Evolution and maintenance of male care: is increased paternity a neglected benefit of care?

In many animals, including birds, fish, and insects, the degree of male care often (but not always; Sheldon, 2002) correlates positively with the level of paternity assurance, across species (Møller and Birkhead, 1993; Smith, 1997; Wright, 1998), within species (Dixon et al., 1994; Sheldon et al., 1997), and even within individuals (Freeman-Gallant, 1996; Fu et al., 2001). The established explanation for this relationship is that high confidence of paternity is a prerequisite for male care to evolve because a male should be selected to care only for his own offspring (Alexander and Borgia, 1979; Smith, 1997; Trivers, 1972; Wright, 1998). However, males providing care may enjoy better success in sperm competition, or in avoiding it, and the resulting increase in paternity can be sufficient to explain the evolution and maintenance of male care. This scenario results in the same positive relationship between male care and paternity, but with reversed causality.

Although males of most animals provide no parental care, there are numerous examples of species in which males do supply care. Paternal care is particularly widespread in fishes, insects, and birds but also occurs in some mammals and anurans (Clutton-Brock, 1991; Reynolds et al., 2002; Ridley, 1978). Here I deal with male care of eggs or young in the widest sense, specifically including both pre- and postzygotic care. Such care can range from nutritious contributions to female gametes and nest building to brood defense and provisioning of young (Clutton-Brock, 1991). Before proceeding, I want to clarify my use of two important terms. First, I favor the term male care to paternal care, simply because care is not always directed exclusively towards the true genetic offspring of the male. Second, I use the term paternity to refer to the proportion of a brood that is fathered by a male, so paternity can range from zero to one.

The evolution of male care requires benefits exceeding the costs of care to the male. In most discussions, males that provide care are thought to lose out on future mating opportunities, while benefits are assumed to accrue from enhanced offspring fitness (Clutton-Brock, 1991). Although female care evolution is not the topic of this article, it is worth keeping in mind that under an unbiased sex ratio, males and females on average face the same loss of future mating opportunities if parents have identical future prospects of reproduction (Balshine-Earn and Earn, 1998; Kokko and Jennions, 2003; Queller, 1997). Similarly, regardless of which sex provides the care, the two parents share the fitness benefit from better offspring survival equally.

In addition to natural selection, sexual selection has been invoked as a selective mechanism for the evolution of male care through female choice (Hoelzer, 1989; Lotem et al., 1999; Ridley, 1978; Tallamy, 2000; Trivers, 1972; Wagner et al., 1996), which sometimes even results in competition among females for good fathers (Owens et al., 1994; Petrie, 1983). Tallamy (2000) provides an excellent review of how this mechanism is likely to explain the evolution of exclusive male care, with examples from insects and other arthropods. Tallamy argues that because males willing to guard are preferred as mates they enjoy greater mating success than males that are unable or unwilling to guard and that this (better than any natural selection argument) explains why males guard even completely unrelated eggs.

However, my argument is that male care can increase the caring male's paternity among current or future young. In addition to enhanced offspring fitness and mate attraction, this represents a third, largely neglected potential benefit to males of providing care. Under this hypothesis, high paternity is not a prerequisite for male care to evolve but rather an outcome of it. If paternity increases with male care, as frequently may be the case (see later), male care may evolve even if other costs (in terms of lost opportunities for additional matings) and benefits of care (in terms of male mating success or offspring survival) remain unchanged.

Sperm competition, which generally results in a dilution of paternity within each particular brood, is now agreed to be a very powerful selective agent in many species (Birkhead and Møller, 1998). Still, the possibility that male care could be selectively advantageous as an adaptation to reduce risk or intensity of sperm competition has remained virtually unexplored both theoretically and experimentally. In fact, early theoretical work dismissed the idea that paternity would be very important for the evolution of male care based on an assumption that the level of paternity stays the same whether the male exhibits care or not (Maynard Smith, 1982: 127; Werren et al., 1980: 621–622). Although more recent theoretical work suggests a relationship between paternity and male care to be quite plausible, the focus in these studies has been directed at the influence of paternity on male care, rather than the other way around (e.g., Houston, 1995; Kokko, 1999; Møller, 1998; Queller, 1997; Sheldon, 2002; Wade and Shuster, 2002; Westneat and Sherman, 1993; Wright, 1998; but see Knowlton and Greenwell, 1984). Yet there appears to be no theoretical reason why sexual selection could not select for male care through its influence on paternity. This possibility has not yet been investigated. Nevertheless, I believe we can gain important new insights for example by awarding males a higher paternity as a benefit of providing care in new models or as modifications of existing models (such as Harada and Iwasa, 1996; Houston et al., 1997; Kokko, 1999; Werren et al., 1980; Westneat and Sherman, 1993).

A model

A common assumption is that males lose opportunities for additional matings while providing care, compared to males not providing care. If this is true, then A < 1, constituting a cost of care. However, this need not always be the case, as seen for example in studies of the facultatively caregiving wrasse Symphodus tinca, in which care-giving males attract a huge number of spawning females to their nests compared to noncaring males (Warner et al., 1995). Thus, in wrasse and other similar systems A > 1.

An increased paternity as a consequence of care (i.e., B > 1) will provide an advantage to the care strategy. Such an advantage applies only when sperm competition is an issue, that is, when multiple males may gain mating access to the same female. Note that I am specifically rejecting Werren et al.'s (1980) assumption of P_C = P_NC, which has been repeated in textbooks as recently as that by Alcock (1998: 525).

If we follow the common apprehension and assume that C > 1, that is, that “care” is beneficial to the offspring, an increase in paternity is not necessary for care to evolve. However, relaxing such an assumption and setting C equal to 1, we can understand the evolution of carelike behavior such as that of the pine engraver beetle Ips pini. In this species, males remove frass from the tunnels where females lay eggs, and careful study has shown that this behavior increases male paternity assurance without having a noticeable influence on offspring survival (Lissemore, 1997). The Savannah sparrow, Passerculus sandwichensis, provides another similar example (Freeman-Gallant, 1996, 1998; and see later).

A very important consideration is that a single behavior is capable of conveying multiple benefits to an individual (Clutton-Brock, 1991; Simmons, 1995; Wedell, 1994). For example, in many species of fish the male constructs a nest in which he guards and fans the eggs until they hatch. Such nest-building behavior has the potential for positive influences on multiple factors in the model and thus may provide up to three benefits to the male, including (1) increased mating success (e.g., if females base their mate choice on nest appearance), (2) increased paternity share (e.g., if the nest helps the male to defend his paternity against sneaker males), and (3) increased survival of the offspring (e.g., if eggs are easier to defend from egg predators when sheltered by a nest).

Thus, in addition to enhanced offspring fitness and mate attraction, a reduction in sperm competition due to male care may be an important selective factor for male care to evolve. Two clearly testable predictions emerge from this hypothesis. In mating systems where a certain behavior simultaneously increases a male's paternity success and the fitness of his offspring, (a) such a behavior should be more likely to evolve than in a system in which the behavior merely increases offspring fitness, and (b) the optimal level of investment into this behavior should be higher than predicted based on the expected costs and benefits solely in terms of offspring fitness (similar to what Lotem et al., 1999, predicted from what they called “the overlooked signaling component” of parental care). Territoriality in fish might be an example of prediction (a), as this behavior is thought to improve both the territory holder's paternity and the survival of the eggs spawned in his territory, with little or no cost in terms of lost additional mating opportunities (Ah-King et al., 2005). This interpretation is consistent with the observation that territoriality has evolved repeatedly (at least 15 times) in fishes (Ah-King et al., 2005). A testable case and a supporting example for prediction (b) are given under Current Paternity (later).

The relative benefits of male care as described in the model are likely to vary between conspecific males and may depend, for example, on body size, condition, and attractiveness of the individual male. This may lead to alternative mating tactics, which are fairly common in fish. For example, sneaker males have been shown to lower the paternity of pair-spawning males considerably in many fish species, including brook trout, Salvelinus fontinalis (Blanchfield and Ridgway, 1999), sand gobies, Pomatoschistus minutus (Jones et al., 2001), and blue-gill sunfish, Lepomis macrochirus (Fu et al., 2001). In many nest-building species of fish, nest owners are larger than nonnesting sneaker males, and while nest defense may enable large males to enjoy a limited number of relatively secure matings, the smaller sneaker males are constrained to participate in numerous insecure matings (reviewed in Avise et al., 2002; Taborsky, 1998). In the common goby, Pomatoschistus microps, males have been shown to shift from sneaking to guarding as they grow larger, while medium-sized males may act as either sneakers or guarders depending on the size distribution of other males (Magnhagen, 1992). Thus, one and the same male may solve the inequality in the model differently, depending on his age and relative body size.

How male care can influence paternity

The examples listed in this section are meant to illustrate how male care may serve to reduce the risk or intensity of sperm competition in species from various animal taxa and with various mating systems. The effect of care on paternity can be divided into two categories based on whether it influences the paternity of current offspring that receive the care (current paternity) or fertilization success in future matings (future paternity). While some effects of prezygotic care on current paternity are well established, potential effects of postzygotic care on current and future paternity have not yet been considered in a conceptual way. Thus, this idea is a largely untested hypothesis, and these examples are not meant to provide evidence for the hypotheses so much as illustrations of the plausibility of the idea.

Current paternity

Courtship feeding in insects provides a well-understood example of a behavior that can influence paternity and fitness of the same offspring. During mating in many insects, males provide females with a nuptial gift (Thornhill, 1976; Vahed, 1998), which often represents a substantial parental investment in the offspring by the male (e.g., Simmons, 1995). A larger gift not only increases offspring fitness but also results in more time for ejaculate transfer or a longer refractory period (interval before the female is ready to mate again), increasing the probability that the eggs will be fertilized by the donor male (Simmons and Gwynne, 1991; Wedell, 1991, 1994). Thus, by providing a nuptial gift these males gain higher paternity among the offspring in which they invest (Simmons, 1990), a dual benefit that is widely recognized (Simmons, 1995; Wedell, 1994). In relation to aforementioned prediction (b), the optimal size of nuptial gifts that provide such dual benefits to the male would be expected to be greater than would be predicted solely from the benefits of increased offspring fitness.

Nest building by sand gobies, P. minutus, provides another example that may be a bit less obvious. The male builds his nest under a mussel shell, which he covers with sand, and makes a small nest opening out of sand and mucus. A small nest opening has both costs and benefits. If exposed to low levels of dissolved oxygen, males increase both their fanning effort and the size of the nest opening for better oxygenation of the eggs, but in the presence of an egg predator, they trade fanning efficiency for egg protection by reducing the size of the opening, both under low and high levels of oxygen (Lissåker et al., 2003; Lissåker M and Kvarnemo C, unpublished data). However, the nest structure seems to be important for defense not only against egg predators but also against sneaker males (Svensson and Kvarnemo, 2003). Sneaker males quite frequently manage to enter the nest, thus diminishing the benefits of providing care. Of 23 nests collected in the field, sneaking was genetically documented in 52% of the nests, and 22% of the eggs in these sneaked nests were fertilized by a male other than the nest-guarding male (Jones et al., 2001). However, gobies seem to have evolved measures to reduce the problem. Before mating, particularly during courtship, the male often swims upside down in his empty nest, attaching a layer of sperm-containing mucus to the shell where the female will later attach the eggs. This behavior has been documented in several species of gobies (Marconato et al., 1996), including sand gobies (Svensson and Kvarnemo, 2005). In the grass goby, Zosterisessor ophiocephalus, such sperm have been shown to remain fertile for hours or days (Scaggiante et al., 1999). Thus, (1) by providing a nest with a small opening, the male can be expected to better defend his mating success from sneaker males, and (2) by preparing the nest with sperm-containing mucus before spawning, he should both have more time to guard the nest opening and be at an advantage in terms of sperm numbers if sneaking occurs. Thus, in gobies nest building seems to serve not only to increase egg survival but also the current paternity of the caring male. The relative importance of these two aspects of nest building is likely to differ over a male's reproductive cycle, such that the significance of protection against sneaker males gradually decreases and protection against egg predators gradually increases with the number of spawnings. Sand goby males build relatively small nest openings even before any spawning has taken place. The fact that they build only slightly smaller nest openings when housed together with an egg predator (shore crab) than when housed alone, but significantly smaller openings when housed together with two sneaker males (Svensson and Kvarnemo, 2003), is consistent with my aforementioned prediction (b) that the optimal investment into a behavior that increases both paternity success and offspring fitness of a male should be higher than predicted based on benefits of offspring fitness alone. In conclusion, there are several reasons for building a small nest opening despite the extra expenditure in terms of fanning. This behavior cannot be understood fully unless we acknowledge that it is not solely under natural selection to reduce egg predation but also under sexual selection to reduce sneaking success.

Future paternity

The relationship between paternity and male care has been studied extensively in birds (e.g., Dixon et al., 1994; Dunn and Cockburn, 1996; Sheldon and Ellegren, 1998). However, most of these studies have focused on possible adjustments in male care according to natural or experimentally manipulated levels of current paternity as perceived by the male. Yet, as a stimulating exception, Freeman-Gallant (1996) investigated how male care influences paternity in future broods. The Savannah sparrow, P. sandwichensis, which commonly breeds twice during a breeding season, exhibits a high level of female infidelity, with 16–30% of all young sired through extrapair fertilizations. Interestingly, females were shown to respond to males that provide care by increasing their fidelity in subsequent broods. Thus, males that provided more care in the first brood increased their share of paternity in the second brood, while males furnishing less care suffered a reduction in paternity (Freeman-Gallant, 1996). A similar pattern has also been observed in the moustached warbler, Acrocephalus melanopogon (Blomqvist, in press). Consequently, in these species males may actually benefit from provisioning unrelated offspring in earlier broods (Freeman-Gallant, 1997). However, to experimentally test whether care can influence a male's paternity in future broods the level of care must be manipulated, for example, by preventing males to provide care during part of the care period. Such males would be predicted to end up with lower mating and/or paternity success in the next breeding period relative to unrestricted males. To further differentiate between mating and paternity success, the paternity of each male would need to be measured in relation to his mating success (see also later).

The chacma baboon, Papio ursinus, may provide another tentative example of male care influencing the future paternity of the caregiving male. In this species, females often mate with several males (Anderson, 1992). Generally, male care, expressed as carrying of infants, is positively related to their estimated paternity, even though some males carry unrelated infants. Anderson (1992) suggested that females preferentially mate with males that provide care. Thus, in addition to paternal investment, there could be a second function of male care: to secure a high share of future paternity.

In African assassin bugs of the genus Rhinocoris, males care for eggs by guarding them during incubation. The existence of male care in this group of insects has been explained by both sexual selection through female choice (Tallamy, 2000), natural selection through offspring survival, and low costs of male care (Manica and Johnstone, 2004). Nevertheless (given that last-male sperm precedence is by far the most common pattern of fertilization in insects, including Hemiptera, which the assassin bugs belong to; Simmons, 2001: Table 2.3 and Figure 2.3), effects on paternity could provide a further advantage and perhaps influence the degree of male care in this taxon. Males fight for the right to guard an egg mass, regardless of whether they have sired the eggs, presumably because females only leave eggs with males that are already guarding. However, a guarding male only accepts new eggs if he gets to mate with the female first. Because the female lays the eggs immediately after mating, the male is guaranteed to be the last male to mate with her (Thomas, 1994; Tallamy, 2000, and references therein). Assuming last-male sperm precedence, this behavior should benefit his paternity within the brood. Hence, it is conceivable that male care in these assassin bugs is sexually selected through a promotion of both male mating success and paternity success. In summary, the mating systems of chacma baboons and assassin bugs possess important similarities to Savannah sparrows in that a positive effect of male care on paternity of future offspring seems plausible. Such a pattern may be much more common among various taxa than we currently appreciate.

As discussed in the model previously, the number of offspring sired by a male will depend both on his mating success (the number of females he can attract and the number of offspring each female produces) and on his paternity success (the proportion of these offspring that he will actually father). While mating success is determined before or at copulation, paternity success may be the result of both pre- and postcopulatory sexual selection. In addition to male-male competition and sperm competition, female choice is likely to influence both mating and paternity success substantially. Mechanisms allowing a polyandrous female to influence the paternity of her mates include allowing favored males priority access to mate at receptivity (Michl et al., 2002), ejecting sperm from unwanted mates (Pizzari and Birkhead, 2000), or cryptically choosing sperm among competing ejaculates (Eberhard, 1996). Thus, in species with polyandrous females paternity success is clearly a separate issue from mating success, yet many times mating success has been used uncritically as an estimate of paternity success.

CONCLUSIONS

In this article I propose that enhanced paternity may be gained by providing care and could prove to be an important factor for the evolution and maintenance of male care. I also argue that this benefit may accrue whether care is directed at related or unrelated offspring. However, it is imperative to stress that such a paternity benefit for caring males does not exclude other naturally and sexually selected benefits, such as higher offspring survival or increased mating success. Rather, these factors might reinforce each other. The relative importance of the different mechanisms is likely to vary from case to case. This hypothesis of selection for male care through a paternity gain should be applicable to a range of taxa and mating systems as well as to theoretical models. The tentative examples from insects, fishes, birds, and mammals presented in this article are meant to inspire more studies specifically designed to test the predictions of this model. Finally, contrasting between the present hypothesis (idea and model presented in this article) and the one outlined in the introduction that males should be selected only to care for their own offspring, one can make two testable predictions over evolutionary time. First, according to the latter hypothesis, care is expected to have evolved only when males have high paternity, whereas the present hypothesis predicts that male care can evolve also at initially low levels of paternity. Second, if the level of promiscuity and sperm competition increases, thereby diluting paternity, male care is often predicted to be lost (e.g., Kokko, 1999; Møller, 1998). In contrast, the present hypothesis predicts that male care will remain the same or even increase in the face of increased sperm competition as selection for enhanced paternity should be particularly important under such circumstances.

I thank Sami Merilaita for helping construct the model, Malin Ah-King, Ingrid Ahnesjö, Donald Blomqvist, Anne Houde, Adam G. Jones, Hanna Kokko, Sami Merilaita, Geoff A. Parker, Douglas W. Tallamy, Staffan Ulfstrand, and anonymous referees for kindly commenting on the manuscript and Donald Blomqvist for letting me cite unpublished work. I also thank numerous other colleagues for discussing these ideas with me at various stages of forming my thoughts. The Swedish Research Council and the Crafoord Foundation through the Swedish Royal Academy of Sciences financially supported this work.

References