Matthew B. Dugas

Carotenoid supplementation enhances reproductive success in captive strawberry poison frogs (Oophaga pumilio)

Amphibians are currently experiencing the most severe declines in biodiversity of any vertebrate, and their requirements for successful reproduction are poorly understood. Here, we show that supplementing the diet of prey items (fruit flies) with carotenoids has strong positive effects on the reproduction of captive strawberry poison frogs (Oophaga pumilio), substantially increasing the number of metamorphs produced by pairs. This improved reproduction most likely arose via increases in the quality of both the fertilized eggs from which tadpoles develop and trophic eggs that are fed to tadpoles by mothers. Frogs in this colony had previously been diagnosed with a Vitamin A deficiency, and this supplementation may have resolved this issue. These results support growing evidence of the importance of carotenoids in vertebrate reproduction and highlight the nuanced ways in which nutrition constrains captive populations.

Poison frog color morphs express assortative mate preferences in allopatry but not sympatry

The concurrent divergence of mating traits and preferences is necessary for the evolution of reproductive isolation via sexual selection, and such coevolution has been demonstrated in diverse lineages. However, the extent to which assortative mate preferences are sufficient to drive reproductive isolation in nature is less clear. Natural contact zones between lineages divergent in traits and preferences provide exceptional opportunities for testing the predicted evolutionary consequences of such divergence. The strawberry poison frog (Oophaga pumilio) displays extreme color polymorphism in and around the young Bocas del Toro archipelago. In a transition zone between red and blue allopatric lineages, we asked whether female preferences diverged along with coloration, and whether any divergent preferences persist in a zone of sympatry. When choosing among red, blue and phenotypically intermediate males, females from monomorphic red and monomorphic blue populations both expressed assortative preferences. However, red, blue, and intermediate females from the contact zone all preferred red males, suggesting that divergent preferences may be insufficient to effect behavioral isolation. Our results highlight the complexity of behavioral isolation, and the need for studies that can reveal the circumstances under which divergent preferences do and do not contribute to speciation.

Conspicuous and cryptic morphs of a polytypic poison frog differ in reproductive output because of differences in tadpole performance, not parental effort

Predation risk can drive life-history evolution in prey, with high adult mortality favouring the prioritization of current over future reproduction. Populations that evolve or adopt different or differently effective strategies to avoid predation, then, should evolve different life-history strategies. We compared reproductive output, under identical captive breeding conditions, of three allopatric morphs of polytypic poison frog (Oophaga pumilio) that likely experience different predation risk. We predicted that pairs of a well-defended (conspicuous and highly toxic) morph would prioritize future reproduction, and thus reproduce less often than a poorly defended (cryptic and less toxic) morph, while a cryptic but highly toxic lineage would be intermediate. These predictions were generally met: the conspicuous, toxic morph produced fewer juveniles than the cryptic morphs. However, the results of cross-fostering tadpoles among morphs suggested that these differences arose not from the quality of care parents provided, as predicted by life-history theory, but rather from differences expressed in tadpoles. Moreover, all cross-fostered tadpoles were less successful than tadpoles reared by their own parents, perhaps suggesting that parents discriminate against unrelated tadpoles or that parental care and offspring solicitation behaviours have diverged among populations. These results suggest opportunities for comparative studies exploring the entire complexity of the selective landscapes experienced by these polytypic frogs.

The pay-offs of maternal care increase as offspring develop, favouring extended provisioning in an egg-feeding frog.

Offspring quantity and quality are components of parental fitness that cannot be maximized simultaneously. When the benefits of investing in offspring quality decline, parents are expected to shift investment towards offspring quantity (other reproductive opportunities). Even when mothers retain complete control of resource allocation, offspring control whether to allocate investment to growth or development towards independence, and this shared control may generate parent–offspring conflict over the duration of care. We examined these predictions by, in a captive colony, experimentally removing tadpoles of the strawberry poison frog (Oophaga pumilio) from the mothers that provision them with trophic eggs throughout development. Tadpoles removed from their mothers were no less likely to survive to nutritional independence (i.e. through metamorphosis) than were those that remained with their mothers, but these offspring were smaller at metamorphosis and were less likely to survive to reach adult size, even though they were fed ad libitum. Tadpoles that remained with their mothers developed more slowly than those not receiving care, a pattern that might suggest that offspring extracted more care than was in mothers’ best interests. However, the fitness returns of providing care increased with offspring development, suggesting that mothers would be best off continuing care until tadpoles initiated metamorphosis. Although the benefits of parental investment in offspring quality are often thought to asymptote at high levels, driving parent–offspring conflict over weaning, this assumption may not hold over natural ranges of investment, with selection on both parents and offspring favouring extended durations of parental care.

Tadpole begging reveals high quality

Parents can benefit from allocating limited resources nonrandomly among offspring, and offspring solicitation (i.e. begging) is often hypothesized to evolve because it contains information valuable to choosy parents. We tested the predictions of three ‘honest begging’ hypotheses – Signal of Need, Signal of Quality and Signal of Hunger – in the tadpoles of a terrestrial frog (Oophaga pumilio). In this frog, mothers provision tadpoles with trophic eggs, and when mothers visit, tadpoles perform a putative begging signal by stiffening their bodies and vibrating rapidly. We assessed the information content of intense tadpole begging with an experimental manipulation of tadpole condition (need/quality) and food deprivation (hunger). This experiment revealed patterns consistent with the Signal of Quality hypothesis and directly counter to predictions of Signal of Need and Signal of Hunger. Begging effort and performance were higher in more developed and higher condition tadpoles and declined with food deprivation. Free‐living mothers were unlikely to feed tadpoles of a nonbegging species experimentally cross‐fostered with their own, and allocated larger meals to more developed tadpoles and those that vibrated at higher speed. Mother O. pumilio favour their high‐quality young, and because their concurrent offspring are reared in separate nurseries, must do so by making active allocation decisions. Our results suggest that these maternal choices are based at least in part on offspring signals, indicating that offspring solicitation can evolve to signal high quality.

Larval aggression is independent of food limitation in nurseries of a poison frog

Aggression between nurserymates is common in animals and often hypothesized to result from proximate resource limitation. In numerous terrestrial frogs, larvae develop in phytotelmata, tiny water bodies where resources are scarce and competition, aggression, and cannibalism are all common between individuals sharing these nurseries. In some species, mothers provision phytotelm-bound young with trophic eggs, a strategy that compensates for low nutrient availability and could allow mothers to reduce costly aggression and cannibalism among nurserymates. We tested this hypothesis using strawberry poison frog (Oophaga pumilio) tadpoles, staging secondary depositions in arenas occupied by residents that had either been food deprived or fed ad libitum. Resident tadpoles were nearly all aggressive and most killed intruders, but aggression was unrelated to resident food deprivation. Unlike most related frogs studied, O. pumilio residents did not cannibalize their victims. This result supports the hypothesis that proximate food limitation and aggression can be independent.Significance statementAggression and cannibalism are common in the resource-limited nurseries in which many young animals develop. An intuitive hypothesis for this is that individuals are aggressive because they are hungry. Although this hypothesis has usually been supported in birds, we found no support for it in a test we conducted in a terrestrial frog that feeds its young with trophic eggs. Resident tadpoles fed ad libitum were just as aggressive to staged intruders as those that were food-deprived, and were just as likely to kill intruders. Residents did not cannibalize their victims. These results suggest that hunger-independent aggression in young animals is taxonomically widespread, and demonstrates an important ecological constraint on parents: they cannot prevent fatal aggression, they cannot “feed” tadpoles to older siblings, and thus, they cannot use occupied rearing sites.

Commentary: Parental care and the proximate links between maternal effects and offspring fitness

Maternal effects influence the phenotype of offspring through non-genetic mechanisms, and thus are important components of individual life-histories and act as drivers of and/or constraints on phenotypic evolution. A maternal effect common in egg-laying vertebrates is provisioning of the yolk with carotenoids, organic pigments that often color sexual ornaments and are hypothesized to play positive and substantial physiological roles. In a recent study, yolks of great tit (Parus major) eggs were directly supplemented with carotenoids, and the effects on offspring fitness proxies measured (Marri and Richner in Oecologia 176:371–377, 2014a). Nestlings from supplemented broods were heavier early in development and more likely to fledge, but otherwise equivalent to control nestlings. The authors consider in detail the potential physiological mechanisms that might underlie this result, and here I expand on their Discussion by considering a non-exclusive explanation: that parents provided higher quality care to broods that received supplemental carotenoids. I discuss the general non-independence of pre- and post-hatching/parturition maternal effects when parents care for offspring, and then briefly review evidence that carotenoids specifically are tied to the intensity of avian begging displays. Finally, I detail how inclusive fitness opportunities and constraints shape the adaptive landscape in which maternal effects operate, highlighting both theoretical and applied concerns surrounding questions about the adaptiveness of maternal effects.

An experimental test for age-related improvements in reproductive performance in a frog that cares for its young

Reproductive performance often increases with age in long-lived iteroparous organisms, a pattern that can result from within-individual increases in effort and/or competence. In free-living populations, it is typically difficult to distinguish these mechanisms or to isolate particular features of reproduction-influencing outcomes. In captive Oophaga pumilio, a frog in which mothers provide extended offspring provisioning via trophic eggs, we experimentally manipulated the age at which females started breeding and then monitored them across repeated reproductive events. This experiment allowed us to decouple age and experience and isolate maternal care as the proximate source of any differences in performance. Younger first-time mothers produced larger broods than older first-time mothers, but did not rear more offspring to independence. Across repeated reproductive events, maternal age was unassociated with any metric of performance. At later reproductive events, however, mothers produced fewer metamorphs, and a lower proportion of individuals in their broods reached independence. These patterns suggest that performance does not improve with age or breeding experience in this frog, and that eventual declines in performance are driven by reproductive activity, not age per se. Broadly, age-specific patterns of reproductive performance may depend on the proximate mechanism by which parents influence offspring fitness and how sensitive these are to effort and competence.

Morphological correlates of river velocity and reproductive development in an ornamented stream fish

Adaptive phenotypic divergence can arise when environments vary in ways favoring alternative phenotypic optima. In aquatic habitats, the costs of locomotion are expected to increase with water velocity, generally favoring a more streamlined body and the reduction of traits that produce drag. However, because streamlining in fish may come at the cost of maneuverability, the net benefits of drag reduction can differ not only among habitats, but also among individuals (or classes of individuals) that rely on locomotion for different uses (e.g., males vs. females or adults vs. juveniles). We tested these predictions by exploring relationships among river velocity, body streamlining, ornamental fin size, and male reproductive condition in the steelcolor shiner (Cyprinella whipplei), a small-bodied North American cyprinid. Overall, males in peak reproductive condition (defined by the development of sexually dimorphic tubercles) had less streamlined bodies and larger ornamental fins than males in lower reproductive condition or individuals lacking these secondary sexual characters (females and immature males). There was a relationship between river velocity and body streamlining only for males in peak reproductive condition, but it was in the opposite direction of our predictions: these males were less streamlined in faster rivers. We found only weak support for the prediction that ornamental fin size would be negatively associated with river velocity. Overall, these results suggest either that drag is not an important selective pressure in these habitats, or that the sexual selection advantages of a deep body and large fin compensate any natural selection costs for C. whipplei males. This study highlights the often overlooked diversity of selective pressures acting on streamlining in fishes, and can offer novel insights and predictions allowing a more nuanced understanding of fish ecomorphology.

Choosy Cannibals Preferentially Consume Siblings with Relatively Low Fitness Prospects.

When an individual can selfishly cannibalize a relative or altruistically set it free, the benefits of altruism will be positively associated with the relative’s fitness prospects (the benefits it receives from altruism). We tested the prediction that altruism should be preferentially directed toward high-quality relatives using larvae of the New Mexican spadefoot toad (Spea multiplicata), a species in which tadpoles plastically express omnivore and carnivore ecomorphs. In a no-choice design, we presented carnivores with sibling or nonsibling omnivores varying in developmental stage, which is positively associated with survival in this toad’s ephemeral larval environment. There was a significant interaction between relatedness and developmental stage on the probability of cannibalism: carnivores were overall more likely to cannibalize less developed omnivores, but this effect was exaggerated when the potential victim was a sibling. This evidence that altruists favor relatives with high fitness prospects highlights the numerous factors shaping altruism’s payoffs.