How Mate Availability Influences Filial Cannibalism

Volume 91, No. 1 THE QUARTERLY REVIEW OF BIOLOGY March 2016

HOW MATE AVAILABILITY INFLUENCES FILIAL CANNIBALISM

Nicholas D. S. Deal School of Biological Sciences, Monash University Victoria 3800, Australia e-mail: [email protected]

Bob B. M. Wong School of Biological Sciences, Monash University Victoria 3800, Australia e-mail: [email protected]

keywords ﬁlial cannibalism, infanticide, mate choice, parental care, sex ratio, sexual selection

abstract Parents sometimes eat their young to reduce the consequences of brood overcrowding, for nutritional gain, and/or to redirect investment toward future reproduction. It has been predicted that filial cannibalism should be more prevalent when mate availability is high as parents can more easily replace consumed young. Reviewing the available evidence—which comes almost exclusively from studies of paternal caring fish—we find support in some species, but not others. To explain this, we hypothesize that sexual selection against filial cannibalism and/or the tendency to acquire larger broods under conditions of high mate availability discourages filial cannibalism. Additionally, filial cannibalism might occur when mate availability is low to facilitate survival until access to mates improves. Since attractiveness can also influence remating opportunities, we review its effect on filial cannibalism, finding that attractive parents engage in less filial cannibalism. More research is needed to determine if this relationship is a result of individuals showing adaptive plasticity in filial cannibalism based on self-perceived attractiveness, or if the attractiveness of individuals is reduced by their propensity to commit filial cannibalism. More generally, to advance our understanding of how mate availability influences filial cannibalism, future studies should also focus on a wider range of taxa.

Introduction 2012). One especially intriguing form of ILLING one’s own offspring appears infanticide is filial cannibalism, which in- Kto be the antithesis of a good repro- volves parents not only killing, but also eat- ductive strategy, yet such behavior is likely to ing their own offspring. This phenomenon be a significant—but poorly recognized— has been reported in a wide range of taxa, source of mortality among developing juve- including arthropods (Bartlett 1987; Mori niles in many species (Mock 2004; Moreno and Chiba 2009), fish (FitzGerald 1992;

5552.proof.indd 1 Achorn International 02/16/2016 11:13PM 2 THE QUARTERLY REVIEW OF BIOLOGY Volume 91

Manica 2002b; Lindström and St. Mary can offer nutritional benefits to parents 2008), amphibians (Okada et al. 2015), rep- (Candolin 2000a; Kume et al. 2000; Manica tiles (Huang 2008), birds (Tortosa and Re- 2004; Okuda et al. 2004; Gomagano and dondo 1992; Gilbert et al. 2005), and mam- Kohda 2008; Mehlis et al. 2009; Takeyama mals (Bronson and Marsteller 1985; Beery et al. 2013). Furthermore, both can function and Zucker 2012). as brood management. For example, total There has been considerable effort fo- filial cannibalism can enable parents to free cused on understanding why parents engage up valuable space for larger more profitable in filial cannibalism. In this respect, filial can- broods, as seen in Egyptian mouthbrooders nibalism is widely suspected to be adaptive, (Pseudocrenilabrus multicolor; Mrowka 1987) since it would otherwise be selected against and fantail darters (Etheostoma flabellare; if it did not confer fitness benefits to parents Lindström and Sargent 1997). Similarly, (see Lindström and St. Mary 2008 for a de- partial filial cannibalism can enable par- tailed discussion of this point). In particu- ents to remove slower developing offspring lar, much work has focused on identifying or ensure that the number of offspring in the ways in which parents can benefit from a brood does not exceed that which can eating their own young, and under what be supported by parental provisioning, circumstances animals are compelled to do as shown, for instance, in burying beetles so (FitzGerald 1992; Manica 2002b). To this (Nicrophorus tomentosus; Trumbo 1990) and end, two broad classes of nonmutually ex- sand gobies (Pomatoschistus minutus; Klug clusive benefits have been identified. First, and Lindström 2008). In the case of Syrian eating offspring may provide cannibalistic hamsters (Mesocricetus auratus), consump- parents with energy or nutrients (Bronson tion of offspring may even enable parents and Marsteller 1985; Schneider and Wade to manipulate the sex ratio of their broods 1989; Hoelzer 1992; Kraak 1996; Lindström (Beery and Zucker 2012). and Sargent 1997; Kvarnemo et al. 1998; Regardless of how parents benefit from Lindström 1998; Manica 2004; Okuda et al. eating their young, whenever they do so, 2004; Mehlis et al. 2009; Takeyama et al. they are killing off progeny that they might 2013), which can enable the parent to otherwise have reared to directly contrib- produce or better care for other offspring ute to their fitness. Hence, in order to gain (Rohwer 1978; Manica 2002b). Second, in a complete understanding of the circum- common with other forms of infanticide stances under which animals should engage (Mock and Forbes 1995), filial cannibalism in filial cannibalism, we must not only con- could be beneficial in allowing parents to sider the benefits of offspring consumption manage the size or composition of their but also the costs. For parents, whenever broods (Payne et al. 2002, 2004; Creighton the potential benefits derived from eating 2005; Klug et al. 2006; Beery and Zucker young outweigh the cost to replace them, 2012). For example, reducing the number filial cannibalism becomes an effective of young in the brood can be beneficial if strategy. Therefore, the cost of filial canni- offspring survivorship or quality are density- balism will largely be determined by what is dependent (Payne et al. 2002, 2004; Klug required to produce offspring equivalent to et al. 2006). Parents that selectively consume those that were consumed. Included in this low value offspring can also avoid wasting is the cost associated with parental effort time or resources caring for them (Klug and as well as energetic, temporal, and survival Bonsall 2007; Klug and Lindström 2008). costs of remating where necessary. Scientists recognize two forms of filial The costs associated with parental effort cannibalism: total filial cannibalism, where required to replace eaten offspring can all of the young a parent currently has un- help to predict the circumstances under der its care are consumed, and partial filial which filial cannibalism occurs. For exam- cannibalism, where only a fraction of the ple, younger offspring are more likely to fall brood is consumed (Rohwer 1978; Manica victim to filial cannibalism than older off- 2002b). Both of these forms of cannibalism spring, which corresponds to the parental

5552.proof.indd 2 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 3 effort required to replace them (Schwanck more filial cannibalism, as they can more 1986; Petersen and Marchetti 1989; Lavery easily replace young (Okuda and Yanagi- and Keenleyside 1990; Petersen 1990; Man- sawa 1996b; Manica 2002b). This possibility, ica 2002a). Furthermore, within biparental which was raised as a key area for research species, there is evidence to suggest that in the last major review of the topic (Manica fathers have a greater inclination toward 2002b), was, at the time, based on a single filial cannibalism than mothers, which can study (Okuda and Yanagisawa 1996b). Since be partly explained by the greater expen- then, however, there have been several diture required from females to produce more empirical contributions investigating gametes to replace eaten young (Lavery the role of mate availability on filial canni- and Keenleyside 1990; Raadik et al. 1990). balism (Table 1). Similarly, the high incidence of filial can- This review focuses on the prediction nibalism among teleost fish may, at least that filial cannibalism levels are mediated partly, be due to the prevalence of exclu- by mate availability. We begin by summa- sive paternal care in this group, although rizing the empirical findings on the topic detailed studies are required to verify the to illustrate that individuals have been potential link between parent sex and filial observed to respond to changes in mate cannibalism in uniparental species. availability by increasing filial cannibalism The mating effort required to replace rates in some species (Okuda and Yanagi- offspring might also influence a parent’s sawa 1996b; Okuda et al. 2004; Myint et al. incentive to eat them. Here it is likely that 2011a; Takeyama et al. 2013), whereas in the costs of remating are even more vari- other species, filial cannibalism rates re- able than those associated with parental main unchanged (Bjelvenmark and Fors- effort. This is because an individual’s op- gren 2003) or even decrease when pro- portunities to remate are dependent on spective mates are made more accessible environmental conditions—both physical (Pampoulie et al. 2004; Klug et al. 2005). and social. Furthermore, when filial canni- To explain this, we explore a range of po- balism is performed to improve future off- tential mechanisms through which mate spring production or care (as opposed to availability can influence cannibalism. In benefit existing, uneaten young), the suc- particular, we propose mechanisms to ad- cess of this strategy is entirely dependent dress the unexpected findings of reduced on the outcome of remating attempts. For levels of filial cannibalism among animals these reasons, the likelihood and costs of experiencing high mate availability. These remating should be one of the principal include: (1) the possibility that heightened factors mediating when individuals engage mate availability leads to individuals having in filial cannibalism. or expecting to have more young in their There are a number of factors that should broods—and thus greater rewards from car- influence the cost of remating for an indi- ing for such broods discourages filial canni- vidual. Principal among these is the pres- balism under these circumstances; (2) that ence of mature animals of the opposite sex individuals will commit filial cannibalism and their willingness to breed with the in- when mate availability is low to facilitate dividual, which should depend on the op- their survival until times of improved mate erational sex ratio (ratio of sexually active availability; and (3) that committing filial males to fertilizable females at any given cannibalism can deter mates and thus is an time) and the cost of mate search (Emlen inappropriate strategy when many poten- and Oring 1977; Clutton-Brock and Parker tial mates are around. 1992; Kokko and Monaghan 2001), all of Once we have outlined the possible which can broadly be encompassed under mechanisms through which mate availabil- the term “mate availability” (Kondoh and ity could influence filial cannibalism, we Okuda 2002). Thus, it has previously been consider whether the different forms of fil- predicted that animals experiencing high ial cannibalism (partial and total) respond mate availability should, on average, commit differently to mate availability. Since total

5552.proof.indd 3 Achorn International 02/16/2016 11:13PM 5552.proof.indd 4 Achorn International 02/16/2016 11:13PM

TABLE 1 Summary of the results of studies examining the relationship between the availability of mates to parental male fish and the filial cannibalism levels of these males Relationship between level of Males capable Experimental mate availability experienced of brooding provisioning of food by an individual and: multiple clutches to males during Total filial Partial filial Study species Form of care simultaneously? parental care cannibalism cannibalism Reference Cardinalfish (Apogon Mouthbrooder No None + Unexamined Okuda and Yanagisawa doederleini) (1996b) Two-spotted goby Nest-brooder Yes None Not significant Not significant Bjelvenmark and Forsgren (Gobiusculus flavescens) (2003) Sand goby (Pomatoschistus Nest-brooder Yes Daily − Not significant Pampoulie et al. (2004) minutus) Goby (Rhinogobius sp. OR) Nest-brooder Yes None* +** Okuda et al. (2004) Flagfish ( Jordanella floridae) Nest-brooder Yes Daily − Unexamined Klug et al. (2005) Lizard goby (Rhinogobius Nest-brooder Yes*** None Not significant + Myint et al. (2011a) flumineus) Lizard goby (Rhinogobius Nest-brooder Yes*** None Unexamined + Takeyama et al. (2013) flumineus)

*Males were only left to care for their eggs for a small proportion of the brooding period, during which experimenters did not provide food. The effect of different prespawning feeding regimes was examined in this study. **Due to dissection of males prior to the completion of brooding, it is uncertain whether this result applies to total ﬁlial cannibalism, partial ﬁlial cannibalism, or both. ***Typically, however, only single clutches are brooded. March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 5

filial cannibalism can only be seen as an in- quiring the limiting resource. This might vestment into future reproduction, it might be especially important if increased levels be expected that total filial cannibalism in of mate availability lead to the acquisition particular should be favored by high mate and consumption of young becoming a availability. We discuss the empirical support, more viable means of attaining food than or lack thereof, for this hypothesis, as well as traditional foraging. other possible mechanisms through which partial and total filial cannibalism might be affected differently by mate availability. empirical evidence From there, we address how key factors such Although filial cannibalism is known to as the timing of exposure to, and contact occur among all of the major vertebrate with, additional mates might influence filial groups, and to be performed by both males cannibalism. We then broaden our consid- and females showing various forms of pa- eration of how access to mates influences rental care the only available evidence di- filial cannibalism by discussing how an in- rectly testing the effect of mate availability dividual’s attractiveness could affect their on filial cannibalism comes from studies of propensity to devour their young. Follow- fish with exclusive paternal care (Table 1). ing this, we examine the related question The findings from these studies are never- of how the quality of prospective mates theless insightful and show that a diverse influences filial cannibalism, an area that range of responses to altered mate avail- so far has received little attention. Finally, ability can occur. Some studies support at the end of the review, we examine the the orthodox prediction that heightened relationship between filial cannibalism and mate availability will facilitate greater rates mate availability at the macroevolutionary of filial cannibalism (Okuda and Yanagi- scale. sawa 1996b; Okuda et al. 2004; Myint et al. 2011a; Takeyama et al. 2013). For example, in their now classic study, Okuda and The Effect of Mate Availability Yanagisawa (1996b) showed that male car- on Filial Cannibalism at the dinalfish (Apogon doederleini) that had com- Level of the Individual mitted total filial cannibalism were able to The orthodox view is that individuals remate more quickly than males that had will increase the amount of filial cannibal- their broods taken from them by the ex- ism they perform when mate availability is perimenters. This suggests that the canni- high (Okuda and Yanagisawa 1996b; Man- balistic males had above average access to ica 2002b; Bjelvenmark and Forsgren 2003; mates when they ate their young. However, Okuda et al. 2004; Myint et al. 2011a,b). the causal relationship between access to The reasoning behind this is that an ani- mates and filial cannibalism is not entirely mal’s expected future mating ease and suc- clear in this case. It is possible that, rather cess will be greater when many mates are than access to mates having caused males to available. The value of existing young will commit filial cannibalism, the act of eating therefore be lower, since they are more eggs itself may have provisioned cannibal- easily replaced, and so the cost of losing istic males with energy that helped them young from filial cannibalism is lessened. attain additional mates (Manica 2002b) or Following from this, the greater ease and triggered changes in their physiology that likelihood of remating that comes from el- lead to faster remating. Observations by evated mate availability should mean that Takeyama et al. (2002), however, show that parents have the opportunity to reinvest only males with access to a female-biased energy gained from filial cannibalism into operational sex ratio are able to remate future young. Moreover, if the reproduc- quickly after filial cannibalism reinforcing tive success of individuals is limited by ac- the notion that male cardinalfish utilize sur- cess to food rather than mates, then filial plus females to quickly replace cannibalized cannibalism could provide a means of ac- broods. An increase in filial cannibalism

5552.proof.indd 5 Achorn International 02/16/2016 11:13PM 6 THE QUARTERLY REVIEW OF BIOLOGY Volume 91 among individuals with enhanced access to mechanisms favoring a negative mates has also been reported in manipula- association between mate tive experiments on gobies (Rhinogobius sp. availability and Filial cannibalism OR). Specifically, Okuda et al. (2004) found Increased Expectation of Brood Size that male gobies paired with two gravid fe- with High Mate Availability Leads males consumed more eggs than males that to Reduced Filial Cannibalism were paired with a single gravid female. Because males offered an additional mate In fish, it is well established through consumed more eggs, even when they suc- field correlates and brood size manipula- ceeded in spawning with only one female, it tions that individuals are often more likely is likely that mate availability itself, and not to completely consume small broods than merely extra eggs from an additional mate, large broods (Kramer 1973; Schwanck 1986; caused the increase in cannibalism. Mrowka 1987; Petersen and Marchetti 1989; Other studies, by contrast, have found no Lavery and Keenleyside 1990; Petersen effect of mate availability on either total 1990; Petersen and Hess 1991; Forsgren et al. (Bjelvenmark and Forsgren 2003; Myint et al. 1996; Lindström and Sargent 1997; Man- 2011a) or partial filial cannibalism (Bjelven- ica 2002a; Pampoulie et al. 2004; Lissåker mark and Forsgren 2003; Pampoulie et al. and Kvarnemo 2006; Myint et al. 2011a; 2004). For instance, a study of two-spotted but see Payne et al. 2003). This is because gobies (Gobiusculus flavescens), which created the care provided by most fish, usually egg a high mate availability treatment (by expos- guarding and fanning, is considered to be ing brooding males to gravid females) and essentially “nondepreciable” (sensu Altmann low mate availability treatments (by exposing et al. 1977; Blumer 1979; Clutton-Brock brooding males to other males or no other 1991; Smith and Wootton 1995; but see fish), found no effect of mate availability on Klug et al. 2006). In other words, for each filial cannibalism (Bjelvenmark and Fors- unit of parental expenditure, the benefit gren 2003). received by each member of the brood re- Indeed, several studies have found that mains largely unchanged with increasing heightened mate availability can even de- brood size. As a consequence, larger broods crease the frequency of filial cannibalism. should offer a better payoff to the parent Specifically, in both sand gobies (P. minu- than smaller broods. Indeed, the expected tus) and flagfish ( Jordanella floridae), males number of young surviving from a small that are exposed to females while brood- brood could be so low that the parent may ing have been shown to completely con- actually be better off eating the entire brood sume their broods less often than brood- rather than caring for the young and endur- ing males that are not exposed to females ing the associated costs (Rohwer 1978; Pe- (Pampoulie et al. 2004; Klug et al. 2005). tersen and Marchetti 1989; Manica 2002b). Clearly, the orthodox prediction of height- It is important to note, however, that strictly ened mate availability leading to increased nondepreciable care is not a precondition filial cannibalism does not accord with for this brood size effect. However, cost per cases where a greater access to poten- young raised in large broods should be less tial suitors has been associated with un- than that of smaller broods. changed or even reduced rates of filial can- The tendency for parents to completely nibalism (Bjelvenmark and Forsgren 2003; consume small broods might result in in- Pampoulie et al. 2004; Klug et al. 2005). In creased total filial cannibalism by individu- the next section we outline some potential als experiencing low mate availability. This mechanisms through which heightened ac- is because, in species where males can care cess to mates could elicit reduced levels of for the young of multiple females at once, cannibalism. low mate availability may result in males

5552.proof.indd 6 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 7 acquiring smaller broods due to access to Myint et al. 2011a). In accordance with fewer females. Consequently, under such this, we note that it is only total filial canni- a scenario, total filial cannibalism may be- balism that has been observed to decrease come more prevalent, as it tends to occur in response to elevated access to mates in response to small brood size. It should (Pampoulie et al. 2004; Klug et al. 2005). be noted, however, that this phenomenon Moreover, the expected brood size mech- should only occur if males can expect mate anism relies on the assumption that males availability to increase in the future, since will avoid consuming small broods where males should only commit total filial canni- additional mates are likely to add young to balism of small broods if they can expect to them. Therefore, this mechanism is only attain larger broods in the future. relevant for those species in which males Although the effect of brood size might brood the eggs of many females at once. In- explain a negative association between fil- deed, those species that decrease total filial ial cannibalism and mate availability in the cannibalism in response to mate availabil- field, it cannot explain experimental results ity do have males that care for the young that have explicitly controlled for brood of multiple females at once (Pampoulie size (e.g., Pampoulie et al. 2004; Klug et al. et al. 2004; Klug et al. 2005). By contrast, 2005; Lindström and St. Mary 2008). How- among those species where males increase ever, it is conceivable that even when brood filial cannibalism in response to mate avail- size is controlled for, males may base their ability, males never or only rarely brood the filial cannibalism decisions on their expected young of multiple females at once (Okuda brood size. Consider that a male experienc- and Yanagisawa 1996b; Okuda et al. 2004; ing elevated mate availability may expect to Myint et al. 2011a; Takeyama et al. 2013). have a greater number of clutches added to This suggests that elevated mate availability his brood. If males respond to the expec- might favor increased total filial cannibal- tation of large brood size with a reduced ism in species where males expect to care tendency to engage in total filial canni- for the young of only one female at a time, balism, as they do when they actually have whereas for species in which males often large broods, then high mate availability care for offspring from several females at should result in reduced levels of total filial once, elevated mate availability could pro- cannibalism. mote males to continue to engage in brood To recap, we have outlined two logical care since the likelihood of obtaining a mechanisms through which the increase in large brood is increased. expected future reproductive success that corresponds to high mate availability might influence filial cannibalism. From the or- Filial Cannibalism as a Strategy thodox position, animals should be more to Survive Times of Low Mate willing to commit filial cannibalism when fu- Availability/Mating Success ture mating is likely, since consumed young Another mechanism that could favor a are easily replaced. Paradoxically, an ele- negative association between filial canni- vated likelihood of additional mating might balism levels and mate availability involves also favor reduced total filial cannibalism, the use of filial cannibalism as a strategy since the possibility of adding more young to survive times of low mate availability. It to a brood could increase the incentive to has previously been suggested that eating provide care. Differences between these young may provide crucial nutrition, and two mechanisms do seem to accord with the can also free a parent from the demands available empirical evidence. For example, of engaging in costly parental behaviors the expected brood size mechanism is only (Petersen and Marchetti 1989; Smith and relevant for cases of total filial cannibalism, Wootton 1995). Filial cannibalism has the because it is total and not partial filial can- potential, therefore, to improve parental nibalism that tends to be committed in re- survival and should be employed to pro- sponse to small brood size (Manica 2002b; long survival where the expected increase

5552.proof.indd 7 Achorn International 02/16/2016 11:13PM 8 THE QUARTERLY REVIEW OF BIOLOGY Volume 91 in fitness from surviving longer is greater operating against filial cannibals is the pref- than the fitness loss associated with con- erence of females to deposit their eggs in suming young. As a consequence, when ac- nests that already contain eggs. This prefer- cess to mates varies over time, under some ence has been reported in a number of taxa circumstances, it may befit an individual (e.g., assassin bug, Rhynocoris tristis, Thomas to perform filial cannibalism when mates and Manica 2005; harvestman, Pseudopu- are scarce. This strategy would facilitate crolia sp., Nazareth and Machado 2010), the survival of the individual until mates but is particularly well documented in fish become more abundant or accessible and (Ridley and Rechten 1981; Marconato and should be favored if the fitness benefits Bisazza 1986; Sikkel 1988; Unger and Sar- of surviving longer are greater for those gent 1988; Knapp and Sargent 1989; Kraak individuals experiencing low (rather than and Videler 1991; Goldschmidt et al. 1993; high) mate availability. Following this, when Kraak and Groothuis 1994; Forsgren et al. many mates are available, filial cannibalism 1996; Manica 2010; also reviewed in Reyn- may be reduced, so that animals can max- olds and Jones 1999), where males some- imize reproductive gains during the time times steal or adopt eggs from other males of peak mate availability. Importantly, this to use to attract mates (Rohwer 1978; Unger mechanism should only act where mate and Sargent 1988; but see Östlund-Nilsson availability to individuals can increase with 2002). Intriguingly, it has been proposed the passage of time. Such variation could be that female preference to oviposit alongside predictable, as might occur with breeding other eggs may even be a counterstrategy to seasons, aging, and growth. Alternatively, male filial cannibalism in some species, as it variation could arise from stochastic events, protects eggs via a dilution effect and since such as the movements of mates into and the likelihood of total filial cannibalism de- out of an animal’s territory. creases as brood size grows (Rohwer 1978; Kraak 1996; Kraak and Weissing 1996; Lind- ström 2000). For example, female sand go- Sexual Selection Against Filial Cannibals bies (P. minutus) prefer egg-tending males Facilitates Low Cannibalism Rates and, as a result, gain direct benefits through During Heightened Mate Availability reduced filial cannibalism (Forsgren et al. In some species, sexual selection may act 1996), as do female Mediterranean blen- against filial cannibalism. This may arise di- nies (Aidablennius sphynx ; Kraak and Videler rectly from potential mates avoiding mating 1991; Kraak and Groothuis 1994; Kraak with known filial cannibals or indirectly as a 1996). Consequently, filial cannibalism may consequence of other forms of mate choice. promote the evolution of this egg-laying For example, in the sand goby (P. minutus) strategy, thus potentially explaining why it has been suggested that females might ac- the two behaviors so commonly co-occur tively avoid males that appear to have eaten within species (Kraak and Weissing 1996; young (Lindström and Kangas 1996). It has Lindström 2000). Females, of course, may further been theorized that, in some spe- choose egg-tending males for other reasons, cies, females might use “test eggs” to avoid including those benefits associated with mating with males with a predilection for mate choice copying (Gibson and Höglund offspring consumption (Manica 2010). In- 1992), as well as for protection from other deed, female scissortail sergeants (Abudef- egg predators through the dilution effect duf sexfasciatus) sometimes deposit small and because offspring in large broods may numbers of eggs within the nests of males, elicit greater parental effort from their fa- and return a short time later to assess the thers (Sargent 1988; Jamieson 1995). How- care provided to their eggs before deciding ever, regardless of the female motivation for whether or not to commit a full clutch to favoring egg-tending males, this phenome- the attendant male (Manica 2010). non should result in males that engage in Perhaps the most widespread phenom- total filial cannibalism being less successful enon that could result in sexual selection at attracting mates over the short term.

5552.proof.indd 8 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 9

The likelihood that females directly or in- turn, should provide further disincentive for directly avoid mating with males that have males to engage in filial cannibalism during committed filial cannibalism has implica- times of high mate availability. tions for the patterns of filial cannibalism displayed in relation to mate availability. Specifically, since filial cannibalism may be effects of mate availability on total aversive to females, it could be that males and partial filial cannibalism are reluctant to engage in this behavior Considering studies of species with exclu- unless their access to prospective mates is sive paternal care in which males can tend sufficiently high that, even after deterring the clutches of multiple females simulta- some mates through offspring consump- neously, there is an emerging pattern for to- tion, their reproductive success is not lim- tal and partial filial cannibalism to respond ited by access to females. This might be differently to changes in mate availability expected to occur in species such as the sig- (Table 1). For instance, experimental stud- nal blenny (Emblemaria hypacanthus), where ies have shown that heightened mate avail- male reproductive success can be limited by ability increases the intensity of partial filial the amount of space available for egg depo- cannibalism in lizard gobies (Rhinogobius sition within their gastropod shell nesting flumineus; Myint et al. 2011a; Takeyama et al. site (Hastings 1992). 2013) but decreases the incidence of to- Female aversion to mate with males en- tal filial cannibalism in flagfish ( J. floridae) gaging in filial cannibalism might also have and sand gobies (P. minutus ; Pampoulie the opposite effect: discouraging males et al. 2004; Klug et al. 2005). The latter is from consuming their young when access surprising given that total filial cannibalism to mates is high. Such a relationship might can only be a successful strategy when the arise if males strategically engage in filial parent is able to produce a new brood, the cannibalism most often when few females likelihood of which presumably increases are around, either because this means that (rather than decreases) with heightened ac- the number of mates that are deterred cess to mates. By contrast, partial filial can- by cannibalism is minimized, or because nibalism could be beneficial for an individ- males may be able to covertly engage in ual even if another brood is not produced filial cannibalism with few females around (Payne et al. 2002, 2004; Creighton 2005; to detect it. Klug et al. 2006; Beery and Zucker 2012). Future investigations might also benefit So why might total filial cannibalism be from considering whether female choosi- negatively associated with mate availabil- ness varies with mate availability to males, ity while partial filial cannibalism shows a and how this influences filial cannibalism. positive association? This observation, of For instance, if the availability of mates to course, may simply reflect the low number males is elevated as a result of a decrease of studies on this topic. However, plausi- in the operational sex ratio, then females ble biological explanations also exist. As may become less discriminating in their previously discussed, it is total filial canni- mate choice—and thus males might be per- balism (and not partial filial cannibalism) mitted to engage in more filial cannibalism that parents might avoid during times of under such conditions. By contrast, if the high mate availability since they expect availability of mates to males is elevated more clutches to be added to their brood. as a consequence of more frequent male- Similarly, if, as previously suggested, males female encounters (as might occur when forego filial cannibalism at times of peak population density increases, or mate mate availability so as not to deter potential search becomes safer), instead of a change mates, then these males might benefit most in the operation sex ratio, we could expect from avoiding total filial cannibalism. This females to become more choosy (Pomian- is because female mate choice strategies kowski 1987; Real 1990; Slagsvold and Dale are likely to penalize total filial cannibal- 1991; Milinski and Bakker 1992). This, in ism more so than partial filial cannibalism.

5552.proof.indd 9 Achorn International 02/16/2016 11:13PM 10 THE QUARTERLY REVIEW OF BIOLOGY Volume 91

Furthermore, if filial cannibalism is em- experimental considerations ployed as a strategy to survive times of low mate availability, as we hypothesized earlier, Timing of Presentation of then total filial cannibalism should proba- Prospective Additional Mates bly be favored as it can provide nutrition The timing of mate exposure can have and free the parent from the demands of important implications when trying to brood care, whereas partial filial cannibal- uncover the effects of mate availability ism only does the former. on fil ial cannibalism. As highlighted re- A reduction in total filial cannibalism cently by Myint et al. (2011a), studies that and increase in partial filial cannibalism found an increase in filial cannibalism might also be expected to follow a rise in with heightened mate availability had ad- mate availability if parents are able to gain ditional mates presented prior to spawn- enough energy through partial filial canni- ing (Okuda and Yanagisawa 1996b; Okuda balism that they no longer need to engage et al. 2004; Takeyama et al. 2013). By con- in total filial cannibalism under such condi- trast, when differences in mate availability tions. Evidence for this is, however, currently between treatments were only manipulated weak, as no study has concurrently reported after spawning, this effect was not observed an increase in total filial cannibalism and a (Bjelvenmark and Forsgren 2003; Pampou- decrease in partial filial cannibalism as a re- lie et al. 2004; Klug et al. 2005). Further- sult of heightened mate availability. Mathe- more, Myint et al. (2011a) demonstrated matical models that can isolate mechanisms that only exposure to additional females through which mate availability affects filial prior to spawning, and not after, elicited cannibalism may be useful in determining elevated levels of filial cannibalism in male which mechanisms are necessary to explain lizard gobies (R. flumineus). The generality different effects of mate availability on total of these results, however, remains unclear and partial filial cannibalism. because brooding male R. flumineus often Interestingly, there is currently only one close off their nest entrance after spawning species, the cardinalfish A. doederleini, in (Myint et al. 2011a,b). Thus, in contrast to which individuals have been reported to many other species that potentially could engage in total filial cannibalism more fre- rely on postspawning mate availability cues, quently when their access to mates is in- male R. flumineus may be adapted to re- creased (Okuda and Yanagisawa 1996b). spond only to prespawning cues (Myint et al. Perhaps the most important difference be- 2011a). Still, perhaps there is a more gen- tween this species and those that have so far eral theoretical explanation as to why pre- been found to reduce total filial cannibal- spawning (and not postspawning) mate ism in the face of elevated mate availability availability cues trigger filial cannibalism. is the form of parental care. In the case of For example, maybe prespawning cues of the latter, males typically have the ability to mate availability more accurately reflect tend the young of multiple females simulta- the level of mate availability an individ- neously. By contrast, A. doederleini is a pater- ual will experience after consuming their nal mouthbrooder in which males are con- young because postspawning cues are mis- strained to caring for the young of a single leading since the individual possesses young female at a time (Okuda et al. 1997). Thus, that might attract additional female atten- male A. doedeleini must consume any eggs tion (see discussion of female preferences they are brooding before they can gain im- for egg-tending males in the section, Sexual mediate access to others (Okuda and Yan- Selection Against filial Cannibals Facilitates agisawa 1996b). Clearly, further research on Low Cannibalism Rates During Heightened other mouthbrooders would help to redress Mate Availability). To further clarify the the strong bias toward studies on nest brood- importance of prespawning cues, future ers with simultaneous polygyny and confirm studies are needed that examine the effects how brooding style interacts with the effect of prespawning mate availability in species of mate availability on filial cannibalism. that have previously been shown not to al-

5552.proof.indd 10 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 11 ter their filial cannibalism in response to to propose that the effect of mate avail- postspawning mate exposure. ability on filial cannibalism might interact with the effect of food availability. In this respect, filial cannibalism could potentially Role of Physical Contact with be more prevalent under conditions where Prospective Additional Mates potential mates are abundant, but food is It has also been suggested that the failure scarce. In particular, if limited food sup- of some studies to find a positive effect of plies (rather than access to mates) restricts mate availability on filial cannibalism could the number of young that can be acquired be due to experimental setups that prevent and reared, then the consumption of some physical contact with any additional mates offspring could provide parental males with offered (Okuda et al. 2004). For example, a means to capitalize on abundant access to Bjelvenmark and Forsgren (2003) and mates and acquire more food. Under such Pampoulie et al. (2004) found that the circumstances, extra nourishment pro- presence of females did not increase rates vided to males by filial cannibalism might of filial cannibalism. However, in both stud- enable them to rear a greater number of ies, brooding males were physically isolated young than would have been possible with- from stimulus females with transparent out engaging in cannibalism. barriers. By contrast, males were found to Examining experimental studies that display greater rates of filial cannibalism have investigated the effect of mate avail- when they had the opportunity to physically ability on filial cannibalism, we see high spawn with additional females by Okuda mate availability is associated with inflated et al. (2004). Similarly, in a study of the levels of filial cannibalism in studies where cardinalfish A. doederleini, where high mate parental males were not fed during paren- availability was found to be associated with tal care (Table 1; but see Bjelvenmark and filial cannibalism, males were studied in the Forsgren 2003). By contrast, high mate natural environment with no artificial sep- availability is associated with decreased aration from additional mates (Okuda and cannibalism rates in studies where food Yanagisawa 1996b). Recent work on the liz- was provided to males during parental care ard goby (R. flumineus), however, has shown (Table 1). If this pattern is borne out by that physical contact is not always neces- further research, it could suggest that in- sary to elicit an effect (Myint et al. 2011a; dividuals respond differently to changes in Takeyama et al. 2013). Furthermore, work mate availability according to their access on the flagfish ( J. floridae) demonstrates to food. It should be noted, however, that that when direct mate contact is allowed, the decisions by researchers about whether high mate availability can still be found to provide males with food during paren- to reduce filial cannibalism levels (Klug tal care are probably related to whether et al. 2005). It therefore seems unlikely that the parental care behavior of their study physical contact with mates is responsible species restricts parental feeding opportu- for the failure of some studies to find high nities under natural conditions. Therefore, mate availability to increase filial cannibal- the apparent pattern in experimental find- ism levels. ings might reflect interspecies differences rather than behavioral plasticity of individuals with respect to access to food (see the Food Availability section, The Macroevolutionary Effect of An additional component of experimen- Mate Availability on filial Cannibalism). In tal design worthy of consideration in future particular, species in which parental males experiments is the provisioning of food to have restricted access to food as a conse- parents. As noted in the introduction, one quence of brood care (because males en- important function of filial cannibalism gage in mouthbrooding or hold themselves can be the acquisition of nutrients or en- up inside their nest during parental care; ergy for parents. It is reasonable, therefore, see Okuda and Yanagisawa 1996b; Myint

5552.proof.indd 11 Achorn International 02/16/2016 11:13PM 12 THE QUARTERLY REVIEW OF BIOLOGY Volume 91 et al. 2011a; Takeyama et al. 2013) could be gren 1997; Lehtonen and Lindström 2007) more likely to respond to increased access and the three-spined stickleback (Gasteros- to mates by engaging in greater levels of teus aculeatus; Candolin 2000a,b). Preferred filial cannibalism. males that engage in costly courtship dis- Ideally, to investigate a potential interac- plays have also been shown to cannibalize tion between the effect of food and mate fewer eggs in the stream goby (Rhinogobius availability on individual behavior, a crossed brunneus; Takahashi and Kohda 2004) and experimental design examining each of the bicolor damselfish (Stegastes partitus; these factors in a single species would be Knapp and Kovach 1991). In the next sec- employed. However, only a single study, tion, we consider mechanisms that might conducted by Okuda et al. (2004), has so far explain the negative association between an taken this approach. Although they found individual’s attractiveness and their propen- that both poor condition (from being food sity to engage in filial cannibalism. restricted) and elevated access to mates are associated with increased levels of filial cannibalism in male gobies, no interaction why are attractive individuals between these two factors was detected. less prone to filial cannibalism Nonetheless, further investigation of this Several processes could underlie the neg- potential interaction is needed, especially ative association between filial cannibalism since the power to detect an interaction in and attractiveness. The relationship could the aforementioned study was low (Okuda arise without any direct causal link between et al. 2004). the two factors. For instance, poor body condition, foraging capabilities, or access to resources may lead to individuals becoming relationship between unattractive, while at the same time compel- offspring consumption and ling them to commit filial cannibalism for attractiveness of the cannibal nutritional gain. In terms of a causal rela- So far, we have examined studies that tionship, it is plausible that the propensity manipulate mate availability by altering the of individuals to eat their young can directly number of gravid females to which males influence their attractiveness. Thus, mate are exposed. Yet, just as the physical pres- choice based on cues or signals that pre- ence of the opposite sex may influence the dict an individual’s likelihood of engaging likelihood of an individual’s future repro- in cannibalism may explain why cannibals duction, so too should the willingness of are less attractive. The best evidence for potential suitors to mate with the individ- this involves male expression of epigamic ual. An individual’s perceived attractiveness behavior. For example, studies have shown could therefore influence its tendency to that females often prefer males that court commit filial cannibalism. Here, one possi- intensely, or in energetically demanding bility is that attractive parents could exploit circumstances, because such males have their heightened access to mates by engag- superior energy reserves or efficient metab- ing in greater levels of cannibalism. Mean- olisms—and are therefore less likely to eat while, unattractive males might engage in their young (Knapp and Kovach 1991; Taka- less filial cannibalism since they have lower hashi and Kohda 2004). expected future reproductive opportunities The reverse causality is also possible: at- on account of their low sex appeal. How- tractiveness can potentially influence an ever, this does not appear to be supported individual’s access to mates and, in so do- by the literature, with evidence suggesting, ing, affect their inclination to engage in in fact, that attractive males are less likely cannibalism. For instance, already attrac- to eat their young. For example, males with tive males may avoid eating their young, as preferred phenotypes have been shown to doing so may reduce their attractiveness in bring a greater proportion of eggs to hatch- the short term and, thus, their competitive ing in both the sand goby (P. minutus; Fors- advantage over rivals. Unattractive males,

5552.proof.indd 12 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 13 by contrast, might eat their offspring in the cost of losing young from filial canni- anticipation of becoming more attractive balism against the benefits they can obtain in the future. Indeed, for such males, re- from additional mating opportunities, this sources acquired from filial cannibalism could encourage them to eat their young could actually be used to directly improve where doing so enables them to attract their future attractiveness. This has been higher quality mates. For example, filial shown in the cardinalfish (A. doederleini), cannibalism could occur if it provides the where younger males perform filial canni- cannibal with the resources needed to at- balism to fuel growth to a larger size, which tract a better quality mate or to invest in makes them more appealing to females in their young. Here, a study of the lizard future mating attempts (Okuda et al. 1997; goby (R. flumineus) shows that males can in- Takeyama et al. 2002). deed distinguish between potential mates, It is currently unclear which direction and engage in elevated levels of filial can- of causality is most important in explain- nibalism only when exposed to gravid (as ing the negative association between filial opposed to nongravid) females (Takeyama cannibalism and attractiveness. It is worth et al. 2013). However, the effects of other noting, however, that male signaling of egg differences in mate quality are yet to be hatching success appears to be particularly investigated. important among egg guarding ectotherms, There are several reasons why parents suggesting that filial cannibalism propensity might be selective over which mates they is likely to affect attractiveness (Møller and will consume some or all of their young to Jennions 2001). However, results of studies gain access to. Certainly, for males, willing- in which nest size has been manipulated ness to consume existing young could be suggest that males with small nests—which contingent on the number of young that a could be less attractive to females—commit prospective female mate would produce. In greater rates of filial cannibalism (Okuda particular, in the absence of other benefits, et al. 2004; Pampoulie et al. 2004; Klug et al. we might expect that fathers will only com- 2006; but see Björk and Kvarnemo 2012), mit filial cannibalism to gain access to an thus indicating that males may alter their additional mate where the subsequent mat- cannibalistic tendencies based on their self- ing produces a greater number of young perceived attractiveness. Other explanations, than were consumed. However, there are such as reduced ventilation or increased other reasons why parents (of either sex) egg density, might also explain this obser- might be choosy about which potential vation, therefore further experiments ex- mates are worth consuming young for a amining the effect of attractiveness manip- chance to reproduce with. In particular, ulation are needed. only high-quality mates may be acceptable as they produce either genetically superior or better resourced progeny. Similarly, it Effect of Prospective Mate Quality might be that parents base their decision on an Individual’s Tendency to on whether to cannibalize existing young Commit Filial Cannibalism on the genetic compatibility of a new pro- Although there is a growing interest in spective mate. how the presence of additional mates af- The behavior of potential suitors could fects an individual’s tendency to commit also be influential in driving parents to filial cannibalism, we know far less about commit filial cannibalism. In some species, how the quality of these mates might influ- prospective mates are known to kill the ence the cannibal’s behavior. It seems rea- offspring of caring parents so as to coerce sonable to suppose that in some situations, the parents into breeding with them (Hrdy when a brooding parent encounters extra 1974; Palombit 2015). A range of coun- mates, the quality of these mates could af- terstrategies have evolved to help parents fect the likelihood and extent of filial can- protect their young from this fate, or lessen nibalism by the parent. If parents trade off the cost when it occurs (Palombit 2015).

5552.proof.indd 13 Achorn International 02/16/2016 11:13PM 14 THE QUARTERLY REVIEW OF BIOLOGY Volume 91

One potential counterstrategy (the Bruce parent to defend their brood from the new effect), well documented among a range of prospective mate influence the likelihood rodent species (among other mammals), of cannibalism. involves female termination of pregnancy in response to encountering unfamiliar males (Bruce 1959; Labov 1981; Becker The Macroevolutionary and Hurst 2008; Roberts et al. 2012). An Effect of Mate Availability analogous strategy available to parents car- on Filial Cannibalism ing for young after parturition or ovipo- So far, we have focused on how an in- sition would involve parents eating their dividual’s access to mates, and the qual- own young before infanticidal prospective ity of these mates, influences the amount mates are able to kill the young. This would and form of filial cannibalism they should allow the parent to recover some nutri- commit; hereafter, we refer to this as the tional resources from their offspring before individual-level effect of mate availability. In remating (Labov 1981). There is some ev- this section, we discuss macroevolutionary idence to suggest that this counterstrategy patterns of filial cannibalism. Specifically, is employed by parental male sticklebacks we consider how between-species differ- (FitzGerald and van Havre 1987). How- ences in mate availability might modulate ever, as far as we are aware, it remains to the relative rates of filial cannibalism of be investigated whether such a strategy also species. We refer to this as the species-level exists in mammalian taxa, such as rodents, effect of mate availability. In other words, which display the Bruce effect (e.g., Bruce the individual-level effect of mate availabil- 1959; Heske and Nelson 1984; Hackländer ity represents phenotypic plasticity, whereby and Arnold 1999; Pillay and Kinahan 2009; the amount of filial cannibalism performed Marashi and Rülicke 2012) and engage in by an individual varies according to prevail- high levels of filial cannibalism (Day and ing environmental conditions experienced Galef 1977; Bronson and Marsteller 1985; by them, in particular their access to mates. but see Weber et al. 2013). By contrast, the species-level effect of mate It may be difficult to differentiate be- availability refers the influence that the tween filial cannibalism as a counterstrat- overall accessibility of mates to members of egy to avoid infanticide and filial cannibal- one sex within a species has on the aver- ism that is a strategy to facilitate trading up age rate of filial cannibalism engaged in by to a higher quality mate (cf. Becker and that sex. Hurst 2008 for a similar discussion regard- Some may argue that since the amount ing the adaptive significance of the Bruce of filial cannibalism occurring within a spe- effect). In both cases, the value of the exist- cies is merely the sum of the actions of all ing clutch relative to that of the expected individuals in the species, then the species- clutch may influence the parent’s choice to level effects of mate availability should simply commit filial cannibalism. This is because reflect the individual-level effect. However, the trading-up hypothesis relies on a more it is possible that mate availability and filial valuable brood being attained from the cannibalism rates have different relation- new clutch, while for the infanticide coun- ships at the species level and at the indi- terstrategy, parents may be willing to put vidual level. For instance, when comparing themselves at greater risk to defend more a group of species, it might be that the valuable clutches, as seen for example in average level of access to mates and filial cichlids (Aequidens coeruleopunctatus) and cannibalism are positively correlated. How- bluegill sunfish (Lepomis macrochirus; Car- ever, each of these species could conceivably lisle 1985; Coleman et al. 1985). One infor- be composed of individuals that are more mative difference between filial cannibal- likely to consume their young when they ism as a counterstrategy to infanticide, and experience lulls in mate availability. Accord - that used to trade up, would be that only in ingly, we believe that future work will bene- the former would the capacity of the caring fit from explicitly differentiating between

5552.proof.indd 14 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 15 species- and individual-level effects of mate research on this genus would be desirable availability. to elucidate more details about the mech- On a macroevolutionary scale, the cur- anism through which high mate availabil- rent view is that within a sex, but between ity in A. doederleini favors its relatively high species, there will be a positive association cannibalism rate. One possible mechanism between filial cannibalism levels displayed is that filial cannibalism imposes minimal by the sex and mate availability experienced costs on A. doederleini males, as high mate by the sex (Okuda 1999a, 2000; Kondoh and availability means they are able to replace Okuda 2002). The currently accepted basis lost young easily. Another nonmutually ex- for this prediction is that organismal-level clusive mechanism is that male A. doederleini selection will favor greater rates of filial can- commit more filial cannibalism to compen- nibalism by individuals of a given sex within sate for the greater amounts of time and species where that sex experiences relatively energy they spend mouthbrooding (Okuda high mate availability. This would be because and Yanagisawa 1996a; Okuda 1999a, 2000), replacing eaten young should be less costly with the increase in time spent mouthbrood- where mate availability is high. Therefore, ing resulting from more frequent matings with the effective cost of filial cannibalism that accompanies increased access to mates. lessened, it should become more common. Kondoh and Okuda (2002) developed Moreover, it can be argued that, as mate the only model that sets out to determine availability is increased to a given sex, access how mate availability influences filial canni- to mates could become less of a constraint balism at the species level. They modeled a on that sex’s reproduction. Accordingly, population of exclusive paternal carers with other factors, including energetic demands, filial cannibalistic males that care for a set might begin to limit reproduction in that number of clutches per brood. A game the- sex. Consequently, selection may favor an oretic approach was used to determine an increased tendency toward performing filial evolutionary stable strategy for the number cannibalism among the sex where access to of clutches cannibalized per brood. Consis- mates is high, thereby allowing members of tent with empirical studies of Apogon cardi- this sex to attain resources for more repro- nalfish, they found that, at the species level, duction and exploit the availability of mates. high mate availability facilitates increased Field surveys of various species of pa- filial cannibalism. The value of this model ternal mouthbrooding cardinalfish of the to the development of the field cannot be genus Apogon provide evidence for mate underestimated. Nonetheless, as is nearly availability having a species level effect always the case with modeling, certain sim- on the incidence of filial cannibalism. It plifying assumptions were made that could has been found that males of both Apogon influence the results. Extension of this niger (Okuda 1999a) and Apogon notatus model could help determine if mechanisms (Okuda 2000) consume the clutches they that might influence the individual-level are brooding less frequently than males effects of mate availability (especially those of A. doederleini (Okuda and Yanagisawa listed in the section, Mechanisms Favoring 1996b; Okuda et al. 1997). Interestingly, A. a Negative Association Between Mate Avail- niger (Okuda 1999a) and A. notatus (Okuda ability and filial Cannibalism) also lead to 1999b, 2000) both come from populations species-level effects. At present, this remains in which there is a male bias in the opera- a largely unresolved question. In particular, tional sex ratio and adult sex ratio, whereas it would be instructive to develop models these are both female biased for A. doeder- that include temporal and spatial fluctua- leini throughout most of the breeding sea- tions in mate availability, sexual selection son (Okuda and Yanagisawa 1996b). Thus, against cannibals, and brood size effects observations of Apogon cardinalfish support on parental care and offspring survival. the prediction that high mate availability at Furthermore, Kondoh and Okuda’s (2002) the species level will result in high average model assumes that filial cannibalism is a rates of filial cannibalism. However, further genetic, fixed strategy without phenotypic

5552.proof.indd 15 Achorn International 02/16/2016 11:13PM 16 THE QUARTERLY REVIEW OF BIOLOGY Volume 91 plasticity (Okuda et al. 2004). Developing Thomas and Manica 2003), and biparen- filial cannibalism models in which animals tal care (e.g., burying beetle, Nicrophorus can respond to their environment with a vespilloides; Bartlett 1987). Furthermore, conditional strategy would be more realis- maternal (e.g., Egyptian mouthbrooder, P. tic and, most importantly, provide insights multicolor ; Mrowka 1987), biparental (e.g., into both species-level and individual-level convict cichlid, Amatitlania nigrofasciata; effects of mate availability on cannibalism Lavery and Keenleyside 1990), and even (Okuda et al. 1997, 2004; Takeyama et al. noncaring (e.g., green razorfish, Xyrichtys 2002). Although such models have been splendens; Nemtzov and Clark 1994) fish created (see, for example, Sargent 1992; species also offer opportunities to study Sargent et al. 1995), they have not yet been filial cannibalism among species with more used to substantially investigate mate avail- varied parental care systems. In addition to ability effects. Furthermore, models could this, the significance of filial cannibalism in be useful in gaining insight into whether other vertebrate lineages requires greater partial and total filial cannibalism levels attention. The consumption of offspring is are each affected differently by changing increasingly being reported in birds (Par- mate availability at the species level; an sons 1971; Bortolotti et al. 1991; Tortosa area of particular interest as Kondoh and and Redondo 1992; Gilbert et al. 2005; So- Okuda (2002) only addressed partial filial laro and Sarasola 2012; Franke et al. 2013), cannibalism in their model. Clearly, more and is also well documented in rodents investigations into the species-level effects and, to a lesser extent, other mammalian of mate availability on filial cannibalism taxa (Day and Galef 1977; Bronson and are needed to test the predictions of Kon- Marsteller 1985; Braastad 1987; Cockburn doh and Okuda’s (2002) model and verify 1994; but see Weber et al. 2013). In this the findings in a wider range of taxa. Ex- respect, future studies involving a wider perimental evolution with manipulations range of taxa will be important in provid- to mate availability should provide further ing robust tests of the existing interpreta- insight into species-level effects. The chal- tions of the relationships between mate lenge is to find a species that displays ad- availability and filial cannibalism. equate levels of filial cannibalism and an The reduced cost of replacing young appropriately short life cycle. when many mates are available does seem to trigger elevated levels of filial cannibalism in some species (Okuda and Yanagi- Conclusions sawa 1996b; Okuda et al. 2004; Myint et al. To date, all of the studies that have ex- 2011a; Takeyama et al. 2013). However, amined the effect of mate availability on in other species, there seems to be no ef- filial cannibalism have been carried out on fect of mate availability (Bjelvenmark and fish with exclusive paternal care. It is un- Forsgren 2003), or even a decrease in filial derstandable why this bias in the literature cannibalism in response to elevated mate exists, as fish with exclusive paternal care availability (Pampoulie et al. 2004; Klug show relatively high levels of filial cannibal- et al. 2005). To explain this, we suggest one ism, and because males are likely to expe- or all of several alternative mechanisms may rience greater variation in mate availability. be involved. The first mechanism is that Nonetheless, there is an obvious need for when the availability of mates is high, indi- research on more varied study systems in viduals are able to gain large numbers of regards to both phylogeny and life history. young, or expect to do so. As a consequence There are a number of invertebrate species of this, individuals may avoid committing that appear to show high enough levels of total filial cannibalism, which is usually per- filial cannibalism to make studies feasible formed in response to having a small brood in species with maternal (e.g., maritime for which the cost of caring outweighs the earwig, Anisolabis maritime ; Miller and Zink reproductive gain. A second explanation for 2012), paternal (e.g., assassin bug, R. tristis ; why some animals have an increased pro-

5552.proof.indd 16 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 17 pensity for filial cannibalism when mate is needed to determine whether this is a availability is low is that filial cannibalism consequence of females preferring males may be used as a strategy by parents to facili- that are unlikely to commit filial cannibal- tate their survival until times of higher mate ism or whether attractive males avoid filial availability. Finally, it could be that mate cannibalism as a consequence of increased choice against individuals that have recently access to mates. Similarly, future research engaged in filial cannibalism encourages should investigate whether parents take parents to avoid eating their own young into account the quality of prospective when many potential mates are around as mates when deciding to eat their young. doing so could lead to large costs in lost mat- Specifically, consideration should be given ing opportunities. Future research testing to the possibility that such behavior may the veracity of these mechanisms is needed. represent parents attempting to trade up One interesting emerging trend is that to higher quality mates or, alternatively, to when a negative association between filial avoid infanticide or offspring predation by cannibalism and mate availability is re- potential mates. ported, it involves total filial cannibalism Finally, it is important to distinguish the (Pampoulie et al. 2004; Klug et al. 2005). difference between the effect of mate avail- Contrastingly, when a positive association ability at the level of the individual and at is reported at the individual level, partial the level of the species. The former rep- filial cannibalism is more often affected resents behavioral plasticity, which enables (Myint et al. 2011a; Takeyama et al. 2013). individual animals to alter their filial can- This pattern suggests that the aforemen- nibalism levels according to their circum- tioned mechanisms, which drive reduced stances, and is not equivalent to the latter. filial cannibalism when mate availability is The species-level effect describes macro- high, may act more strongly on total filial evolutionary patterns where the evolution cannibalism. Nonetheless, further research of the filial cannibalism rate of a species is investigating how total and partial filial can- influenced by availability of mates within nibalism are affected differently by mate that species. So far there have been very availability is required. few studies investigating these phenomena Although researchers have focused largely and more are needed, especially because on the role that physical access to addi- the species-level effect on partial filial can- tional mates plays in determining filial cannibalism has not been empirically studied. nibalism, compelling areas for future studies involve investigating the role of parent acknowledgments attractiveness as well as the quality of potential future mates. Findings from studies We thank Anmei Vuong, David Chapple, members of that have measured attractiveness (or traits the School of Biological Sciences Postgraduate Writ- ing Group, and the anonymous reviewers for com- that confer it) suggest that attractive males ments on drafts of this manuscript. We also acknowl- commit the least filial cannibalism (Knapp edge financial support from The Holsworth Wildlife and Kovach 1991; Forsgren 1997; Cando- Research Endowment—ANZ Trustees Foundation (to lin 2000a,b; Takahashi and Kohda 2004; Nicholas David Shalders Deal) and The Australian Re- Lehtonen and Lindström 2007). Research search Council (to Bob B. M. Wong).

REFERENCES

Altmann S. A., Wagner S. S., Lenington S. 1977. Two Signals in Vertebrates 11, edited by J. L. Hurst, R. J. models for the evolution of polygyny. Behavioral Beynon, S. C. Roberts, and T. D. Wyatt. New York: Ecology and Sociobiology 2:397–410. Springer. Bartlett J. 1987. Filial cannibalism in burying beetles. Beery A. K., Zucker I. 2012. Sex ratio adjustment by Behavioral Ecology and Sociobiology 21:179–183. sex-speciﬁc maternal cannibalism in hamsters. Becker S. D., Hurst J. L. 2008. Pregnancy block from Physiology & Behavior 107:271–276. a female perspective. Pages 141–150 in Chemical Bjelvenmark J., Forsgren E. 2003. Effects of mate

5552.proof.indd 17 Achorn International 02/16/2016 11:13PM 18 THE QUARTERLY REVIEW OF BIOLOGY Volume 91

attraction and male-male competition on paternal terosteidae: Pisces). Behavioral Ecology and Sociobi- care in a goby. Behaviour 140:55–69. ology 20:125–128. Björk J. R., Kvarnemo C. 2012. Mechanisms behind Forsgren E. 1997. Female sand gobies prefer good fa- size-assortative nest choice by sand goby males in thers over dominant males. Proceedings of the Royal the absence of intrasexual competition. Animal Be- Society B: Biological Sciences 264:1283–1286. haviour 83:55–62. Forsgren E., Karlsson A., Kvarnemo C. 1996. Female Blumer L. S. 1979. Male parental care in the bony sand gobies gain direct benefits by choosing males fishes. Quarterly Review of Biology 54:149–161. with eggs in their nests. Behavioral Ecology and Socio- Bortolotti G. R., Wiebe K. L., Iko W. M. 1991. Canni- biology 39:91–96. balism of nestling American kestrels by their par- Franke A., Galipeau P., Nikolaiczuk L. 2013. Brood re- ents and siblings. Canadian Journal of Zoology 69: duction by infanticide in peregrine falcons. Arctic 1447–1453. 66:226–229. Braastad B. O. 1987. Abnormal behaviour in farmed Gibson R. M., Höglund J. 1992. Copying and sexual silver fox vixens (Vulpes vulpes L.): tail-biting and selection. Trends in Ecology & Evolution 7:229–232. infanticide. Applied Animal Behaviour Science 17: Gilbert W. M., Nolan P. M., Stoehr A. M., Hil G. E. 376–377. 2005. Filial cannibalism at a house finch nest. Wil- Bronson F. H., Marsteller F. A. 1985. Effect of short- son Bulletin 117:413–415. term food deprivation on reproduction in female Goldschmidt T., Bakker T. C. M., Feuth-de Bruijn E. mice. Biology of Reproduction 33:660–667. 1993. Selective copying in mate choice of female Bruce H. M. 1959. An exteroceptive block to preg- sticklebacks. Animal Behaviour 45:541–547. nancy in the mouse. Nature 184:105. Gomagano D., Kohda M. 2008. Partial filial cannibal- Candolin U. 2000a. Changes in expression and hon- ism enhances initial body condition and size in esty of sexual signalling over the reproductive life- paternal care fish with strong male-male competi- time of sticklebacks. Proceedings of the Royal Society tion. Annales Zoologici Fennici 45:55–65. B: Biological Sciences 267:2425–2430. Hackländer K., Arnold W. 1999. Male-caused failure Candolin U. 2000b. Increased signalling effort when of female reproduction and its adaptive value in survival prospects decrease: male-male competi- alpine marmots (Marmota marmota). Behavioral Ecol- tion ensures honesty. Animal Behaviour 60:417–422. ogy 10:592–597. Carlisle T. R. 1985. Parental response to brood size in Hastings P. A. 1992. Nest-site size as a short-term con- a cichlid fish. Animal Behaviour 33:234–238. straint on the reproductive success of paternal Clutton-Brock T. H. 1991. The Evolution of Parental fishes. Environmental Biology of fishes 34:213–218. Care. Princeton (New Jersey): Princeton University Heske E. J., Nelson R. J. 1984. Pregnancy interruption Press. in Microtus ochrogaster: laboratory artifact or field Clutton-Brock T. H., Parker G. A. 1992. Potential re- phenomenon? Biology of Reproduction 31:97–103. productive rates and the operation of sexual selec- Hoelzer G. A. 1992. The ecology and evolution of tion. Quarterly Review of Biology 67:437–456. partial-clutch cannibalism by paternal Cortez dam- Cockburn A. 1994. Adaptive sex allocation by brood selfish. Oikos 65:113–120. reduction in antechinuses. Behavioral Ecology and Hrdy S. B. 1974. Male-male competition and infanti- Sociobiology 35:53–62. cide among the langurs (Presbytis entellus) of Abu, Coleman R. M., Gross M. R., Sargent R. C. 1985. Pa- Rajasthan. Folia Primatologica 22:19–58. rental investment decision rules: a test in bluegill Huang W.-S. 2008. Predation risk of whole-clutch filial sunfish. Behavioral Ecology and Sociobiology 18:59–66. cannibalism in a tropical skink with maternal care. Creighton J. C. 2005. Population density, body size, Behavioral Ecology 19:1069–1074. and phenotypic plasticity of brood size in a bury- Jamieson I. 1995. Do female fish prefer to spawn ing beetle. Behavioral Ecology 16:1031–1036. in nests with eggs for reasons of mate choice copy- Day C. S. D., Galef B. G., Jr. 1977. Pup cannibalism: ing or egg survival? American Naturalist 145:824– one aspect of maternal behavior in golden ham- 832. sters. Journal of Comparative and Physiological Psychol- Klug H., Bonsall M. B. 2007. When to care for, aban- ogy 91:1179–1189. don, or eat your offspring: the evolution of paren- Emlen S. T., Oring L. W. 1977. Ecology, sexual selec- tal care and filial cannibalism. American Naturalist tion, and the evolution of mating systems. Science 170:886–901. 197:215–223. Klug H., Lindström K. 2008. Hurry-up and hatch: FitzGerald G. J. 1992. Filial cannibalism in fishes: why selective filial cannibalism of slower developing do parents eat their offspring? Trends in Ecology & eggs. Biology Letters 4:160–162. Evolution 7:7–10. Klug H., Chin A., St. Mary C. M. 2005. The net effects FitzGerald G. J., van Havre N. 1987. The adaptive of guarding on egg survivorship in the flagfish, Jor- significance of cannibalism in sticklebacks (Gas- danella floridae. Animal Behaviour 69:661–668.

5552.proof.indd 18 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 19

Klug H., Lindström K., St. Mary C. M. 2006. Parents brood care on filial cannibalism in the sand goby. benefit from eating offspring: density-dependent Behavioral Ecology and Sociobiology 42:101–106. egg survivorship compensates for filial cannibal- Lindström K. 2000. The evolution of filial cannibalism ism. Evolution 60:2087–2095. and female mate choice strategies as resolutions Knapp R. A., Kovach J. T. 1991. Courtship as an honest to sexual conflict in fishes. Evolution 54:617–627. indicator of male parental quality in the bicolor Lindström K., Kangas N. 1996. Egg presence, egg damselfish, Stegastes partitus. Behavioral Ecology loss, and female mate preferences in the sand 2:295–300. goby (Pomatoschistus minutus). Behavioral Ecology 7: Knapp R. A., Sargent R. C. 1989. Egg-mimicry as a 213–217. mating strategy in the fantail darter, Etheostoma fla- Lindström K., Sargent R. C. 1997. Food access, brood bellare: females prefer males with eggs. Behavioral size and filial cannibalism in the fantail darter, Ecology and Sociobiology 25:321–326. Etheostoma flabellare. Behavioral Ecology and Sociobiol- Kokko H., Monaghan P. 2001. Predicting the direc- ogy 40:107–110. tion of sexual selection. Ecology Letters 4:159–165. Lindström K., St. Mary C. M. 2008. Parental care and Kondoh M., Okuda N. 2002. Mate availability in- sexual selection. Pages 377–409 in Fish Behaviour, fluences filial cannibalism in fish with paternal edited by C. Magnhagen, V. A. Braithwaite, E. care. Animal Behaviour 63:227–233. Forsgren, and B. G. Kapoor. Enfield (New Hamp- Kraak S. B. M. 1996. Female preference and filial can- shire): Science Publishers. nibalism in Aidablennius sphynx (Teleostei, Blenni- Lissåker M., Kvarnemo C. 2006. Ventilation or nest de- idae); a combined field and laboratory study. Be- fense—parental care trade-offs in a fish with male havioural Processes 36:85–97. care. Behavioral Ecology and Sociobiology 60:864–873. Kraak S. B. M., Groothuis T. G. G. 1994. Female pref- Manica A. 2002a. Alternative strategies for a father erence for nests with eggs is based on the presence with a small brood: mate, cannibalise or care. Be- of the eggs themselves. Behaviour 131:189–206. havioral Ecology and Sociobiology 51:319–323. Kraak S. B. M., Videler J. J. 1991. Mate choice in Manica A. 2002b. Filial cannibalism in teleost fish. Bi- Aidablennius sphynx (Teleostei, Blenniidae); females ological Reviews 77:261–277. prefer nests containing more eggs. Behaviour 119: Manica A. 2004. Parental fish change their cannibal- 243–266. istic behaviour in response to the cost-to-benefit Kraak S. B. M., Weissing F. J. 1996. Female preference ratio of parental care. Animal Behaviour 67:1015– for nests with many eggs: a cost-benefit analysis of 1021. female choice in fish with paternal care. Behavioral Manica A. 2010. Female scissortail sergeants (Pisces: Ecology 7:353–361. Pomacentridae) use test eggs to choose good fa- Kramer D. L. 1973. Parental behaviour in the blue thers. Animal Behaviour 79:237–242. gourami Trichogaster trichopterus (Pisces, Belontii- Marashi V., Rülicke T. 2012. The Bruce effect in Nor- dae) and its induction during exposure to varying way rats. Biology of Reproduction 86:17, 1–5. numbers of conspecific eggs. Behaviour 47:14–32. Marconato A., Bisazza A. 1986. Males whose nests con- Kume G., Yamaguchi A., Taniuchi T. 2000. Filial canni- tain eggs are preferred by female Cottus gobio L. balism in the paternal mouthbrooding cardinalfish (Pisces, Cottidae). Animal Behaviour 34:1580–1582. Apogon lineatus: egg production by the female as Mehlis M., Bakker T. C. M., Frommen J. G. 2009. the nutrition source for the mouthbrooding male. Nutritional benefits of filial cannibalism in three- Environmental Biology of fishes 58:233–236. spined sticklebacks (Gasterosteus aculeatus). Natur- Kvarnemo C., Svensson O., Forsgren E. 1998. Paren- wissenschaften 96:399–403. tal behaviour in relation to food availability in the Milinski M., Bakker T. C. M. 1992. Costs influence se- common goby. Animal Behaviour 56:1285–1290. quential mate choice in sticklebacks, Gasterosteus Labov J. B. 1981. Pregnancy blocking in rodents: aculeatus. Proceedings of the Royal Society B: Biological adaptive advantages for females. American Natural- Sciences 250:229–233. ist 118:361–371. Miller J. S., Zink A. G. 2012. Parental care trade-offs Lavery R. J., Keenleyside M. H. A. 1990. Filial canni- and the role of filial cannibalism in the maritime balism in the biparental fish Cichlasoma nigrofascia- earwig, Anisolabis maritima. Animal Behaviour 83: tum (Pisces: Cichlidae) in response to early brood 1387–1394. reductions. Ethology 86:326–338. Mock D. W. 2004. More than Kin and Less than Kind: The Lehtonen T. K., Lindström K. 2007. Mate compatibil- Evolution of Family Conflict. Cambridge (Massachu- ity, parental allocation and fitness consequences setts): Belknap Press of Harvard University Press. of mate choice in the sand goby Pomatoschistus Mock D. W., Forbes L. S. 1995. The evolution of pa- minutus. Behavioral Ecology and Sociobiology 61:1581– rental optimism. Trends in Ecology & Evolution 10: 1588. 130–134. Lindström K. 1998. Effects of costs and benefits of Møller A. P., Jennions M. D. 2001. How important are

5552.proof.indd 19 Achorn International 02/16/2016 11:13PM 20 THE QUARTERLY REVIEW OF BIOLOGY Volume 91

direct fitness benefits of sexual selection? Naturwis- Okuda N., Ito S., Iwao H. 2004. Mate availability and senschaften 88:401–415. somatic condition affect filial cannibalism in a pa- Moreno J. 2012. Parental infanticide in birds through ternal brooding goby. Behaviour 141:279–296. early eviction from the nest: rare or under- Östlund-Nilsson S. 2002. Does paternity or paternal reported? Journal of Avian Biology 43:43–49. investment determine the level of paternal care Mori H., Chiba S. 2009. Sociality improves larval and does female choice explain egg stealing in growth in the stag beetle figulus binodulus (Cole- the fifteen-spined stickleback? Behavioral Ecology optera: Lucanidae). European Journal of Entomology 13:188–192. 106:379–383. Palombit R. A. 2015. Infanticide as sexual conflict: Mrowka W. 1987. Filial cannibalism and reproductive coevolution of male strategies and female counter- success in the maternal mouthbrooding cichlid strategies. Cold Spring Harbor Perspectives in Biology fish Pseudocrenilabrus multicolor. Behavioral Ecology 7:a017640. and Sociobiology 21:257–265. Pampoulie C., Lindström K., St. Mary C. M. 2004. Myint O., Takeyama T., Okuda N., Ohnishi N., Kohda Have your cake and eat it too: male sand gobies M. 2011a. Mate availability facilitates cannibalistic show more parental care in the presence of female behaviour in a nest brooding fish: effects of timing partners. Behavioral Ecology 15:199–204. during the brood cycle. Behaviour 148:247–264. Parsons J. 1971. Cannibalism in herring gulls. British Myint O., Tsujimoto H., Ohnishi N., Takeyama T., Birds 64:528–537. Kohda M. 2011b. Mate availability affects female Payne A. G., Smith C., Campbell A. C. 2002. Filial choice in a fish with paternal care: female counter- cannibalism improves survival and development strategies against male filial cannibalism. Journal of of beaugregory damselfish embryos. Proceedings of Ethology 29:153–159. the Royal Society B: Biological Sciences 269:2095–2102. Nazareth T. M., Machado G. 2010. Mating system and Payne A. G., Smith C., Campbell A. C. 2003. The ef- exclusive postzygotic paternal care in a Neotrop- fect of clutch size on whole clutch cannibalism in ical harvestman (Arachnida: Opiliones). Animal the beaugregory damselfish. Journal of fish Biology Behaviour 79:547–554. 62:955–958. Nemtzov S. C., Clark E. 1994 Intraspecific egg preda- Payne A. G., Smith C., Campbell A. C. 2004. A model tion by male razorfishes (Labridae) during broad- of oxygen-mediated filial cannibalism in fishes. cast spawning: filial cannibalism or intra-pair para- Ecological Modelling 174:253–266. sitism? Bulletin of Marine Science 55:133–141. Petersen C. W. 1990. The occurrence and dynamics of Okada S., Fukuda Y., Takahashi M. K. 2015. Paternal clutch loss and filial cannibalism in two Caribbean care behaviors of Japanese giant salamander An- damselfishes. Journal of Experimental Marine Biology drias japonicus in natural populations. Journal of and Ecology 135:117–133. Ethology 33:1–7. Petersen C. W., Hess H. C. 1991. The adaptive signifi- Okuda N. 1999a. Female mating strategy and male cance of spawning synchronization in the Carib- brood cannibalism in a sand-dwelling cardinalfish. bean damselfish Stegastes dorsopunicans (Poey). Animal Behaviour 58:273–279. Journal of Experimental Marine Biology and Ecology Okuda N. 1999b. Sex roles are not always reversed 151:155–167. when the potential reproductive rate is higher in Petersen C. W., Marchetti K. 1989. Filial cannibalism females. American Naturalist 153:540–548. in the cortez damselfish Stegastes rectifraenum. Evo- Okuda N. 2000. Interspecific differences in male can- lution 43:158–168. nibalistic behavior between two sympatric cardi- Pillay N., Kinahan A. A. 2009. Mating strategy predicts nalfishes (Pisces: Apogonidae). Journal of Ethology the occurrence of the Bruce effect in the vlei rat 18:5–10. Otomys irroratus. Behaviour 146:139–151. Okuda N., Yanagisawa Y. 1996a. Filial cannibalism by Pomiankowski A. 1987. The costs of choice in sexual mouthbrooding males of the cardinal fish, Apogon selection. Journal of Theoretical Biology 128:195–218. doederleini, in relation to their physical condition. Raadik T. A., Bourke D. W., Clarke M. F., Martin A. A. Environmental Biology of fishes 45:397–404. 1990. Behaviour and reproductive success of pairs Okuda N., Yanagisawa Y. 1996b. Filial cannibalism in and lone parents in the convict cichlid Heros nigro- a paternal mouthbrooding fish in relation to mate fasciatus. Animal Behaviour 39:594–596. availability. Animal Behaviour 52:307–314. Real L. 1990. Search theory and mate choice. I. Mod- Okuda N., Takeyama T., Yanagisawa Y. 1997. Age- els of single-sex discrimination. American Naturalist specific filial cannibalism in a paternal mouth- 136:376–405. brooding fish. Behavioral Ecology and Sociobiology Reynolds J. D., Jones J. C. 1999. Female preference 41:363–369. for preferred males is reversed under low oxygen

5552.proof.indd 20 Achorn International 02/16/2016 11:13PM March 2016 MATE AVAILABILITY AND FILIAL CANNIBALISM 21

conditions in the common goby (Pomatoschistus mi- Solaro C., Sarasola J. H. 2012. First observation of crops). Behavioral Ecology 10:149–154. infanticide and cannibalism in nest of chimango Ridley M., Rechten C. 1981. Female sticklebacks pre- caracara (Milvago chimango). Journal of Raptor Re- fer to spawn with males whose nests contain eggs. search 46:412–413. Behaviour 76:152–161. Takahashi D., Kohda M. 2004. Courtship in fast water Roberts E. K., Lu A., Bergman T. J., Beehner J. C. 2012. A currents by a male stream goby (Rhinogobius brun- Bruce effect in wild geladas. Science 335:1222–1225. neus) communicates the parental quality honestly. Rohwer S. 1978. Parent cannibalism of offspring and Behavioral Ecology and Sociobiology 55:431–438. egg raiding as a courtship strategy. American Natu- Takeyama T., Okuda N., Yanagisawa Y. 2002. Seasonal ralist 112:429–440. pattern of filial cannibalism by Apogon doederleini Sargent R. C. 1988. Paternal care and egg survival mouthbrooding males. Journal of fish Biology 61: both increase with clutch size in the fathead min- 633–644. now, Pimephales promelas. Behavioral Ecology and So- Takeyama T., Namizaki N., Kohda M. 2013. Mate ciobiology 23:33–37. availability accelerates male filial cannibalism in a Sargent R. C. 1992. Ecology of filial cannibalism in nest brooding fish: effects of number and fecun- fish: theoretical perspectives. Pages 38–62 in Can- dity of females. Behavioral Ecology and Sociobiology nibalism: Ecology and Evolution Among Diverse Taxa, 67:421–428. edited by M. A. Elgar and B. J. Crespi. Oxford Thomas L. K., Manica A. 2003. Filial cannibalism in an (United Kingdom): Oxford University Press. assassin bug. Animal Behaviour 66:205–210. Sargent R. C., Crowley P. H., Huang C., Lauer M., Neer- Thomas L. K., Manica A. 2005. Intrasexual competition gaard D., Schmoetzer L. 1995. A dynamic program and mate choice in assassin bugs with uniparental for male parental care in fishes: brood cycling and male and female care. Animal Behaviour 69:275–281. filial cannibalism. Behaviour 132:1059–1078. Tortosa F. S., Redondo T. 1992. Motives for parental Schneider J. E., Wade G. N. 1989. Effects of maternal infanticide in white storks Ciconia ciconia. Ornis diet, body weight and body composition on infan- Scandinavica 23:185–189. ticide in Syrian hamsters. Physiology & Behavior Trumbo S. T. 1990. Regulation of brood size in a bury- 46:815–821. ing beetle, Nicrophorus tomentosus (Silphidae). Jour- Schwanck E. 1986. Filial cannibalism in Tilapia mariae. nal of Insect Behavior 3:491–500. Journal of Applied Ichthyology 2:65–74. Unger L. M., Sargent R. C. 1988. Alloparental care in Sikkel P. C. 1988. Factors influencing spawning site the fathead minnow, Pimephales promelas: females choice by female garibaldi, Hypsypops rubicundus prefer males with eggs. Behavioral Ecology and Socio- (Pisces, Pomacentridae). Copeia 1988:710–718. biology 23:27–32. Slagsvold T., Dale S. 1991. Mate choice models: can Weber E. M., Algers B., Hultgren J., Olsson I. A. S. cost of searching and cost of courtship explain 2013. Pup mortality in laboratory mice—infanti- mating patterns of female pied flycatchers? Ornis cide or not? Acta Veterinaria Scandinavica 55:83. Scandinavica 22:319–326. Smith C., Wootton R. J. 1995. The costs of parental care in teleost fishes. Reviews in fish Biology and Associate Editor: Daniel T. Blumstein fisheries 5:7–22. Handling Editor: John J. Wiens

5552.proof.indd 21 Achorn International 02/16/2016 11:13PM The Quarterly Review of Biology Complimentary Offer and Page Charge Form

Authors of articles, commentaries, and lead reviews are entitled to 10 complimentary copies of the journal issue or a year’s subscription (or renewal) free of charge.

Author(s): Article title: Page length of article:

Complimentary Offer Please check one: ( ) 10 free copies of the journal issue or ( ) one free year’s subscription or renewal

Page Charges Authors who attribute support to funding agencies will be expected to meet charges of $40 per journal page for each of the first 20 journal pages, and $80 for each journal page in excess of 20. Authors not supported by funding agencies and authors of invited articles are eligible for 20 free journal pages (about 50 double-spaced typewritten manuscript pages or their equivalent in bibliographic or illustrative matter). When such articles are longer than 20 journal pages, authors will be charged $80 for each page above the allotment of 20 free journal pages. These charges will be used for direct support of the issue in which the article appears. For online-only material, authors are eligible for 6 free double-spaced manuscript pages, but authors with funding will be charged $20 for every 3 manuscript pages in excess of 6.

In order to determine the length of your article, please count the number of paged galleys. The cost of author alterations is chargeable to the author and will be invoiced after the article is published, along with the cost of page charges.

Ship To: Bill To: (for page charges)

Email: Telephone: Fax:

Please return this form to The Quarterly Review of Biology, E-5340 Melville Library, Stony Brook University, Stony Brook, NY 11790; [email protected]; Fax: 631 632 9282. Thank you.