[Evaluation of variability of duration of pregnancy in mares]

In a set of mares of English Thoroughbred horse of the Napajedla stock the gravidity length in the time period from 1880 to 1972 was evaluated. The variability of the gravidity length was evaluated in sets which were put together by combination of two age groups with a division of the studied time period into ten-year stages. Variance and sample means of the gravidity lengths in these sub-groups are significantly different in the period under study. In order to eliminate the action of non-genetic influences qualifying the variability of the gravidity length, and to increase the objectivity of their evaluation it is proposed not to calculate with absolute values of the gravidity length, but on the contrary with transformed values: (xi - x) divided by s which would be computed for corresponding year of the foaling period. Another methodical approach to exclude the variability of the gravidity length depending on seasonal influences was based on the use of correction factors which were determined for individual months of foaling, on the basis of the sinusoid form of the curve of the gravidity length in the course of the calendar year. When observing the influence of the age of sires on the gravidity length, the shortening of the gravidity length in dependence on their higher age has been suggested, however this relation was significant only sporadically. The submitted work is a part of a larger theme referring, to the reproduction of horses.