PRECISION OF SAMPLING TECHNIQUES IN DETERMINING MICROSCOPIC FOLLICULAR POPULATION IN GILTS

The precision of estimating populations of atretic and nonatretic antral follicles by sampling 10, 20 and 25% of follicles in serially sectioned ovaries was determined. From records on 18 374 gilt follicles (Data 1), three new data sets (Data 2, 3 and 4, generated by attributing a detection probability of 0.1, 0.2 and 0.25 to each of the follicles, were compared with Data 1. Estimates of the follicular populations obtained from Data 2, 3 or 4 differed from the values of Data 1 only when there were fewer than five follicles of a given size on the ovary. The coefficient of variation of the estimated numbers of follicles from values of Data 1 was below 32 ± 19% for Data 2 and 3 when there were more than 50 follicles in a class. For Data 4, however, it was 8 ± 2% or below except for classes containing less than five follicles. Data 4, but not Data 2 and 3, was as efficient as Data 1 in detecting differences in follicle distribution between breeds, age and type of ovary (P < 0.05), except for follicular classes containing less than five follicles. Thus, follicular populations, as determined by sampling 25% of the total number of follicles, were consistent with the actual distribution if five or more follicles were present in a class. Key words: Follicles, sampling, ovary, gilts