.21 ± 0.28 -0.19 ± 0.22 0.17 ± 0.03 0.08 ± 0.04 
Chick production 0.15 ± 0.28 -0.16±024 0.07 ± 0.04 0.03 ± 0.05 
Hatchability 0.15 ± 0.34 0.13 ± 0.28 -0 01 ± 0.04 -0.01 ± 0.05 
·--~-
Chest circumference (B) x 
Tail circumference (B) 0.93 ± 0.06 - 0.12 ± 0.04 0.32 ± 0.04 
Live weight (E) 0.83 ± 0.14 0.95 ± 0.08 0.11±0.04 0.48 ± 0.03 
Chest circumference (E) 108±0.15 0.92 ± 0.09 0 14 ± 0.04 0.45 ± 0.03 
Tail circumference (E) 0.88 ± 0.10 - 0.03 ± 0.04 0.28 ± 0.04 
Egg production 0.09 ± 0.37 0.01 ± 0.20 0.01 ± 0.04 0.02 ± 0.05 
Chick production 0.16±0.37 -0.01 ± 0.22 -0.06 ± 0.04 -0.01 ± 0.05 
Hatchability -0.15 ± 0.49 0.01 ± 0.22 -0 05 ± 0.05 -0.04 ± 0.05 
--
Tail circumference (B) x 
Live weight (E) 0.94 ± 0.08 - 0.03 ± 0.04 0.31 ± 0.04 
Chest circumference (E) 0.87 ± 0.10 - -0 01 ± 0.04 0 23 ± 0.04 
Tail circumference (E) 0.93 ± 0.05 - 0.04 ± 0.04 0.35 ± 0.04 
Egg production 0.25±0.17 - 0.05 ± 0.04 0.09 ± 0.04 
Chick production 0.26 ± 0.17 - 0.03 ± 0.04 0.08 ± 0.04 
Hatchability 0.21 ± 0.24 - -0.00 ± 0.04 0.04 ± 0.05 
Live weight (E) x 
Chest circumference (E) 0.80±016 1.05±0.10 0.14 ± 0.04 0.48 ± 0.03 
Tail circumference (E) 0.94 ± 0.06 - 0.11 ± 0.04 0.42 ± 0.04 
Egg production -0.12 ± 0.31 -0.11 ± 0.21 -0 04 ± 0.04 -0 08 ± 0.04 
Chick production -0.14 ± 0.30 -0.18 ± 0.23 -0.10 ± 0.04 -0.12 ± 0.04 
Hatchability -0.26 ± 0.37 -0.08 ± 0.22 -0.12 ± 0.04 -0.12 ± 0.04 
--~-
Chest circumference (E) x 
Tail circumference (E) 0.85±0.11 - 0.11±0.04 0.31 ± 0.04 
Egg production 0.03 ± 0.37 0.03 ± 0.23 -0.02 ± 0.04 000±0.04 
Chick production 0.11 ± 0.37 -0.07 ± 0.25 -0 04 ± 0.04 -0 02 ± 0.04 
Hatchability -0 11 ± 0.46 -0.14 ± 0.25 -0.02 ± 0.05 -0.05 ± 0.04 --
Tail circumference (E) x 
Egg production 0.10±017 - -0.06 ± 0.04 -0.01 ± 0.05 
Chick production --0.01 ± 0.16 - -0 04 ± 0.04 -0 03 ± 0.05 
Hatch ability -0.10±0.23 - -0.06 ± 0.05 -0 05 ± 0.05 
Egg production x 
Chick production 1.16 ± 0.11 0 71 ± 0.11 0.75 ± 0.02 0.78 ± 0.02 
Hatch ability 0.80 ± 0.26 0.24 ± 0.25 0.22 ± 0.04 0.29 ± 0 04 
-- --
Chick production x 
Hatchability 076±0.17 0.82 ± 0.11 ___0_ 7_7 ±_ 0_.09_ __ 0.74 ± 0.02 :._____ 
--
Female permanent environment not s1gnif1cant in both traits 
130 
DISCUSSION 
Data description 
Previous studies on ostriches have reported coefficients of variation of 11 to 12 % for live weight at 
slaughter (Cloete et al., 1998; Bunter et al., 1999) and between 12 and 13 % for mature live weight 
(Van Schalkwyk et al., 1996; Lambrechts et al., 1999). These values are relatively low compared to 
coefficients of variation for weight in other species (-20%), perhaps due to selective recording of 
breeding animals only. Chest and tail circumferences of two-year old females had coefficients of 
variation of slightly below 10 % (Lambrechts et al., 1998). Low coefficients of variation for these traits 
could also suggest a developmental or structural optimum in these body dimensions at a given live 
weight. 
Ostrich reproductive traits have previously been shown to be highly variable. Coefficients of variation 
ranged between 45 and 52 % for annual egg production and between 61 and 81 % for annual chick 
production (Van Schalkwyk et al., 1996; Cloete et al., 1998; Lambrechts et al., 1999; Bunter et al., 
2001). Deeming (1996) reported a coefficient of variation of 58 % for weekly egg production, and of 
66% for hatchability for breeding ostriches in the United Kingdom. Despite the observed !arge 
coefficients of variation, the present data still largely conformed to normality, with low levels of 
skewness and kurtosis. Similar conclusions were made by Van Schalkwyk et al. (1996) and Bunter 
(2002), on smaller subsets of the present data. 
Systematic effects 
No comparable results pertaining to the effects of age on ostrich female live weight at the 
commencement or cessation of breeding were found in the literature The present results suggest that 
ostrich females are fairly mature in terms of live weight when they are mated for the first time at 2 years 
of age, as live weight generally did not increase much with breeding age. In poultry, Leghorn pullets 
are expected to increase from 1250 to 1350 g at the commencement of laying to 1650 g at peak 
production to ensure sufficient reserves for the production process (Summers, 1995). Expressed as a 
percentage of the initial live weight, it indicates increases of between 22 and 32 %. Ostrich females 
obviously do not achieve anything remotely in this order, but it has to be conceded that the production 
process in ostriches are much less intensive than in commercial poultry. In other species where 
females are maintained over consecutive breeding seasons, live weight generally increases markedly 
with age until a relatively stable mature live weight is attained. In females maintained for a longer 
period, live weight would subsequently decline. Further studies on the effect of age on the live weight 
of ostrich females are indicated, including guidelines pertaining to the growing out of replacement birds 
as well as further studies on the nutrition of breeding birds. Brand et al. (2003) found that live weight 
was reduced in breeder birds maintained on energy concentrations of 7 5 and 8.5 MJ/kg ME, while egg 
production was only impaired on the former diet. 
Previous results pertaining to egg and chick production of ostrich females showed a significant 
increase from 2 to 3 years of age (Cloete et al., 1998; Bunter 2002). Two-year-old females took longer 
to commence laying than older hens, resulting in a shorter duration of lay in the study of Bunter (2002) 
More ( 1996) correspondingly related low productivity levels in ostriches kept in Queensland (Australia) 
to a large proportion of young non-laying females in the breeding operations Peak egg and chick 
production levels were achieved at 8-9 years of age (Cloete et al., 1998, Bunter, 2002), when 
reproduction was recorded at higher ages than in the present study. These levels of peak production 
were followed by a general decline in reproductive performance at older ages, which was more 
pronounced for chick than it was for egg production. Bunter (2002) reported divergence for egg 
production and chick production at older ages. A corresponding tendency was evident in the present 
study. The age trend obtained for hatchability in the present study refutes a previous allegation that 
this trait is largely independent of female age (Cloete et al., 1998). It is, however, in close agreement 
with a more recent trend reported by Bunter (2002), involving a larger data base, and may be related to 
changes in egg weight and shell quality with female age, which ultimately influence hatchability of eggs 
under a constant incubation environment. 
From recent results, there is little motivation for ostrich females to be kept in the breeding flock for 
more than 10-11 years (Bunter, 2002). Although the oldest age groups in the latter study were still 
capable of a good egg production, chick production declined with female age above 1O  to 11 years of 
age. Average egg and chick weights also showed a decline for females aged 10 years and older. 
Bunter (2002) suggested that this could lead to higher chick mortality levels in commercial breeding 
systems. In this regard, it has been demonstrated that lighter day-old chicks were more likely to 
succumb before one month of age than heavier contemporaries (Cloete et al., 2001 ). Apart frorn 
considerations pertaining to flock performance, too many age groups will affect genetic progress 
deleteriously, with an increase in the generation interval (Cloete et al., 1998). 
Estimates of genetic parameters 
Estimates of h2 for ostrich live weight at a slaughter age of approximately 14 months ranged between 
0.17 and 0.45 (Bunter et al., 1999; Bunter, 2002). The only previous h2 estimates for mature live 
weight in ostric~es were derived from the current data set, involving fewer records (Cloete et al., 
2002b). Estimates for h2 and c2 thus agreed with those reported in the present study for both initial and 
final live weight. Previous estimates for the repeatability of mature live weight ranged from 0.61 to 0. 72 
(Van Schalkwyk et al., 1996; Lambrechts et al., 1999), corresponding well with estimates derived from 
the present data set (0.69 for initial live weight and 0.52 for final live weight). Combining initial and final 
live weight in a three-trait analysis with reproduction traits resulted in slightly more animal variance 
132 
.. 
being partitioned to h2 in the case of final live weight Information pertaining to the initial live weight 
recordings probably resulted in the slightly higher h2 estimate in this instance. No other estimates were 
available in the literature for comparison with the h2 or c2 of linear body measurements. 
Previous estimates of h2 and repeatabilities for parameters of egg and chick production were derived 
from part of the data set used for the present investigation (e.g. Bunter et al, 2001a, Bunter, 2002, 
Cloete et al., 2002b). Thus, estimates reported here are basically similar to those previously reported, 
confirming moderate heritabilities (h2 - 0.13-0.22) for egg and chick production traits. Parameter 
estimates involving the three-trait analyses including egg or chick production were generally in 
agreement with single trait estimates. The current study also confirms a significant service sire effect 
(ss2 = o.22) on hatchability, but suggests a low heritability (0.10) and only moderate repeatability (0.31) 
for this trait Previously estimated variance ratios for hatchability were 0.01 for h2, 0.35 for female 
permanent environment (c2), and 0.13 for variation due to service sire (ss2) (Bunter, 2002). 
When results from previous studies are considered, a generally higher proportion of the phenotypic 
variation seemed to be partitioned to h2 in the present study, at the expense of lower c2 estimates. 
Changes in genetic parameters were, however, not significantly different to previous estimates due to 
fairly large standard errors. An increase in h2 as more data become available was not entirely 
unexpected, since the pedigreed portion of the population and the number of parents with reproduction 
records in the data increases all the time (Bunter, 2002). When h2, c2 and ss2 variance ratios (where 
applicable) were summed to provide repeatability estimates, these estimates were 0.41 for egg 
production, 0.45 for chick production and 0.53 for hatchability. All estimates closely resembled 
previous estimates ranging from of 0.37 to 0.47, 0.51 to 0.55 and 0.38 to 0.56 respectively (Van 
Schalkwyk et al., 1996; Cloete et al., 1998; Lambrechts et al., 1999). Heritability estimates for ostrich 
reproductive traits were furthermore in accordance with estimates for poultry reviewed from the 
literature (Gowe and Fairfull, 1995). When the marked genetic improvement in the reproduction of 
poultry are considered, it is clear that similar successes could be expected from well-constructed 
breeding programs involving ostriches. 
The confirmed influence of service sire on the reproductive performance of companion females agrees 
with results reported by Bunter et al. (2001a), Bunter (2002) and Cloete et al. (2002b). It can thus now 
be accepted that the influence of service sire upon egg production of females is limited. Service sires 
do, however, affect chick production by exerting an influence on the fertility of eggs (Bunter, 2002) and 
thereby influencjng the hatchability records for individual females 
Correlations 
High to very high genetic correlations were previously reported for live weight of slaughter ostriches 
between six and approximately 14 months of age (Bunter el al, 1999; Bunter, 2002). No genetic 
133 
-,,.. 
correlations between ostrich live weight and linear body measurements were found in the literature. It 
is, however, generally accepted that ostrich live weight can be fairly accurately predicted from functions 
derived from linear body measurements (Bezuidenhout and Van Schalkwyk, 1996; Deeming et al, 
1996). The high genetic and permanent environmental correlations between the same measurements 
made at the commencement and cessation of breeding are readily explicable, as they are the same 
traits but measured at different times during the reproductive cycle. 
It is generally accepted that reproduction and growth traits are unfavourably correlated in small 
domestic poultry species (Marks, 1995; Gowe and Fairfull, 1995). Phenotypic correlations in earlier 
studies by Van Schalkwyk and Cloete (1996) and Van Schalkwyk et al. (1996) seemed to support the 
argument that the same applies in ostriches. It has, however, been argued that these relationships in 
mixed age flocks could have resulted from unproductive birds becoming fatter with an increase in age. 
Phenotypic correlations derived from data on two-year old females (where confounding age-
productivity relationships are not present) also did not conform to this generalization (Lambrechts et al., 
1998). The present study confirms that there are no marked unfavourable genetic correlations 
between female live weight at the beginning of breeding and the measures of reproduction assessed 
here. Phenotypic correlations between initial female live weight and egg production in this study (0.08) 
were similar to previous estimates of 0.04 (Van Schalkwyk et al., 1996; Lambrechts et al., 1999). 
Corresponding correlations for chick production in the literature were generally opposite in sign (-0.04 
and -0.07) but of similar magnitude to that of the present study (0.03). The phenotypic correlation 
between initial live weight and hatchability of -0.16 estimated by Van Schalkwyk et al (1996) was 
much larger in magnitude than the present correlation, which is based on a much larger and more 
informative data set. 
Estimates of genetic correlations between live weight recorded at the end of the breeding season and 
reproductive traits were negative in sign, but not significantly different from zero. This change in sign 
compared to live weight at the commencement of breeding is difficult to explain, given the near unity 
genetic correlation between live weight recorded at the respective stages. It has to be considered that 
sampling issues could have caused this result. The lack of an influence due to the inclusion of both 
initial and final live weight together with egg or chick production on the genetic correlations of final live 
weight with reproduction in three-trait analyses does not support this line of reasoning. From these 
analyses it appears that the negative genetic covariance between final live weight and reproduction is 
real. Much more data are, however, required for the estimation of genetic correlations with sufficiently 
low standard errprs to confirm or refute the contentions listed above. 
On the female PE level, live weight was generally negatively related to reproduction. These 
correlations were somewhat smaller when assessed in three-trait analyses. It could be argued that the 
negativ,:; sign of these covariance components could be related to the demands of egg production on 
the metabolism of the breeding female being carried over to subsequent seasons. In other species, 
134 
correction for previous reproduction were found to impact on the magnitude of the female permanent 
environmental correlation between live weight and reproduction (Cloete et al., 2004). In the study of 
L.ambrechts et al. (1999), it was also shown that live weight change within a breeding season was 
negatively related to egg (-0.20) and chick production (-0.17) on a phenotypic level. Overall, the 
derived genetic and permanent environmental correlations resulted in a low but significant negative 
association of reproductive traits and hatchability with final live weight on the phenotypic level. 
Genetic correlations between chest circumference and reproductive traits tended to be similar to those 
between live weight and reproductive traits, but of lower magnitude. Phenotypic correlations of chest 
circumference with reproductive traits were essentially zero and consistent with those available in the 
literature (Van Schalkwyk and Cloete, 1996; Lambrechts et al, 1998). On a genetic basis, there was 
evidence of a stronger positive relationship between tail circumference and reproductive traits, 
although significance could not be demonstrated. Initially, significant negative phenotypic correlations 
of this trait measured at the beginning of the breeding season with reproductive traits were thought to 
hold promise with regard to the prediction of the reproductive capacity of ostrich females (Van 
Schalkwyk and Cloete, 1996). These correlations were -0.20 tor egg production, -0 12 for chick 
production and -0.22 for hatchability. Phenotypic correlations in the present study were, however, 
much lower in magnitude and different in sign in the majority of cases. The differences in the sign and 
magnitude of the phenotypic correlations between studies are most likely to be associated with 
differences in the age structure of the breeding flock. During the time Van Schalkwyk and Cloete 
( 1996) conducted their study, most breeding females were older than 1O  years, with ages exceeding 
20 years in some instances. Breeding females in the present flock were mostly younger than 12 years 
of age. When only two-year old females were considered, there was little evidence of a strong 
phenotypic relationship between tail circumference and reproductive performance (Lambrechts et al., 
1998). This can be ascribed to the two-year old females becoming sexually mature later in a breeding 
season, resulting in the number of eggs produced being generally low because they come into lay later 
than their older contemporaries. 
The high correlations between reproductive traits were not unexpected. Previous studies suggested 
high correlations between additive genetic (Bunter, 2002; Cloete et al., 2002b), and female permanent 
environmental (Bunter et al., 2001 b; Bunter, 2002; Cloete et al., 2002b) effects, culminating generally 
in moderate to high phenotypic correlations (Van Schalkwyk et al., 1996; Lambrechts et al., 1999; 
Bunter, 2002). Lower environmental and phenotypic correlations between egg production and 
hatchability wer? due to the influence of service sire effects on the level of egg fertility. Service sire 
solutions for infertility were shown to be unfavourably related to chick production, suggesting that 
service sire infertility or mate incompatability deleteriously affected the chick production of companion 
females (Bunter et al., 2001b, Cloete et al., 2002b). The favourable correlation between service sire 
effects for hatchability and chick production in the present study supports this line of reasoning. 
135 
CONCLUSIONS 
Significant genetic parameters for egg and chick production of commercial ostriches were evident in 
this study. Genetic progress in reproductive traits thus seems feasible in this livestock species. When 
it is considered that the production of leather and meat constitutes the outputs of any commercial 
ostrich operation (Cloete et al., 1998; 2002a), it is evident that reproductive performance is of 
paramount importance in a commercial breeding program. Comparatively high repeatability estimates 
also suggest that worthwhile progress is also achievable in the current flock (Bunter et al., 2001 a). 
Moreover, reproduction does not seem to be unfavourably related to live weight prior to breeding on a 
genetic or phenotypic level. This result seems to open the avenues for the development of a dual-
purpose ostrich strain, where emphasis may be placed on selection for both growth and reproductive 
traits. Other traits, like hide attributes, still need to be considered for the development of a breeding 
strategy that addresses all significant traits in the breeding goal. 
It is important to state that the material used for the present study, as well as the bulk of research cited, 
was essentially derived from the same resource population. The results may not necessarily be robust 
for extension to other populations and regions (Bunter, 2002; Cloete et al., 2002a). There is an urgent 
need for similar studies on other resource populations to confirm or refute the results obtained from the 
present study. 
REFERENCES 
BEZUIDENHOUT AG & VAN SCHALKWYK SJ, 1996. Prediction of body weight and skin area from various 
body measurements in South African ostriches. In: Improving our understanding of ratites in a 
farming environment. Ed: D.C. Deeming. Proc. 1st /nternat. Ratite Cont., 27-29 March 1996, 
Manchester, England. pp. 5-6. 
BRAND Z, BRAND TS, BROWN CR & VAN SCHALKWYK SJ, 2000. Preliminary results of the dietary protein 
and energy levels on the production of female breeding ostriches. S. Afr. J Anim. Sci. (Suppl) 
30: 19-20. 
BRAND Z, BRAND TS & BROWN CR, 2003. The effect of energy and protein levels on production in 
breeding female ostriches Brit. Poult. Sci. 44: 598-606. 
BUNTER KL, 2002. The Genetic Analysis of Reproduction and Production Traits Recorded for Farmed 
Ostriches' ( Struthio came/us). Ph.D. dissertation, University of New England, Armidale, 
Australia. 
BUNTER KL, CLOETE SWP & VAN SCHALKWYK SJ, 1999. Significant genetic parameters for egg, chick 
and Juvenile weight traits in ostriches. Proc. Assoc. Adv. Antm. Breeding Gen. 13: 476-479. 
136 
BUNTER KL, CLOETE SWP, VAN SCHALKWYK SJ & GRASER H-U, 2001a. Factors affecting reproduction In 
farmed ostriches. Proc. Assoc. Adv. Anim. Breeding Gen. 14: 43-46. 
BUNTER KL, CLOETE SWP, VAN SCHALKWYK SJ & GRASER H-U, 2001 b. Relationships between 
measures of reproductive performance in farmed ostrich hens. Proc. Assoc. Adv. Anim. 
Breeding Gen. 14: 47-50. 
CLOETE SWP, BUNTER KL & VAN SCHALKWYK SJ, 2002a. Progress towards a scientific breeding 
strategy for ostriches. Proc. lh World Congr. Gen. Appl. Livestock Prod. 30: 561-568, 18-23 
August, Montpellier, France. 
CLOETE SWP, BUNTER KL, VAN SCHALKWYK SJ & BRAND Z, 2002b. Genetic parameters for live weight 
and reproduction in ostrich females. Proc. 1st Joint Congr Grassland Soc. S. Afr. and S. Afr. 
Soc. Anim. Sci., 13-16 May 2002, Christiana Aventura, p. 81. 
CLOETE SWP, GILMOUR AR, OLIVIER JJ & VAN WYK JB, 2004. Genetic and phenotypic trends and 
parameters in reproduction, greasy fleece weight and live weight in Merino lines divergently 
selected for multiple rearing ability. Austr. J. Exp. Agric. 44(3), in press. 
CLOETE SWP, LAMBRECHTS H, PUNT K & BRAND Z, 2001. Factors related to high levels of ostrich chick 
mortality to 90 days after hatching in an intensive rearing system. J. S. Afr. Vet. Assoc. 72: 
197-202. 
CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998. Ostrich breeding - progress towards a 
scientifically based strategy. Proc. 2nd lnternat. Scient. Ratite Congr., 21-25 September 1998, 
Oudtshoorn, pp. 55-62. 
DEEMING DC, 1996. Production, fertility and hatchability of ostrich ( Struthio came/us) eggs on a farm in 
the United Kingdom. Anim. Sci. 63: 329-336. 
DEEMING DC & AR A, 1999. Factors affecting the success of commercial incubation. In: The Ostrich: 
Biology, Production and Health, pp. 275-292. Ed: Deeming, D.C. GABI Publishing, 
Wallingford, Oxon, United Kingdom. 
DEEMING DC, SIBL Y RM & MAGOLE IL, 1996. Estimation of ostrich mass and condition from body 
measurements. In: Improving our understanding of ratites in a farming environment. Ed: D.C. 
Deeming. Proc 1st lnternat. Ratite Conf., 27-29 March 1996, Manchester, England, pp. 6-7. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3. NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
GILMOUR AR, THOMPSON R & CULLIS BR, 1995. Average information REML, an efficient algorithm for 
variance parameter estimation in linear mixed models. Biometrics 51: 1440-1450. 
GowE RS & FAfRFULL RW, 1995. Breeding and genetics of egg laying chickens. In: Poultry 
Production, pp. 435-456. Ed P Hunton, World Animal Science: Volume 9, Sub-series C: 
Production System Approach, Elsevier, The Netherlands. 
LAMBRECHTS H, CLOETE SWP & BRAND Z, 1999. Live weight as an indicator of egg and chick 
production in commercial ostrich ( Struthio came/us) breeding systems Proc. Annual Congr. S. 
Afr Soc. Anim. Sci. 37· 41-42. 
137 
LAMBRECHTS H, CLOETE SWP, VAN SCHALKWYK SJ & BRAND Z, 1998. Phenotypic relationships of body 
measurements with egg production of 2-year old ostriches. Proc. 2nd lnternat. Scient. Ratite 
Congr., 21-25 September 1998, Oudtshoorn, pp. 72-74. 
LAMBRECHTS H, SWART D, CLOETE SWP & GREYLING JPC, 2002. Preliminary results on teasing and 
flushing practices for breeding ostriches. Proc. 2nd World Ostrich Congr., 26-29 September 
2002, Warsaw, Poland, Ed: J.O. Horbanczuk, pp. 209-210. 
MARKS HL, 1995. Genetics of growth and development. In: Poultry Production, pp. 157-188. Ed: P 
Hunton, World Animal Science: Volume 9, Sub-series C: Production System Approach, 
Elsevier, The Netherlands. 
MORE SJ, 1996. The performance of farmed ostrich hens in Australia. Prev. Vet. Med. 29: 107-120. 
PETITTE JN & DAVIS G, 1999. Breeding and genetics. In: The Ostrich· Biology, Production and Health, 
pp. 275-292. Ed: Deeming, D.C. CABI Publishing, Wallingford, Oxon. United Kingdom. 
SUMMERS JD, 1995. Nutrition and poultry reproduction. In: Poultry Production, pp. 311-358. Ed: P 
Hunton, World Animal Science: Volume 9, Sub-series C: Production System Approach, 
Elsevier, The Netherlands. 
VAN SCHALKWYK SJ, 1998. Improvement of fertility and hatchability of artificially incubated ostrich eggs 
in the Little Karoo. M.Sc. thesis, Rhodes University, Grahamstown, South Africa. 
VAN SCHALKWYK SJ & CLOETE SWP, 1996. Body measurements and appearance of ostriches in 
relation to reproduction. In: Improving our understanding of ratites in a farming environment 
51 
Ed: D.C. Deeming. Proc. 1 lnternat. Ratite Conf., 27-29 March 1996, Manchester, England. 
pp. 81-83. 
VAN SCHALKWYK SJ, CLOETE SWP, BROWN CR & BRAND Z, 2000. Hatching success of ostrich eggs in 
relation to setting, turning and angle of rotation. Brit. Poult. Sci. 41: 46-52. 
VAN SCHALKWYK SJ, CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations for 
live weight and reproduction in commercial ostrich breeding pairs. Brit. Poult. Sci. 37: 953-
962. 
VERBYLA AP, CULLIS BR, KENWARD MG & WELHAM SJ, 1999. The analysis of designed experiments 
and longitudinal data using smoothing splines. J. R. Stat. Soc., Series C 48: 269-311. 
138 
Chapter 10 
Semen quality of ostrich males (Struthio came/us var. 
domesticus) in relation to fertility in a commercial 
breeding system 
Abstract 
Phallus traits and semen quality were determined in 421 ostrich males, before the onset and at the end of breeding, 
and for four consecutive breeding seasons. Phalluses could be extracted in 386 males. Aspermic ejaculates, and 
urine and faeces contamination were recorded in 72 and 64 cases, respectively. Asperm1c eiaculates were more 
prevalent before the onset of breeding. Phallus length and circumference were normally distributed, and moderately 
repeatable (0.30 and 0.28, respectively) and correlated on a permanent environment, environmental, and phenotypic 
level. Phallus size tended to increase with an increase in age, and was positively correlated to sperm concentration. 
Sperm concentration and percentage abnormal spermatozoa were exceedingly variable. Sperm concentration 
increased with an increase in age, but tended to decline in males aged 6 years and older. The percentage of 
abnormal sperm per ejaculate was not influenced by age, while sperm concentration was moderately repeatable 
(0. 18), and the percentage abnormalities lowly repeatable (0. 14). Ejaculate colour and viscosity increased linearly 
with an increase in age, and ejaculate colour tended to be lowly repeatable (0.05). Viscosity of ejaculates obtained 
from ostrich males subjected to flushing management practices, were higher than that obtained from males not 
subjected to flushing management practices. Sperm concentration was positively and significantly correlated with 
fertility at the end of breeding. Fertility of 75% and higher was reported for males between 5 and 8 years of age, and 
tended to decline in males older than 8 years. The most common sperm abnormalities included damage to the 
acrosome, swollen heads, and broken and loose tails. Aggregations of sperm were observed in samples obtained 
from different males, and possibly represent a form or mechanism of sperm competition in ostriches. No reliable 
male trait could be identified that can be used as a reliable indicator of male fertility in commercial breeding flocks. 
Traits that can aid in the development of a semen quality index that can be used in commercial breeding program to 
select potentially fertile ostrich males, need to be identified. The role of endocrine and behavioural mechanisms, 
such as sperm competition and cryptic female choice, that may contribute to an individual's reproductive success, 
also warrants further investigation. 
INTRODUCTION 
Puberty in the avian male can be defined as the age when spermatozoa first appear in the 
ejaculate. Puberty and sexual maturity, however, is not synonymous. Sexual maturity is defined 
as the time when testis growth reaches its maximum, and the number and quality of spermatozoa 
is at a peak (Howarth, 1995). Ostrich males are considered sexually mature at approximately 3-4 
years of age, with factors such as breed and time of the season when hatched, influencing the 
onset of sexual maturity (Hicks-Aldredge, 1994; Smith et al., 1995; Horbariczuk and Sales, 1999). 
Spermatogenic activity, however, has been reported for prepubertal ostriches, with 
139 
spermatogenesis undergoing seasonal changes as observed in mature breeding ostrich males 
(Madekurozwa et al., 2002) . 
The ostrich male has an intromittant phallus which lies folded up into the proctodeum of the cloaca. 
The ostrich phallus is characterised by a dorsal sulcus or groove that runs on the dorsal surface, 
and the absence of a urethra. The dorsal groove directs the semen into the cloaca of the female 
during copulation (Fowler, 1991). In the mature ostrich male, phallus length varies from 20cm 
when flaccid , to 40 cm when erect. Kreibich and Sommer (1995) reported phallus lengths of 
between 25 to 39 cm for immature ostrich males. When erect, the ostrich phallus projects from the 
cloaca in a ventro-cranial curve, and deviates to the left due to the presence of asymmetrical 
fibrolymphatic bodies (McCracken, 2000; Fowler, 1991). 
Several factors may influence a male's fertility, and these may include genetic, behavioural and 
environmental aspects. Soley and Roberts ( 1994) determined that ostrich spermatozoa are of the 
sauropsid type that are characteristic of non-passerine birds, and in general resembles chicken, 
turkey, guinea fowl , tinamou and budgerigar spermatozoa. A variety of sperm defects are known 
that adversely affect fertility in domestic animals, with defects well documented by Oettle and Soley 
(1988), Hafez (1987), Marquez and Ogasarawa (1975), Saeki and Brown (1962), and Saeki (1960). 
Various studies have reported parameters used for ostrich semen evaluation (Soley et al., 1996; 
Hemberger, 1996; Irons et al., 1996; Irons, 1995; Bertschinger et al., 1992; Hicks, 1990; Von 
Rautenfeld, 1977). Studies performed by Soley (1992; 1996) and Von Rautenfeld (1977) also 
reported on the histology of the ostrich male reproductive tract and ostrich sperm morphology. 
Very little, however, is known about the relationship between phallus traits and various ejaculate 
parameters, and the relationship with fertility of eggs produced by companion females. This study 
investigated the relationship between male phallus traits and semen quality, and the fertility of eggs 
produced by companion females under commercial breeding conditions, as well as the influence of 
age, season, and management on the phallus and sperm traits measured in ostrich males. 
MATERIALS AND METHODS 
Experimental animals 
Ostrich males maintained in a breeding flock (n=136 breeding pairs) at the Little Karoo Agricultural 
Development Centre outside Oudtshoorn, South Africa, were used in the study. Management of 
the breeding flock have been documented by Van Schalkwyk et al. (1996), and Bunter and Graser 
(2000) . The ages of the males used in the study ranged between 2 and 10 years. During the 
second and third year of the study, the breeding pairs were subjected to flushing and teasing 
management practices (Materials and Methods, Chapter 2) . 
140 
Measurement of phallus length and circumference 
Phallus length and circumference were measured with a tape measure. The phallus was everted 
from the proctodeum and extended distally, and phallus length was obtained by measuring the 
distance (in cm) from the opening of the dorsal groove {A, Figure 1) to the distal tip of the phallus 
(B, Figure 1) . Phallus circumference was obtained by measuring the circumference (in cm) , as 
indicated by C and D (Figure 1) . 
Figure 1 
Figure 1 and Plate 1. The drawing on the left represents an ostrich phallus, with A, B, C and D indicating the 
measurement points for phallus length and circumference (adapted from Jensen et al., 1992). Plate 1 
represents a post-mortem photograph of an ostrich phallus, indicating the deviation of the phallus to the left-
hand side when everted (photograph taken from the dorsal side). 
Semen collection and evaluation 
Due to the lack of sexual imprinting in the ostrich males used in this study, the forced massage 
method, as described by Soley (1992) , with a minor modification, was used to collect the semen 
samples. Each male was hooded and injected with 5 IU of oxytocin 5 minutes prior to stimulation 
and collection (Bertschinger et al., 1992). In several cases it was found that the reaction period of 
5 minutes was too short to allow collection to be satisfactory performed. Instead of administering a 
second dose of oxytocin, the period between injection and collection was extended to 7-10 minutes 
and this proved to be sufficient to allow successful collection in most cases. Semen samples were 
obtained frail) ostrich males before the onset and at the end of the 1999/2000, 2000/2001 , 
2001/2002 and 2002/2003 breeding seasons. 
Macroscopic evaluation 
Semen samples were macroscopically evaluated on a scale of 0-4 for colour, and on a scale of 1-5 
for viscosity (Table 1). Ejaculate colour ranged from creamy white (4) to almost transparent (0) , 
and ejaculate viscosity ranged from thick creamy (5) to watery (1 ). Viscosity and colour of 
141 
ejaculates were only evaluated during the 2000/2001, 2001/2002 and 2002/2003 breeding 
seasons. 
Table 1. A scale for the macroscopic assessment of colour and viscosity score of ostrich semen samples. 
TRAIT SCORE TRAIT SCORE 
Colour Viscosity 
Creamy white colour 4 Thick creamy 5 
Creamy colour 3 Creamy 4 
Light grey 2 Thin creamy 3 
Light brown 1 Milky 2 
Almost transparent 0 Watery 1 
After macroscopic evaluation, nigrosin-eosin (1-% eosin and 5% nigrosin in 3% sodium citrate) 
smears were prepared by adding 1 drop of the staining reagent to 1 drop of semen. The sperm-
nigrosin-eosin droplet was gently mixed and after approximately a minute a drop of the mixture was 
transferred to a clean microscopic slide and a smear was made. Each smear was allowed to air-
dry and stored for later analysis. 
Microscopic evaluation 
Spermatozoa with the head gently curved, crescent-shaped or convoluted, and that formed a 
complete unit with the mid-piece and tail, were regarded as normal, viable spermatozoa 
(Hemberger et al., 2001; Irons et al., 1996; Soley and Roberts, 1994; Soley, 1992). Spermatozoa 
exhibiting any of the following defects were regarded as abnormal: 
-) Swollen heads, i.e. bulb-headed spermatozoa 
-) Broken head-mid-piece junctions, i.e. crooked neck spermatozoa 
-) Damage to the acrosome, i.e. translucent head where nucleus could be observed under 
segments of remaining acrosome 
-) Damage to the tails, i.e. bent or broken tails 
-) Dead spermatozoa, i.e. spermatozoa completely stained by the staining reagent 
Male fertility was not directly determined, and therefore the fertility of eggs produced by a male's 
companion female was regarded as the fertility of that specific breeding male. Only breeding pairs 
where females produced 5 or more settable eggs, were considered in the calculation of fertility. 
Fertility of eggs produced by companion females was expressed as a percentage, and calculated 
according to the following equation: 
Fertility (%) = [(number of eggs set-number of infertile eggs)/number of eggs set] X 100 
The analySIS™ program (Version 3.00; Soft Imaging System GmbH, Germany) was used to 
determine the number of sperm per ejaculate, and to evaluate sperm morphology. Nine fields were 
counted and averaged to obtain a representative count for each smear Counts were performed at 
X200 magnification, and examination of abnormalities were performed at X400 magnification. 
142 
Statistical analysis 
Data obtained on an all-or-none scale (i.e. whether it was possible to extract or not extract a 
phallus, contamination of semen with urine or faeces or no contamination, and asperm'1c or no 
2 
aspermic samples) were expressed as proportions and analysed by standard Chi procedures 
(Siegel, 1956). Proportions at the commencement of breeding were compared to those at the 
cessation of breeding, using 2 X 2 analysis of variance. Analyses of the phallus traits and semen 
quality were complicated by the fact that the same individuals were sampled repeatedly. The 
covariance arising from the repeated sampling of the same animal was accounted for by the 
inclusion of random animal effects in the mixed models used during the analyses. In view of the 
fact that only 192 to 232 observations, depending on the trait considered, were available for 
analysis, it was not attempted to partition the between animal effects into direct genetic and 
permanent environmental components. 
ASREML (Gilmour et al., 1999) was used for the analyses. The software is suitable for the 
analysis of mixed models, and allows for the estimation of variance components of selected 
random effects, as well as the prediction of least squares means for specific fixed effects. Fixed 
effects included in the present study included breeding season, management practice (flush 
feeding and/or teasing), and stage of the breeding season when ejaculates were obtained (before 
onset of or end of breeding). Trends with regard to animal age were longitudinal by definition and 
modelled by using cubic splines (Verbyla et al., 1999). The cubic splines consisted of a fixed linear 
component and a random non-linear component depicting random deviations conforming to a 
smooth trend. Initially, random deviations from linearity not conforming to a smooth trend were 
also fitted, but this component was found to be not significant (P2:0.05) for any of the variables. The 
first analyses involved fitting various combinations of fixed effects, random spline components and 
interactions between them to obtain an operational model. When this model was established, 
random animal effects were added. Between and within animal variance components were used 
for the estimation of repeatability for the respective phallus traits and sperm characteristics. Two-
trait models were subsequently fitted to obtain all relevant between animal, environmental and 
phenotypic correlations between the phallus and sperm traits. 
RESULTS 
It was possible to extract the phalluses of 368 (87.4%) of the 421 males that were sampled during 
the study period, i.e. between 1999 and 2002. No difference was found between males sampled 
successfully at the beginning (86%) Or end (89%) of the breeding season (0.86 VS. 0.89 
respectively). Of the 368 males that were sampled, 64 semen samples (17.4%) were discarded 
due to contamination with urine and/or faeces, with no difference between the proportions at the 
beginning or end of the breeding season. Of the remaining 304 individuals that were sampled, 72 
(23.7%) were found to be aspermic. These proportions differed (P:<00.01) according to the time of 
the breeding season (0.26 vs. 0.13 at the beginning and end of breeding, respectively). 
143 
In the calculation of the descriptive statistics of the anatomical and sperm traits, only data recorded 
for males where both the phallus could be retrieved and an ejaculate with sperm could be obtained, 
were used. Phallus length and circumference were normally distributed, with coefficients of 
variation of 14.1 and 11.3 %, respectively (Table 2). Sperm concentration and the percentage of 
abnormal sperm showed considerable variation. The transformation of sperm concentration to the 
natural logarithm scale assisted to stabilise the variance for this trait. The percentage of abnormal 
sperm was also transformed, for similar reasons, using the arc sine transformation prior to analysis 
(Snedecor and Cochran, 1967). Visually assessed sperm traits (viscosity and colour) were 
normally distributed and fairly variable, with coefficients of variation just exceeding 30%. The 
fertility of eggs produced by companion females varied considerably, ranging between 0 and 100%. 
Table 2. Phallus traits and semen quality parameters (mean and SD) measured in ostrich males during four 
consecutive breeding seasons. 
TRAIT N Mean (SD) Range CV(%) 
Anatomical traits 
Phallus length (cm) 228 270 (3.8) 13 - 35 14.1 
Phallus diameter (cm) 228 18.6 (2.1) 12 - 24 11.3 
Sperm traits 
4
Sperm concentration per ml (X10 ) 232 830 (978) 26.7 - 6588.3 1.18 
Abnormalities (%) 232 23.5 (22.7) 1.0 - 84.4 96.6 
Viscosity (score) 151 4.01 (1.26) 1-5 31.4 
Colour (score) 151 2.84 (1 05) 0-4 36.9 
Fertility(%) 171 74.2 (26 8) 0 - 100 36.1 
Influence of management practice and period of collection 
Almost all phallus and sperm traits were unaffected by the flush feeding and teasing management 
practices. The only exception was the viscosity of ejaculates that were significantly (P~0.05) 
increased by flush feeding (Table 3). Phallus traits varied significantly (P~0.05) between years, 
with phallus measurements obtained at the end of breeding being bigger (P:o;0.01) than 
measurements obtained before the onset of breeding (Table 3). 
Sperm concentration tended to be the highest during the first and last years of the study, and lower 
prior to the onset of breeding (Table 3). The percentage of abnormalities was also considerably 
lower (P~0.01) during 1999 than in any of the other years. The percentage of abnormal sperm also 
tended to be lower prior to the onset of breeding than at the end of a breeding period (Table 3). 
The viscosity ,of ejaculates did not differ between breeding seasons, with viscosity of ejaculates 
obtained prior to the onset of breeding being lower (P~0.01) than viscosity of ejaculates obtained at 
the end of breeding. The colour scores for ejaculates obtained prior to the onset of breeding, 
however, was higher (P~0.01) than that of ejaculates obtained at the end of breeding. The colour 
scores assigned to ejaculates were highest (P~0.01) during the 2000/2001 breeding season, and 
lower during the 2001/2002 and 2002/2003 breeding seasons. 
144 
Table 3. Means (± SE) of phallus length (PL), phallus circumference (PC), and sperm traits measured in 
ostriches, as influenced by management practice, breeding season, and time of collection. 
PHALLUS TRAITS SPERM TRAITS 
PARAMETER 
PL (cm) PC (cm) Sperm concentration Abnormalities Viscosity Colour 
Overall mean 26.9±0.4 18.9±0.2 6.3±0.2 (528.5X1 O') 24.8±2.0 (17.6) 4.1±0.1 2.6±0.1 
Management practice 
Flushing ns ns ns ns * ns 
4
No 26.4±0.6 18.9±0.3 6.3±0.2 (550.0X10 ) 24.1±2.7 (16.7) 3.9±0.2 2.7±0.2 
Yes 27.4±0.4 19.0±0.2 6.2±0.2 (512.9X10
4
) 25.4±1.9 (18.4) 4.3±0.1 2.5±0.1 
Teasing ns ns ns ns ns ns 
4
No 26.8±0.6 18.6±0.3 6 3±0.2 (523.2X10 ) 25.5±2.7 (18.5) 3.9±0.2 2.6±0.1 
Yes 27.1±0.4 19.3±0.3 6.3±0.2 (539.2X10
4
) 24.0±2.0 (16.5) 4.3±0.2 2.6±0.1 
Breeding season * ** ns ** ns ** 
1999/2000 28.2±0.7 17.6±0.4 6.4±0.3 (584.1X10
4
) 10.2±3.0 (3.1) ---- ----
2000/2001 26.3±0.4 19.0±0.2 6.2±0.2 (473.4X10
4
) 28.2±2.1 (22.3) 3.8±0.1 3.4±0.1 
4
2001/2002 25.4±0.5 18.3±0.3 6.2±0.2 (492.7X10 ) 32.7±2.2 (29.2) 3.9±0.1 2.3±0.1 
2002/2003 27.8±0.5 20.8±0.6 6.4±0.4 (584.1 X10
4
) 28.0±4.6 (22.0) 4.5±0.4 2.0.±.0  3 
When collected ** ** ns ns ** 
Before onset of breeding 25.6±0.5 18.0±0.3 6 2±0.2 (473.4X104) 25.9±2.2 (19.1) 3.6±0.2 2.8±0.1 
At the end of breeding 28.3±0.6 19.9±0.3 6.4±0.2 (595.9X104) 23.6±2.4 (16.0) 4.6±0.2 2.4±0.1 
Note: The sperm concentration and percentage abnormal1t1es values represent the transformed In- and arcsine values, 
respectively (non-transformed values in brackets): ns = Not significanl (P20.05); • = Significant (P,;0.05); ** = Highly 
significant (P,;0.01) 
Influence of age on phallus and sperm traits 
The average phallus length and circumference measured in two-year old males was 26.7±0.7 cm, 
compared to 27.1±0.5 cm in 10 year old males (Figure 2). This difference in phallus circumference, 
however, was found to be not significant. 
29.0 
27.0 
E 
~ 
Ql 25.0 
() 
c: 
.Q..l  
2 23.0 
E 
:J 
.(.).  21.0 
:i:..c
3 
..  
Cl 19.0 
c: 
Ql 
..J 
17.0 
15.0 
2 3 4 5 6 7 8 9 10 
Age (years) 
@_P.hallus_l_engt_ti_ D Phallus .9~~~erenc_ij 
Figurn 2. The influence of age on phallus length and circumference (±SE) measured in breeding ostrich 
males. 
145 
Sperm concentrations generally increased (Ps;0.05) with an increase in male age to reach a 
maximum in five-year old ostrich males (Figure 3), and declined in males aged six years and older. 
.E...  
Q) 
a. 
U) 
..... 6.5 
0 
.... 
Q) 
.c 
E .El 
::J .!!! 
c: 
..... ::J 6.0 () 
0 Ill 
E ·w 
..c: .... 
.... Q) 
·c o.. 
Ill 
Cl 5.5 
E. 
i.i..i.  
.:.:.J.  
zIl l 5.0 
2 3 4 5 6 7 8 9 10 
Age (years) 
Figure 3. The effect of age on the sperm concentration (±SE) of ejaculates obtained from breeding ostrich 
males. 
Age groups did not differ in terms of the percentage of abnormal sperm (Figure 4). 
32.0 
E 
[ 30.0 
i"ii'  
E 
0 28.0 
c: 
.D 
Ill 
.~....  26.0 
0 
Ql 
c: 
"iii 
u 24.0 
< 
22.0 
2 3 4 5 6 7 8 9 10 
Age (years) 
Figure 4. The influence of male age on the percentage of abnormalities 1n ejaculates obtained from breeding 
ostrich males. 
Subjective scores of ejaculate viscosity and colour increased (P:00.05) linearly with an increase in 
age (Table 4). 
146 
Table 4. Relationship between age, and colour and viscosity scores (±SE) recorded for ostrich semen 
samples 
Age (years) Colour score Viscosity score 
[Scale=1-4] [Scale=1-5] 
2 1.9±0.2 3.9±0.2 
3 1.9±0.2 3.9±0.2 
4 2. 1±0.2 4.1±0.2 
5 2.2±0 2 4.1±0.1 
6 2.3±0. 1 4.2±0.2 
7 2.3±0. 1 4.2±0.2 
8 2.4±0. 1 4.3±0.2 
9 2.5±0.2 4.4±0.2 
10 2.5±0.2 4.4±0.2 
Fertility of eggs produced by companion females generally increased when the females were 
paired with males aged 3 years and older (Figure 5). Fertilities of 75% and higher were reported 
for males between 5 and 8 years of age. Fertility tended to declined below 75% in ostrich males 
older than 8 years. 
75.0 
~ 
~ 
~ 
:e 
Q) 
u. 
65.0 
2 3 4 5 6 7 8 9 10 
Age (years) 
Figure 5. The effect of male age on the fertility of eggs produced by companion females. 
Phallus length and circumference measurements were moderately repeatable (Table 5), and with 
the exception of number of spermatozoa per ejaculate, none of the estimates for sperm traits 
exceeded a level of more than double the corresponding standard error. The residual or 
environmental component of the phenotypic variance was consistently higher for all traits. The 
between animal variance for viscosity score went to the boundary of parameter space (zero), and is 
thus not presented. 
147 
Table 5. Variance components and repeatability (t) estimates for phallus traits and semen parameters of 
breeding ostrich males. 
VARIANCE PHALLUS SPERM 
COMPONENTS Length Circumference Concentration Abnormalities Colour 
Between males 3.339 0.969 0.268 29.72 0.037 
Environment 7.992 2.513 1.214 186.48 0.732 
Phenotype 10 331 3.482 1.492 216.20 0.769 
t±SE 0.30±0 08 0.28±0.08 0.18±0.08 0 14±0.08 0.05±0.11 
Phallus length and phallus circumference was significantly (P:s:0.05) correlated on a permanent 
environment (0.67), environmental (0.50), and phenotypic (0.55) level (Table 6). Phallus length 
and circumference were also positively correlated (P:s:0.05) with sperm concentration. Correlations 
of phallus traits with abnormalities and subjective colour score were generally low. Sperm 
concentration could not be related to colour score, and was closely related to the percentage of 
sperm abnormalities per ejaculate. 
Table 6. The correlation of phallus length and circumference with semen parameters of breeding ostrich 
males. 
CORRELATION 
TRAIT 
Between males Environment Phenotypic 
Phallus length X 
Phallus circumference 0 67±0.15 0.50±0 07 0.55±0.05 
Sperm concentration 0 20±0.28 0 12±0 09 0.14±0.07 
Colour -0.43±0.71 0.06±0.11 -0.01±0.09 
Abnormalities 010±0.33 -0 14±0 09 -0.09±0.07 
Phallus circumference X 
Sperm concentration -0 12±0.30 0.17±0.09 0.10±0.07 
Colour 0.00±0.66 -0.01 ±0 11 -0.01 ±0.08 
Abnormalities -0 01±0.33 -0 09±0 09 -0.07±0.07 
SQerm concentration X 
Colour 0.42±1.16 -0.15±011 -0.10±0.08 
Abnormalities -1.01±0.14 -0.69±0 05 -0.74±0.03 
Colour X 
Abnormalities -0 68±149 0.06±0 10 0.01±0.08 
Repeatability estimates for fertility of the ejaculates obtained before the onset and at the end of 
breeding were highly repeatable (0.40±0.12 and 0.41 ±0.12, respectively). The repeatability 
estimates for'fertility with colour score was 0.39±0.12 and 0.40±0.12, for the period before and at 
the end of breeding, respectively. 
The correlations reported for phallus and sperm traits with fertility were insignificant in most cases. 
Phalius length was negatively correlated (P:s:0.05) with fertility when assessed prior to the onset of 
breeding, on a phenotypic level (Table 7). The only exception was the correlation of sperm 
148 
concentration with fertility, on an environmental level and measured at the end of breeding (Table 
7). Fertility tended (P:S:0.10) to be negatively correlated with the percentage of sperm abnormalities 
on a permanent, environmental , and phenotypic level. In most cases, higher correlations were 
reported at the end of a breeding season (Table 7). Fertility tended (P:S:0.10) to be positively 
correlated with sperm concentration only on an environmental and phenotypic level, with higher 
correlations at the end of breeding. Fertility tended (Ps0.10) to be negatively correlated with sperm 
concentration on a permanent environmental level, and negatively correlated with colour of 
ejaculates on a permanent environmental and phenotypic level before onset of breeding , and on a 
permanent environmental level at the end of breeding (Table 7). Similarly, ejaculates with a higher 
colour score tended (P:S:0.1 O) to be more fertile when assessed prior to the onset of breeding , and 
also on both an environmental and phenotypic level when assessed at the end of breeding (Table 
7) . 
Table 7. The correlations of phallus traits and semen parameters with fertility of eggs produced by companion 
females, determined for ejaculates obtained before the onset of and at the end of breeding for 
ostrich males. 
PARAMETER AND TIME OF CORRELATION WITH FERTILITY 
COLLECTION Permanent environment Environment Phenotypic 
Prior to breeding 
Phallus length -0 28±0.30 -0.15±0.14 -0.19±0.09* 
Phallus circumference -0 04±0.27 -0.27±0.14 -0.18±0.09 
Sperm concentration/ml -0 05±0.35 0.15±0.15 0.08±0.10 
% Abnormalities -0 57±0.34 -0.29±0.18 0.05±0.12 
Colour -0.28±0.32 0.13±0.24 -0.06±0.1 3 
End of breeding 
Phallus length -0.35±0.41 0.20±0.1 3 0.04±0.09 
Phallus circumference 0.04±0.26 0.07±0.14 0.05±0 09 
Sperm concentration/ml -0.18±0.39 0.36±0 14* 0.18±0.09 
% Abnormalities -0 25±0.46 -0.12±0.16 -0.14±0.10 
Colour -0.36±1.21 0.26±0.18 0.1 4±0.11 
* : Ps:0.05 
Microscopic sperm abnormalities 
Plates 2 and 3 represent micrographs depicting normal morphology of ostrich spermatozoa, at 
X200 and X400 magnification, respectively. 
149 
Plate 2. A micrograph depicting morphologically normal ostrich sperm (X200 magnification) . 
Plate 3. A micrograph depicting morphologically normal ostrich sperm (X400 magnification) . 
Sperm abnormalities observed during evaluation were sperm with acrosomal damage {Plate 4; 
X400 magnification), swollen heads {Plate 5; X400 magnification), and damage to tails {Plate 6; 
X400 magnification). 
150 
.. . -, .' . 
•.  ... • ..  
Plate 4. A micrograph depicting acrosomal damage to ostrich sperm (X400 magnification) . 
Plate 5. A micrograph depicting ostrich sperm with swollen heads (arrows) (X400 magnification). 
151 
Plate 6. A micrograph depicting ostrich sperm with abnormal morphology, i.e. broken and loose tails (X 400 
magnification) . 
An interesting observation during the microscopic semen evaluation, was the agglutination of 
spermatozoa. Hundreds of spermatozoa clumped together to form "clusters" of sperm, and 
spermatozoa in the vicinity of such clusters appeared to be orientated with their heads towards 
such clusters. Plates 7 and 8 depict micrographs of agglutinated sperm (X200 and X400 
magnification, respectively). 
Plate 7. A micrograph depicting the agglutination of ostrich sperm (X200 magnification). 
152 
Plate 8. A micrograph depicting the agglutination of ostrich sperm (X400 magnification) . 
DISCUSSION 
Phallus traits 
Phallus length and circumference measured ranged between 13-35cm and 12-24cm, respectively. 
The minimum phallus length reported in this study is shorter than that reported by Kreibich and 
Sommer (1995) and Fowler (1991 ). Kreibich and Sommer (1995) reported a phallus length of 
25cm for males of 16-18 months, and of 29-39 cm for males older than 2 years of age, and Fowler 
(1991) reported phallus lengths of up to 40cm when erect. 
Phallus size differed significantly between the respective years, being the largest during the first 
year of the study. A higher proportion of young males that were included in the breeding 
population during the second year of the study possibly attributed to this significant difference in 
phallus size. Larger phalluses measured at the end of breeding can be possibly be ascribed to the 
fact that several males were still sexually passive or inactive before being paired for breeding. 
Ostrich males tend to come into production later than females, and circulating levels of 
gonadotrophins and steroid hormones may have been too low to result in phallus development 
before the onset of breeding (Deeming and Bubier, 1999). 
Phallus length was significantly correlated with phallus circumference on all levels of assessment. 
Phallus size was also positively correlated with sperm concentration. In waterfowl that also 
possess an intromittant copulatory organ , both testes size and phallus length were related to a 
higher frequency of extra-pair copulations in males (Coker et al., 2002). 
153 
Sperm concentration and percentage of abnormalities 
Sperm concentration and the percentage of sperm abnormalities per ejaculate were extremely 
variable in this study. Sperm concentrations ranged between 26.7 X 104 mL-1 and 6588.3 x 
104 .mL-1 . Rozenboim et al. (2003), Hemberger et al. (2001), Irons et al. (1996), and Hicks (1990) 
also reported large variation in sperm concentration of ostrich semen samples. Hemberger et al. 
6
(2001) reported sperm concentrations of between 8.9 and 78.1 X 10 .µL 1 , Irons et al. (1996) sperm 
concentrations of between 1.67 and 636 X 106 sperm.ml 1, and Hicks (1990) sperm concentrations 
of between 2.5 and 3.2 X 10 9.ml 1. In the study of Irons et al. (1996), some samples had too few 
sperm to count. This was also the case in this study. Sperm concentrations of samples collected 
by means of the funnel, vacuum, and tube methods, were 0.66±0.14, 2.35±0.26, and 2.13±0.27 X 
109.ml-1, respectively (Rozenboim et al., 2003). YaJie et al. (2001) found sperm concentrations of 
9
1.57 X 10 .ml 1. Malecki et al. (1996a) reported an average sperm concentration of 3.34±0.14 
X10 9.ml-1 for male emus, and a range of 0.87 to 4.67X109.mL 1. In frizzled Hungarian ganders, 
sperm concentrations also showed extreme variation, i.e. 0.26 to 2 25X106 µ1--1, with 84% live 
spermatozoa and 37% abnormalities (Almasi et al., 2002). Sperm concentration was moderately 
repeatable, whereas the repeatability of percentage abnormalities did not reach significance. 
Sperm concentration was negatively correlated with percentage abnormalities, on a permanent 
environment, environment, and phenotypic levels. 
Generally, higher sperm concentrations were reported for males aged five to eight years, with a 
tendency for sperm concentration to decline in males aged 9 years and older. Sperm 
concentrations tended to be higher at the end of breeding, which is contradictory to that observed 
by Bertschinger et al. (1992). Bertschinger el al. (1992) reported high sperm concentrations at the 
beginning of breeding, with increased sperm concentrations associated with the peak breeding 
season and elevated testosterone levels. Rozenboim et al. (2003) on the other hand, found semen 
characteristics to show no seasonal pattern, with peak sperm concentrations reported during June-
July in Israel. In Northern pintail ducks, semen production were correlated with testosterone 
concentrations, with highest levels associated with the peak mating period (Penfold et al., 2000). 
Bah et al. (2001) also reported a positive correlation between sperm concentration, total sperm 
count, and semen volume in the domestic fowl. 
The percentage of abnormalities was highly variable and ranged between 1.0 and 84.4%. 
Hemberger et al. (2001) reported 5-26% abnormal and 4-28% dead spermatozoa, Rozenboim et 
al. (2003) 5.8±1.8%, 4.7±1.2%, and 7.1 ±1.7% abnormalities for the funnel, vacuum, and tube 
collection methods, respectively, and YaJie el al. (2001) 13.3% percentage abnormalities. Age did 
not significa'ntly influence the percentage of sperm abnormalities per ejaculate. The percentage 
abnormalities were significantly lower during the first year of the study, however, possibly resulting 
from a higher proportion of sexually mature males in the breeding population The percentage 
abnormalities were also lower before the onset of breeding than at the end of a breeding season. 
154 
Colour and viscosity 
Ejaculate colour and viscosity were variable, and comparable to that described by Hemberger et al. 
(2001) and Irons et al. (1996). Ejaculates collected before the onset of breeding were awarded 
higher colour scores than ejaculates obtained at the end of breeding. A lymph-like fluid of cloaca! 
origin, also known as transparent fluid, contributes to semen volume during copulation in poultry, 
and is derived from lymph folds adjacent to and caudal to the ejaculatory papillae (Howarth, 1995; 
Jensen el al., 1992). It is possible that this transparent fluid contributed to a limited extent to 
ejaculate volume due to the absence of copulation in the period before the onset of breeding, 
compared to ejaculates obtained at the end of a breeding season when this contribution is 
expected to be more extensive due to a higher mating frequency. Assigned semen colour scores 
also differed between the respective years, being the highest during the second year of the study. 
Male age significantly influenced the colour scores of ejaculates, with ejaculates from young males 
receiving the lowest colour scores. The darker ejaculates produced by the younger males can 
possibly be attributed to the absence of sexual activity or a lower mating frequency. Ejaculate 
colour were lowly repeatable. 
Ejaculates obtained from ostrich males receiving a flushing diet were more viscous than ejaculates 
from males not fed a flushing diet. The viscosity of ejaculates produced at the end of breeding was 
higher than that of ejaculates produced before the onset of breeding. Age had a significant 
influence on the viscosity of ejaculates, improving with an increase in age. According to Irons et al. 
( 1996), viscosity (consistency) of ejaculates obtained from ostrich males is related to the presence 
of sperm granules, and not to sperm concentration. 
Correlation of phallus and sperm traits with fertility 
The correlations of phallus measurements and sperm traits with fertility were insignificant in almost 
all cases. The only exception was sperm concentration, that was positively (P:s:0.05) correlated to 
fertility when estimated at the end of a breeding season. This is contradictory to the absence of a 
seasonal effect on the fertilisation rate of ostrich spermatozoa reported by Malecki and Martin 
(2003). In the latter study, the low fertilisation rates appeared to be a result of a lack of synchrony 
between timing of laying and sperm production, and a lack of mating. Fertility tended to be 
negatively correlated with percentage abnormalities, with higher correlations at the end of breeding. 
The fertility of ejaculates with a high colour score tended to be higher when assessed before the 
onset and at the end of breeding. Repeatability estimates of the correlation of fertility with assigned 
colour scores were also moderately repeatable. Repeatability of sperm traits estimated for 
beginning ~f season ejaculates were moderately to highly repeatable, in contrast to that of 
ejaculates obtained at the end of a breeding season. 
Fertility is, however, not always associated with high sperm numbers in an ejaculate. In sandhill 
cranes (Grus canadensis pratensis), higher fertility was associated with male sandhill cranes that 
155 
produced ejaculates of an inferior quality (Chen et al., 2000). Chen et al. (2000) suggested that 
males that bred successfully with their mates produced e1aculates with lower sperm counts. Bahr 
and Bakst (1987) found no correlation between male libido and fertility in domestic chickens, and a 
tendency for frequent copulators to produce asperm1c ejaculates. 
Fertility increased with an increase in age, and declined in males 6 years and older. Fertilities of 
75% and higher were recorded for males aged 5 to 8 years. Bunter and Graser (2000), however, 
found ostrich males aged 4-6 years produced more fertile ejaculates, and did not consider the 
fertility of eggs produced by companion breeding females. When the age of the companion female 
was also taken into account in their study, ostrich males aged 3 to 9 years were found to have 
higher fertility levels than younger and older males (Bunter and Graser, 2000). Soley et al. (1991) 
reported fertilities ranging between 3 and 55% for males ranked according to the fertility of their 
breeding companions. Bramwell et al. (1996) reported a decrease in the fertilising ability of 
spermatozoa with an increase in age in chickens. 
Fertilities observed in this study was lower than that reported for eggs produced after artificial 
insemination (Al) in ostriches and emus (Hemberger, 1996; Malecki et al., 1996b). The average 
fertility of eggs produced 3-5 days after Al in ostriches was 83.2%, with a hatchability of 78.6% 
(Hemberger, 1996). Fertility of emu eggs produced after Al was 89.9%, with a hatchability of 
84.6% (Malecki et al., 1996b). Malecki and Martin (2003) reported fertilities in ostrich eggs being 
89.4±3.4% and 94.4±3.1 % during two consecutive breeding seasons, with ranges of 78.6 to 98.2%, 
and 64 to 100%, respectively. 
Sperm morphology 
The sperm in the semen samples obtained in this study exhibited typical characteristics previously 
reported for ostrich spermatozoa (Baccetti et al., 1991; Soley, 1989, 1993; Soley and Roberts, 
1994; Hemberger, 1996). Abnormalities observed in this study included bulb-headed spermatozoa, 
crooked-neck spermatozoa (i.e. with a broken mid-piece junction), spermatozoa with damage to 
the head (i.e. damage to the acrosome), and broken tails. These abnormalities are consistent with 
abnormalities observed by Irons et al. (1996). Hemberger (1996), and Hemberger et al. (2001 ). 
Another abnormality observed in samples in this study is that of translucent sperm, which was also 
reported by Soley ( 1992). Cytoplasmic droplets (Soley et al., 1996) were not observed in 
ejaculates obtained from the breeding ostriches in this study. Lukaszewicz! et al. (2000) observed 
similar abnormalities in semen samples obtained from ganders during an entire breeding season. 
The most frequent type of abnormalities observed in frizzled Hungarian geese included different 
types of acr6somal aberrations and spermatozoa with enlarged nuclei (Almasi et al., 2002) Tail, 
head, and acrosome and mid-piece defects, were the most common defects observed in turkey 
semen samples (Alkan et al., 2002). 
156 
Aggregation of spermatozoa 
An interesting observation during the microscopic evaluation of the ejaculates in this study was the 
presence of sperm 'clusters' or aggregations of spermatozoa in some ejaculates Aggregation of 
spermatozoa has also been reported in quail, where spermatozoa tend to form clusters in the vas 
deferens (Fujihara et al., 1989). Seminal plasma in the male fowl is characterised by a high 
concentration of glutamate, with glutamic acid being the principle amino acid in the oviduct and vas 
deference (Barna and Boldizsar, 1996; Barna et al., 1996; Howarth, 1995). Glutamate has an 
inhibitory effect on the respiratory rate and metabolism of spermatozoa, thus maintaining them in a 
relative state of quiescence within the male and female reproductive tracts (Howarth, 1995). In the 
male quail, a white frothy or foamy secretion is produced by the dorsal proctodeal gland and is 
ejected during ejaculation and transferred to the female cloaca during mating. Quail spermatozoa, 
when exposed to this foamy secretion, are freed, and it thought that this may play an important role 
in the motility and sperm transport in the oviduct (Fujihara et al., 1989). 
Some species, i.e. Desmognathine salamanders and the common wood mouse (Apodemus 
sylvaticus), also show a form of sperm aggregation in their reproductive tracts. Female 
salamanders have cloaca! spermathecae that store sperm before oviposition (Sever and Hamlett, 
1998). During storage in the spermathecae, sperm are arranged in parallel arrays or may have no 
patterned arrangement. Sever and Hamlett (1998) suggested that the aggregations of sperm may 
present a form of sperm competition in salamanders. Thus, sperm from another male may be 
excluded from the spermathecae or initial ejaculates are then pushed further into the 
spermatheca's secretory matrix, thereby minimising or eliminating spermatozoa from the any 
previous matings' chance of fertilising an egg. Altruistic behaviour of spermatozoa have been 
observed in the common wood mouse, and were thought to be a mechanism to allow spermatozoa 
of a given male to gain an advantage when inter-male sperm competition is intense (Moore et al., 
2002). 
Adelie penguins are considered to be the first evidence of an avian species that allocates 
ejaculates, like certain insect and fish species (Hunter et al., 2000). Male Adelie penguins withhold 
ejaculates from their social partners, and are more likely to transfer sperm during extra-pair 
copulations than during pair copulations (Hunter et al., 2000). Female feral fowl were found to be 
biased in terms of sperm retention, and when inseminated by subordinate males, differentially 
ejected the ejaculates of the subordinate males and retained that of the most dominant males 
(Pizzari and Birkhead, 2000). 
Whether this ability of ostrich spermatozoa to aggregate presents a form of sperm competition or 
mechanism to preserve sperm integrity and viability in ostriches is unsure and warrants further 
investigation. The presence of a proctodeal gland and possible contribution of its secretion to the 
preservation of sperm integrity have not been determined in ostriches, and also warrants further 
research. 
157 
CONCLUSIONS 
Semen collection in birds can be performed by means of voluntary ejaculation, forced massage, or 
electro-ejaculation (Rozenboim et al., 2003; YanBo et al., 2001; Irons et al., 1996; Malecki el al., 
1996a; Hicks, 1990; Tan, 1980). The successful application of the forced massage semen 
collection method in this study was, however, hampered by several factors. An inability to retrieve 
the phallus in several cases, the level of stress that accompanies the procedure, urine and faeces 
contamination, and also the production of aspermic ejaculates, limits the practical implementation 
of this collection method. It is thus possible that birds that produce highly fractious samples may 
be discriminated against, for ejaculates that are obtained may not be a physiological representative 
sample of the animal (Irons et al., 1996; Hicks-Aldredge, 1994). Rozenboim et al. (2003) and 
Yan Bo et al. (2001) investigated various stress- and restraint-free methods for semen collection in 
ostriches. These methods, however, involved either the training of males to ejaculate in an artificial 
vagina, or the interruption of the mating process. Given the current nature of ostrich farming 
systems in South Africa, and the lack of sexual imprinting in ostrich males, the application of the 
latter methods are of limited value. 
Sperm concentration was positively and significantly correlated with fertility. The study, however, 
failed to identify any other male trait that could be used as a reliable indicator of female fertility 
under commercial breeding conditions. Sperm concentration and male phallus traits were found to 
be repeatable in ostrich males. Phallus traits together with sperm concentration, ejaculate colour 
and viscosity increased with an increase in male age. Fertility of ejaculates similarly increased with 
an increase in male age. The phallus traits and semen quality parameters may also show genetic 
variation, as is the case in other domesticated species. Flush feeding increased the viscosity of 
ejaculates. 
It is important to understand the extent to which sperm characteristics contribute to the 
insemination success of an ostrich male in order to improve fertility and overall reproduction 
efficiency in commercial farming systems. Traits therefore need to be identified that can aid in the 
development of a semen quality index, as used in the commercial poultry industries (Parker et al., 
2000). Such an index may potentially be used to reliably indicate the fertility of individual males for 
the potential use in breeding programs. 
REFERENCES 
' 
ALKAN S, BARAN A, AZDAS OB & EVECEN M, 2002. Morphological defects in turkey semen. Turk 
Veterinerlik ve Hayvancilik Dergisi 26(5): 1087-1092. 
ALMASI A, VARGA A, BARNA J, BOGENFURST F & MOLNAR M, 2002. Preliminary study on 
spermatological characteristics of frizzled Hungarian ganders. Acta Agraria 
Kaposvariensis 6(2): 289-292. 
158 
BACCETII B, BURRINI AG & FALCHETII E, 1991. Spermatozoa and relationships in Palaeognath 
birds. Biol. Cell. 71: 209-216. 
BAH GS, CHAUDHARI SUR & AL-AMIN JD, 2001. Semen characteristics of local breeder cocks in the 
Sahel region of Nigeria. Revue d'Elevage et de Medecine Veterinaire des Pays Tropicaux 
54(2): 153-158. 
BAHR JM & BAKST MR, 1987. Poultry. In: Reproduction in farm animals, 5th ed., pp. 379-398. Ed.: 
ESE Hafez, Lea & Febiger, Philadelphia. 
BARNA J & BOLDIZSAR H, 1996. Motility and agglutination of fowl spermatozoa in media of different 
amino acid content and pH value in vitro. Acta Vet. Hung. 44(2): 221-232. 
BARNA J, BOLDIZSAR H & ZSOLNAl-HARCZI I, 1996. Amino acid secretion of the hen's oviduct during 
the egg formation cycle. Acta Vet. Hung. 44(2); 233-241. 
BERTSCHINGER HJ, BURGER WP, SOLEY JT, DE LANGE JH, 1992. Semen collection and evaluation in 
the male ostrich. Proc. S. Afr. Vet. Assoc. Biennial Vet. Congr., 7-10 September, 
Grahamstown, South Africa, pp. 154-158. 
BRAMWELL RK, McDANIEL CD, WILSON JL & HOWARTH B, 1996. Age effect of male and female 
broiler breeders on sperm penetration of the perivitelline layer overlying the germinal disc. 
Poult. Sci. 75(6): 755-762. 
BUNTER K & GRASER H-U, 2000. Genetic evaluation for Australian ostriches - a report for the Rural 
Industries Research and Development Corporation. Publication no. 00/153, Rural 
Industries Research and Development Corporation, Australia. 
CHEN G, GEE GF, NICOLICH JM & TAYLOR JA, 2000. The effects of semen collection on fertility in 
captive, naturally fertile, sandhill cranes. Proc. fjh North Am. Crane Workshop, 11-14 
January, Albuquerque, New Mexico, USA, pp. 185-194. 
COKER CR, McKINNEY F, HAYS H, BRIGGS sv & CHENG KM, 2002. lntromittant organ morphology 
and testis size in relation to mating system in waterfowl. Auk 119(2): 403-413. 
DEEMING DC & BUBIER NE, 1999. Natural and Captive Behaviour. In: The ostrich - biology, 
production and health, pp. 83-104. Ed.: DC Deeming, CABI Publishing, CAB International, 
Wallingford, Oxon, United Kingdom. 
FOWLER ME, 1991. Comparative clinical gross anatomy of ratites. J. Zoo Wild/. Med. 22(2): 204-
227. 
FUJIHARA N, NISHIYAMA H & KOGA 0, 1989. Physiological characteristics of semen and 
spermatozoa in the Japanese quail. Jap. Poult. Sci 26: 379-380. 
GILMOUR AR, CULLIS BR, WELHAM SJ & THOMPSON R, 1999. ASREML - Reference manual. NSW 
Agriculture Biometric Bulletin No. 3, NSW Agriculture, Orange Agricultural Institute, Forest 
Road, Orange 2800, NSW, Australia. 
HAFEZ ESE, 1987. Reproduction in farm animals, 51h ed. Lea and Febiger, Philadelphia. 
HEMBERGER Y, 1996. Spermatological examination and artificial insemination of ostriches. Proc. 151 
lnternat. Ratite Conf, 27-29 March 1996, Manchester, United Kingdom, pp. 155-157. 
HEMBERGER MY, HOSPES R & BOSTEDT H, 2001. Semen collection, examination and spermiogram 
in ostriches. Reprod. Dom. Anim. 36: 241-243. 
159 
HICKS KD, 1990. Male reproduction. Proc. 1990 Ostrich Workshop, Texas A&M University, 7 April, 
pp. 3-5. 
HICKS-ALDREDGE K, 1994. Ratite reproduction. American Ostrich pp.18-19, 21, 38, 41, 45, 48, 57-
58. 
HORBANCZUK JO & SALES J, 1999. Reproduction is a crucial problem in ostrich breeding. World 
Poult. 15(5): 28-30. 
HOWARTH B, 1995. Physiology of reproduction: the male. In: Poultry production, World Animal 
Science series, volume 9, pp. 243-270. Ed.: P Hunton, Elsevier Publishers, Amsterdam, 
Netherlands. 
HUNTER FM, HARCOURT R, WRIGHT M & DAVIS LS, 2000. Strategic allocation of ejaculates by male 
Adelie penguins. Proc. R. Soc. Land B Biol. Sci. 267(1452): 1541-1545. 
IRONS PC, 1995. Ostrich reproductive research: quantum physics or back to the drawing board? 
Proc. 6h Ann. Congr. Livestock f-lealth Prod. Group S. Afr. Vet. Assoc., June, Warmbaths, 
South Africa, pp. 140-148. 
/RONS PC, BERTSCHINGER HJ, SOLEY JT & BURGER WP, 1996. Semen collection and evaluation in 
the ostrich. Proc. 1st Internal. Ratite Conf, 27-29 March 1996, Manchester, United 
Kingdom, pp. 157-159. 
JENSEN JM, JOHNSON JH & WEINER ST, 1992. Husbandry and medical management of ostriches, 
emus and rheas. Wildlife and Exotic Animal TeleConsultants, College Station, Texas, 
USA, ISBN: 0-9626069-8-7. 
KREIBICH A & SOMMER M, 1995. Ostrich farm management. Landwirtschaftsverlag GmbH, 
MOnster-Hiltrup. 
LUKASZEWICZL E, FURUTA H, XI Y-M & FUJIHARA N, 2000. Comparative study on semen quality of 
one- and two-year old ganders during the entire reproductive season. Asian J. Andra/. 2: 
139-142. 
MADEKUROZWA MC, CHABVEPI TS, MATEMA s & TEERDS KJ, 2002. Relationship between seasonal 
changes in spermatogenesis in the juvenile ostrich (Struthio came/us) and the presence of 
the LH receptor and 3beta-hydroxysteroid dehydrogenase. Reprod. 123(5): 735-7 42. 
MALECKI IA & MARTIN GB, 2003. Sperm supply and egg fertilization in the ostrich (Struthio 
came/us). Reprod. Domest. Anim. 38(6): 429-435. 
MALECKI IA, MARTIN GB & LINDSAY DR, 1996a. Semen collection and production of spermatozoa in 
the male emu. Proc. 1st lnternat. Ratite Conf, 27-29 March 1996, Manchester, United 
Kingdom, pp. 159-160. 
MALECKI IA, O'MALLEY p & MARTIN GB, 1996b. Length of fertile period in the female emu 
(Dromaius novaehollandiae). Proc. 13th fnternat. Congr. Anim. Reprod., 30 June to 4 July, 
Sydney, Australia, p. 13. 
MARQUEZ BJ· & OGASAWARA GX, 1975. Scanning electron microscope studies of turkey semen. 
Poult. Sci. 54: 1139-1143. 
McCRACKEN KG, 2000. The 20-cm spiny penis of the Argentine Lack Duck (Oxyura vittata). Auk 
117(3): 820-825. 
160 
MOORE H, DVORAKOVA K, JENKINS N & BREED W, 2002. Exceptional sperm cooperation in the wood 
mouse. Nature 418: 174-177. 
OETILE EE & SOLEY JT, 1988. Sperm abnormalities in the dog: a light and electron microscopic 
study. Vet. Med. Rev. 59: 28-70. 
PARKER HM, YEATMAN JB, SCHULTZ CD, ZUMWALT CD & MCDANIEL CD, 2000. Use of a sperm 
analyzer for evaluating broiler breeder males. 2. Selection of young broiler breeder 
roosters for the sperm quality index increases fertile egg production. Poult. Sci. 79(5): 771-
777. 
PENFOLD LM, WILDT DE, HERZOG TL, LYNCH W, WARE L, DERRICKSON SE & MONFORT SL, 2000. 
Seasonal patterns of LH, testosterone and semen quality in the Northern pintail duck (Anas 
acuta). Reprod. Fett. Dev. 12(3-4): 229-235. 
PIZZARI T & BIRKHEAD TR, 2000. Female feral fowl eject sperm of subdominant males. Nature 
(Land) 405(6788): 787-789. 
ROZENBOIM I, NAVOT A, SNAPIR N, ROSENSTRAUCH A, EL HALAWANI ME, GVARYAHU G & DEGEN A, 
2003. Method for collecting semen from the ostrich (Struthio came/us) and some of its 
quantitative and qualitative characteristics. Brit. Poult. Sci. 44(4): 607-611 
SAEKI Y, 1960. Crooked-necked spermatozoa in relation to fertility in the artificial insemination of 
fowl. Poult. Sci. 39: 1354-1361. 
SAEKI Y & BROWN Kl, 1962. Effect of abnormal spermatozoa on fertility and hatchability in the 
turkey. Poult. Sci. 41: 1096-1100. 
SEVER OM & HAMLETI WC, 1998. Sperm aggregations in the spermatheca of female 
Desmognathine salamanders (Amphibia: Urodela: Plenodontidae). J. Morphol. 238: 143-
155. 
SIEGEi_ S, 1956. Non-parametric statistics for the behavioural sciences. McGraw-Hill Book Co. 
(Inc), New York. 
SMITH WA, CILLIERS SC, MELLETI FD & VAN SCHALKWYK SJ, 1995. Ostrich production - a South 
African perspective. Proc. Alltech's 11th Annual Symp., pp. 175-197. Ed.'s: TP Lyons & 
KA Jacques, Nottingham University Press. 
SNEDECOR GW & COCHRAN WG, 1967. Statistical methods - 6th edition. Iowa State University, 
USA 
SOLEY JT, 1996. Differentiation of the acrosomal complex in ostrich (Struthio came/us) spermatids. 
J. Morphol. 227: 101-111. 
SOLEY JT, 1993. Ultrastructure of ostrich (Struthio came/us) spermatozoa: Transmission electron 
microscopy. Onderstepoort J. Vet. Res. 60(2): 119-130. 
SOLEY JT, 1992. A histological study of spermatogenesis in the ostrich (Struthio came/us). Ph.D. 
thesis, University of Pretoria, Pretoria, South Africa. 
SOLEY JT, 1989. Transmission electron microscopy of ostrich ( Struthio came/us) sperm. Proc. 
Electron Microsc. Soc. S. Afr. 19: 145-146. 
SOLEY JT, BERTSCHINGER HJ, ELS HJ & BURGER WP, 1996. The morphology and incidence of 
retained cytoplasmic droplets in ostrich spermatozoa. Proc. 1st lnternat. Ratite Cont., 27-29 
March 1996, Manchester, United Kingdom, pp. 16-18 
161 
SOLEY JT, BERTSCHINGER HJ, ROBERTS DC, DE LANGE JH & BURGER WP, 1991 Correlation 
between semen quality and egg fertility in ostriches. Seminar on Ostrich Research, 17-18 
July, Oudtshoorn, South Africa, pp. 45-47. 
SOLEY JT & ROBERTS DC, 1994. Ultrastructure of ostrich (Struthio came/us) spermatozoa. II. 
Scanning electron microscopy. Onderstepoort J Vet. Res. 61 (3): 239-246. 
TAN NS, 1980. The frequency of collection and semen production in Muscovy ducks. Brit. Poult. 
Sci. 21 (4); 265-272. 
VAN SCHALKWYK SJ' CLOETE SWP & DE KOCK JA, 1996. Repeatability and phenotypic correlations 
for body weight and reproduction in commercial ostrich breeding pairs. Brit. Poult Sci. 37: 
953-962. 
VERBYLA AP, KENWARD BR & WELHAM SJ, 1999. The analysis of designed experiments and 
longitudinal data by using smoothing splines. Appl. Statist. 48(3): 269-311. 
VON RAUTENFELD BO, 1977. Mitteilungen zur Kunstlichen Besamung, Geschlechts- und 
AltersBestimmung beim Strauss (Struthio came/us australis, GURNEY). Der praktische 
Tierartzt. 5: 359-364. 
YAJIE J, YANBo Y & WuZ1 D, 2001. Study on ostrich semen characteristics and semen storage at 
low temperature. J Econ. Anim 5(3): 49-54. 
YANBO Y, YAJIE J & WUZI D, 2001. Methods for ostriches semen collection. J. Econ. Anim. 5(2): 
32-35. 
162 
Chapter 11 
Management practices and performance indicators of 
reproduction in ostriches - conclusions and 
recommendations 
South Africa produces approximately 60% of the world's total number of slaughter ostriches. The 
emphasis on producing primarily for export markets and high input costs, together with a volatile 
exchange rate, necessitates that commercial farmers produce slaughter ostriches as cost-
effectively as possible. The reproductive efficiency of breeding ostriches, i.e. the number of fertile 
eggs and chicks produced under commercial breeding conditions, plays an integral role in 
determining the success of a commercial ostrich farming enterprise. Commercial ostrich production 
systems, however, are characterised by low reproductive performances and a large variation in egg 
production, as well as high chick mortalities. Identification of poor or non-producing breeding 
individuals in commercial breeding systems is hampered by the fact that breeding ostriches are 
maintained in large free-ranging flocks and the communal nesting system of laying ostrich females. 
The emphasis on the cost-effective production of slaughter ostriches formed the basis of this study. 
The aim was to determine possible management practices to improve the overall reproduction 
efficiency of breeding ostriches under commercial farming conditions. Behavioural, physiological 
and anatomical parameters were also investigated to determine their potential as performance 
indicators to allow the identification and selection of breeding ostriches with a better reproductive 
performance under commercial farming conditions. The findings are summarized in the sections 
below, with recommendations on the practical implementation of these findings under commercial 
farming conditions to improve the overall reproduction efficiency of breeding ostrich males and 
females. 
Management of breeding ostriches to improve seasonal reproductive performance 
The response of the breeding ostriches in the study clearly demonstrated a photoperiod-dependent 
breeding strategy. Peak production was recorded during June to January. Breeding ostrich males 
and females are traditionally joined for an 8-month breeding period at the end of May, which is 
approximately one month before the winter solstice in the southern hemisphere. Joining breeding 
ostriches before the winter solstice, i.e. when still in a photorefractory state, had the effect that the 
reproductive cycles of breeding ostrich males and females were not yet synchronised, resulting in a 
longer period to first laying and a lower seasonal reproductive performance. Joining ostriches after 
the winter solstice synchronised the reproductive activities of the breeding males and females 
shortened the time to the onset of reproduction after being paired and improved the overall 
reproductive performance of the breeding birds. 
163 
In cases where a breeding program necessitates the joining of breeding birds before the winter 
solstice, certain management practices can be applied to initiate and synchronize the reproductive 
cycles of breeding ostrich males and females. A pre-season rest, together with flush feeding and 
teasing, can be used to initiate and synchronize the reproduction in breeding ostrich males and 
females. Breeding ostriches without a pre-breeding rest period performed worse throughout a 
breeding season in all the reproductive traits recorded. Maintaining ostrich males and females in 
single-sex flocks during the non-breeding season and reintroducing males to neighbouring camps 
two weeks before being paired for a breeding season, significantly improved egg production of the 
breeding ostriches. The potential of flushing and teasing to synchronize the reproductive cycles of 
breeding ostriches, however, appears to be limited by the time of the year when breeding ostriches 
are paired for breeding. Pairing breeding ostriches when photoperiod is sufficiently long to 
stimulate reproduction may be necessary to ensure the successful implementation of these 
management practices. 
Breeding ostriches may experience a quiescent period in reproduction during breeding seasons. 
Subjecting breeding ostriches that experience a decline in production to a mid-season rest was 
successful in reinitiating reproduction and improving egg and chick production of breeding pairs 
(subjected to flushing and teasing) during the post-rest period. Egg and chick production was 
significantly improved by 52.5% and 44.0%, respectively, during the post-rest period. The potential 
of a mid-season breeding rest to reinitiate the reproductive cycles of breeding ostriches, however, 
is determined by the state of photosensitivity of the birds. Implementing the management practice 
too late during a breeding period when the breeding birds have already started to become less 
sensitive, limits the reaction to any management practice to reinitiate reproduction in both the 
males and the females. 
Intra-season rest periods also proved a means to overcome the photoperiodic control of the 
reproductive cycles of breeding ostriches paired before the winter solstice. This may allow a 
commercial ostrich farmer to ensure the timely production of slaughter ostriches according to 
market demand. Intra-season rest periods were effective in synchronizing the reproductive cycles 
of breeding males and females and improving overall reproductive performance during the post-
rest periods. The effectiveness of this management practice, however, is limited by the time of the 
breeding period when the breeding birds are subjected to a breeding rest. Enforced breeding rest 
periods are only effective in reinitiating reproductive activities during the time of the year when 
ostriches are still photosensitive and have not yet entered a state of photorefractoriness 
Ultrasonography as a management tool to identify reproductively healthy, and early producing 
ostrich female's: 
In domestic fowl, high producing birds have a very long prime sequence early in lay, and the length 
of the prime laying sequence might be a good indicator of reproductive efficiency of a female. The 
extreme variation in the egg production potential reported for breeding ostrich females, i.e. between 
164 
0 and 121 in an 8-month breeding period, motivated the study to determine the potential of 
ultrasound scanning (diagnostic imaging) as a management tool to identify actively and early 
producing ostrich females. 
Ultrasound scanning can be used to identify females that come into lay early in a breeding season, 
and progress can be made in commercial breeding flocks through the selection of ostrich females 
with well-developed follicular hierarchies at the beginning of a breeding season. The correlation 
reported for early season scanning and overall egg production suggests that females can be 
selected on partial production records tor use in breeding programs. The duration of lay after the 
commencement of the breeding season was strongly associated with reproductive success. 
Ultrasound can thus be used to identify ostrich females with a late onset of lay and lower 
production success under commercial breeding conditions. The respective ultrasound parameters 
were closely related to short-term egg production on a genetic level, indicating the ability of the 
technique to identify healthy reproductive ostrich females. The heritability reported for follicle 
classification score was low but significant Females below 2 years of age proved to be too young 
to be included usefully in breeding programs. 
Partial egg production records for a given season, however, may not be a reliable indicator of the 
ability of a female if the commencement of lay is delayed for a specific female or if the age of the 
female is unknown. The age of ostrich females needs to be taken into account when selection 
decisions are based on phenotypic data. Failure to do so may discriminate against young ostrich 
females. Management practices to initiate and synchronise reproduction in ostrich females (e.g. 
flush feeding and teasing) will be indispensable for the practical application of diagnostic 
ultrasonography in the broader industry. 
Management of large commercial ostrich breeding flocks, in terms of stocking rate and male:female 
ratio 
Ostriches are the main farming interest in the Little Karoo, and approximately 65% of the South 
African ostrich breeding population are found in this region. Ostrich breeding flocks are 
concentrated on relatively small areas, and through trampling have the largest effect on the natural 
vegetation. The influence of stocking rate and male:female ratio was therefore investigated on a 
commercial ostrich farm outside Ladismith in the Western Cape Province, South Africa, to 
determine the influence of these two factors on the overall reproductive performance under 
intensive breeding conditions. 
The findings 1ndicated that ostriches can be maintained at higher stocking rates than is currently 
advised for ostriches under intensive flock breeding conditions, i.e. 1 breeding bird per 1O ha for an 
8-month breeding period. This has important implications in terms of the intensification of ostrich 
farming, especially in areas that are characterised by vegetation that are exposed to the trampling 
effect of ostriches. A stocking rate of 114 birds ( 1M :2F) per 1O  000 m2 performed the best in terms 
165 
of all reproductive traits. Increasing the stocking rate in smaller 625m2 (0 13ha) camps resulted in 
an overall improvement in reproduction, possibly because of more frequent sexual interactions 
between breeding ostrich males and females 
A trade-off was observed in terms of the reproduction traits, i.e. when average egg production per 
female was improved, fertility and hatchability was compromised. Breeding pairs, trios and quads 
2
could be maintained on smaller areas (312.5 to 625 m /bird; 208.3 to 416.7 m2/bird; and 156.3 
m2/bird, respectively) than normally prescribed with acceptable production levels. Although the 
number of quads in the study was relatively small, it indicates that harem size under intensive 
conditions may be as high as three females per male, which supports the polygamous nature of 
ostrich males. Maintaining commercial ostrich breeding flocks at too high stocking rates will, 
however, impact negatively on the normal reproductive behavioural patterns of breeding birds, and 
may not be beneficial for the overall reproductive performance and well-being of ostriches under 
intensive farming conditions. 
Ostrich male aggression and shin colour in relation to the reproductive performance of companion 
females 
Breeding ostrich males were evaluated on a monthly basis in terms of territorial aggression and 
shin colour, and this was then related to the egg production performance of their companion 
females. Territorial aggression was moderately heritable, and male shin colour Jowly heritable in 
breeding ostrich males. Territorial aggression was positively correlated with egg production 
performance of companion females. Male aggression and shin colour was also highly correlated 
with egg production performance of companion females. However, a negative genetic correlation 
of male aggression with female egg production performance warrants further studies to quantify the 
extent to which excessive aggression contributes to overall aggression exhibited by ostrich males. 
The possible relationship of corticosterone with male aggression may aid in the understanding of 
ostriches' ability to cope with stress related to an intensive breeding environment, and warrants 
further investigation. 
Serum hormone levels, in relation to male aggression and shin colour, and female egg production 
performance 
The study presents information on the relationship between serum levels of testosterone, LH, 
progesterone, and prolactin, and male aggression and shin colour, as well as the egg production 
performance of ostrich females measured at the beginning and end of breeding seasons. In 
addition, the ·study also presents novel information on sexual dimorphism in adult breeding 
ostriches. 
The serum levels of the respective hormones were extremely variable, with prolactin being the only 
exception. Serum testosterone and LH varied on a permanent environment level, and gender and 
166 
time of the year influenced serum prolactin and progesterone levels. Animal permanent 
environment contributed to the variation observed for live weight, LH, testosterone, aggression, and 
egg production performance. The only significant correlation was reported for LH and egg 
production of females on an environmental level. Low serum levels of testosterone and high levels 
of aggression observed at the end of breeding implies that factors other than L.H contribute to the 
maintenance of territorial aggression at the end of a breeding season. 
The fact that blood samples were collected throughout the day in the study and not at a specific 
time, i.e. during the mornings or in the afternoons; may have affected the LH concentrations 
measured in this study. It is known that LH secretion follows a diurnal pattern in female birds, and 
it is possible that peak levels were measured in some females and low levels in others, which may 
have contributed to the observed variation. Future studies should aim at determining base line 
concentrations of L.H, testosterone, estradiol, progesterone, as well as prolactin, in ostriches during 
non-breeding and breeding seasons and also on a diurnal basis. This may provide a better insight 
on the endocrine control of reproduction in both males and females, and how this is linked to 
reproductive behaviour exhibited by both genders. 
Female live weight and body measurements, in relation to reproductive performance 
Literature on genetic parameters of growth and reproduction in ostriches are scarce. This is in stark 
contrast to experience with small domestic poultry species, where extensive knowledge is readily 
available for growth and development and egg production of commercial stocks. The study 
investigated the potential of the live weight and body measurements of breeding ostrich females, 
as selection criteria to improve reproduction performance in commercial breeding systems. 
The heritabilities reported for live weight, chest circumference, and tail circumference varied 
between 0.12 and 0.38. The heritabilities reported for egg production, chick production, and 
hatchability ranged between 0.10 and 0.23. The estimates for female permanent environment 
effects were 0.18 for egg and chick production, and 0.21 for hatchability. Service sire exerted 
significant effects on hatchability and chick production. Egg and chick production was highly 
correlated on a genetic and phenotypic level, while hatchability was positively related to chick 
production on a service sire level. No adverse relationship was found between live weight, body 
measurements and reproductive traits. 
Relatively high repeatability estimates previously reported for ostriches suggest that worthwhile 
progress is achievable in commercial ostrich breeding flocks. The reproductive performance of 
breeding ostrich females does not seem to be negatively related to live weight before breeding on a 
genetic or phenotypic level. These findings imply the potential development of a dual-purpose 
ostrich strain, where emphasis may be placed on selection for both growth and reproductive traits. 
Significant genetic parameters for egg and chick production of commercial ostriches were evident 
in tilis study, and genetic progress in reproductive traits thus seems feasible for ostriches. 
167 
Reproductive traits of ostrich males in commercial breeding systems. 
Ostrich males are sexually mature at approximately 3 to 4 years of age, with factors such as breed 
(e.g. SA Black vs. Kenyan Red or Zimbabwean Blue) and time of the season when hatched, 
influencing the onset of sexual maturity. Spermatogenic activity, however, has been reported for 
prepubertal ostriches, with spermatogenesis undergoing seasonal changes as observed in mature 
breeding ostrich males. The study investigated the relationship between phallus and sperm traits, 
and fertility under commercial breeding conditions, as well as the influence of male age, season, 
and management on these traits. 
The study failed to identify a male trait that can be used as a reliable indicator of fertility in ostriches 
under commercial flock-mating conditions. Male phallus traits and sperm concentration was found 
to be repeatable in ostrich males. These traits may also exhibit genetic variation, as found in other 
domesticated species. Sperm concentration was moderately repeatable and sperm abnormalities 
lowly repeatable. Phallus measurements were positively correlated with sperm concentration. 
Phallus measurements and sperm concentration increased with an increase in male age. Fertilities 
of 75% and higher were reported for males between 5 and 8 years of age. Although the study 
failed to identify a reliable indicator of male fertility, correlations reported for phallus measurements 
and sperm concentration were in the right direction. Future studies including a larger number of 
males may aid in establishing reliable indicators to quantify and qualify the reproductive fitness of 
ostrich males under intensive breeding conditions. 
Little is known regarding the mechanisms determining the reproductive success of ostrich males 
after copulation and insemination, and how these mechanisms interact with one another and with 
selective mechanisms occurring before insemination, i.e. mate choice and mate acquisition. It is 
important to understand the extent to which sperm characteristics contribute to insemination 
success in order to improve reproduction efficiency in commercial farming systems. Traits need to 
be identified that can aid in the development of a sperm quality index including e.g. sperm 
concentration, viscosity and fertilising ability as criteria, that may potentially be used to reliably 
indicate the fertility of individual males. 
To conclude: 
The implementing of flush feeding and teasing management practices will aid in the 
synchronization of the reproductive cycles of breeding ostriches, thus enabling the commercial 
ostrich farmer to optimize the reproductive potential of breeding ostrich males and females under 
intensive breeding conditions. The use of these management practices, together with ultrasound 
scanning, will aid in the accurate identification of and consequent selection of breeding ostrich 
females with an early onset of production. Results from the study indicated that reproductive 
success is in part determined by an early onset of reproduction, implying that the identification and 
seiection of breeding ostriches with an early onset of production will subsequently lead to an overall 
168 
improvement and optimization of the reproductive efficiency and thus reproductive performance of 
breeding ostriches under intensive breeding conditions. Improving the reproductive performance of 
breeding ostrich males and females will potentially lead to an increase in the number of day-old 
chicks produced per female, thereby decreasing the overall costs of producing a day-old chick, as 
well as contribute to decreasing the feed costs of a commercial ostrich breeding operation. 
The observed high degree of genetic variation in the commercial ostrich breeding population 
implies that considerable genetic improvement is feasible through the identification of breeding 
ostriches with an improved reproductive performance, given the high repeatability estimates 
previously reported for ostriches under flock-mating conditions. It is, however, important to keep in 
mind the long-term goal of a breeding program, i.e. selection must be performed but without 
compromising production under commercial conditions. 
Unravelling the neuroendocrine mechanisms behind the flexibility in reproductive timing in ostriches 
may aid in the understanding of why ostrich males and females differ in terms of basic 
reproduction-related mechanisms, e.g. why some come into production earlier than others. This 
will aid in determining the influence of domestication on the ability of ostriches to adapt to and cope 
with stressful breeding environments, and also potentially assist the commercial ostrich farmer in 
the selection of individuals with a better reproduction efficiency under commercial conditions. 
169 
SUMMARY 
South Africa is the world's largest ostrich producer, producing approximately 60% of the world's 
total number of ostriches slaughtered annually. Almost 90% of all ostrich products (i.e. leather, 
meat and feathers) generated are exported. High input costs, together with fluctuations in the 
exchange rate, places commercial ostrich producers under constant pressure to produce slaughter 
birds as cost-effectively as possible. The characteristic low reproductive performance of 
commercial ostrich breeding systems, as reflected in the large variation in the numbers of eggs 
produced, and high chick mortalities, contribute to the emphasis on the cost-effective production of 
slaughter ostriches. The identification of poor or non-producing breeding ostriches are hampered 
by a communal nesting system and the free-range conditions of commercial ostrich breeding 
systems. Different management practices, and behavioural and physiological aspects were 
studied to determine their potential to identify and select breeding ostrich males and females with a 
higher reproductive performance, and thereby improving overall reproductive performance of 
breeding ostriches under intensive farming conditions. Maintaining ostrich males and females 
separately as single-sex flocks during the non-breeding resting period improved the overall 
reproductive performance of the breeding ostriches. The reproductive cycles of breeding ostriches 
can successfully be manipulated to ensure the timely production of chicks and thus slaughter birds. 
The successful implementation of intra-season breeding rest periods to synchronize and thus 
improve overall reproductive performance is however, determined by the time of the breeding 
season when the breeding rest is enforced. Breeding ostriches can be maintained at stocking 
rates higher than those currently applied. However, too high stocking densities are detrimental to 
the well-being of the breeding ostriches under intensive conditions. A trade-off was observed in 
terms of egg production and hatchability, with hatchability decreasing as average egg 
production/female increased. The egg production performance of ostrich females was significantly 
correlated with serum LH levels. Territorial aggression exhibited by breeding ostrich males did not 
exhibit significant genetic variation to serve as an indirect selection criterion for egg production in 
companion breeding females. The territorial aggression observed in the breeding ostrich males, 
however, showed genetic variation, thus indicating that the temperament of ostriches may be 
changed through genetic selection, potentially improving their ability to adapt better to intensive 
husbandry practices. Ultrasound scanning of follicle development can be used to identify ostrich 
females with an early onset of production and selection of females on partial reproduction records 
will improve overall reproduction efficiency of commercial breeding systems Significant 
heritabilities were reported for follicle score during scanning in ostrich females. No negative 
correlation was reported for female live weight, body measurements, and egg production. Although 
no reliable .in dicator was identified to quantify the reproductive fitness of ostrich males under 
intensive conditions, correlations reported for phallus and semen quality traits with male fertility 
were in the right direction. 
170 
To conclude: 
The implementing of flush feeding and teasing management practices will aid in the 
synchronization of the reproductive cycles of breeding ostriches, thus enabling the commercial 
ostrich farmer to optimize the reproductive potential of breeding ostrich males and females under 
intensive breeding conditions. The use of these management practices, together with ultrasound 
scanning, will aid in the accurate identification and consequent selection of breeding ostrich 
females with an early onset of production Results from the study indicated that reproductive 
success is in part determined by an early onset of reproduction, implying that the identification and 
selection of breeding ostriches with an early onset of production will subsequently lead to an overall 
improvement and optimization of the reproductive efficiency and thus reproductive performance of 
breeding ostriches under intensive breeding conditions. Improving the reproductive performance of 
breeding ostrich males and females will potentially lead to an increase in the number of day-old 
chicks produced per female, thereby decreasing the overall costs of producing a day-old chick, as 
well as contribute to decreasing the feed costs of a commercial ostrich breeding operation 
The observed high degree of genetic variation in the commercial ostrich breeding population 
implies that considerable genetic improvement is feasible through the identification of breeding 
ostriches with an improved reproductive performance, given the high repeatability estimates 
previously reported for ostriches under flock-mating conditions. It is, however, important to keep in 
mind the long-term goal of a breeding program, i.e. selection must be performed but without 
compromising production under commercial conditions. 
Unravelling the neuroendocrine mechanisms behind the flexibility in reproductive timing in ostriches 
may aid in the understanding of why ostrich males and females differ in terms of basic 
reproduction-related mechanisms, e.g. why some come into production earlier than others. This 
will aid in determining the influence of domestication on the ability of ostriches to adapt to and cope 
with stressful breeding environments, and also potentially assist the commercial ostrich farmer in 
the selection of individuals with a better reproduction efficiency under commercial conditions. 
Keywords: breeding ostriches, reproduction efficiency, flush feeding, teasing, stocking density, male 
territorial aggression, serum hormone levels, ultrasound scanning, semen quality 
171 
OPSOMMING 
Suid-Afrika is die wereld se grootste produsent van volstruise en produseer ongeveer 60% van die 
wereld se totale getal volstruise wat jaarliks geslag word. Die Suid-Afrikaanse volstruisbedryf is 
primer 'n uitvoerbedryf - ongeveer 90% van alle volstruisprodukte (leer, vleis en vere) word 
uitgevoer. Hoe insetkostes, tesame met 'n fluktuerende wisselkoers, plaas kommersiele 
volstruisprodusente onder gedurige druk om slagvolstruise so koste-effektief as moontlik te 
produseer. Kommersiele volstruisproduksiestelsels word gekenmerk deur 'n Jae reproduksie-
doeltreffendheid, soos weerspieel in 'n groat variasie in eierproduksie, asook hoe 
kuikenmortaliteite Die identifisering van individuele volstruismannetjies en -wyfies met lae 
reproduksiedoeltreffendhede word bemoelik deur die feit dat volstruiswyfies in meer as een nes 
kan le, asook deur die ekstensiewe tot semi-intensiewe omstandighede en troppparingstelsels 
waaronder kommersiele volstruisbroeikuddes aangehou word. Die potensiaal van sekere 
bestuurspraktyke en gedrags- en fisiologiese aspekte om aktief-reproduserende volstruise te 
identifiseer en sodoende die reproduksiedoeltreffendheid van 'n broeikudde te optimaliseer, is in 
die studie ondersoek. Die fisiese en visuele skeiding van mannetjies en wyfies tydens die 
russeisoen het 'n positiewe invloed op die sinkronisasie van die reproduksiesiklusse van die 
volstruise gehad, met 'n betekenisvolle verbetering in seisoenale reproduksieprestasie. Die 
reproduksiesiklusse van broeivolstruise kon suksesvol deur middel van gedwonge binne-seisoen 
rusperiodes gesinkroniseer en verbeter word. Die sukses van die bestuurspraktyk om reproduksie 
te sinkroniseer en te verbeter, word egter bepaal deur die tyd van die jaar wanneer die 
bestuurspraktyk ge'fmplementeer word. Volstruisbroeikuddes kan teen hoer digthede as wat tans 
deur die toepaslike riglyne voorgeskryf word, aangehou word. Met 'n toename in die gemiddelde 
eierproduksie per wyfie, het die uitbroeibaarheid van eiers afgeneem. Volstruisbroeitroppe wat by 
te hoe digthede aangehou word, kan ook 'n negatiewe invloed op die normale reproduksiegedrag 
van volstruise he. Eierproduksie van volstruiswyfies was betekenisvol gekorreleerd met serum LH-
vlakke. Territoriale aggressie by volstruismannetjies het daarop gedui dat temperament by 
volstruise oorerflik is, wat impliseer dat volstruise geneties selekteer kan word om beter by 
intensiewe volstruisboerderypraktyke aan te pas. Territoriale aggressie het egter geen waarde as 
'n indirekte seleksiemaatstaf waarvolgens volstruiswyfies vir verhoogde eierproduksie geselekteer 
kan word nie. Ultraklankskandering kan suksesvol gebruik word om aktief-reproduserende 
volstruiswyfies te identifiseer, asook om wyfies met 'n hoer reproduksiepotensiaal tydens 'n 
broeiseisoen uit te soek. Geen negatiewe korrelasie is gevind tussen die liggaamsmassa, 
Jiggaamsmates en eierproduksiepotensiaal van volstruiswyfies nie. Alhowel geen betekenisvolle 
maatstaf van semenkwaliteit by volstruismannetjies ge'fdentifiseer is nie, was die korrelasies 
positief en in d.ie  regte rigting . 
Om saam te vat: 
Die implementering van prikkelvoeding- en koggelpraktyke sal bydra tot die sinkronisasie van die 
reproduksiesiklusse van broeivolstruise en so die kommersiele volstruisboer in staat sal stel om die 
172 
reproduksiepotensiaal van sy broeivolstruise onder intensiewe toestande te verbeter. Die 
toepassing van hierdie bestuurspraktyke, tesame met die gebruik van ultraklankskandering, sal die 
identifisering en gevolglike seleksie van broeiwyfies wat vroeg in le kom, moontlik maak. 
Bevindinge van die studie het aangedui dat reproduksiesukses deels deur 'n vroee aanloop van 
e1erproduksie bepaal word, wat beteken dat die identifisering en seleksie van wyfies wat vroeg in le 
kom, sal lei tot die algehele verbetering en optimering van die reproduksiedoetreffendheid en -
prestasie van broeivolstruise onder intensiewe toestande 'n Verbetering in die 
reproduksieprestasie van volstruise sal potensieel die aantal dagoud kuikens per wyfie 
geproduseer verhoog, wat weer die koste om 'n dagoud kuiken te produseer, sal verlaag. Deur 
nie-produserende indiwidue in kommersiele broeitroppe te identifiseer, sal ook wesenlik tot 
aansienlike besparings in voerkoste bydra. 
Die hoe mate van genetiese variasie in die kommersiele volstruisbevolking impliseer dat daar 
aansienlike genetiese vordering gemaak kan word deur die identifisering van volstruise met 'n 
verhoogde/verbeterde reproduksieprestasie, gesien teen die hoe herhaalbaarhede wat reeds vir 
volstruise in tropparingstelsels gerapporteer is. Dit is egter belangrik die langtermyn doelwit van so 
'n seleksieprogram altyd nagestreef word sander om produksie in kommers1ele sisteme in te boet 
nie. 
Die ontrafeling van die neuro-endokriene meganismes wat 'n rol in die reproduksiesiklusse van 
volstruise speel, sal bydra tot verklarings waarom volstruismannetjies en -wyfies verskil in terme 
van basiese reproduksiemeganismes (bv. hoekom sommige vroeer in le kom as ander, ens.). Dit 
sal ook bydra om die invloed van domestikering op die aanpassingsvermoe van volstruise om 
potensiele stresvolle toestande te hanteer en dus beter onder intensiewe kommersiele 
boerderyomstandighede te vaar, te verstaan. Die seleksie van volstruise met 'n verbeterde 
aanpassingsvermoe vir kommersiele boerderypraktyke sal potensieel tot 'n algehele verbetering in 
reproduksiedoeltreffendheid bydra. 
Sleutelterme: broeivolstruise, reproduksiedoeltreffendhe1d, prikkelvoeding, koggelpraktyk, tropgrootte, 
territoriale aggressie, serum-hormoonvlakke, ultraklankskandering, semenkwaliteit 
173 
... 
Appendix 
Prolactin assay 
Radio-iodination materials and procedure 
Chicken prolactin iodination buffers 
Tris/Tween 20 
-7 1.Sg sodiumchloride 
-7 2.42g Tris or 3.14g Tris/HCI 
-7 0.1 g sodiumazide 
-7 200µ1 Tween 20 
-7 Make up to 200ml with glass-distilled water 
0.3M sodium phosphate buffer 
-7 3 68g Na2HP04 + 0.64g NaH2P04.2H20 
-7 Make up to 1 OOmL with glass-distilled water 
-7 Adjust pH to 7.5 
Chloramine-T 
-? 5mg chloramine + 1 Oml 0.3M Na-phosphate buffer 
-7 Do not use metal spatula 
-7 Do not mix more than 10 min before use 
2% Bovine serum albumin (BSA) buffer 
-7 O 2g BSA+ 10ml 0.3M phosphate buffer 
Columns 
PD-10 - equilibrated with Tris/Tween 
G-100 - equilibrated with Tris/Tween 
Iodination procedure: 
Add 25µ1 of glycerol to each of 15 tubes (for the collection from the PD-10 column) 
Add 1 Oml of the phosphate buffer to the chloramine-T. 
Aliquot 14µ1 from prolactin stock into eppendorf tubes 
Add 16µ1 of the phosphate buffer to the recombinant chicken prolactin. 
Add 5µ1 of the [1251]-Nal to the prolactin 
Add 1 OµI of the chloramine-T solution to the prolactin tube. 
• Vortex and i,ncubate at room temperature for 60 seconds. 
Stop the reaction by the addition of 200µ1 of the 0.3M phosphate buffer. 
• Apply the mixture immediately to the PD-10 column and allow to drain in. 
• Wash out the reaction tube with 250µ1 of the phosphate buffer and apply to the column. 
• Apply fresh Tris/Tween 20 buffer to the column and collect 11 drop fractions into each of the 15 tubes with 
added glycerol. 
175 
• Monitor the tubes for radioactivity - the first peak should be around tubes 6-8 or earlier, and the second 
peak around tubes 12 or 13. 
Mix the solution in the peak tubes with the glycerol and pool the fractions. 
• Allow the G100 column to drain down to the top gel surface. 
• Apply the pooled radio-active fraction, and gently top up the fluid with Tnsffween 20 to the top of the 
column, then attach either a pump or gravity feed to the column. 
• Collect 1 ml fractions and monitor for radio-activity - depending on the column the prolactin should come 
off at around fraction 40. 
• Dilute the peak fractions collected in 1·1 with the 2% BSA in phosphate buffer and store in aliquotes at-
700C until use. 
Radio-immuno assay media and procedure 
Basic prolactin buffer (PRL buffer) 
0.1 M Na-phosphate buffer, pH 7.5 900mL 
0.9% NaCl 9.0g NaCl 
• 10mM Na2EDTA 3.72g Na2EDTA 
• 0.1 % sodium azide 1.0g NaN3 
• 0.1% BSA 1.0g BSA 
-7 Adjust the pH to 7.0 with 1M  NaOH and make up to 1l itre 
First antibody 
-7 Number 31/1, 1 :6000, 50µ1 per tube, for 1200 tubes 
-7 = 59ml prolactin buffer+ 1m l stock of first antibody (1 :100) 
Tracer 
-7 50µ1 per tube, for 1200 tubes 
-7 =6 0ml prolactin buffer and stock trace to> 12000 counts per 50µ1 
Normal rabbit serum (NRS) 
-7 50µ1 per tube, for 1200 tubes 
-7 =1 00µ1 rabbit serum + 59.9ml prolactin buffer 
Second antibody 
-7 50µ1 per tube, for 1000 tubes 
-7 = 30ml donkey serum (doneky 5 batch 2@ 1: 2) + 30ml prolactin buffer 
Notes: 
• Run samples, B-zero, quality control pools, and standards at 100µ1 with no added buffer. 
• Non-specific; binding tubes (NSB) get 150µ1 buffer 
• Add all reagents at 50µ1 
• For pour-off buffer, use 6% polyethyleneglyco\ (PEG) at 1m l 
• Stored stock for standard curve is 5µm per vial, and working stock of 12500ng.ml'1 per 5ml vial 
• Top standard is made up by bringing 10µ1 working stock up to 1ml 1n PRl buffer, thus= 125ng.ml·1 
176 
The prolactin assay was set up in triplicate in 12 X 75mm polypropylene tubes (tubes for standard curve from 
tube 6 to 15) 
-7 1 tube (total count) (empty) 
-7 1 tube (NSB, blank) (200µ1 PRL buffer) 
-7 3 tubes for zeros (100µ1 PRL buffer) 
-7 1 tube containing 125ng/ml standard (500µ1 standard + 500µ1 PRL buffer) 
-7 1 tube containing 62.5ng/ml standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 tube containing 31 25ng mr standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 tube containing 15 625ng.mr standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 tube containing 7.8125ng.m1· standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 tube containing 3.9ng.mr standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 tube containing 1.95ng m1· standard (500µ1 standard+ 500µ1 PRL buffer) 
1 
-7 1 tube containing 0.977ng.m1· standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 tube containing 0.4883ng.mr standard (500µ1 standard + 500µ1 PRL buffer) 
1 
-7 1 lube containing 0.24ng.mr standard (500µ1 standard + 500µ1 PRL buffer) 
-7 From here on, unknown samples were aliquoted in duplicate and at 100µ1 per tube for assay 
Testosterone assay 
Gelatine phosphate buffer (GPB) 
-7 = 1.0g gelatin + 1 litre PBS 
First antibody 
-7 100µ1at1 :80000dilution in GPB 
Tracer 
-7 Counts = 15000dpm 
Normal rabbit serum (NRS) 
-7 100µ1 rabbit serum (1 :500) in GPB with 0 1M EDTA 
Second antibody 
-7 Donkey 6 batch 5 - 1 30 
On day 1, the assay was set up in 1O mm glass tubes and in triplicate (tubes for standard curve from tubes 5-
13) as follows: 
-7 1 tube (total count) (empty) 
-7 1 tube (NSB, blank) (300µ1 GPB) 
-7 3 tubes for zefos (200µ1 GPB) 
-7 1 tube containing 16ng/ml standard (25µ1 standard + 175µ1 GPB) 
-7 1 tube containing Sng/ml standard (25µ1 standard + 175µ1 GPB) 
-7 1 tube containing 4ng.m1· 1 standard (25µ1 standard + 175µ1 GPB) 
-7 1 tube containing 2ng.mr1 standard (25µ1 standard + 175µ1 GPB) 
-7 1 tube containing 1 ng.m1· 1 standard (25µ1 standard + 175µ1 GPB) 
177 
-7 1 tube containing 0 5ng ml- 1 standard (25µ1 standard+ 175µ1 GPB) 
-7 1 tube containing 0.25ng.mr 1 standard (25µ1 standard+ 175µ1 GPB) 
1 tube containing 0.0625ng.ml 1 -7 standard (25µ1standard+175µ1 GPB) 
Quality controls were prepared in duplicate by using serum collected from 3 ostrich males and 4 ostrich 
females at a commercial ostrich farm. A set of tubes were made up for each respective ostrich male and 
female, and were prepared as follows: 
-7 1 tube containing 25µ1 serum (25µ1 made up to 200µ1 with GPB) 
-7 1 tube containing 20µ1 serum (20µ1 made up to 200µ1 with GPB) 
-7 1 tube containing 1 OµI serum (10µ1 made up to 200µ1 with GPB) 
-7 1 tube containing 5µ1 serum (5µ1 made up to 200µ1 with GPB) 
-7 1 tube containing 2.5µ1 serum (2.5µ1 made up to 200µ1 with GPB) 
-7 From here on, 1 tube per unknown sample (25µ1 serum made up to 200µ1 with GPB) 
LH assay 
Iodination of RC-AEl-s-1 in chlorarnine-T 
• 1.5 ml Eppendorf sample cup 
• PH--10 (Pharmiacia column) or 2g Sephadex G25 medium and 1X10 glass column with top 
Prepare on day of use: 
O. SM sodium phosphate buffer 
-7 6.13% Na2HP04 + 1.06% NaH2P04.2H20 
-7 Adjust to pH 7.5 
0. 1M  phosphate gel 
-7 1.226% Na2HP04 + 0.212% NaH2P04.2H20 + 0.25% gelatine 
-7 Adjust pH to 7.5 
0. 1M  phosphate-BSA 
-7 1.226% Na2HP04 + 0.212% NaH2P04.2H20 + 1.1% bovine serum albumin (BSA) 
-7 Adjust pH to 7.5 
Chloramine-T 
-7 5mg chloramrne-T in 10ml 0.1M phosphate 
-7 Adjust pH to 7.5 
Sodium metabisu/fite 
-7 10mg Na-metabisulfite in 100ml 0.1M phosphate 
-7 Adjust pH to 7.5 
178 
Potassium iodide 
-7 1g Kl in 10mL0.1M phosphate 
-7 Adjust pH to 7.5 
Procedure for iodination of PRC-AEl-s-1 in chloramine-T 
Remove sealed tip of PD-10 column and allow sotrage solution to run to waste. 
• Wash column with 1ml Kl then 15ml 0.25% gelatine-phosphate buffer. 
OR 
• Pour G25 into the glass column (after pre-soaking in 0.25% gelatine-phosphate buffer overnight) and 
allow to pack down 
• Add the 0.1 M phosphate buffer (pH 7.5) to the chloramine-T, sodium-metabisulfite, and Kl 
Place an agitator in the sample cup (small piece of paper clip) and start the magnetic stirrer underneath 
Add 1 OµI of the PRC-AEl-s-1 (1 25µg made up in water). 
Add 5µ1 0.5mCi Na 1251 
• Add 20µ1 chloramine-T and vortex 
Incubate for 60 seconds 
Stop reaction with 100µ1 Na-metabisulfite 
• Add 200µ1 Kl 
Drain the column down to the gel surface and place the contenl of reaction vessel on the column. 
Rinse the reaction vessel with 250µ1 Kl and add to the column. The column effluent from these two steps 
can be allowed to run to waste. 
The [1 251]PRC-AEl-s-1 is eluted from the column by using 0.25% gelatine-phosphate buffer 
0.5ml fractions are collected into glass tubes containing 0.5ml of 1%  BSA-phosphate buffer 
The [1251]PRC-AEl-s-1 is eluted first from the column, in fractions 5-7 from a PD-10 column 
• The free iodine is eluted last in fractions 12-18 of a PH-1 o column 
Radio-immunoassay media and procedure 
Basic LH buffer 
-7 0.04M phosphate buffered saline (0 15M NaCl) 
-7 Prepare 1 llitre - need 350µ1 per tube 
• 0.5M phosphate buffer 80ml 
• 0.15M NaCl 8.759 NaCl 
• 0.01M EDTA 2.92g EDTA 
• 0.1 % sodium azide 1.0g NaN3 
• 0.1% BSA 1.0g BSA 
-7 Bring to 900ml in double-distilled H20 
-7 Adjust the pH to 7.0 with 1M NaOH 
-7 Make up to 1 L 
179 
0.5M phosphate buffer (1 litre) 
-7 For assay diluent and iodination 
~ = 613g Na2HP04 + 10.6g Na2HP04.2H20 
-7 Bring up to 900ml with double distilled H20 
-7 Adjust the pH to 7.0 with 1M  NaOH 
-7 Make up to 1 litre 
First antibody 
-7 50µ1 per tube, 1:17K 
-7 = 98.823ml basic LH buffer+ 1.177ml stock of first antibody@ 1 :200 
Tracer 
-7 50ml per tube 
-7 = 100µ1 basic LH buffer+ stock trace (>1 OOOOcounts per 50µ1) 
Normal rabbit serum (NRS) 
-7 50µ1 per tube 
-7 = 99.75ml basic LH buffer+ 250µ1 rabbit serum (1400) 
Second antibody 
-7 50µ per tube 
-7 83.33ml basic LH buffer+ 16.67mL donkey serum (donkey 6 batch 7, 1 :6) 
On day 1, the LH assay was set up in triplicate (tubes for standard curve from tube 5-13) as follows: 
-7 1 tube for total count (leave empty) 
-7 1 tube (blank) (50µ1 diluent) 
2 tubes for zero standard (200µ1 diluent + 50µ1 anti 3/3) 
1 tube containing 5ng PRC-AEl-s-1.ml-1 (200µ1 standard+ 50µ1 anti 3/3) 
1 tube containing 2.5ng PRC-AEl-s-1.ml- 1 (200µ1 standard+ 50µ1 anti 3/3) 
1 tube containing 1.25ng PRC-AEl-s-1.ml1 (200µ1 standard+ 50µ1 anti 3/3) 
1 tube containing 0.625ng PRC-AEl-s-1 ml1 (200µ1 standard+ 50µ1 anti 3/3) 
1 tube containing O 312ng PRC-AEl-s-1.ml 1 (200µ1 standard+ 50µ1 anti 3/3) 
1 tube containing 0.156ng PRC-AEl-s-1.ml 1 (200µ1 standard + 50µ1 anti 3/3) 
1 tube containing 0.078ng PRC-AEl-s-1.ml-1 (200µ1 standard + 50µ1 anti 3/3) 
1 tube containing 0 039ng PRC-AEl-s-1.ml 1 (200µ1 standard + 50µ1 anti 3/3) 
1 tube containing 0.019ng PRC-AEl-s-1.ml1 (200µ1 standard + 50µ1 anti 3/3) 
1 tube containing 20µ1 of cockerel pool (20µ1 made up to 200µ1 with RIA diluent+ 50µ1 anti 3/3) 
1 tube containing 40µ1 of hen pool (40µ1 made up to 200µ1 with RIA diluent+ 50µ1 anti 3/3) 
From here on , 1 tube per unknown sample (200µ1 of sample + 200µ1 of RIA diluent) 
Notes: 
11 Use basic buffer to dilute samples 
• Run samples at 25-50µ1 - make up to 200µ1 with basic buffer 
• For pour-off buffer (1 ml) use 6% PEG 
180