The data were evaluated separately for each animal in
each year of collection. The results were analyzed in an iterative
process of excluding values that were higher than the
mean plus 2 standard deviations until basal values were
determined. The progesterone (high salivary progesterone)
and nonprogesterone (basal salivary progesterone) phases
were determined. A cyclic hormonal pattern was considered
when the presence of both phases was observed in the
same sample period, while the maintenance of basal salivary
progesterone levels during an entire sampling period
was considered an acyclic pattern. The interval between
two peaks of salivary estradiol followed by a rise in progesterone
was considered a complete cycle for the calculation
of cycle length.