Behav Ecol Sociobiol (2006) 59: 606–613 DOI 10.1007/s00265-005-0087-7

ORIGINA L ARTI CLE

James H. Bandoli

Male spottail darters (Etheostoma squamiceps) do not use chemical or positional cues to discriminate between sired and foster eggs Received: 17 January 2005 / Revised: 15 July 2005 / Accepted: 1 September 2005 / Published online: 25 October 2005 # Springer-Verlag 2005

Abstract Filial cannibalism is hypothesized to allow nestguarding males to recoup energy lost during nest defense. Males in many species of fishes occasionally defend broods containing both sired and foster eggs due to shifts in nest site ownership or cuckoldry. Such males are predicted to consume primarily foster eggs if the ability to discriminate among eggs exists. In a previous laboratory study, male spottail darters (Etheostoma squamiceps) consumed significantly more foster eggs than sired eggs, suggesting the existence of a mechanism for discrimination using chemical or positional cues. This discrimination mechanism in the spottail darter was tested by creating nest sites with half sired eggs and half foster eggs (n=15), or with all sired eggs with half positionally relocated (n=9). Males defended these nests for 2 days in the field, and cannibalism was determined by counting eggs lost during that interval. Neither foster eggs nor relocated sired eggs were consumed in greater numbers or percentages of the original brood sizes than the unchanged sired eggs, indicating that male spottail darters do not discriminate between foster and sired eggs by olfaction or position. Alternatively, the hypothesis that males treat all eggs in nest sites in which they have spawned as sired explains the results of both studies. Keywords Percidae . Egg defense . Filial cannibalism . Alloparental care

Introduction Filial cannibalism describes the phenomenon whereby a nest-guarding male consumes some or all of the eggs he is Communicated by K. Lindström J. H. Bandoli (*) Department of Biology, University of Southern Indiana, Evansville, IN 47712, USA e-mail: [email protected] Tel.: +1-812-4641792 Fax: +1-812-4651052

guarding (Rohwer 1978). It occurs widely in teleost fishes (Manica 2002), and Rohwer (1978) hypothesized that filial cannibalism is an adaptive investment in future reproduction, allowing a nest-guarding male to offset some of the energetic costs of egg defense by consuming sired eggs, thereby increasing his chances of future spawning. This hypothesis predicts that levels of filial cannibalism increase as male condition decreases, but tests have yielded equivocal results. Supplemental feeding or improved energetic condition in males decreased filial cannibalism in male scissortail sergeants Abudefduf sexfasciatus (Manica 2004), bluegill sunfish Lepomis macrochirus (Neff 2003a), and the common goby Pomatoschistus microps (Kvarnemo et al. 1998), but not in the sand goby Pomatoschistus minutus (Lissåker et al. 2003) or the fantail darter Etheostoma flabellare (Lindström and Sargent 1997). In many fish taxa in which males guard nest sites visited by females, turnover of nest-tending males due to death or displacement by other males leads to episodes of alloparental care, in which males defend eggs they did not sire (Wisenden 1999). This may be enhanced in species in which females prefer to spawn with males that are already defending eggs (Rohwer 1978), possibly causing some males to adopt eggs as a tactic to attract females (Unger and Sargent 1988; Sargent 1989). Additionally, mean relatedness between the guarding male and his brood may be decreased by cuckoldry via “sneaker” males or males mimicking females—strategies that, when successful, allow satellite males to fertilize eggs in an otherwise guarded nest (Taborsky 1994). Nest site and egg defense can be costly (Lindström 1998; Östlund and Ahnesjö 1998; Jones and Reynolds 1999); therefore, males should decrease the energy spent on defense as the proportion of the brood made up of foster eggs increases. Sargent (1989) found that male fathead minnows (Pimephales promelas) defending foster eggs decreased care (egg rubbing) and defense against a perceived egg predator compared to males defending sired eggs. Similarly, male bluegill sunfish (L. macrochirus) decreased parental effort (defense) in response to perceived decreases in paternity due to the presence of sneaker males (Neff and Gross 2001; Neff 2003b).

607

Cannibalism of sired eggs, while potentially increasing future reproductive success, carries the cost of decreased current reproductive success through loss of offspring, since each carries one half of the guarding male’s genes. Cannibalism of foster eggs provides the same benefit, but without this cost. Males with decreased confidence of paternity due to exposure to potential cuckolders such as sneakers are expected to increase egg cannibalism—a behavior seen in male scissortail sergeants (Manica 2004), but not in the common goby (Svensson et al. 1998). Additionally, if males can discriminate between sired and foster eggs, they should selectively consume foster eggs. Tests of this prediction have shown discrimination against foster eggs in male fathead minnows (Sargent 1989), and Loiselle (1983) found that primitively custodial male desert pupfish (Cyprinodon macularius) consumed fewer sired eggs than alien eggs. However, male threespine stickleback (Gasterosteus aculeatus) do not discriminate between sired and alien eggs (FitzGerald and van Havre 1987). Similarly, a genetic analysis of eggs in the stomach of one male tessellated darter (Etheostoma olmstedi) guarding both foster and sired eggs found only sired eggs in his stomach despite the fact that foster eggs made up 60% of his brood (DeWoody et al. 2001). The spottail darter (Etheostoma squamiceps), a percid fish found in small streams in southern Illinois, western Kentucky, and southwestern Indiana, USA (Page et al. 1992), spawns in cavities under rocks. Males defend these cavities, and visiting females attach eggs to the ceilings in monolayers (Page 1974). Several females may contribute clutches to a brood, and broods typically contain between 100 and 800 eggs, but can exceed 1,500 eggs, depending on the size of the nest site (Page 1974; Bandoli et al. 1991). The ability of a male to acquire and defend a nest site is positively related to body size (Bandoli 1997). While neither sneaking nor female mimicry by males has been documented in this species, alloparental care undoubtedly occurs as a result of displacement of males during competition for nest sites or as a result of occasional death of a guarding male. In a laboratory experiment (Bandoli 2002), males defending nest sites were collected from the field and allowed (1) to defend their original nest sites and sired eggs (parent males), or (2) to defend a new nest site containing foster eggs (alloparent males). Additionally, some of the alloparent males had been defending sired eggs immediately before they were given the new nests and eggs (experienced), while others had been defending empty nest sites (inexperienced). Over 3 days, all males consumed some of the eggs, but parent males consumed significantly fewer eggs than did alloparent males. This result raises the question of how male spottail darters recognized the status (sired or foster) of the eggs they defended. One possibility is that males used chemical cues, and several studies have shown that kin recognition in fish involved olfaction (Myrberg 1975; McKaye and Barlow 1976; Brown and Brown 1996). In laboratory experiments, Loiselle (1983) showed that desert pupfish can discriminate between sired and alien eggs using only chemical cues. Neff and Sherman (2003) demonstrated that

male bluegill sunfish used chemical cues to discriminate between sired and foster fry, but did not discriminate between sired and foster eggs. Sargent (1989) also found that male fathead minnows, a species with a breeding system similar to that of spottail darters, can discriminate between foster and sired eggs, although chemical cues were not confirmed as the basis of discrimination. Sargent (1989) noted that, since experimental males received the foster eggs before being allowed to spawn, positional and age differences between sired and foster eggs could have served as cues. Another possibility is that male spottail darters used egg presence and/or position on the ceilings of nest sites as cues. In the spottail darter study (Bandoli 2002), parent males would not have detected a difference in egg number or position, while alloparent males were exposed to all “new” eggs (inexperienced) or to eggs in different positions (experienced). A third possibility is that males treated all eggs in any nest in which they have spawned as sired. All eggs in the “new” nest sites that the alloparent males received were foster eggs and were therefore cannibalized at a higher rate. To distinguish between these possibilities, a new field study was conducted in which eggs on one side of spottail darter nest sites were either (1) moved to change their position only (repositioned sired egg sites) or (2) swapped with eggs from another nest site, changing both position and parentage (foster egg sites). In both cases, approximately half of the eggs in each nest site were relocated or replaced, while eggs on the other side of the nest site remained unchanged in position and paternity. If males use olfactory cues to discriminate among eggs, egg losses due to filial cannibalism should be higher among swapped eggs than unchanged eggs at foster egg nest sites, with no differences in egg loss between relocated and unchanged eggs among nest sites containing only repositioned sired eggs. If darters use egg position to determine paternity, filial cannibalism should be higher for the manipulated eggs compared to unchanged eggs in sites containing either repositioned sired eggs or foster eggs.

Materials and methods Artificial nest sites (15-cm sections of ceramic field tiles) have been used successfully as spawning sites for spottail darters (Bandoli et al. 1991). For this study, each tile was modified by the addition of two 13×8-cm sheets of clear acetate positioned on the underside of the tile and held in place with rubber bands (Fig. 1). Acetate sheets were lightly sanded to roughen the surface to improve egg adhesion. A pilot study in 2001 showed that males readily defended tiles with acetate sheets, and eggs deposited by females on the surface of the acetate were not displaced when the sheets were removed from the tiles. In late March and early April 2003, 35 modified tiles were placed in three tributaries of Bayou Creek in Vanderburgh County, IN, USA; 30 tiles were placed in two of these tributaries in 2004. Tiles were placed approximately 10 m apart to minimize the chance that a male

608

Fig. 1 Drawing of the underside of a ceramic field tile nest site showing the position of the two acetate sheets on which eggs were deposited

could simultaneously guard more than one nest site. Tiles were checked weekly for eggs and guarding males, which were easily captured using a hand net. Eggs in nests were assumed to have been sired by the guarding male. While the presence of foster eggs in spottail darter nests is certainly possible due to the death or displacement of the guarding male, high levels of nest site fidelity observed in this species over several weeks (Bandoli 1997) suggest that these events are relatively infrequent. This assumption will be addressed in more detail in the “Discussion.” Foster egg nest sites were established by swapping eggs between nest sites. Egg swapping occurred when: (1) at least two tiles at a locality were found to have eggs deposited on both acetate sheets; (2) there were >50 eggs on each sheet; (3) eggs were in a relatively early developmental state and, therefore, were not going to hatch during the experiment; and (4) the male guarding each tile was captured. Eggs were swapped by removing one sheet (together with its adhering eggs) from each tile and swapping them between tiles. Broods were photographed both before and after the eggs were swapped. Guarding males were measured (standard length) and marked with a small clip on the upper lobe or the lower lobe of the caudal fin and returned to their original tiles, usually within 30 min of capture. The combination of standard length and caudal fin clip allowed individual recognition of all males. Repositioned sired egg nest sites met conditions (2)–(4) (above) and were established by removing, reversing, and reattaching one of the acetate sheets and the attached eggs (henceforth relocated eggs). Eggs on the other side of the nest site were unchanged. In spottail darters, each female tended to deposit her clutch in a discrete location, such that eggs on acetate sheets were not spread uniformly across the sheet. Reversing the sheet therefore produced a significant change in egg position on that side of the nest site without changing paternity, allowing assessment of male response to positional changes in sired eggs. All nest sites used met conditions (2)–(4) (above). Manipulated tiles were revisited 48 h after the eggs were swapped. Guarding males were captured and identified by standard length and caudal fin clip, and broods were

photographed. Males were then returned to their nest sites; no male was used more than once in this study. In the laboratory, all eggs were counted from projected slides of brood photographs. Side-by-side projections of broods before and after the 48-h test interval allowed the identification and exclution of new eggs added during the interval (n=14 of 43 trials). Trials in which it was difficult to distinguish between original eggs and eggs added during the test interval were excluded from the analysis. All eggs lost during the 48-h test interval were assumed to have been consumed by the guarding male. Support for this assumption comes from (1) laboratory observations of significant levels of filial cannibalism by male spottail darters (Bandoli 2002), and (2) the tendency of guarding males to remain with their eggs in the field (personal observation), thereby decreasing opportunities for egg consumption by predators. This assumption is addressed further in the “Discussion.” Since this study employed naturally occurring broods of eggs, there was variation between tiles in initial brood size, which could influence the degree of filial cannibalism by the guarding male. Therefore, paired t tests were used to compare differences in egg numbers and percentages across nest sites. Since individual males may vary in the overall number of eggs they consume, correlation was used to compare the percentage of sired and foster eggs consumed (foster egg nest sites) or the percentage of relocated and unchanged eggs consumed (repositioned sired egg nest sites) across males. Depending on normality, Student’s t tests or Mann–Whitney tests were used to compare initial conditions and egg consumption means between foster egg and repositioned sired egg trials. Results are expressed as mean ± 1 standard error (SE); all statistical analyses were performed using SigmaStat 2.0 (Jandel Scientific, 1995).

Results Of 22 repositioned sired egg trials (the position of sired eggs on one side of the tile was changed) initiated, nine yielded valid data (2 of 5 in 2003 and 7 of 17 in 2004). Trials were invalidated by heavy rainfall or other disturbances that either moved tiles or prevented their relocation, or by our inability to recapture the guarding male. In eight trials, new eggs were added during the 48-h interval. In three of these trials, the positions of the new eggs were discrete enough to exclude them from eggs present initially; these trials were included in data analysis. The mean initial brood size was 590.7±64.2 and was similar to the mean of 586.0 eggs found in a previous study using tiles without acetate sheets (Bandoli 1997). Following the removal and reattachment (relocation) of one of the acetate sheets, the number of unchanged eggs ðX ¼306:3
52:9Þ was similar across trials to the number of relocated eggs ðX ¼284:3
41:1Þ (paired t test, t8=0.31, P=0.76). The mean number of unchanged and relocated eggs consumed during the 48-h test interval was 81.0±27.0 and 78.9±18.5, respectively, and similar numbers of un-

609

changed and relocated eggs were consumed across individual trials (paired t test, t8=0.09, P=0.93). (Due to variation in total egg consumption across trials, the power of this test is low; the probability of detecting a 20-egg difference in egg consumption between unchanged and relocated eggs, which is equivalent to 25% of the mean egg consumption, is only 0.125.) Percentages of unchanged and relocated eggs consumed (X ¼25:0
4:5 and 25:1
3:5; respectivelyÞ were similar across trials (paired t test, t8=0.05, P=0.96). (The power of this test is considerably higher; the probability of detecting a 6.25% difference in egg consumption, which is analogous to a 20-egg difference in egg consumption between unchanged and relocated eggs, is 0.571.) The total number of eggs consumed in the nine valid trials ranged from 59 to 426; there was a positive but nonsignificant correlation across trials between the numbers of unchanged and relocated eggs consumed (Spearman’s rank correlation coefficient, rs=0.47, P=0.19, df=8). However, there was a strong correlation between the percentages of unchanged and relocated eggs consumed (Spearman’s rank correlation coefficient, rs=0.82, P=0.004, df=8; Fig. 2). Of 21 foster egg trials (one acetate sheet with sired eggs was replaced by one containing foster eggs), 15 yielded valid data, all in 2003. In six trials, new eggs were added during the 48-h interval. In four of these trials, the positions of the new eggs were discrete enough to exclude them from eggs present initially; these trials were included in the data analysis. Across the 15 trials, the number of sired eggs on the removed acetate sheet (X ¼240:7
31:1) was similar to the number of foster eggs (X ¼276:9
34:6) that replaced them (paired t test, t14=−1.27, P=0.22). After replacement, the mean brood size was 583.5±62.8, which was not

significantly different from the mean initial brood size in the relocated egg trials (590.7±64.1; two-sample t test, t22=0.07, P=0.94). Following egg replacement, the number of foster eggs (X ¼276:9
34:6 ) was similar to the number of sired eggs (X ¼306:7
42:5) across trials (paired t test, t14=0.66, P=0.52). The mean number of sired and foster eggs consumed during the 48-h test interval was 104.5±20.6 and 90.3±20.5, respectively. Levels of sired and foster egg consumption across trials were similar (paired t test, t14=0.68, P=0.50). (Variation in egg loss across trials significantly reduces the power of this test; the probability of detecting a 25-egg difference in consumption between sired and foster eggs, which is equivalent to 25% of the mean egg consumption, is 0.204.) Percentages of original sired and foster eggs consumed (X ¼ 39:9
8:2% and 40:6
8:2%;respectively) were similar across trials (paired t test, t14=0.16, P=0.88). (The power of this test is considerably higher; the probability of detecting a 10% difference in egg consumption, which is analogous to a 25-egg difference between unchanged and relocated eggs, is 0.631.) Guarding males did not appear to discriminate between sired and foster eggs. Moreover, despite individual variations in total egg consumption, individual males consumed similar numbers and percentages of sired and foster eggs. Spearman’s rank correlation coefficients comparing levels of sired and foster egg consumption were significant for both the number of eggs consumed (rs=0.52, P=0.04, df=14) and the percentage of eggs consumed (rs=0.81, P<0.001, df=14; Fig. 3). The valid repositioned sired egg and foster egg trials did not differ in mean guarding male size (standard length),

Fig. 2 Sired egg consumption by guarding male spottail darters in repositioned egg trials. At each nest, eggs on one side were unchanged, while those on the other side were relocated within that side by reversing the acetate sheet to which they were attached. Egg consumption is expressed as a percentage of the initial broods of unchanged and relocated eggs at each nest site lost during 2 days. The diagonal line represents equal percentage consumption in both categories. Despite variation among males in egg consumption, males consumed similar percentages of eggs in each category (Spearman’s rank correlation coefficient, rs=0.82, P=0.004)

Fig. 3 Sired and foster egg consumption by guarding male spottail darters in foster egg trials. At each nest, eggs on one side were unchanged, while those on the other side were replaced by eggs from a different nest site. Egg consumption is expressed as a percentage of the initial broods of unchanged (sired) and replaced (foster) eggs at each nest site consumed during 2 days. The diagonal line represents equal percentage loss in both categories. Despite variation among males in egg consumption, males consumed similar percentages of eggs in each category (Spearman’s rank correlation coefficient, rs=0.81, P=0.001)

610 Table 1 Comparisons of means (± SE) of guarding male standard length (SL), initial brood size, and egg consumption between repositioned sired egg and foster egg trials. Only valid trials are included (see text) Parameter

Pa

Trials Repositioned sired Foster eggs eggs (n=9) (n=15)

Guarding male SL 67.8±2.0 (mm) Initial brood size following 590.7±64.2 egg manipulation Total number of eggs 159.9±40.5 consumed Percentage of initial brood 25.3±3.9 consumed

67.4±1.7

0.89

583.5±62.8 0.94 194.8±35.6 0.49 39.3±7.8

0.20

a

P value was from Student’s t tests or, when normality was violated, from Mann–Whitney test

initial brood size (after egg manipulation), total egg consumption, or percentage of the initial brood consumed (Table 1). There was a significant positive correlation between male size (standard length) and premanipulation brood size across combined valid trials (Spearman’s rank correlation coefficient, rs=0.50, P=0.01, df=23). Egg consumption across all valid trials was not related to male size (rs=0.12) or to initial postmanipulation brood size (rs=0.14). The mean number of eggs consumed across all valid trials was 181.7±26.2, ranging from 47 to 512.

Discussion Males guarding nest sites in which half of the eggs were repositioned by reversing the acetate sheet to which they were attached did not appear to discriminate between unchanged and relocated sired eggs. Individual males consumed similar percentages of the initial broods in each category, suggesting that the position of eggs within the nest site does not act as a cue for egg ownership; however, it is possible that, despite the tendency of females to deposit eggs in distinct clusters, reversing an acetate sheet did produce enough of a change in egg position for the guarding male to detect a difference. If males use olfaction to discriminate between sired and foster eggs, males guarding nest sites in which half of the eggs are not sired by the guarding male are predicted to consume more foster eggs than sired eggs; however, this was not the case. While these males varied considerably in the extent of egg consumption, levels of foster and sired egg consumption were generally similar at each nest site, both in terms of the absolute number of eggs and the percentage of the initial sired and foster broods consumed. Males consumed similar total numbers of eggs in repositioned sired egg and foster egg trials, suggesting that the presence of unrelated eggs did not alter overall levels of filial cannibalism.

Kin recognition via chemical cues has been demonstrated in many species of fishes (Blaustein et al. 1987), and recognition requires the production of a chemical cue that is unique to a kin group. In salmonids, both urine (Moore et al. 1994) and bile acids (Zhang et al. 2001) have been implicated as chemical cues that alter the behavior of conspecifics; however, in both studies, odorants were collected from adult fish. It is possible that these or similar chemical cues are either not produced in early embryos or produced in quantities that are insufficient for use as effective chemical cues. This may explain why male bluegill sunfish are able to distinguish between sired and unrelated fry, but not eggs (Neff 2003a; Neff and Sherman 2003), and why adult female cichlids can discriminate among embryos only when they are close to hatching (Myrberg 1975). Eggs in the first third of their development were used in the current study to minimize the possibility of hatching during the experiment; this may have also limited the production of chemical cues by the embryos. However, early-term eggs were also used in a previous laboratory study in which guarding male spottail darters discriminated against foster eggs (Bandoli 2002). In that study, males given their original nest sites and sired eggs consumed significantly fewer eggs than did alloparent males given a new nest site containing foster eggs, implying a mechanism for discrimination that the current study failed to detect. A possible resolution to this contradiction is the hypothesis that male spottail darters do not discriminate among eggs within a nest site in which the male has spawned. All males used in the current study had spawned in that nest site prior to any egg manipulation. Under this rule, all eggs are treated equally regardless of subsequent changes in relatedness or relative position. In the previous study, only the males that received their original nest sites and sired eggs had spawned in that nest site and therefore treated all eggs as sired. The alloparent males had either not spawned prior to receiving the new nest site and foster eggs, or had spawned in a different nest and therefore treated all eggs as unrelated. Discrimination in the previous study may therefore have been based on the familiarity of the nest site rather than of the eggs within. Equal treatment of all eggs in a nest in which the male spawned and therefore has at least some sired eggs may have parallels in passerine birds in which parents recognize nests but not individual offspring. In both tree swallows (Tachycineta bicolor; Whittingham et al. 1993) and western bluebirds (Sialia mexicana; Dickinson 2003), males did not reduce offspring provisioning when paternity confidence was reduced as long as they had some mating opportunities with the females during egg laying. However, replacement males that arrived during incubation and had no genetic interest in the clutch behaved differently. Replacement male tree swallows killed the nestlings (Robertson 1990), and replacement male western bluebirds frequently failed to provide food for the nestlings (Dickinson and Weathers 1999). Males in these species behaved as if following the “all-or-none” rule (Dickinson and Weathers 1999)—provisioning at normal levels as long

611

as they had some possibility of genetic representation in the nestlings, and ignoring or killing them when their likelihood of paternity was zero. Increased egg cannibalism by male spottail darters defending eggs at nest sites in which they have not spawned and therefore have no paternity may indicate their use of a similar all-or-none rule. Males guarding sired eggs did not differ from males guarding foster eggs in the number of eggs consumed; however, there was considerable variation in the number and percentage of eggs consumed across males in both categories. This variation was not explained by variation in male size or initial postmanipulation brood size, and may be due to variation in availability of alternate food (Neff 2003a) or male condition (Kvarnemo et al. 1998; Manica 2004). Alternatively, some variation in levels of filial cannibalism may be due to microbial egg contamination. Guarding male darters may consume infected eggs to limit microbial infection (Winn 1958), and nests with higher infection rates may necessitate more egg consumption. Additionally, if guarding male spottail darters secrete antimicrobial compounds in epidermal mucus as do males of the closely related fringed darter (Etheostoma crossopterum; Knouft et al. 2003), variation among males in the amount or effectiveness of the secreted antimicrobial compound may contribute to the variation in the need to consume infected eggs. One vital assumption of this study is that males actually sired the eggs that were initially found in their nests. If males frequently change nest sites after spawning, then mean relatedness between guarding males and eggs would be reduced, leading to increased alloparental care. Alloparental care has been documented in two species of darters: the tessellated darter (E. olmstedi) and the striped darter (Etheostoma virgatum). In the tessellated darter, Constantz (1985) observed that males frequently moved between nest sites (rocks), with larger males displacing smaller males from limited breeding sites and abandoning them when most of the available surfaces for egg deposition had been filled. Smaller males would frequently occupy these abandoned nests, tending the eggs and attempting to attract additional females. Constantz (1979) also reported two instances of possible female mimicry by subordinate male tessellated darters as a tactic to increase egg fertilizations. These observations suggest a strong potential for alloparental care, but a genetic analysis of eggs and guarding males in a different population of E. olmstedi found evidence of foster eggs in only 3 of 16 nests sampled (DeWoody et al. 2000). Porter et al. (2002) also used genetic analysis to document alloparental care in male striped darters. The extent of alloparental care varied, as some foster eggs were found in 4 of 8 nests in one population, but only in 1 of 11 nests in another. Striped darters are in the same subgenus as spottail darters (Page 1983), and competition for limited nest sites may produce some opportunities for alloparental care in both species. However, there is no evidence that either species has the high rate of internest movement seen

in male E. olmstedi. Rather, evidence suggests that the degree of movement by male spottail darters between nest sites is limited. In a field study, Bandoli (1997) found that 9 of 14 uniquely marked E. squamiceps males in southern Indiana were consistently found at the same nest sites for several weeks; three of these males defended multiple consecutive broods without changing nest sites. In the current study, the same male that was initially guarding the tile was recaptured at that tile on 32 of 35 instances in which the tile was not significantly disturbed or buried during the intervening 2 days. At only one of these tiles was the original male replaced by a different male. These observations do not rule out the possibility of the cuckoldry of nest-guarding males—however, neither do they support the level of internest movement of males that would significantly reduce the mean paternity of guarded eggs at a significant proportion of nests. A second critical assumption is that egg loss during the 2-day trial period was due to filial cannibalism. The streams where this study was conducted contain several invertebrate and vertebrate egg predators, and several years of observation have shown that abandoned eggs quickly disappear. In the fringed darter (E. crossopterum), egg loss to predators reaches 100% in 24 h following the removal of the guarding male (Knouft and Page 2004). While guarding male spottail darters occasionally leave the nest to forage, they are rarely absent for longer than a few minutes; males initially missing from their nests are often captured there 5–10 min later (Bandoli, personal observation). Therefore, opportunities for egg predation would seem to be limited. Furthermore, filial cannibalism has been documented in this species in the laboratory. Even when supplemental food was provided, males consumed more than 200 sired eggs or more than 300 foster eggs during 3 days in the laboratory (Bandoli 2002)—a rate similar to the mean of 181.7 eggs consumed in 2 days by males in this study. It therefore appears likely that the majority of egg loss in this study was due to filial cannibalism. In conclusion, a previous laboratory study in which male spottail darters were given either their own nest sites and eggs or an alien nest site containing foster eggs found that males consumed more foster eggs than sired eggs, implying a mechanism for discrimination. The current study found that neither egg position in the nest nor chemical cues based on relatedness to the guarding male are used as discrimination cues. A possible solution to the apparent contradiction is that males treat all eggs the same in nests in which they have previously spawned, regardless of any positional or chemical variation. This would explain the higher rates of filial cannibalism in males defending foster eggs in the previous study, since these males were defending nest sites in which they had not spawned. This is not to say that discrimination between foster and sired eggs in a mixed brood would not be beneficial to the guarding male, and the apparent inability of spottail darters to use chemical or positional cues in these situations implies either a relatively recent history of alloparental care, a

612

lack of reliable chemical cues in eggs, or a cost of evolving egg discrimination mechanisms exceeding benefits. Acknowledgements T. Hayes, A. Jackson, M. Walls, and S. Loveless provided valuable assistance in the field experiments. B. Wilding provided statiscal advice. Comments by three anonymous reviewers greatly improved an earlier version of this manuscript. Partial funding for this research was provided by a grant from the Indiana Academy of Science. This research was conducted with approval from the University of Southern Indiana Institutional Animal Care and Use Committee (permit no. 101).

References Bandoli JH (1997) Factors influencing reproductive success in male spottail darters (Etheostoma squamiceps, Pisces, Percidae). Proc Indiana Acad Sci 106:145–157 Bandoli JH (2002) Brood defense and filial cannibalism in the spottail darter (Etheostoma squamiceps): the effects of parental status and prior experience. Behav Ecol Sociobiol 51:222–226 Bandoli JH, Lanigan JT III, Sheckles TA (1991) Reproduction in the spottail darter in Indiana: use of artificial nest sites. Proc Indiana Acad Sci 100:65–75 Blaustein AR, Bekoff M, Daniels TJ (1987) Kin recognition in vertebrates (excluding primates): mechanisms, junctions and future research. In: Fletcher DJC, Michener CD (eds) Kin recognition in animals. Wiley, Toronto, pp 333–357 Brown GE, Brown JA (1996) Kin discrimination in salmonids. Rev Fish Biol Fish 6:201–219 Constantz GD (1979) Social dynamics and parental care in the tessellated darter (Pisces: Percidae). Proc Acad Nat Sci Philadelphia 131:131–138 Constantz GD (1985) Alloparental care in the tessellated darter, Etheostoma olmstedi (Pisces: Percidae). Environ Biol Fish 14: 175–183 DeWoody JA, Fletcher DE, Wilkins SD, Avise JC (2000) Parentage and nest guarding in the tessellated darter (Etheostoma olmstedi) assayed by microsatellite markers (Perciformes: Percidae). Copeia 2000:740–747 DeWoody JA, Fletcher DE, Wilkins SD, Avise JC (2001) Genetic documentation of filial cannibalism in nature. Proc Natl Acad Sci U S A 98:5090–5092 Dickinson JL (2003) Male share of provisioning is not influenced by actual or apparent loss of paternity in western bluebirds. Behav Ecol 14:360–366 Dickinson JL, Weathers WW (1999) Replacement males in the western bluebird: opportunity for paternity, chick-feeding rules, and fitness consequences of male parental care. Behav Ecol Sociobiol 45:201–209 FitzGerald GJ, van Havre N (1987) The adaptive significance of cannibalism in stickleback (Gasterosteidae: Pisces). Behav Ecol Sociobiol 20:125–128 Jones JC, Reynolds JD (1999) Costs of egg ventilation for male common gobies breeding in conditions of low dissolved oxygen. Anim Behav 57:181–188 Knouft JH, Page LM (2004) Nest defense against predators by the male fringed darter (Etheostoma crossopterum). Copeia 2004: 915–918 Knouft JH, Page LM, Plewa MJ (2003) Antimicrobial egg cleaning by the fringed darter (Perciformes: Percidae: Etheostoma crossopterum): implications of a novel component of parental care in fishes. Proc R Soc Lond B Biol Sci 270:2405–2411 Kvarnemo C, Svensson O, Forsgren E (1998) Parental behavior in relation to food availability in the common goby. Anim Behav 56:1285–1290 Lindström K (1998) Energetic constraints on mating performance in the sand goby. Behav Ecol 9:297–300

Lindström K, Sargent RC (1997) Food access, brood size and filial cannibalism in the fantail darter, Etheostoma flabellare. Behav Ecol Sociobiol 40:107–110 Lissåker M, Kvarnemo C, Svensson O (2003) Effects of a low oxygen environment on the parental effort and filial cannibalism in the male sand goby, Pomatoschistus minutus. Behav Ecol 14:374–381 Loiselle PV (1983) Filial cannibalism and egg recognition by males of the primitively custodial teleost Cyprinodon macularius californiensis Girard (Atherinomorpha: Cyprinodontidae). Ethol Sociobiol 4:1–9 Manica A (2002) Filial cannibalism in teleost fish. Biol Rev 77:261– 277 Manica A (2004) Parental fish change their cannibalistic behavior in response to the cost-to-benefit ratio of parental care. Anim Behav 67:1015–1021 McKaye KR, Barlow GW (1976) Chemical recognition of young by the midas cichlid, Cichlasoma citrinellum. Copeia 1976:276– 282 Moore A, Ives MJ, Kell LT (1994) The role of urine in sibling recognition in Atlantic salmon, Salmo salar (L.) parr. Proc R Soc Lond B Biol Sci 255:173–180 Myrberg AA Jr (1975) The role of chemical and visual stimuli in the preferential discrimination of young by cichlid fish Cichlasoma nigrofasciatum (Gûnther). Z Tierpsychol 37:274–297 Neff BD (2003a) Paternity and condition affect cannibalistic behavior in nest-tending bluegill sunfish. Behav Ecol Sociobiol 54:377–384 Neff BD (2003b) Decisions about parental care in response to perceived paternity. Nature 422:716–719 Neff BD, Gross MR (2001) Dynamic adjustment of parental care in response to perceived paternity. Proc R Soc Lond B Biol Sci 268:1556–1559 Neff BD, Sherman PW (2003) Nesting recognition via direct cues by parental male bluegill sunfish (Lepomis macrochirus). Anim Cogn 6:87–92 Östlund S, Ahnesjö I (1998) Female fifteen-spine sticklebacks prefer better fathers. Anim Behav 56:1177–1183 Page LM (1974) The life history of the spottail darter, Etheostoma squamiceps, in Big Creek, Illinois and Ferguson Creek, Kentucky. Ill Nat Hist Surv Bull 33:275–295 Page LM (1983) Handbook of darters. TFH Publications, Inc. Ltd., Neptune City, NJ Page LM, Ceas PA, Swofford DL, Buth DG (1992) Evolutionary relationships within the Etheostoma squamiceps complex (Percidae: subgenus Catonotus) with descriptions of five new species. Copeia 1992:615–646 Porter BA, Fiumera AC, Avise JC (2002) Egg mimicry and alloparental care: two mate-attracting tactics by which nesting striped darters (Etheostoma virgatum) males enhance reproductive success. Behav Ecol Sociobiol 51:350–359 Robertson RJ (1990) Tactics and counter-tactics of sexually selected infanticide in tree swallows. In: Blondel J, Gosler A, Lebreton JD, McCleery R (eds) Population biology of passerine birds, an integrated approach. Springer, Berlin Heidelberg New York, pp 381–390 Rohwer S (1978) Parent cannibalism of offspring and egg raiding as a courtship strategy. Am Nat 112:429–440 Sargent RC (1989) Alloparental care in the fathead minnow, Pimephales promelas: stepfathers discriminate against their adopted eggs. Behav Ecol Sociobiol 25:379–385 Svensson O, Magnhagen C, Forsgren E, Kvarnemo C (1998) Parental behavior in relation to the occurrence of sneaking in the common goby. Anim Behav 56:175–179 Taborsky M (1994) Sneakers, satellites, and helpers: parasitic and cooperative behavior in fish reproduction. Adv Study Behav 23:1–100 Unger LM, Sargent RC (1988) Alloparental care in the fathead minnow, Pimephales promelas: females prefer males with eggs. Behav Ecol Sociobiol 23:27–32

613 Whittingham LA, Dunn PO, Robertson RJ (1993) Confidence of paternity and male parental care: an experimental study in tree swallows. Anim Behav 46:139–147 Winn HE (1958) Observations on the reproductive habits of darters (Pisces—Percidae). Am Midl Nat 59:190–212

Wisenden BD (1999) Alloparental care in fishes. Rev Fish Biol Fish 9:45–70 Zhang C, Brown SB, Hara TJ (2001) Biochemical and physiological evidence that bile acids produced and released by lake char (Salvelinus namaycush) function as chemical signals. J Comp Physiol B Metab Transp Funct 171:161–171