Male animals in pair bonds face a challenge regarding the paternity of offspring when their mate’s sexual behavior includes extra-pair copulations (“EPCs”). The ample indications of female sexual agency in nominally monogamous bird species (Brouwer and Griffith, 2019) are compelling evidence of this challenge. A male bird may have options to mitigate the risk of extra-pair fertilizations (“EPFs”) that may result from EPCs, including mate-guarding during his mate’s fertile period (e.g., Beccardi et al., 2023), adjustments in amount and timing of within-pair copulations (“WPCs”; e.g., Gill et al., 2019), retaliatory aggression (e.g., Valera et al., 2003), and—after egg laying—withholding parental care from offspring (e.g., Schroeder et al., 2016). Females may conceal their extra-pair activity to avoid these countermeasures (e.g., Wagner, 1992), especially when males have similar or greater physical power than their mate, a common case in birds.

Nazca boobies (Sula granti), a colony-nesting seabird with obligate bi-parental care, provide an unusual opportunity to evaluate this sexual conflict when females have unconstrained sexual agency. Females are 13–17% heavier than males (Apanius et al., 2008; Howard et al., 2021), with corresponding larger skeletal measurements (Anderson, 1993; Howard et al., 2021) and stronger bite force (Rebol and Anderson, 2022). Divorce is common and typically controlled by females (Maness and Anderson, 2008). While pre-laying EPCs are common in this study’s population in Galápagos—often during a brief, conspicuous liaison at a neighboring nest site in view of the regular male partner—males invariably perform affiliative pair-bonding behaviors with the female when she returns to him and do not retaliate (Anderson et al., 2025).

The males’ tolerance of EPCs is explained by the greater physical strength of females, the higher aggression that females show in at least some interactions (Tarlow et al., 2003; Anderson et al., 2004), and by the constraints imposed by the strongly male-biased adult sex ratio (Townsend and Anderson, 2007). Specifically, the relative scarcity of females and frequent divorce limit males’ breeding opportunity, so alienating a female that he has attracted, or abandoning a suspect clutch, are costly options. Mate attraction and pair-bond consolidation are lengthy processes of several weeks that require each male to attend his colony nest site where these occur (Nelson, 1978; Maness and Anderson, 2008). Returning to EPCs, sperm competition (Birkhead and Montgomerie, 2020) via copulation is apparently the only option to protect paternity; copulation also occurs at each male’s nest site. Thus, we hypothesize that males rely on frequent colony attendance before and during their mate’s fertile period to achieve temporal overlap with the female when she returns to the colony between foraging trips. This facilitates mate attraction and pair-bonding, and enables WPCs whenever the female permits them. However, foraging is necessary, and requires leaving the colony.

Nazca boobies are pelagic seabirds, foraging far from the colony. GPS tags on large samples of birds show that foraging absences are long, typically spanning multiple consecutive days and the intervening nights, with pursuit of prey occurring almost exclusively in daylight (Zavalaga et al., 2012; McKee et al., 2023). Thus, each bird spends only some days and nights in the colony, trading off two classes of location-based benefit: nutritional self-maintenance at sea and reproductive activities in the colony. Each sex receives reproductive benefits from colony attendance early in the season by establishing and building pair bonds, and by copulating to fertilize ova. Males have additional sexually-selected incentives to attend the colony: to maintain possession of their nest sites, which are required to attract a female (Nelson, 1978; Maness and Anderson, 2008), to interact with the limited pool of females, and to maximize WPC opportunities in the context of sperm competition. After egg laying, either parent incubates the clutch, alternately freeing the other to forage. Pair-bonding activities before the clutch is started occur during both day and night. Nocturnal sharing of a nest site is a core pair-bonding behavior (DA, unpublished data) and is more frequent at night because colony attendance, and thus temporal overlap with potential mates, is higher. Copulation is almost exclusively diurnal (Anderson et al., 2025).

We predict that each bird’s colony attendance increases as egg laying approaches, but that males attend more than females during pair-bonding and during the female’s fertile period(s), presumably ~24 h before egg laying (Howarth, 1974). Nazca boobies may lay a second egg 2–10 days after the first (Anderson, 1989). We assume that males do not know whether or when a second egg will appear, and predict that males will attend the colony more than females after the first egg’s appearance — despite this species’ alternating biparental incubation (Nelson, 1978) — mainly to protect paternity in a possible second egg via WPC opportunities with their now-established mate.

We studied Nazca boobies at Punta Cevallos, Isla Española, Galápagos Islands, Ecuador (1°23’S, 89°37’W; Apanius et al., 2008). Breeding at this colony is seasonal, with most egg laying from October to December. Because most boobies are at sea during daylight (Anderson and Ricklefs, 1987; Anderson et al., 2004), we used nighttime attendance to assess each individual’s expression of the foraging vs. mating effort tradeoff before, during, and shortly after clutch initiation. At the beginning of each of four breeding seasons (2014 to 2017) we recorded presence/absence of each bird with band resight surveys at 2000 h local time (roughly 45 min after nightfall) in a subarea of the colony where each adult had a uniquely numbered and easily readable plastic leg band. This population is tolerant of human presence and habituated to our methods, allowing data collection without disturbance, and detection probability (Burnham and Anderson, 2002) of these marked birds is high (0.99–1.00; Supplementary Table 1). Nighttime fidelity of breeding birds to this subarea of the colony is virtually 100% (see Supplementary Material).

Nocturnal attendance is an effective index of each individual’s overall attendance at the colony, involving a single survey of the focal subarea per 24 h solar cycle. One survey per night is adequate to capture all nocturnal attendance, because surveys repeated throughout the night give essentially identical results (Supplementary Table 1). At this time of year, each adult that has returned to the colony spends all night on the ground at or beside a single recognizable nest site, with few exceptions, either alone or with an opposite-sex adult, simplifying survey logistics. Most foraging trips include more than one (usually complete) entire daylight period and any intervening nights, with departures from the colony concentrated at dawn and arrivals at dusk (Anderson and Ricklefs, 1992). At this time of year roughly half of nights are spent at sea (Figure 1). Consequently, this band-resight survey method precisely captures nighttime attendance (which varies markedly; Supplementary Figure 1), and provides a proxy for attendance in adjacent daylight hours. In summary, single nighttime surveys exploited the schedule characteristics of Nazca booby foraging absences to provide a labor-minimizing estimate of overall attendance.

We evaluated the probability of nighttime attendance during three parts of the breeding cycle involving pair-bonding and copulation. Capital letters indicate formal names that we have assigned to each period and to predictors of attendance (e.g., “Sex”). The Pre-laying Period of each breeding pair was from 30 to 3 nights before their first egg appeared (Figure 1). The 5-night Laying Period, with the night following the first daytime detection of a new clutch and the preceding and subsequent nights; and the Early Incubation Period, from the third to tenth nights after the first egg is laid. The Early Incubation Period includes the laying of any second egg (Anderson, 1993). The Main Incubation Period is nights 11–43, completing the expected time to hatching of the first egg (Anderson, 1993), and providing baseline attendance patterns when pair-bonds and paternity have been established.

The boundary between the Pre-laying Period and the Laying Period (our night -2) falls one night before the female’s fertilization window is assumed to close (on roughly night -1). The ~24 h after ovulation when the ovum is isolated from sperm by deposition of the egg’s other components (Howarth, 1974) probably begins between nights -2 and -1 in most Nazca booby egg formations. The boundary between the Laying Period and the Early Incubation Period is placed somewhat arbitrarily but acknowledges that in some seabird species females go to sea shortly after the effort of laying (e.g., Tickell, 1968; Quillfeldt et al., 2019). Thus, the Laying Period includes immediate physiological and behavioral precursors, and immediate consequences, of oviposition. In this demarcation of time, each bird’s schedule was scaled before analysis to the day on which its first egg was recorded in daily nest monitoring: the subsequent night is that bird’s Night 0, one night earlier is Night -1, etc. Although a total of 327–367 adults were monitored nightly in each early breeding season, our analyses for each of the three Periods concern the subset that started a clutch during our data collection and satisfied other criteria (see Supplementary Material).

In a model selection framework conducted in R (Version 4.3.2; R Core Team, 2023), we used generalized linear mixed-effects models (lme4 package, version 1.1-31; Bates et al., 2015) with a binomial error distribution and a logit link function to predict sex-specific nightly attendance probability, separately for each of the three Periods. Initially, the full model for each Period used fixed effects Night (each individual’s night in relation to clutch initiation), Sex, and their interaction. Crossed random effects of Bird Identity and Year, and their interaction, were also included initially in all models (Table 1). Bird Identity is the bird’s unique leg ring ID. Night was a continuous predictor except for models of attendance during the Laying Period, when Night was categorical and acknowledged that females might be suddenly absent, and male mates correspondingly present to incubate, just after laying (e.g., Tickell, 1968; Quillfeldt et al., 2019) in an attendance pattern inconsistent a continuous time predictor. Supplementary Table 3 provides full descriptions of each predictor. A Period’s model set comprised the most complex model.

and all subsets of that model. Models were compared using Akaike’s Information Criterion, corrected for small sample sizes (AICc). The top model (lowest AICc) best explained variation in the data. We considered all models falling within 2 AICc units of the top model to be highly supported, and between 2 and 4 AICc units to receive meaningful support, unless they were more complex versions of a nested top model (Burnham and Anderson, 2002; Arnold, 2010). We report these complex models with uninformative additional predictors (Arnold, 2010) but ignore them in inference.

For the Pre-laying Period, the models that included both Sex and Night had equivalent explanatory power and were separated from competing models by more than 56 AICc units (Table 1a). We used the simpler model m2 for inference. The effect size of Sex in m2 indicated that females had substantially lower nighttime attendance than males (β = -0.487, 95% CI: -0.605, -0.369), a 39% reduction in the odds of attendance (Odds Ratio = 0.61, 95% CI: 0.55, 0.69; Supplementary Table 5a). Additionally, nighttime attendance increased steadily as the Laying Period approached, with the odds of attendance increasing by approximately 4% per night (β = 0.038, 95% CI: 0.034, 0.042; Odds Ratio = 1.04, 95% CI: 1.03, 1.04; Supplementary Table 5a). These results indicate higher nighttime attendance by males and an increasing probability of nighttime attendance by each sex as the Laying Period approaches (Figure 1a).

For the Laying Period, the model including Sex, Night, and their interaction (model m6) was the top model, with no other model within 212 AICc units (Table 1b). The effect size of Sex in model m6 showed that on Night -2, females had lower nighttime attendance than males (β = -0.309, 95% CI: -0.454, -0.164), a 27% decrease in the odds of attendance (Odds Ratio = 0.73, 95% CI: 0.64, 0.85; Supplementary Table 5b). However, strong Sex by Night interactions revealed night-specific differences in attendance, most notably the near absence of males on Night -1, an 85% decrease in the odds of attendance relative to Night -2 (Odds Ratio = 0.15, 95% CI: 0.11, 0.19; Supplementary Table 5b). These results indicate that males’ nighttime attendance was similar to or higher than that of females, except on the night preceding egg laying (night -1), when most males were absent (Figure 1b). Males’ attendance on night -1 was an anomaly for either sex.

For the Early Incubation Period, models m11 and m12 were each highly supported by the data, and m14 received meaningful support (Table 1c). Although the ΔAICc for m14 was slightly larger than 2, we considered m14 as the top model, having similar explanatory power to the more complex m11 and m12 in this nested model set (Burnham and Anderson, 2002; Arnold, 2010). The fact that the three models including Sex hold essentially all model weight is notable. The top-ranked model lacking Sex (m13) was >10 AICc units different from m14, again supporting an effect of Sex. The effect size of Sex in model m14 indicated consistently lower nighttime attendance by females during early incubation (β = -0.300, 95% CI: -0.455, -0.145), a 26% decrease in the odds of attendance (Odds Ratio = 0.74, 95% CI: 0.63, 0.87; Supplementary Table 5c). The predicted values of the top model m14 (Figure 1c) show higher nighttime attendance for males, and the attendance probability of each sex varied little across the Period (Figure 1c).

Further details of results (e.g., full parameter estimates for top models) are in Supplementary Material Tables 5, 6.

We hypothesized that Nazca boobies benefit from attending the colony during the early breeding season, despite the possible nutritional cost of time away from the ocean, but males should benefit more than females. We found that males generally spend more nights in the colony, and not at sea, than females do at these times (Figure 1). For weeks during the Pre-laying Period, and then during the Early incubation Period, male attendance probability consistently exceeded that of females by approximately 0.15. On the two nights during the Laying Period after his mate started their clutch his probability of attendance jumped to 0.27–0.37 higher than hers. The remarkably precise exception to this pattern was the only night in our study period (night -1 in Figure 1) between the fertilization windows of the first and possible second eggs, when attendance probabilities reversed: the probability for males was 0.32 lower than that for females. Males, 1/3 of whom are excluded from breeding each year, have more to lose than females do in this species from the attendance-related activities that we focused on: nest defense, mate attraction, and WPCs. Accordingly, breeding males in this study incurred any cost of high attendance, and less time foraging, more than females did. This contrast between the sexes almost disappeared after the clutch was complete: the difference in attendance during the Main Incubation Period (nights 11 and after) was a small fraction of that in Figure 1’s timeline of pair formation and paternity establishment (Supplementary Material Table 4, Supplementary Figure 3).

Feeding sites and reproductive sites are spatially separated for seabirds, creating an unavoidable tradeoff. If male Nazca boobies foraged more efficiently than their mates, then higher attendance might not actually induce excess nutritional costs for males that are rooted in their management of the tradeoff. However, males apparently forage less efficiently than females (Anderson and Ricklefs, 1992; Howard et al., 2021; McKee et al., 2023), exacerbating the consequences of forgoing foraging opportunities. Male breeders may escape the tradeoff when they can: this is suggested by the anomalously low attendance of breeding males on the single night in our timeline when pair maintenance and paternity are irrelevant (Figure 1b). The dip in male attendance on night -1 was surprising to us because that occasion in individual breeding schedules is spread over many calendar nights at the population level (that is, breeding is asynchronous), and so was never observable in calendar-based behavior data. Many females did not attend the colony on the nights after their egg was first detected in daytime monitoring, apparently reflecting a quick daytime visit to lay, enabled by high attendance by their mates on nights 0 and 1 to cover the egg’s incubation. The lower attendance of females around laying matches observations from some other seabirds (e.g., Tickell, 1968; Quillfeldt et al., 2019), and also supports recouping body condition after relief from mating effort.

Other factors that we have not considered may also contribute to sex differences in colony attendance early in the breeding cycle. A prominent suggestion for seabirds (e.g., Siddiqi-Davies et al., 2025) is that females must assemble resources for egg production at sea, and males do not (in species without courtship feeding, like Nazca boobies). Certainly this applies to Nazca boobies, but probably offers no explanation for the higher male attendance as much as a month before clutch initiation (Figure 1a). Spatial sexual segregation at sea (e.g., Reyes-González et al., 2021) could also affect colony attendance in seabirds, although years of GPS tracking have not revealed spatial segregation by sex in Nazca boobies (McKee et al., 2023). We focused a priori on social factors that should affect male Nazca boobies acutely, given the unusual sexual agency of females in mate choice and copulation (Anderson et al., 2025). While the generally high attendance of males compared to females could be consistent with selective influences that are not related to social interactions of the sexes, the anomalously low attendance in the short window without paternity risk seems to be attributable to a single influence: confidence of paternity.

To our knowledge, this is the first evidence, albeit correlational, from a socially monogamous bird species of a tradeoff for males between self-maintenance and mating effort. For male Nazca boobies, the high attendance pays off in the form of high WPC frequency in the days preceding the first egg (Anderson et al., 2025). Ultimately, the EPF rate is estimated as 0, despite the frequent EPCs (Anderson and Boag, 2006; Anderson et al., 2025). Protecting paternity and forming pairs may involve other tradeoffs also, like between mate-guarding and pursuit of EPCs (Westneat et al., 1990). The tradeoff with self-maintenance that Figure 1 suggests should be accentuated in the case of Nazca boobies by the male-biased sex ratio and the complete sexual agency of females. Nonetheless, the spatial separation of food from the breeding site applies to all seabirds and, to a degree, to any bird species in which mating effort compromises nutrition.

Finally, we note that the unexpected attendance dip of males may indicate an awareness of when the pair’s first egg will appear. We do not know of other data bearing on this point in birds. How the male could become informed is unknown. Voluntary, purposeful signaling by the (well-informed) female could be involved, because the female should benefit during his incubation stints from any condition recovery that the male accomplishes during his night -1 absence.