Pelvic area as a tool for selection of beef replacement heifers

The objective of this study was to examine the usefulness of pelvimetry in the selection of first-calf heifers to reduce the incidence of dystocia. Data from 345 Angus (AN) and 112 Polled Hereford (PH) heifers bom during the years 1987-1991 and their subsequent calving records were utilized within this analysis. Preselection of heifers was based upon weaning weight (Angus, 246 kg and Polled Hereford, 243 kg). Prebreeding measures included: weaning weight, weaning hip height, yearling weight, yearling hip height, yearling pelvic area (categorization: smallest 25%, intermediate 50% and largest 25%), and frame score. Postbreeding measures included: precalving weight, precalving hip height and precalving pelvic area (smallest 25%, intermediate and largest 25%). A dystocia score was assigned to each birth of 328 AN and 104 PH calves (1 = no assistance, 2 = hand pull, 3 = mechanical assistance and 4 = caesarean). In addition, calf birth weight, sex of calf and date of birth were recorded for each calving. Data analyzed within the statistical (single group) model indicated that calf birth weight was the most important variable associated with dystocia score for both AN and PH heifers (20 and 21%, respectively, of the variation in dystocia). Yearling pelvic height and width accounted for 11% and yearling pelvic area for only 8% of the variation in dystocia score within the AN heifers. However, within the PH heifers, 21% of the variation in dystocia was explained by yearling pelvic height and width. Analysis of the AN data set within the prediction model indicated sire birth weight EPD, dam birth weight, weaning hip height and yearling hip height as the most significant discriminating variables within the categorization of dystocia (1, 2 or 4). Though these four variables were highly significant (P \u3c .05), the Average Canonical Square Correlations (ACSC) per variable only ranged from 0.048 toO.099. Cross validation was utilized in testing the models predictive ability and resulted in an average error rate of 58%. Sire birth weight EPD was the only significant variable identified within the PH data set (ASCCof .04). Sire birth weight EPD, dam birth weight, weaning hip height and yearling hip height were utilized within the cross validation procedure for the PH heifers and attained an average error rate of 79%. In conclusion, results of the statistical model indicated that nonpelvic size measures can be useful in explaining the variation in dystocia without the use of yearling pelvic measurements. However, yearling pelvic measures within the PH heifers did account for significantly larger amounts of variation than models including only nonpelvic measures. In addition, predictive analysis of the data set indicated that nonpelvic size measures were more useful than yearling pelvic height, width and area. Consequently, variables included in these data sets were unable to accurately predict and classify heifers within a dystocia level