A grading system for Medicago sativa hay in South Africa

English: The purpose of this study was to develop a national grading system for lucerne hay in South Africa by identifying the most appropriate sampling, as well as sample handling and preparation procedures, the most accurate NlRS - nutrient calibrations and an accurate, cost effective quality model. Six hundred lucerne hay samples were obtained from different cuttings at different times in the seasons and from different commercial irrigation areas (sites) in South Africa. The 600 samples were scanned and screened through a NlR Systems Model 5000 monochromator (Foss). One hundred and sixty- eight samples representing the spectral characteristics of the South African lucerne hay population were selected and chemical analysed. The variation in nutritive value of South African lucerne hay was evaluated as an initial study. The highest moisture content recorded (13.54%) was safely below the critical moisture level of 16% for effective storage. The coefficient of variation (CV) ranged from 1.2% for dry matter (DM) up to 66.2% for acid detergent fibre-crude protein (ADF-CP). The average ash content was 12.97% (7.3 to 29.5%), indicating soil contamination. Relatively high average values were recorded for calcium (Ca) (l.35%), potassium (K) (2.53%) and iron (Fe) (874 ppm). The fibre fractions varied as follows: acid detergent fibre (ADF) (2l.26 to 47.28%), neutral detergent fibre (NDF) (28.89 to 65.93%), lignin (4.32 to 16.25%), cellulose (16.29 to 36.44%) and hemicellulose (5.26 to 19.86%). The mean NOMD for both 24 and 48hr (69.26 and 73.19% DM, respectively), was representative (CV = ± 8%) of the lucerne hay population. Crude protein (CP) (average = 20.7%DM) consists of 76.9% true protein. According to ADFCP, 6% of the samples were heat damaged. A second study was conducted to determine the effect of the grinding procedure on the moisture and CP content of the ground sample. Variance of analyses revealed significant (P<0.0001) differences in moisture concentration between ground (CV = 16.1%) and unground (CV = 27.4%) samples ranging from 14.7 up to 4l.1% (of the unground sample). However, the grinding process had a non-significant (P>0.05) influence on the CP content of the final ground product. Even though r2 between measured moisture results on unground samples and values predicted by electronic moisture tester (EMT) seems to be significantly high (r = 0.79; P<0.0001), individual predicted values for higher moisture samples (>10%) failed to accurately predict moisture content around the moisture area of critical concern (16%) and higher. In a third study, the accuracy of near infrared reflectance spectroscopy (NIRS) to predict chemical and digestibility parameters was investigated. Values for r2 and ratio of prediction to deviation (RPD) used as estimates of calibration accuracy for chemical and digestibility parameters were considered as follows: good for DM (r2 = 0.97; RPD = 4.84), CP (r2 = 0.97; RPD = 4.57), ADF (r2:;; 0.95; RPD;;;; 3.97), NDF (r2 <= 0.0.95; RPD = 3.99), lignin (r2 = 0.94; RPD = 3.61), ash (r2 = 0.93; RPD = 3.12) and chloride (Cl) (r2 = 0.95; RPD = 3.74); intermediate for neutral detergent fibre-crude protein (NDF-CP) (r2 = 0.91; RPD = 2,96), sugar (r2 = 0.91; RPD = 2.82), in vitro organic matter digestibility at 24 hr (IVOMD24) (r2 = 0.90; RPD = 2.84) and NOMD at 48hr (IVOMD48) (r2 = 0.89; RPD = 2.70); and low (RPD<2.31 ) for soluble protein (SP), acid detergent fibre-crude protein (ADF-CP), fat, starch, neutral detergent fibre digestibility (NDFD) and the macro minerals (Ca, P, Mg, P, Na and S). The results recorded in the present study indicated that the NIRS technique is acceptable for DM, CP, ADF, NDF, lignin, ash and Cl analysis and for inclusion in quality models. Milk yield (MY) derived from the CNCPS model, by replacing the average lucerne hay in a complete diet with the rest of the 168 samples, was used as a criterion to evaluate and/or develop models for lucerne hay quality grading. The best single predictor of MY was the ADF content of lucerne hay, which explained 67% of the measured variation. The relatively poor performance of CP (r2 = 0.04) and other protein related parameters (r2< 0.25; ACP, ADF-CP, NDF-CP and SP) in predicting MY suggest that protein content of lucerne hay is an unreliable indicator of lucerne hay quality. The developed empirical equation named lucerne milk value (LMV), including ADF, ash and lignin (Y = 64.18 - 0.23ADF - 0.53ash - 0.90lignin), accurately predict MY. Application of the theoretically-based summative TDNlig model of Weiss et a/. (1992), using lignin to determine truly digestible NDF, explained almost all of the variation in MY (r2 = 0.98). However, several of its components were poorly predicted by NIRS and therefore, not suited for quality assessment. Current available models for assessing lucerne hay quality revealed lower accuracies, especially when protein was included in the model. The results of the present study clearly indicated that large variations occur in the energy and protein composition as well as the utilisation of nutrients in South African lucerne hay. This emphasises the need for a rapid and accurate quality evaluation system for lucerne hay in practice. The developed LMV proved to be a practical, simplistic, economical and accurate quality evaluation model for commercial application.