Our results confirm the two-year breeding pattern and breeding seasonality (Qi et al. 2008). Focal females who had viable offspring from the previous breeding season (i.e., from March 2015 to June 2015) did not conceive during the study period, while those who did not reproduce previously conceived and delivered. Hence, we categorized the focal females into two groups, according to their receptivity. Group A included 6 receptive females, and Group B contained 6 unreceptive ones.
Menstrual cycle synchrony
Because conception and pregnancy disrupt normal menstrual cycles, causing temporal elevations in gonadal estrogen levels, we ran tests for examining menstrual synchrony only in unreceptive females. Among them, the mean length of menstrual cycles in the mating season was 24.8 ± 3.3 days, with only one peak, indicating a short receptive window. As revealed by the ANOVA test, the dates of peaks in estradiol (E2) levels show a within-unit similarity but between-group differences (SQ unit: P<0.05, n=9; WF unit: P<0.05, n=6). In other words, females in the same unit have synchronized ovulatory cycles, suggesting highly overlapped ovulation timings within OMUs.
[Table 1 here][Figure 1 here]
Temporal variations in gonadal estrogen levels
The hormonal profiles of receptive females differ markedly from unreceptive ones, confirming the two-year breeding cycle and breeding seasonality. Specifically, in Group A females, the estradiol (E2) levelsduring the mating season were significantly greater than that in the breeding season (mating season, 454792.43 ± 329795.27 pg/g, mean ± SD, n=192, vs. breeding season, 78140.23 ± 23802.80 pg/g, mean ± SD, n=119; Mann-Whitney U test, U=474, P<0.01), so was the progesterone (P) levels (mating season, 482.29 ± 339.14 ng/g, mean ± SD, n=193, vs. breeding season, 121.37 ± 91.41 ng/g, mean ± SD, n=123; Mann-Whitney U test, U=2272, P<0.01). Hence, among receptive females, gonadal estrogen levels elevated in the mating season but plunged back to the baseline levels once the breeding season started, confirming the observed seasonal breeding pattern.
In addition, the changes in gonadal estrogen levels predicted conception and birth events. In the mating season, after the first peak of gonadal estrogen levels, three Group A females showed signs of pregnancy (e.g., enlarged abdomen); and before the subsequent breeding season, they all delivered offspring. Successful delivery resulted in a sudden drop in both estradiol and progesterone levels (estradiol: 78140.23 ± 23802.80 pg/g, mean ± SD, n=119, progesterone: 121.37 ± 91.41 ng/g, mean ± SD, n=123).
Group B females, however, did not reproduce throughout the study period likely because they were in interbirth intervals. Specifically, compared with group A females, during the mating season group B females had significantly lower levels of both estradiol and progesterone (Mann-Whitney U test, estradiol: U=1135, P<0.01; progesterone: U=41852, P<0.01). But once the breeding season started, this difference disappeared (Mann-Whitney U test, estradiol: U=7628, P>0.01; progesterone: U=4105, P>0.01). This contrast suggests a lack of temporal fluctuation in gonadal estrogen levels in Group B females, whose hormonal levels stayed around the baseline levels across the study periods (estradiol: mating season, 80759.30 ± 24658.55 pg/g, mean ± SD, n=216 vs. breeding season, 74417.43 ± 11163.95 pg/g, mean ± SD, n=120; Mann-Whitney U test, U=4298, P>0.01) (progesterone: mating season, 94.45 ± 63.88 ng/g, mean ± SD, n=231, vs. breeding season, 65.12 ± 38.79 ng/g, mean ± SD, n=105; Mann-Whitney U test, U=10249, P>0.01).
[Figure 2 here]
Solicitation
We observed 498 incidences of solicitation, which were all initiated by females (301 in SQ unit and 197 in WF unit). Likewise, the temporal variation in solicitation also confirms breeding seasonality and two-year breeding pattern. Solicitations concentrated in the mating season (N=493) but were rarely observed in the breeding season (N=5).
In the mating season, receptive females solicited at a higher frequency than unreceptive ones (solicitation events per day: group A, WF unit, 1.01 ± 1.87, mean ± SD, n=156; SQ unit, 1.82 ± 2.84, mean ± SD, n=153, vs. group B, WF unit, 0.22 ± 0.92, mean ± SD, n=156, SQ unit, 0.18 ± 0.64, mean ± SD, n=153, Mann-Whitney test, WF unit: U=8586, P<0.01; SQ unit: U=6626, P<0.01). However, in the breeding season, all subjects solicited little, regardless of their receptive statuses. Hence, solicitations were performed mainly by females who were in estrous.
Among receptive females, the conflicts for mating opportunities were not observed. The absence of mating conflict likely resulted from a higher-than-expected OSR. In both units, among the six females, only three were receptive while the other half had reproduced in the previous year. This was not likely a mere accident but resulted from conflict-avoidance strategies. Although our result could not confirm this possibility, another potential strategy was also observed.
Within units, females avoided soliciting frequently on the days when others were soliciting at peak frequencies. On the days when copulation through solicitation was observed, on average, the residential males copulated primarily with only one receptive female from their units (79%), while the rest accounted for only 18% and 3% of all copulations (N=3). Because no dyadic aggression was observed, the dominance hierarchy among focal females was either very weak or absent. Hence, among females in the same units, dominance status could not account for this difference.
[Table 2 here] [Figure 3 here][Figure 4 here]
Copulation and testosterone level variation
Besides, during the mating season when the copulation frequency peaked (4 to 6 times on average), we did not observe a temporal elevation in testosterone levels. In fact, in the residential males fecal testosterone levels were significantly higher in the breeding season than in the mating season (mating season: 1.174 ±0.775ng/g, mean ±SD, n=14, vs. breeding season: 3.221 ±2.181g/g, mean ±SD, n=19; t=-3.78, df=23.736, P<0.01).
This is not surprising. First, the males did not copulate with all three receptive females at a high frequency simultaneously. On a single day, copulation was concentrated in only one receptive female (accounting for 79% of all copulations). Hence, generating a higher level of testosterone for enhancing sexual motivation and spermatogenesis may not be necessary. Second, compared with the breeding season, the lower testosterone levels in the mating season likely represent a natural drop resulting from repeated copulations and ejaculations that happened earlier before defecating. This is because we collected fecal samples from 1300 to 1500, the period after most copulations were observed. Obtaining morning fecal testosterone data by non-invasive method was difficult because most monkeys defecate during this period.