Hormonal Monitoring of Reproductive Status in Monogamous Wild Female Owl Monkeys (Aotus azarai) of the Argentinean Chaco

Abstract

The Neotropical owl monkeys (Aotus spp.) are a good model for evaluating the hypothesis that monogamy may arise if female reproductive cycles limit the mating potential of males. To evaluate this hypothesis, we first needed to assess the feasibility of using fecal sampling for monitoring the reproductive status of females. We collected fecal samples (n = 242, from 7 females) from wild adult Aotus azarai females in the Gran Chaco forests of Argentina during 3 years. Fecal estrone-1-glucuronide (E₁C) and pregnenadiol-3-glucuronide (PdG) tended to rise in parallel during the luteal phase. The average cycle length was 22 ± 3 days (n = 5 females, 10 cycles). We identified 2 conceptive cycles and characterized the E₁C and PdG profiles of 2 pregnancies. This report is the first of its kind on wild female owl monkeys. Despite the difficulties in sample collection and processing in the field and providing a species-specific validation in the laboratory, we show that fecal samples from A. azarai can be used for monitoring female reproductive status and function.

Introduction

A traditional hypothesis for explaining the evolution and maintenance of monogamy in primates suggests that monogamy was a default social system imposed upon males when the spatial or temporal distribution of females made it difficult for single males to simultaneously defend access to more than one mate [Emlen and Oring, 1977; Rutberg, 1983; Van Schaik and Van Hoof, 1983]. In this theoretical context, it is proposed that social monogamy may arise if females exhibit synchronous reproductive cycles that limit the number of females that a male can fertilize. Even if not synchronous, it is possible that in a socially monogamous species with intensive care, the combination of pregnancy and postpartum amenorrhea generates a very brief period when females are receptive.

A thorough evaluation of a hypothesis that considers the reproductive cycles of females requires that demographic and behavioral information on reproductive events be collected together with adequate information on the females’ hormonal status. The development of methods for the collection and processing of field samples over the last two decades [Ziegler and Wittwer, 2005] has allowed significant progress in the monitoring of the reproductive ecology of wild primates [Ziegler et al., 1997; Valeggia, 2007; Van Belle et al., 2009; Ziegler et al., 2009; Pappano et al., 2010]. The Neotropical owl monkeys (Aotus spp.) are a good model for evaluating hypotheses on the evolution of monogamy, but there have been to date no field studies to characterize the reproductive cycles of this small, sexually monomorphic, arboreal primate.

Owl monkeys are consistently found in small, socially monogamous groups composed of 1 adult heterosexual pair, 1 infant, and 1 or 2 individuals of smaller size [Fernandez-Duque, 2011]. In an Azara's owl monkey (Aotus azarai) population of the Argentinean Chaco, females produce 1 offspring a year, with the majority of births occurring between October and November [Fernandez-Duque et al., 2002]. Infants nurse intensively until 5 months of age, but they are still observed on the nipple when they are 8 months old [Rotundo et al., 2005; Huck and Fernandez-Duque, in press]. Male parental investment is unusually high in owl monkeys, with the male carrying the infant over 80% of the time [Fernandez-Duque et al., 2009]. The reproducing pair occupies a relatively small and exclusive territory (4–10 ha), and within their territory they forage primarily for fruits, but also obtain significant amounts of leaves, flowers and insects [Van der Heide et al., in press]. Social groups interact aggressively with other groups at territory boundaries or with individuals who, after dispersing from their groups, are ranging solitarily among territories.

We present here the results of a preliminary study of hormonal changes during the reproductive cycle in wild A. azarai females. We evaluated the feasibility of using fecal sampling for monitoring the reproductive status (e.g. pregnant, cycling) of wild A. azarai females. Owl monkeys do not show any visual signs of ovarian cycling [Bonney et al., 1979, Bonney et al., 1980; Wolovich and Evans, 2007], therefore it is necessary to rely on changes in hormone levels to ascertain their reproductive status. To our knowledge, this is the first report of a characterization of female reproductive cycles from a free-ranging population.

Methods

Study Area and Population

The area of study is located in the Guaycolec Ranch, 25 km from the city of Formosa in the Argentinean Gran Chaco of South America (58°11′ W, 25°58′ S). The region consists of a matrix of grasslands dotted with patches of forests and transected by gallery forests that grow along rivers. In the area, owl monkeys inhabit the gallery forests that grow along the Pilagá and Guaycolec Rivers, as well as the patches of forests dispersed in the grasslands. Approximately 20 social groups have been monitored, at least monthly, since 1997 providing background information on the timing of birth seasons, mating seasons and dispersal patterns [Fernandez-Duque et al., 2002; Fernandez-Duque, 2009].

Fecal Sample Collection and Processing

We collected 242 fecal samples from 7 reproductive females during 2005, 2008 and 2009. We collected samples, on average, every 3.5 days (range: 1–7 days) between February and August of each year. Two of the females were sampled in all 3 years, 1 female was sampled in 2005 and 2008, and the remaining 4 were sampled once in different years. Our sample collection and processing methods followed established protocols [Ziegler and Wittwer, 2005; Pappano et al., 2010]. Briefly, fecal droppings were collected as soon as the individual defecated; the specimens were transferred to a screwcap plastic tube, which was then labeled and placed in an insulated bag with a cold pack until processing, usually within a few hours. At the field station, we mixed individual samples thoroughly, and 0.1 g of fecal material was weighed and transferred to a tube containing 5 ml of a 1:1 ethanol:distilled water solution. Tubes were vortexed for 5 min and then centrifuged at high speed for 10 min to separate the solid material. A 2-ml aliquot of the supernatant was transferred to solid-phase extraction columns (Prevail C18 from Alltech®), which were then stored at −20°C until shipment to the Reproductive Ecology Laboratory at the University of Pennsylvania, where they were stored again at −20°C until processed.

Hormonal Assays

Once at the Reproductive Ecology Laboratory, samples were eluted from the solid-phase extraction columns with 2 ml of ethanol and evaporated in a 50°C water bath under a constant stream of air. Once dry, the samples were reconstituted in 1 ml of ethanol and stored at −20°C until assaying. On the day of the assay, samples were dried again and reconstituted in phosphate buffer and assayed for estrone-1-glucuronide (E₁C) and pregnenadiol-3-glucuronide (PdG) using enzyme conjugates and antisera provided by Coralie Munro at the Clinical Endocrinology Laboratory at UC Davis. The E₁C and PdG enzyme immunoassays were validated for owl monkey fecal samples by determining parallelism and accuracy. Serial dilutions of a pooled sample (n = 4) paralleled the standard curve with no statistically significant difference in slope. The intra- and interassay coefficients of variation for E₁C and PdG were 6.5 and 20.8, and 12.3 and 24.8%, respectively. Hormone values are provided as nanograms per gram of feces for E₁C and as micrograms per gram of feces for PdG.

Data Analysis

We used the criterion of progesterone-defined ovulation to identify ovulatory cycles. We considered changes in PdG levels of more than 50% above baseline as evidence that the cycle was ovulatory. Baseline levels were calculated as an average of the follicular phase levels for each individual. PdG levels were either consistently low or exhibited a clear peak, with few instances of hormonal profiles exhibiting intermediate progesterone activity. We estimated cycle length by two methods: (1) counting the days between two consecutive E₁C peaks (for females with more than 2 cycles, we computed the average cycle length) and (2) counting the days between the nadir (the lowest) points in the profile. We report both cycle length estimations. We estimated the possible conception window period by backcounting the published estimates of pregnancy duration in Aotus spp. (121–133 days [Wolovich et al., 2008]) from a known birth date [Ziegler and Sousa, 2002]. Then, the conceptive cycle leading to the pregnancy was identified by looking at the cycle that most closely preceded the window of conception. We calculated gestation length range as the number of days from the progesterone-defined ovulation day to the range of days for the birth of an infant. Hormonal values for pregnant females were identified by the same backcounting method.

Results

Ovarian Cycles

Fecal E₁C and PdG followed a similar pattern, both rising almost in parallel during the luteal phase. Fecal E₁C, a conjugated form of estrone, did not show a follicular surge prior to ovulation. To illustrate a general pattern that was observed in all females, figure 1a shows E₁C and PdG changes in 1 female during a 2-month period. The hormonal profile of a second female who gave birth at the end of November 2007 is shown in figure 1b. The profile, which starts when she was approximately 3 months into the postpartum period, suggests 2 complete nonconceptive ovulatory cycles and a partial one at the end of the sampling period.

Fig. 1.

a, b Fecal E₁C and PdG excretion profile of 2 adult females in 2 different years of data collection (a 2005, b 2008).

One female contributed 4 nonconceptive cycles, 1 female contributed 3 cycles, 2 females contributed 2 cycles each and 1 female contributed 1 cycle. The overall average cycle length was 22.2 ± 2.7 days (range 18–25 days), when computed from E₁C peak to E₁C peak (n = 5 females, 12 cycles) and 24.0 ± 1.6 days (range 20–26 days), when computed from nadir to nadir (n = 5 females, 10 cycles). The average cycle length per female is presented in table 1. Mean fecal E₁C and PdG levels characteristic of ovarian cycles and pregnancy are presented in table 2.

Table 1.

Average cycle length per female calculated as number of days between successive E₁C peaks (P2P) and number of days between successive nadir samples (N2N)

Female ID	Ovulatory cycles contributed n	Average P2P cycle length ± SD days	Average N2N cycle length ± SD days
Eva	4	22.5 ± 2.9	22.2 ± 1.3
Duquesa	3	21.3 ± 3.1	22.7 ± 2.3
Celina	2	25 ± 1.4	24
Mony	2	24 ± 0.0	25
Estrella	1	18	23
Overall average		22.2 ± 2.7	24.0 ± 1.6

Table 2.

Mean values (± SD) of fecal E₁C and PdG characteristic of ovarian cycles and gestation

	Cycles n	Females n	Mean E1C ± SD ng/g feces	Mean PdG ± SD μg/ feces
Ovarian cycles
Intercycle troughs	10	5	6.9 ± 3.5	1.1 ± 0.8
Postovulatory peak	12	5	49.9 ± 14.1	18.1 ± 9.1
Pregnancy
First days		1	51.6 ± 27.6	6.4 ± 8.3
Midpregnancy		2	43.6 ± 20.2	5.2 ± 2.8

Conception and Gestation

We identified 2 conceptive cycles and we characterized the E₁C and PdG profile of 2 pregnancies. Figure 2a shows a possible window of conception around June 30, 2009, by a female who gave birth between October 30 and November 6. The estimated gestation period for this female is 120–126 days. The profile in figure 2b suggests a conception around August 8, 2005, for a second female who gave birth between December 20 and January 14. Given the relatively longer time between observations of this second female, we estimated her gestation length using information on the developmental stage of the infant when first observed. This resulted in an estimated gestation length of approximately 132–140 days.

Fig. 2.

a, b Fecal E₁C and PdG excretion profile of 2 adult females in 2 years (a 2009, b 2005) showing conceptive cycles. Arrows indicate the possible timing of conceptions, based on backcounting from the approximate date of birth of an infant.

Two females were well into their pregnancies during the time when they were sampled (fig. 3). In both cases, fecal E₁C levels show a slight, but sustained increase during pregnancy, while PdG remains at basal levels. The first female (fig. 3a) gave birth on September 18; counting back 133 days from the birth date, this profile covers from approximately day 46 through the last day of pregnancy. The second female (fig. 3b) gave birth on October 6, and, thus, this profile covers from approximately day 46 through day 113 of pregnancy.

Fig. 3.

a, b Fecal E₁C and PdG excretion profile of 2 adult females during pregnancy in 2005.

Discussion

We have presented the results of a preliminary, yet solid first attempt at characterizing the reproductive biology of free-ranging owl monkeys. The primary contribution of this technical report is showing the feasibility of characterizing the reproductive biology of wild A. azarai females, including the detection of ovarian function, conception and pregnancy. As seen in other Neotropical primates [Shideler et al., 1994; Ziegler et al., 1996b], fecal PdG and E₁C, metabolites of progesterone and estrogen, respectively, appear to cycle synchronously in A. azarai. During a cycle there is a peak in both fecal PdG and E₁C, and there appears to be no preovulatory rise in fecal E₁C during the follicular phase. Thus, fecal estrone conjugates appear to reflect luteal function.

Our determination of cycle length using the number of days between E₁C peaks and nadirs generated estimates which were close to the one described for captive A. azarai (25 days [Wolovich et al., 2008]), and in the highest end of the range of other Aotus spp. [Fernandez-Duque, 2011]. Estimates of gestation length, based on visual inspection of the profiles, were within the range obtained from captive animals (133 days [Hunter et al., 1979]; 121 days for A. azarai [Wolovich et al., 2008]). However, it is worth noting that fecal E₁C and PdG levels do not rise above cycling levels as seen in urinary hormone profiles. We are unable to provide an explanation for this finding, but it is likely that this pattern is related to steroid conjugation and/or excretion dynamics during pregnancy.

Although we found good evidence of ovarian cyclicity using glucuronide conjugates of estrone and pregnanediol, the synchronous cycling of these metabolites, together with the lack of a clear rise following conception, prompts us to suggest that future studies using fecal samples to monitor reproductive cycles should include the analysis of free estrone and progesterone. It is also important to highlight the complex picture of reproductive hormone metabolism and excretion reported for Neotropical monkeys [Ziegler et al., 2009]. Each species, even within the same genus, seems to differ in the way and the timing these steroid hormones are metabolized in the liver (and other organs) and then excreted in feces [Ziegler et al., 1996a].

Given the technical nature of this report, we deem it important to discuss the difficulties and limitations of our study so that researchers interested in using fecal samples for monitoring reproductive function in Aotus spp. in the field may avoid some of the pitfalls we have encountered and can capitalize on our experience. Collecting small fecal pellets from free-ranging small arboreal monkeys represents the first challenge. However, to date, it is the only available noninvasive biological medium since collecting blood or salivary samples would require capturing the animals, and collecting urinary samples is not feasible because the urine stream seldom reaches the ground. Second, collecting fecal samples is extremely time consuming, requiring field assistants to wait several hours for the focal monkey to defecate. This results in a trade-off between frequency of collection and sample size. Do we sample a few individuals quite frequently or increase the sample size at the cost of reducing the frequency of collection? In our case, the target frequency was to collect 3–4 samples per week per female. This was not always feasible given the difficulties in locating the monkeys, finding the feces of the target individual on the ground and sometimes missing the defecation bout. It is then possible that our sampling may have resulted in missed hormonal peaks and some error in the calculation of cycle and pregnancy length.

In a comparative perspective, the reproductive characteristics of female owl monkeys we have described conform to the general pattern described for other cebid taxa. Quite interestingly, many aspects of the reproductive biology of owl monkeys are strikingly similar to the ones reported for titi monkey females [Valeggia et al., 1999]. Females in both genera give birth to singletons once a year, following an almost identical gestation period (Callicebus 128.6 ± 4.4 days, range 122–135 [Valeggia et al., 1999]). The length of the ovarian cycle (22–24 days) and the typical profile of reproductive hormones are also similar in the two taxa. In nature, both genera live in small groups consisting of a monogamously mated pair and younger animals [Fernandez-Duque, 2011; Norconk, 2011] and males of both genera provide obligate and intensive paternal care [Fernandez-Duque et al., 2009]. Are the similarities in their reproductive biology driving the equally similar patterns of socially monogamous relationships and biparental care? Could data on reproductive biology offer some insights to the current debate on the phylogenetic relatedness of these two genera [Schneider and Rosenberger, 1996]?

Although we are still far from understanding the evolution and maintenance of monogamy in owl monkeys, the data presented here contribute to a larger attempt at evaluating the hypothesis that the spatial and temporal distribution of female reproductive opportunities may regulate social monogamy. The hormonal profiles we have presented provide evidence that the monitoring of reproductive cycles can, and should, be part of any attempt at examining the mechanisms regulating monogamy in owl monkeys.

Our results lay the necessary methodological foundation to develop studies that can examine if the temporal distribution of reproductive opportunities may function as an effective female strategy to limit the male's reproductive potential. Given the information we have presented, supported by information from captivity, it seems that owl monkey females may have a pregnancy of 4.2–4.5 months, wean their infants at approximately 6 months and have a cycle length of 22–24 days. In A. azarai, which shows such a narrow seasonal reproductive pattern [Fernandez-Duque et al., 2002; Fernandez-Duque, 2012] we should expect a window of only a few ovulatory cycles in each year. However, we did find a case in which the female started cycling 3 months after giving birth (March) and did not conceive until 4 months later, raising the possibility of several nonconceptive cycles. This is a possibility that should be explored in future studies. Females may be further limiting the males’ opportunities by relatively high levels of same-sex aggression in females that may prevent other reproductive females from joining the group. Finally, females may be reducing the opportunities that males have of seeking other females by requiring a relatively prolonged courtship prior to mating, as well as by synchronizing their reproductive cycles with other females. The possible ovulatory suppression of daughters in the group, the synchronicity of females’ reproductive cycles, a demand of a prolonged courtship and female-female aggression may limit the male's access to many females, therefore promoting a monogamous system.

We hope that the information provided here will be applicable to the understanding of the reproductive characteristics of other Aotus species, as well as to a broader understanding of the reproductive biology of other New World primates.