Scrotal circumference, bodyweight and semen characteristics in growing dairy-breed natural-service bulls in Tasmania, Australia.

To provide herd managers with a set of decision rules allowing them to predict the likelihood that a juvenile bull is ready for Bull Breeding Soundness Evaluation (BBSE), or breeding, if bodyweight and scrotal circumference are known. This was a longitudinal study following two groups of young pasture-fed Holstein and Jersey bulls from northwest Tasmania, Australia. Individual scrotal circumference, bodyweight and semen characteristics were recorded at 6-8 weekly intervals, from 6-18 months of age. Classification and regression tree analyses were used to predict the probability that a bull had ≥70% normal sperm morphology based on scrotal circumference and bodyweight measurements. Overall 1,661 scrotal circumference and bodyweight measurements were obtained, and 518 semen samples from 356 bulls were assessed for sperm morphology, from 16 examination sessions that took place between 29 May 2015 and 17 August 2016. Classification and regression tree analyses generated a decision tree for Holstein bulls with four node endpoints, and for Jersey bulls with three node endpoints. Diagnostic test performance showed that for Holstein bulls, using the node endpoints of scrotal circumference ≥27 cm and bodyweight ≥349 kg, 98% had ≥70% normal sperm (positive likelihood ratio 10.4; 95% CI = 2.7-41), and using the node endpoints of scrotal circumference ≥27 cm and bodyweight between 282-349 kg, 89% had ≥70% normal sperm (positive likelihood ratio 1.6; 95% CI = 0.9-2.6). For Jersey bulls, using the node endpoints of bodyweight ≥259 kg and scrotal circumference ≥29 cm, 88% had ≥70% normal sperm (positive likelihood ratio 3.4; 95% CI = 1.6-7.0). This study provides a set of relatively simple decision rules based on bodyweight and scrotal circumference measurements that allows herd managers to assess the likelihood that juvenile bulls are ready for BBSE or breeding. BBSE: Bull breeding soundness evaluation; BRT: Boosted regression tree.