Ex-situ avian sex skews: determinants and implications for conservation

Introduction

Sex ratios, that is, the ratio of males to females in a population, are considered a critical parameter influencing population growth and can undermine demographic potential when a skew to a particular sex goes beyond an optimal threshold (Lambertucci et al., 2012; Székely et al., 2014). Sex skews occur across a broad range of in- and ex-situ avian taxa and age cohorts, and are particularly prevalent in small and threatened populations (Donald, 2007; Morrison et al., 2016). It has recently been observed that sex ratios in wild plant and animal populations are changing as a response to anthropogenic effects such as global climate change, habitat loss, urbanisation and invasive species (Capdevila et al., 2020; Katzner et al., 2020; Schacht et al., 2022). To understand this trend and successfully conserve threatened species, population managers will require greater insight into population sex ratios, and how key drivers such as breeding systems and sex allocation may respond at the individual and population level (Undin & Castro, 2022). For example, even small sex skew variations from parity adversely affect ex-situ flamingo (Phoenicopterus spp.) populations, not only by limiting the number of breeding pairs, but via nest failures due to the disruptive behaviour of unpaired birds (Mooney et al., 2023).

Avian sex ratios may be calculated at various stages of development i.e., at fertilisation (primary sex ratio), at hatch (secondary sex ratio) or as the adult sex ratio (tertiary sex ratio). The critical parameter when investigating the effects and conditions of skewed sex ratios, however, is the effective or operational sex ratio (OSR). The OSR is defined as the number of reproductively active males to females in a population, thus providing the most accurate assessment of population viability through reproductive output. Most commonly, the OSR is observed at the individual level, where one female may produce more sons or daughters, but can also be calculated at the population level.

Avian sex ratios rarely occur at parity for many physiological and environmental reasons (Donald, 2007). Regardless of the cause, sex skews may reduce the effective population size via an imbalance in the number of males or females available for breeding, which is a critical parameter of population growth, sustainability and adaptive potential in wildlife (Frankham, 1995). When the effective population size becomes too low, there may be a trend towards increased homozygosity through inbreeding, loss of genetic variability and consequent genetic drift (Frankham, 1995; Charlesworth, 2009). In addition to genetic factors, skewed sex ratios may also have severe effects on population dynamics and social behaviour of individuals in small or closed populations (Tschumi et al., 2019) and may exacerbate population declines and undermine conservation efforts. Observations of the causative factors of sex ratio variation can be made at various developmental levels; however, ultimately start with the primary sex ratio.

The physiological drivers behind avian sex allocation at the primary sex ratio level lay with the female. In birds, females are the heterogametic sex, carrying both a Z and a W sex chromosome (males ZZ), so theoretically, sex allocation is under maternal control (Alonso-Alvarez, 2006; Dijkstra et al., 2010; Major & Smith, 2016). The mechanisms for variation in subsequent life stages are usually due to sex-biased mortality ((Potvin & MacDougall-Shackleton, 2010; Székely et al., 2014); also see reviews by Donald (2007), Payevsky (2021)).

Although a skewed sex ratio at any developmental stage may create short-term, proximate complications at the individual level, fundamental ecological theories suggest it will even out when viewed at the population level (Trivers & Willard, 1973; Frank, 1990). The precise mechanisms for this are unknown; however, some may correlate with species’ mating strategies (Nunney, 1993; Charlesworth, 2009). Polygamous species are expected to exhibit a stronger tendency towards population sex skews compared to monogamous species, with population fitness decreasing more rapidly in monogamous species than in species exhibiting polygamy due to the limitation of mates (Emlen & Oring, 1977; Frankham, 1995). In addition to sex-specific mortality, sex ratio bias has also been attributed to numerous other causative factors, including sex-specific within-clutch laying orders (Alonso-Alvarez, 2006), or the size (Anderson et al., 1993b) or rate of maturation of offspring (i.e., either the slower to mature or those with the best chance of reproducing in the same season are laid first) (Krebs et al., 2002; Dijkstra et al., 2010). Other influences include the attractiveness of a mate (Burley, 1981), weather patterns (Brooke et al., 2012), causative factors of declining threat status (i.e., small population size, alee effects) (Donald, 2007; Morrison et al., 2016) and deteriorating maternal condition throughout the laying period (Nager et al., 1999). If a predictable hypothesis for modelling sex allocation theory in birds is to be formed, it will require looking beyond theoretical studies of adaptive significance to include trends in genetics, ontogeny and phylogeny (Frank, 1990), and advancing our knowledge of the mechanisms of avian sex determination, species-specific mating strategies and various factors of an individual’s multidimensional environment. Data from ex-situ populations may inform the conservation management of species in the wild and in zoo settings.

The ex-situ environment removes many hurdles of observing sex ratios and sex allocation in a wild setting by enabling direct access to offspring and life history data at all stages of development. Additionally, it can provide insight from longitudinal management and expertise at a global level to align theories and predict patterns that may be extended to related taxa (Sheldon, 1998; Fisher & Owens, 2004; Mooney et al., 2023). These data may also be incorporated into population models, inform experimental manipulations and highlight potential problems for conservation programs (Taylor & Parkin, 2008; Conde et al., 2013). The success of ex-situ programs; however, is limited by multiple factors including institutional participation (i.e., the number of enclosures available), knowledge of species biology, and importantly, population sex ratios. Published accounts of avian ex-situ populations at genus and species level regularly report a sex skew to male (Taylor & Parkin, 2008; Ferrie et al., 2013; Stojanovic et al., 2018; Mooney et al., 2023); however this pattern may be shifting to more even sex ratios according to recent global analyses on avian offspring in ex-situ holdings (Machado & Miller, 2023).

If we are to effectively integrate the management of species between in- and ex-situ environments, understanding broad patterns of sex ratios in captive populations will increase management options through informing whether adaptive alterations to both environmental and social factors that influence sex allocation and fine-scale conditions during critical stages of ontogeny need to be considered (Wedekind, 2012). The ability to augment sex ratios will facilitate the strategic supplementation of additional males or females (Lenz, Jacob & Wedekind, 2007; Wedekind, 2012) to wild populations as required in threatened species recovery programs such as the orange-bellied parrot (Neophema chrysogaster) (Stojanovic et al., 2018; Troy & Lawrence, 2021), regent honeyeater (Anthochaera phrygia) (Heinsohn et al., 2022) and Andean condor (Vultur gryphus) (Lambertucci et al., 2013).

The conservation management of threatened species requires a multi-disciplinary approach based on a foundation of accurate, accessible data to inform adaptive management processes (Conde et al., 2013; Schwartz et al., 2017). The Zoological Information Management System (ZIMS) (Species360, 2023) is a sophisticated, real-time global records database of 22,000 animal species under human care and has been collecting data since 1974. More than 1,300 registered zoos and aquaria from 102 countries use the system to manage data on species’ life histories including births, deaths, parentage, movements, health and reproductive history. These data are shared globally by registered ZIMS members to achieve best practice in animal management and conservation goals (Species360, 2023). The increasing use, and success, of integrated management strategies (Schwartz et al., 2017) have seen the expansion of the database to link information about animals that transition between in-situ (in the wild) and ex-situ (in zoological facilities) environments. The mass of data collected on these species to date forms a rich resource for conservation research, including areas of reproductive biology, life history, ontogenesis, diagnosis and treatment of illness, nutrition, behaviour and the formation of biomaterial banks. Reports generated from ZIMS data present an unparalleled opportunity to understand animal reproductive biology in a controlled setting without the experimental manipulation of threatened species (Farquharson, Hogg & Grueber, 2021). Despite the quality and accessibility of this data, there is very little existing literature using this resource to observe sex ratio variation in birds on a global scale (Machado & Miller, 2023).

Using ZIMS data, we investigated the prevalence of sex skews in global avian populations. We used quantitative modes of enquiry to investigate the global proportion of males in 30 orders of birds. Five of these orders were chosen to examine in greater detail how the proportion of males related to (i) IUCN threat status, (ii) broad-level phylogenetic relationships and (iii) how the proportion changed at multiple scales (institutional to global). Three of these five focus orders were further selected to investigate the effect of housing within or outside of their natural range on the proportion of males. The findings in this study provide the first broadscale examination of ex-situ sex ratios and will contribute to the field of integrated species management.

Materials and Methods

To examine global patterns in sex ratios, all current (as at 01/03/2023) ex-situ avian holding records were exported from the ZIMS global resources portal, excluding domestics, breeds and varieties. All ambiguous taxonomy (i.e., listed only at genus level, or listed too broadly, such as ‘ducks, geese and swans’) and obsolete taxonomy (i.e., historical data from superseded taxonomic classification, as compared with the Birds of the World list as of the date 23/03/2023) were removed. Data listings were also removed if there were no known males or females (i.e., all individuals listed as unknown sex) or if there was only a single institution listing for that species. This resulted in a final dataset with 30 orders from 39,964 institutional listings. All subsequent analyses used institutional listings as a sample; herein described as a holding. Data regarding the holdings were limited to species (as well as genus and order), IUCN listing, holding mnemonic, holding location, number of males, number of females and number of ‘other’ (unknown) sex. Any further detail regarding the origin (e.g., wild-caught, captive-bred) of the individuals in the holding or the type of housing (e.g., indoor, outdoor aviaries etc.) were not included for this broadscale analysis. It is likely that most institutions reported the sexes of individuals at maturation, and that this study consequently examines adult sex ratios.

The sex ratio was calculated for each holding as a proportion of males to females, i.e., (M/(M+F)). These proportions were then used to calculate the mean global sex ratio (i.e., from all holdings within each of the 30 orders) as an exploration of the data. All subsequent analyses, however, utilised the full dataset using each institutional holding as an individual sample.

For the full dataset, significant differences in sex ratios (using each holding as an individual sample) were tested between (a) orders and (b) genera, to provide a broad indication of taxonomic signature on sex skews. That is, we used sex ratio (presented as a decimal indicating the proportion of males) of each holding as a response variable, and (a) order, and (b) genus as predictor variables to understand if sex ratios differed between these higher taxonomic groups. Jamovi version 2.2.5.0 (The jamovi project, 2023) was used for all inferential data analyses. Data were checked for skewness and normality (Shapiro-Wilk), with parametric or non-parametric equivalents applied where appropriate. A two-way analysis of variance (ANOVA) with Tukey’s post-hoc pairwise comparison with Bonferroni correction was employed to examine whether significant differences in the proportion of males were present between the predictor variables of taxonomic grouping (genus and order). Five focal orders were chosen for additional statistical analysis due to the quality of data, coverage of a wide variety of species, and global holding trends (i.e., ‘popular’ groups). They included the Struthioniformes (ratites), Psittaciformes (parrots and cockatoos), Columbiformes (pigeons and doves), Passeriformes (perching birds) and Coraciiformes (Kingfishers and allies). For these analyses, threat categories were included as a predictor variable, as set out by the International Union for the Conservation of Nature (IUCN, 2023) and included least concern, near threatened, vulnerable, endangered, critically endangered and extinct in the wild. To determine the effects of housing a species within or outside of their natural range, in-situ distribution as listed by the IUCN (2023) was compared to the country of ex-situ holding (ZIMS) in the Struthioniformes, Columbiformes and Psittaciformes. These orders were chosen due to their broad global distribution within the database and the presence of previously identified sex skews.

As with the previous analyses, data were checked for skewness and normality, with parametric or non-parametric equivalents applied when appropriate. A two-way ANOVA with Tukey’s post-hoc pairwise comparison were employed to examine whether differences in the proportion of males were present between IUCN threat status categories. An independent t-test was used to examine if the proportion of males significantly differed when holdings were inside or outside a species’ natural range.

To confirm the validity of our results, we calculated global sex ratios for (a) all orders and (b) all genera using a dataset which only included species with 100 holdings or more. We then compared these to the full dataset, to examine whether small token holdings were contributing to any observed skews in the above analyses.

Finally, in order to account for phylogenetic relationships and avoid pseudoreplication, we performed a Bayesian mixed model in R (R Core Team, 2021) using the package brms (Bürkner, 2017). First, we created a phylogenetic tree using our dataset through birdtree.org (Jetz et al., 2012), using the Ericson All Species source (10,000 trees with 9,993 operational taxonomic units (OTUs) each), and selecting a maximum likelihood tree using the phangorn package (Schliep, 2011). We then ran the mixed model, testing IUCN status and holding institution effects on the proportion of males at a species level, with the phylogenetic random effect included. We used uninformative priors, two chains, a warmup of 1,000 runs and an iteration of 3,000 runs for this analysis.

For the purposes of this study, populations were considered skewed if they deviated from the expected Mendelian sex ratio of 1:1 (Fisher, 1930). Given that observed proportions rarely conform to exact parity due to natural variation and dataset limitations, we applied a buffer of ±0.005 to account for minor fluctuations (Donald, 2007). Populations were categorized as male-skewed if the proportion of males was ≥0.505 and female-skewed if the proportion of males was ≤0.494. ZIMS data were current as of 23 March 2023.

Results

Proportion of males in global avian orders and genera

Overall, we found that 61% of all genera and 80% of all orders were skewed to male, 13% of genera and 16% of orders were skewed to female and 7% of orders and 8% of genera were at parity (Fig. 1 and Table S1). In the validation stage, where only species with 100 or more institutional holdings were analysed, percentages were found to be more even with genera consisting of 69% male skewed, 17% female skewed and 13% at parity; and orders of 68% male skewed, 16% female skewed and 8% at parity. Our ANOVA results showed that taxonomic rank affected the proportion of males at the genus level (p < 0.001); however, not at the level of order (p > 0.05) (Table 1). Orders that were skewed to female also showed greater variability between genera.

Table 1:

ANOVA test results showing the effect of, and interaction between, taxonomic grouping (genus and order) on the proportion of males in the global holdings of 30 avian orders.

Source	df	F	Sig.
Corrected model	829	1.983	<0.001
Intercept	1	4,507.427	<0.001
Order	2	0	1
Genus	774	1.798	*<0.001
Order * Genus	26	0.021	1

DOI: 10.7717/peerj.19312/table-1

Notes:

a. R squared = 0.040 (Adjusted R squared = 0.020).

Asterisk indicates significant result. df, degrees of freedom; F, F statistic.

Discussion

An initial objective of this research was to observe sex skews and avian sex ratio variation in global ex-situ populations. With respect to this, we found that although ex-situ sex ratios varied significantly at the institutional level, the orders were closer to parity at a global scale. This was a consistent factor throughout the 30 orders. Most striking was the substantial and unexpected finding in the disparity between the large proportion of orders with sex ratios skewed to male compared to female. Although there is evidence that because female Aves are heterogametic, there is a tendency for sex ratios to be male-biased (Pipoly et al., 2015), it is interesting to see the variability between groups within the class. The proportion of males held in each order consistently showed statistically significant variability, which in large-bodied charismatic avian families, such as some raptors (Accipitriformes, Falconiformes and Strigiformes) and parrots (Psittaciformes) should be interpreted with caution as members of these orders are also held singly for their sexually dimorphic characteristics, free-flight shows or other in-house education or entertainment purposes (Wilkinson, 2000; Brereton & Brereton, 2020), rather than with an intention to breed (Frynta et al., 2010). This is obviously an important factor in understanding the causative factors of ex-situ sex skews as it may explain, for example, variations in annual reproduction. Very little was found in the literature to further analyse and compare our results of broadscale sex ratio variation in ex-situ avian populations, which is surprising given the abundance and quality of data available. However, we can look to studies of wild birds, and accounts of individual species both in- and ex-situ to give our findings context and advance our understanding of sex ratio variation in ex-situ avian populations.

In sequence of increasing proportion of males, the basal avian order Struthioniformes (ratites) had the lowest mean proportion of males, followed by four taxonomically linked orders under the raptorial banner—the Cariamiformes (seriemas), Accipitriformes (hawks, eagles, vultures and kites), Cathartiformes (new world vultures) and Strigiformes (owls) (Jarvis et al., 2014; McClure et al., 2019; Méndez et al., 2022). A notable omission in this sequence is the taxonomically separated raptor order Falconiformes (falcons and caracaras), which is skewed to male. There are similarities between our broadscale findings and published accounts (Clutton-Brock, 1986; Xirouchakis et al., 2023). The majority of raptor orders exhibit reverse sexual size dimorphism whereby females are heavier than males (Massemin, Korpimäki & Wiehn, 2000; McDonald et al., 2005) so their skew to female contravenes empirical sex allocation theories. This direction of population sex skew to the larger sex also accords with an order with perhaps the most extreme example of sexual size dimorphism in the Aves, although this time to males, the bustards (Otidiformes).

Ingraldi (2005) postulated that a male bias in the usually female-skewed Accipitriformes was due to environmental stress. This is interesting as the release of the stress hormone corticosterone in the maternal female can trigger brood sex ratio adjustment and bias offspring sex ratios to male (Schoech et al., 2009; Geffroy & Douhard, 2019) as also reported in Passeriformes (Gam, Mendonca & Navara, 2011; Pryke et al., 2011) and Galliformes (Pinson et al., 2011). In a captive setting, genotypes and phenotypes of less stressed individuals may be selected for, due to greater fecundity and lower mortality over consecutive generations (Price, 1999; Crates, Stojanovic & Heinsohn, 2023). Such changes may be subtle and add to the complexity of understanding the phenotypic costs of captivity and contradictions of the Fisherian theory of sex allocation such as female-skewed ratios in raptors.

In addition to links in sexual size dimorphism, a difference in maturation age (age at first breeding) is also apparent in the raptorial orders (Daan, Dijkstra & Weissing, 1996) and bustards (Otidiformes) (Alonso et al., 2010), which have pervasive implications in sex allocation theory. For example, in raptors, female biased broods may occur early in the season as they reach sexual maturity earlier than males and thus have a better chance of reproducing first (Alonso et al., 2010; Carmona-Isunza et al., 2017) and the converse in species where males have a faster maturation rate (Daan, Dijkstra & Weissing, 1996). In long-term studies of raptor sex ratios, it was found that although brood sex ratios changed throughout each season (i.e., female biased in early and first hatched chicks and male biased in late and third hatched chicks or vice versa) the sex ratio at population level was closer to parity over time (Dijkstra, Daan & Buker, 1990; Daan, Dijkstra & Weissing, 1996; Xirouchakis et al., 2023). This highlights the importance of communicating the timing and type of sex ratio in the literature, and the use of this knowledge in ex-situ or integrated management programs where only first eggs or clutches may be permitted to hatch per season as a method of limiting production. It is especially relevant for conservation programs of critically endangered species such as the Bengal florican (Houbaropsis bengalensis) which has the added complexity of requiring a minimum captive release to the wild ratio of at least four males to one female for the lek mating strategy to function (Mahood et al., 2021).

Other studies that have found sex biases in captive populations have noted that supplementary feeding of female birds may encourage an excess of male offspring, leading to male biased sex ratios (Clout, Elliott & Robertson, 2002). This is one explanation put forward for a more limited study examining 80 species of Psittaciformes that found an overall male bias in captive holdings (>70% of species; Taylor & Parkin, 2008). Our study found an overall skew to males in Psittaciformes, with a mean sex ratio of 0.55, which was not extreme, and indeed individual holdings showed a wide range of sex ratios, indicating an intermediate result compared with this previous study and that from Machado & Miller (2023), who found that offspring sex ratios of Psittaciformes have historically been male-biased, but since 2000 demonstrate a median ratio of 50:50. This pattern of heavier biases in earlier time periods appeared to be common across many taxonomic groups (Machado & Miller, 2023). We did not delineate between juveniles and adults, nor did we make observations on offspring sex ratios specifically, which may account for our findings showing a moderate sex ratio bias in many taxa, not as skewed as historical datasets but also not predominantly at parity either, as it appears many taxa have become in captive settings (Taylor & Parkin, 2008; Machado & Miller, 2023).

Some of the species demonstrating male bias are cooperative nesting species that engage nest helpers, which is also supported in the available literature (Emlen, Emlen & Levin, 1986). This trend supports Hamilton’s kinship theory (Hamilton, 1967), that is, that parents will invest more in the helping sex where nest helper investment increases with kinship (Brown, 1978; Frank, 1990; Green, Freckleton & Hatchwell, 2016). Parallel to this, is the repayment model where the helping sex repays the breeding pair by way of helping with consecutive broods, and thus become the cheaper sex and support Fisher’s theory of equal investment (Emlen, Emlen & Levin, 1986). This may explain the skews to the helping sex in numerous genera, for example the Coraciiformes (kingfishers and allies) genera Halcyon (M = 0.75) and Merops (M = 0.55), and Passeriformes (perching birds) genera of Malurus (M = 0.75) and Picoides (M = 0.67). This is challenging in a zoo setting as family groups are infrequently kept together post breeding season, possibly further exacerbating the outcome of a male skewed ratio.

Conclusions

The purpose of this study was to gain a better understanding of sex ratio variation in global avian populations. The results show that sex skews to male are more than twice as prevalent in ex-situ populations than with their wild counterparts, and although many parallels were found within our data, the level of variation was often inconsistent with mainstream sex allocation theory. The findings of this investigation do, however, broadly support the Trivers-Willard hypothesis that breeding females will produce the sex which will return the highest reproductive value, which for female birds in good condition is usually male. Taken together, these findings have significant implications for the understanding and management of sex skew variation in ex-situ populations, especially for threatened species and integrated systems of management. Further research incorporating annual births/hatch data, as well as the motivation behind collection holding numbers from institutions would be useful. Additionally, investigations should be undertaken to examine links between primary sex ratio (PSR) and environmental stressors, maternal corticosterone and diet in a conservation or captive breeding context. To develop a better understanding of the social environment, exploring the effects of removing offspring on the sex of consecutive clutches in philopatric, cooperatively breeding species would be beneficial. Future studies are also recommended on the effects of barometric pressure and housing tropical species in temperate climatic regions. Overall, there is abundant room for progress in understanding sex allocation theory in birds, how it differs between taxonomic groups, and how it responds to change in both in-situ and ex-situ environments such that integrated management programs might have the best conservation outcomes.

Supplemental Information