Body size and testicular measurements have been found to be important parameters utilized in breeding soundness evaluation. The present study therefore, aimed at determining the relationship between body weight (BW), body condition score (BCS), testicular length (TL), testicular diameter (TD) and scrotal circumference (SC) in 120 extensively reared Yankasa rams (approximately 30 months old) using linear, quadratic and cubic predictive models. Coefficient of determination (R2), Adjusted R2, the estimate of Mallows’ Cp, RMSE (Root mean squares error) and the parsimony principle (p=number of parameters) were used to compare the efficiency of the different models. Strong Pearson’s correlation coefficients (r = 0.83-0.94; P<0.01) were found between BW, TL, TD and SC. Spearman correlations between BCS and other variables were also highly significant (r = 0.78-0.85; P<0.01). SC was the sole variable of utmost importance in estimating BW, which was best predicted using the cubic model. However, the optimal model for BW prediction comprised TD, SC and BCS with p, R2, Adjusted R2, RMSE and Cp values of 4, 0.948, 0.946, 1.673 and 4.85, respectively. The present findings could be exploited in husbandry and selection of breeding stock for sustainable sheep production especially within the resource-poor farming system under tropical and subtropical conditions.

KEY WORDS: Body size; Body condition; Testicular measurements; Rams; Regression models.