In polygynous birds, where competition among males for access to females is particularly strong, parasite load of males is assessed by females before mating and hence may affect male breeding success [1]–[6]. Self-medication could be one of the mechanisms males use to appear as healthy and vigorous as possible, and thus more attractive to females, but this possible function has not been described.

Self-medication is used by some animals to prevent infections, fight against parasites or pathogens, or improve a variety of suboptimal physiological states [7]–[12]. It involves the consumption of plants, animals or minerals that have a prophylactic or therapeutic effect against disease agents which should lead to an increase in host fitness [7]–[10].

In this context, we examine here the consumption of poisonous insects by great bustards (Otis tarda) during the mating season. The great bustard is a polygynous bird with one of the most strongly skewed male mating success values among birds [13]. Males gather each year at traditional arenas (leks) where they perform elaborate sexual exhibitions directed towards females in order to express their status and condition [14]. Great bustards are among the few birds that feed on blister beetles (Berberomeloe majalis, Physomeloe corallifer; Fig. 1) [15],[16]. These insects are avoided by most animals because they contain cantharidin, a bitter-tasting and highly toxic defensive chemical with high immunogenicity [17] that acts in blister beetles as fungicide and nematocide [18]. Only a few species, such as the spur-winged goose Plectropterus gambensis and the northern leopard frogs Rana pipiens consume blister beetles, with a likely side effect of becoming toxic to predators [19],[20]. Cantharidin is also a well-known aphrodisiac compound that was obtained in the past from a beetle known as Spanish fly [21]. In humans, it causes priapism in men and pelvic congestion in women [22],[23]. Cantharidin-tolerant foragers would benefit from its anti-microbial and anthelminthic properties [24], and thus could enhance their health and their attractiveness to potential mates during the mate selection process.

We hypothesized that cantharidin, obtained by consumption of blister beetles, has a direct effect as an oral anti-microbial and pathogen-limiting compound. To examine this hypothesis, we combined in vitro experiments and field observations. In particular, we (i) evaluated the presence of potential agents of sexually transmitted diseases in male and female faecal samples during the mating season; (ii) tested whether pathogenic microorganisms from bustard faeces were reduced in experimental cultures exposed to dose-dependent cantharidin, and (iii) checked for presence of cantharidin in kidney and liver of great bustards, where it could be excreted or stored. Fulfilment of these evidences would not definitely demonstrate a possible self-medication function of cantharidin in great bustards, but should at least provide partial support for it.

Another appealing hypothesis is that consumption of blister beetles could also enhance the attractiveness of males to females by reducing their parasite load. This would counter-balance the risk of cantharidin poisoning [25] with reproductive benefits. In addition, males could benefit from cantharidin during the display season, when they may be particularly vulnerable to infections due to their presumably depressed immune system associated to their strenuous investment in sexual display [26],[27]. This hypothesis predicts a male-biased blister beetle consumption (i.e., both sexes would consume blister beetles as a healing or prophylactic agent, but males with potential to attract females should consume them also to enhance their body condition and mating opportunities). Previous studies have shown that age, weight, and display effort are the main predictors of male mating success in this species, and that whiskers and neck plumage are reliable indicators of male age and weight [13],[14]. However, the function of the meticulous inspection females perform of the male's cloaca prior to copulation is still unknown. A male's sexual display consists of a series of extravagant body postures and movements that end, when an interested female approaches, with a reiterative and almost obstinate exhibition of the cloaca, which is fully surrounded by pure white feathers that allow an easy detection of possible parasites or their remains [28],[29] (Fig. 2 A–C). Female bustards are extremely choosy during mating, accepting on average only one of ten males attempting to copulate with them [13]. Females closely inspect the male's cloaca, pecking around it as an essential part of mate selection, after which they decide whether to mate or not with the male [30],[31]. Our interpretation of this cloaca pecking is that the female is checking the health status of the male in order to select a copulation partner with genes for tolerance to cantharidin, the consumption of which allows for resistance to pathogens [2], while avoiding immediate disease transmission during copulations (Fig. 2 D–E) [4],[32],[33]. Similarly, by showing the underwing feather shafts and other parts of their ventral plumage (Fig. 2C), which are preferred sites for adults and eggs of mites and lice to stay or lay eggs, great bustards might show to visiting females that they are free of these ectoparasites.

Following this argument, males should consume an optimal amount of blister beetles to prevent and fight against pathogens that affect their health and/or act as agents of sexually-transmitted diseases. The benefits for health of feeding on blister beetles must be balanced against the toxicity of cantharidin, because an excess of blister beetles in the diet can be lethal for a great bustard [25]. To test this, we examined the potential of blister beetles to act against pathogens responsible for sexually transmitted diseases, and any preference for blister beetles by great bustard males. These additional analyses involved (i) determination of cantharidin concentration in blister beetles collected in the study area; and (ii) testing whether there is a sex bias in blister beetle consumption by male and female great bustards.

Great bustards consumed blister beetles, which contain cantharidin, a highly toxic compound that is lethal in high doses [25], but has a demonstrated nematocide and bactericide activity even at low concentrations [53]–[55]. There are various hypotheses that could explain this behaviour, the predictions of which are summarized in Table S1 and discussed in detail below.

The nutritional value hypothesis would explain blister beetle consumption based on their large size, low mobility and short handling time (which make them a highly profitable prey). In spite of the clumped distribution of blister beetles, which would allow bustards to eat more than one at a time, the birds usually consumed only one individual (a maximum of three, Table 2). This limited intake suggests that great bustards are not fully resistant to cantharidin, and indeed the intake of more than a few blister beetles is lethal [25]. Blister beetles may be consumed for nutritional purposes to the limits of their toxicity, however, male-biased blister beetle consumption is not explained through this hypothesis.

The predator deterrence hypothesis suggests that cantharidin would make great bustards appear toxic to their predators, thus helping them to avoid predator attacks, as suggested for the spur-winged goose Plectropterus gambensis, an African bird that also eats blister beetles [19]. In support of this hypothesis, we found cantharidin in the kidneys and livers of great bustards. Against this hypothesis, intoxication of bustard predators like golden eagles Aquila chrysaetus or foxes Vulpes vulpes has never been reported. In several Spanish localities, local people report that great bustards were once hunted for their meat, and this species was sometimes rated as a delicacy [56], with no reported cases of intoxication in humans. Thus, we discard the predator deterrence hypothesis as an explanation of the presence of cantharidin in the kidneys or livers of great bustards.

The self-medication hypothesis proposes a health benefit to the consumer. Great bustards may use cantharidin to fight against disease agents. Following de Roode et al. [7], five conditions must be met for food intake behaviour to qualify as medication:

Our results were consistent with most of Roode et al.'s conditions specified above. Thus, consumption of blister beetles by great bustards could be interpreted as a self-medication behaviour. Moreover, the higher consumption by males suggests an additional function only in this sex. As discussed below, the most plausible function we can think of is its use as a mechanism to appear more healthy to females and thus increase their mating success. Nevertheless, the suggested link between self-medication and sexual selection cannot be definitely tested through manipulative experiments using great bustards for various ethical and logistic reasons: (1) the species is globally endangered and strictly protected; (2) the risk of poisoning adult males in captivity, by allowing them to feed ad libitum on blister beetles; (3) capturing and recapturing wild individuals during courtship to take measurements and samples aiming at testing this prediction would imply increasing their stress level and exposing them to serious risk of death; (4) correlating mating success or pathogen prevalence to cantharidin consumption in individual males is practically impossible due to the low number of effective copulations [13],[30],[31],[61],[62], and to the difficulties mentioned in (3).

As for the higher consumption of blister beetles by males, it was clear from our results. First, the occurrence, abundance and biomass of blister beetles were much higher in the diet of male than female great bustards (Table 2), even after controlling for the sexual difference in body size (Table 2, weighted abundance, weighted biomass). Second, males but not females showed a clear preference for this prey among all invertebrates available (Fig. 4). And third, males selected those of largest size, showing significant large-size selection compared to females only in this (Meloidae) and another family (Tenebrionidae) of their prey spectrum, with much higher differences in average prey size in Meloidae. The absence of a sexual difference in cantharidin content in livers and kidneys of carcasses collected throughout the year did not necessarily contradict the male-biased consumption established through faeces analysis, because the presence of cantharidin in winter carcasses means that it was stored since the previous spring, and during this long time males and females could have made a differential use of the compound. For example, although males ingested more cantharidin than females, it is likely that males need more cantharidin to deal with a higher pathogen and parasite burden than females, since during courtship males surely have a reduced immune resistance associated to their strenuous investment in sexual display [13],[26],[27].

The absence of a sex difference in prevalence of cantharidin-sensitive pathogens in faeces during the breeding season (with the exception of Kocuria sp, which has not been cited as a bacterium producing STDs [63]), supports this conclusion; otherwise the observed male-biased cantharidin intake would have resulted in a higher abundance of pathogens in females.

If males use cantharidin to reduce their intestinal parasite and pathogen load, a clean cloaca could be an honest signal of a low probability of being a carrier of STD agents. Infections with some of the identified pathogens in bustard faeces such as E. coli and Sthaphylococcus sp, or parasite helminths, usually produce diarrhoea symptoms [64], which may be detected by females via visual inspection of the males' cloaca (Fig. 2E). Females probably accept mating with a male only after checking that helminths are not visible (Fig. 2D), and that diarrhoea and odour effects caused by these and other gastrointestinal parasites are absent around its cloaca [63]. Female pecking may induce the protrusion of the male's cloaca to better show the presence or absence of disease agents or their signs. We suggest that it is the importance of selecting the healthiest male that explains the exhaustive inspection of the male's cloaca performed by the female prior to mating, and contributes to the strongly skewed male mating success and extremely choosy mate selection shown by females in this species compared to other lekking birds [13]. Behaviours facilitating male health evaluation are particularly important in strongly polygynous species, in which a female is likely to mate with a male that has already copulated with other females, increasing the risk of disease transmission [65]. The ability of females to recognize and avoid infected males has been observed in other lekking birds [66],[67]. It is not unlikely that self-medication has also evolved in great bustards as a sexual selection mechanism capable of transmitting to females a signal of good resistance to a poisonous compound, in a similar way as other costly secondary sexual traits are exhibited by males of many species, and correlated with their mating success.

Exposure of the cloaca is widespread in pre-copulatory displays of birds and has been related with high probability of transmission of sexual diseases [29]. However, cloaca pecking has only been observed in three species [28],[68],[69]. In these species, it is males who inspect the cloaca of females, contrasting with great bustard, and the behaviour has been explained within a context of high sperm competition, extreme degree of polygynandry, or a high level of extra-pair copulations [29],[33]. Long-lived birds with mating systems characterized by multiple matings are the most likely candidates to have coevolved with virulent STDs [29],[32],[65]. This condition is fulfilled in the great bustard, and also in the great Indian bustard (Ardeotis nigriceps), a species that also consumes blister beetles (Cylindrothorax tenuicollis) and exhibits a similar cloaca inspection behaviour during courtship [70].

Another indirect argument supporting a relationship between self-medication and sexual selection is derived from the temporal scenario for the evolution of blister beetle consumption in great bustards. The intake of blister beetles in other species of the bustard family (Otididae) has been occasionally reported [70],[71] in non-sister branches along their phylogeny [72], suggesting that the potential to tolerate low levels of cantharidin could have evolved at early stages of the evolution of the family. Elaborate courting displays, including exposure of the cloaca, are a common feature for all known populations of great bustards [30],[31],[61], and therefore, sexual selection through female choice was developed in O. tarda before subsequent geographic isolation of populations occurred [73]. On the other hand, consumption of blister beetles by great bustards has been only reported for the Iberian populations [15],[16],[36],[37], suggesting that self-medicating behaviour might have evolved independently in these populations. An independent origin for self-medication in Iberian great bustards is fully congruent with the species' phylogeographic patterns and fossil record, which suggest genetic isolation of the Iberian populations with respect to other European nuclei for at least 200,000 years [73]–[75]. It is also congruent with the geographic distribution of blister beetles of the genera Berberomeloe and Physomeloe, which are present only in the Iberian Peninsula. Beetles of the family Meloidae co-occur with great bustards all over their geographic range [39], but large species of the genus Berberomeloe, are actively and conspicuously wandering at full sunlight in the steppe areas where great bustards perform their courtship display [40],[76]. Under this scenario, the rapid and recent evolution of blister beetle consumption, so far restricted at the population level, could be a direct consequence of the action of strong sexual selection at a local geographic scale.

In conclusion, the hypothesis that self-medication by male great bustards might have evolved in a sexual selection context still needs a definitive demonstration, but its proposal here is based on reasoned arguments and relevant literature (Table S1). We suggest that males use self-medication not only to overcome the pathogen burden during a strenuous display, but also to withstand the close scrutiny of choosy females, showing a clean cloaca as honest signal of health, and so increasing their mating success. The expected relationship between self-medication and subsequent greater mating success, necessary to demonstrate a direct link with sexual selection, is extremely difficult, if not impossible to measure in great bustards at present, but it could be explored in other species. In spite of the universality of sexual selection theory, self-medication has probably been overlooked as a sexually-selected mechanism enhancing male fitness.