Food preferences and related behavior of the browsing sea urchin <Emphasis Type="Italic">Tripneustes gratilla</Emphasis> (Linnaeus) and its potential for use as a biological control agent

The short-spined toxopneustid sea urchin Tripneustes gratilla feeds on a wide variety of algal species and on sea grasses. However, the urchin does show preferences when offered a selection of macroalgal species, which it encounters in nature. Preferences among macroalgae were evident in field-collected urchins exposed to pair-wise tests where the variable was either the consumption rate of the algae or observation of which algal species the urchins chose to touch with their lantern teeth. Exposure of lab-housed urchins to one of five species of macroalgae for 5 months did not seem to alter preferences of urchins in three of the exposure groups, but those exposed to Padina sanctae-crucis seemed to show an enhanced preference for this species when offered a choice of the five species of macroalgae at the end of the exposure period, and those exposed to Gracilaria salicornia seemed to avoid the species when offered the choice of the five species. Perhaps more ecologically important than their preferences were two other observations on these urchins: first, when offered only a single species of algae, the urchins on four of five diets ate the same quantity per day. Second, when simultaneously offered the choice among the five macroalgal species, the urchins consumed more macroalgae per day than when offered only one species. These urchins move about a meter a day. They probably encounter food resources in a relatively coarse-grained fashion and have evolved to eat what is available. Because of their limited movements, their habitat overlap with grazing fishes, their acceptance of a wide variety of macroalgae and their preference for macroalgae, these native urchins are thought to have the potential to serve as biological control agents of alien and invasive macroalgae, which have come to dominate some reef zones normally occupied by corals in Hawaii.     

Shell selection in the hermit crab Pagurus bernhardus

Three shell variables were deemed to be important in the selection of a shell by the hermit crab Pagurus bernhardus (L.). These were the weight, volume and aperture width of a shell. Each of these variables were used in multiple regressions against crab weight for 25 Littorina littorea and 25 Thais lapillus shells chosen when both shell species were together in a tank, and for a further 25 L. littorea and 25 T. lapillus when the two shell species were separate. Principal component analysis was performed on the four groups of 25 shells selected, and multiple-regression equations were calculated using the principal components as the new variables. It was found that crabs chose a shell of suitable general dimensions rather than solely on the basis of one shell variable. The advantages of using statistical techniques developed in this paper over previous approaches to shell selection are discussed.     

Role of food subsidies and habitat structure in influencing benthic communities of shell mounds at sites of existing and former offshore oil platforms

Mussels (Mytilus californianus, M. galloprovincialis) and other organisms sloughed from offshore oil platforms provide a food subsidy to benthic consumers and alter underlying soft bottom habitat by creating hard substrate. The removal of overlying platforms eliminates this food subsidy, but large shell mounds remain. The distribution, abundance, and population characteristics of mobile macroinvertebrates differed among shell mounds beneath existing offshore oil platforms, shell mounds at the former sites of offshore oil platforms, and soft bottom. Predatory and omnivorous echinoderm and mollusk species were more abundant and generally larger on shell mounds under platforms than on shell mounds without platforms. Omnivorous and deposit feeding echinoderms were the most abundant macroinvertebrate taxa sampled on mound-only sites. The brown rock crab (Cancer antennarius), known to have a strong preference for hard substrate, was significantly more abundant on shell mounds, with or without platforms, than adjacent soft bottom sites. Results suggest that the effects of platform removal differed among benthic species according to trophic level, degree of mobility, and substrate preference. Although the shell mound habitat persists after removal of platform structures, species abundance and the composition of the associated benthic community is altered by removal of the platform structure.Communicated by P.W. Sammarco, Chauvin     

Consequences of shell-species preferences for female reproductive success in the hermit crab Pagurus bernhardus

Littoral hermit crabs, Pagurus bernhardus, show a strong preference for Littorina obtusata shells rather than those of Gibbula species. The fitness consequences, in terms of fecundity, for this shell preference is examined for female crabs. Females in the preferred species produced eggs earlier in the season, produced more eggs in the first brood, and produced a second brood more often than did females in the less preferred species. The smaller brood for Gibbula spp. was not a consequence of egg loss from the pleopods due to an unfavourable shape of shell, but rather reflected lower egg production. It is not clear, however, if this differential reproduction is due to direct costs of carrying an unfavourable shell, i.e. the shell impedes reproduction, or whether crabs compete aggressively for favoured shells so that only crabs of low quality inhabit lowquality shells.     

Biochemical genetic variation at a leucine aminopeptidase (LAP) locus in blue (Mytilus galloprovincialis) and greenshell (Perna canaliculus) mussel populations along a salinity gradient

In several marine bivalve species, biochemical genetic variation at a leucine aminopeptidase (LAP) locus is associated with environmental variability, primarily salinity fluctuation. Population genetic variation at a LAP locus was investigated here in two sympatric mussel species (Mytilus galloprovincialis and Perna canaliculus) from three locations along a salinity gradient in Wellington Harbour, New Zealand. The data for M. galloprovincialis and P. canaliculus do not support the hypothesis that the LAP polymorphism in either species is associated with salinity variation among adult mussels. Due to the absence of small mussels among the samples it is not possible to discount the hypothesis that selection acts primarily against juveniles, as it does for M. edulis in Long Island Sound, USA. Wellington Harbour populations of M. galloprovincialis exhibited large and often highly significant heterozygote excesses at the LAP locus, whereas populations of P. canaliculus from the same locations exhibited large and highly significant heterozygote deficiencies. The reason for this inter-specific difference in population structure is unknown. If it is the result of selection, this suggests that selection acts differentially upon the two species, because demographic attributes and reproductive biology are very similar in the two species. For both M. galloprovincialis and P. canaliculus, significant levels of population genetic heterogeneity were recorded among three locations separated by only 8 to 12 km. Neither species exhibited shell length-dependent genetic variation at the LAP locus, suggesting that for these two species the LAP polymorphism is not associated with variation in shell length. Received: 30 December 1996 / Accepted: 6 January 1997     

Potential effects of a non-indigenous predator in its expanded range: assessing green crab, <Emphasis Type="Italic">Carcinus maenas</Emphasis>, prey preference in a productive coastal area of Atlantic Canada

The non-indigenous green crab (Carcinus maenas) is an important predator on bivalve wild beds in coastal areas worldwide. This study explored size-dependent green crab prey preference on American oysters (Crassostrea virginica), blue mussels (Mytilus edulis), and soft-shell clams (Mya arenaria) in a productive coastal system of Atlantic Canada. Using two sizes of prey and three different experimental manipulations, small, medium, and large green crabs were given a choice among these three bivalves, and their daily feeding rates were monitored over the course of 3 days. For both prey sizes, green crabs showed an early feeding preference for soft-shell clams and, only as they declined in numbers, a switch toward mussels and subsequently toward oysters. We found that such changes in the timing (order) of prey preference are related to prey differences in shell thickness, a fairly reliable indicator of prey shell strength.     

Induced anti-predator responses of the green mussel, Perna viridis (L.), on exposure to the predatory gastropod, Thais clavigera Küster, and the swimming crab, Thalamita danae Stimpson

A laboratory experiment was conducted to test whether anti-predator responses could be induced in the green mussel, Perna viridis, on exposure to waterborne cues from two predators, the muricid gastropod, Thais clavigera, and the portunid crab, Thalamita danae, and from damaged conspecifics. There were five treatments in this study. Aquaria of two treatments housed either Thalamita danae or Thais clavigera. Another three treatments housed only intact mussels, broken and intact mussels, or no mussels. No significant differences were found among these five treatments in final shell weight and tissue dry weight. Induced defensive responses were predator-specific. Experimental mussels exposed to waterborne cues of T. clavigera had a higher percentage increase in shell length, height and width, whereas those exposed to waterborne cues of T. danae had a higher percentage increase in shell width and height. Mussels raised in the presence of crabs developed thicker shell at the umbo and lip margin, whereas mussels raised in the presence of gastropods had thicker shell lip. Predator preference for mussels from the control group and for those pre-exposed to waterborne cues of either of the predators was studied. T. clavigera preferentially selected mussels from the control group, whereas no prey preference was found for T. danae, although shell breaking time for mussels exposed to a crab was longer than that for the control. While the present results should be viewed with caution in the absence of a dietary control, this was the first time that prey handling was shown to be affected by defensive changes in the morphology of mussels.Communicated by T. Ikeda, Hakodate     

Shell-size selection by intertidal sympatric hermit crabs

This study evaluated selection for shell size by three species of tropical intertidal hermit crabs, Clibanarius antillensis, C. sclopetarius, and C. vittatus, from species of shells which are frequently used in nature. Crab size and weight were strongly and significantly related to all measured parameters of the selected shells. The strength of these relationships (r² values) depended neither on the crab nor on the shell variables taken into account. The relationships between crab size and the dimensions of the selected shells showed higher r² values than the corresponding relationships with the shells that the crabs had occupied when they were collected (0.482–0.903 in comparison to 0.091–0.652, respectively), indicating that the crabs were occupying sub-optimal shells in nature. Negative allometry was frequently found in the relationships between crab and shell variables, indicating that large crabs select and use proportionally lighter shells than do small crabs. This negative allometry was stronger for the shells used in nature (except for C. antillensis), i.e. larger crabs tended to select heavier shells in the laboratory than in nature. Different allometric relationships were also recorded among the dimensions of shells used in nature and those selected by the hermit crabs in free-access experiments: as shell length increased, the selected shells were heavier and had larger apertures than the shells used in nature. The relationships between crab size and the length and weight of the selected shells did not depend on the species of crab or species of shell, but only on crab size. Therefore, analyses using these variables can be performed without taking the species of crab or shell into account, i.e. data from different crab or shell species can be pooled. The influence of crab and/or shell species was recorded only in the models fitted for aperture length and width, variables which were more related to shell architecture than did shell length or weight. In contrast, if crab weight is used as an independent variable, different crab or shell species can be analyzed together independently of the particular shell parameter. This indicates that crab weight may be less susceptible than crab shield length to shell morphological constraints. Finally, the results indicate that the preferred size of a given shell type chosen by a given hermit crab will depend more on crab size or weight, than on the crab or shell species under consideration, i.e. crab shell-size relationships are not species specific.Communicated by P.W. Sammarco, Chauvin     

Territory quality affects female preference in a Lake Victoria cichlid fish

Mate preferences on male colour have been implicated in generating and maintaining species diversity among haplochromine cichlid fish. Their lek-like mating system suggests that not only male colour but also territory quality is instrumental in mate choice. We assessed the relative importance of territory quality and male colour in mate choice by testing whether territory quality can override the female preference for males of her own colour in the Lake Victoria cichlid genus Pundamilia. First, we showed in experimental groups that the dominant male preferentially monopolised a large tube relative to a small tube. The situation mimics quality difference in rocky crevices that serves as a focal point for male courtship display. Second, in mate-choice tests, Pundamilia nyererei females were allowed to choose between closely related P. nyererei and P. pundamilia males; these species differ strikingly in male nuptial coloration, but little else. We gave either both males the same small tube or one of them a large tube. The preference of P. nyererei females for P. nyererei males in the control situation (where both males had a small tube) was significantly diminished in favour of P. pundamilia males when the latter had the large tube. The results provide experimental evidence that differences in territory quality can override the female preference for males of her own colour. This finding is critical for a recent hypothesis proposing that male competition for mating territories can facilitate the process of sympatric speciation by sexual selection.     

Female mate preference for a costly ornament in male guppies

Female mate preference for dorsal fin length in male guppies (Poecilia reticulata) was investigated. In a dichotomous choice experiment using live males, females preferred males with longer dorsal fins to those with shorter dorsal fins. When the dorsal fin lengths of the two males were reversed by surgical manipulation, the females reversed their preference. To further examine this preference behaviour, a second dichotomous choice experiment was conducted using digitally modified male images that differed only in dorsal fin lengths. In this next experiment, females preferred male images with a longer dorsal fin than those with a shorter dorsal fin. In order to clarify the cost of possessing a long dorsal fin for males, the effect of dorsal fin length on swimming performance of males was also examined with a flow chamber. Male swimming performance was enhanced by surgically shortening the dorsal fin. The result of this study suggests that the long dorsal fin of male guppies is a costly handicap that may have evolved due to female mate choice.     

Genetic population structure of a species' complex of blue mussels (Mytilus spp.)

Blue mussels representing two nominal species (Mytilus trossulus Gould, 1850 and Mytilus galloprovincialis Lamarck, 1819 were collected from 28 intertidal locations along the Pacific coast of the USA in 1990–1991 (total N=1255) and examined for variation at 15 allozyme loci. Twelve samples, mostly from a region of suspected hybridization, were analyzed for variation in seven shell characters. Principal-components analysis of allozyme data revealed three groups based on first principal-component scores, which were identified as M. trossulus, M. galloprovincialis, and hybrids. Canonical discriminant analysis of shell characters was less successful in separating mussels into discrete groups. Each location was characterized for four environmental variables: (1) temperature, (2) salinity, (3) tidal height and (4) degree of exposure to wave action, which were then used as independent variables in a series of multiple-regression analyses, with the proportions of the two species as dependent variables. Temperature and salinity had significant (P<0.05) effects on the macrogeographic distribution of the two species, whereas the effects of height in the tidal zone and degree of wave exposure were not statistically significant. Salinity was found to have a greater influence than temperature on the microgeographic distribution of the two species. M. trossulus was more abundant at locations with lower temperatures and greater salinity variation than M. galloprovincialis. The two species appear to be ecologically distinct, and their genetic integrity is at least partly the result of environmental heterogeneity.     

Resource utilization for decorating in three intertidal majid crabs (Brachyura: Majidae)

 For three spider crabs (Tiarinia cornigera, Micippaplatipes and Pugettia quadridens quadridens), patterns of algal utilization for decorating were compared with the dynamics of algae on an intertidal rocky shore reef where the crabs co-occurred. T. cornigera and P. quadridens quadridens were most abundant from autumn to spring when the dominant algae (Boodleacoacta, Sargassum hemiphyllum, S. thunbergii and Corallina pilulifera) occurred in high coverage, while M. platipes was most abundant from spring to autumn. Monthly change of algae used for decorating was not correlated with algae growing in the crab habitat for T. cornigera, but for M. platipes, it was positively correlated for two algal species, and for P. quadridens quadridens, negatively correlated for one algal species. Each species of the spider crabs used some algal species preferentially for decoration. Decorating preference experiments conducted in the laboratory showed that M. platipes and P. quadridens quadridens exhibited similar preference to their algal utilization in the field, whereas for T. cornigera, algal preference in the experiment differed from utilization in the field. Comparisons between materials used for decoration and gut contents revealed that T. cornigera and M. platipes used algal species differently for decorating and feeding, while P. quadridens quadridens used the same algal species for both decorating and feeding. Different tactics for camouflage are discussed in terms of algal utilizations by the three majid species. Received: 25 September 1999 / Accepted: 22 June 2000     

Extrapair paternity in the swamp sparrow, <Emphasis Type="Italic">Melospiza georgiana</Emphasis>: male access or female preference?

Over the past two decades, the combination of molecular and field methods has revealed considerable variation in the level of extrapair fertilizations among socially monogamous birds. Models predicting extrapair young range in scale from a single population to multiple Orders, and there is no single, unifying theory for these reproductive tactics. We investigated proximate explanations of extrapair fertilizations in two subspecies of the swamp sparrow, Melospiza georgiana georgiana and Melospiza georgiana nigrescens, across a range of social and environmental conditions. The presence of extrapair young was best predicted by the size of two male plumage badges (one correlated with parental care and one with territorial aggression) relative to the badge size of their immediate neighbors, the interaction of these two measures, mean territory size, and the maximum size of the aggression badge among neighbors. The size of the male’s parental care badge (relative to neighbors) was negatively correlated with the probability of lost paternity. The relative size of the aggression badge was positively correlated with the presence of extrapair young when the parental care badge was small and negatively correlated when the badge was large. Controlling for these crown measures, males with larger territories were less likely to suffer losses in paternity. There was no effect of breeding density, breeding synchrony, their interaction, subspecies, or weather during the fertile period on the presence of extrapair young. These results suggest that female preference for males that provide more parental care (or preference for genes that convey this trait) plays a dominant role in extrapair interactions among swamp sparrows. Models based on female assessments of relative mate quality offer a promising explanation of patterns in extrapair fertilizations among bird species.     

Female adaptation to developmental plasticity in male calling behavior

The katydid Neoconocephalus triops exhibits in North America substantial developmental plasticity of male mating calls. The AM rate of the summer calls is significantly faster than that of the winter calls at the same temperature. In the tropics, where N. triops originated, males express only the fast summer-call phenotype. We tested two alternative hypotheses: (1) call plasticity in the population from North America evolved in response to selection by female preference after N. triops colonized North America, or (2) call plasticity evolved before N. triops expanded into North America and its expression in the novel environment led to adaptive change of female preferences. First, we tested whether call plasticity was present in tropical populations of N. triops. Tropical males expressed the winter-call phenotype when reared under winter conditions, indicating that call plasticity did not evolve in response to temperate climates. Second, we compared female preferences among temperate and tropical populations. We found that the temperature dependence of preferred AM rate was significantly steeper in temperate N. triops than in a tropical population of N. triops. Third, we compared temperature dependence of female preference of the N. triops populations to three Neoconocephalus species without call plasticity. Only temperate N. triops had significantly steeper temperature dependence than the other species. This steeper temperature dependence matched female preference to the fast summer call at high temperatures and to the slow winter call at low temperatures in temperate populations. These results support the hypothesis that female preference changed in N. triops in North America to compensate for the plasticity of male calls.     

Large-scale patterns of shell variation in Littorina striata, a planktonic developing periwinkle from Macaronesia (Mollusca: Prosobranchia)

Littorina striata King and Broderip, 1832 is a strictly Macaronesian, intertidal periwinkle with planktonic development. The species displays a high degree of shell variation involving size and sculpture (nodulose vs smooth shells). The present work provides a preliminary account of some aspects of this shell variation on wave-exposed shores over the entire geographical range of the species. Based on morphological patterns observed among other prosobranchs it was predicted that southern specimens of L. littorina should on the average be larger, heavier, more nodulose, and should show more shell repair marks, than northern specimens. These expectations were confirmed for shell size and weight. In contrast, there was no consistent pattern in nodulosity between archipelagos, even though there were differences at much smaller scales. Shell repair marks were more prevalent in northern populations, but this trend was only due to a significant N–S difference among nodulose shells. This is surprising as nodulose shells displayed significantly fewer shell repair marks than smooth shells. These observations were tentatively interpreted as a function of presumed differential N–S patterns of wave action and ambient temperatures. In this context, wave action in Macaronesia seems to increase in the south (contrary to what current theories predict). This atypical situation may confound the interpretation of morphological patterns in L. striata so that firm conclusions cannot be drawn without further experimental work at different spatial scales. Nevertheless, it seems that generalisations about macrogeographic shell morphology patterns, based on interspecific comparisons, are not directly applicable to intraspecific patterns, and may strongly depend on local conditions which make adequate sampling and data treatment very difficult. Received: 10 September 1997 / Accepted: 15 January 1998     

Colonization and early succession on artificial hard substrata by meiofauna

An experiment was undertaken at Farol Island, Brazil, to examine colonization of bare aluminium surfaces by microbes and meiofauna. It was hypothesized that a primary source of meiofaunal colonists was sediment resuspended during upwelling events, two of which occurred during the experiment. Microbial biofilms formed on the experimental substrata within 1 day, and continued to develop throughout the experimental period. Among meiofaunal groups copepods also appeared on the first day, and nematodes on the second. Meiofaunal community structure developed in three main phases: an initial phase of 2 days, characterized by low abundances of copepods; a second phase during the first upwelling period characterized by higher abundances of copepods and also by turbellarians; and a third phase from day 13 onwards characterized by relatively stable abundances of a range of taxa including copepods, cirripedes, nematodes and ostracods. Nematode assemblages also developed in three phases, but with different timings coinciding with upwelling events: an initial phase, from the beginning of the experiment to day 9, characterized by few species and low (or no) abundances; a second phase following the first upwelling characterized by moderate abundances of Chromadorina, Chromadorella, Daptonema and Euchromadora sp. 3; a third phase following the second upwelling period (from day 26 onwards) in which Daptonema disappeared and the assemblage was characterized by moderate to high abundances of Euchromadora (species 1 and 2) and Chromadorella. Although shifts in nematode assemblage structure coincided with upwelling events no evidence was found for sediments being the primary source of colonizers on the aluminium substrata, in contrast to our hypothesis.     

The role of turbidity as a constraint on predator-prey interactions in aquatic environments

Many of the world's most productive aquatic ecosystems usually contain turbid water. Paradoxically, many fish species that live in these habitats are also those that often rely on vision to detect their predators and their prey. For these fish, turbidity will reduce the distance at which predator-prey interactions occur, and there should be a reduction in the opportunity for behavioural modification to control the risk of predation. Under these conditions, we predict that most antipredator behaviour will become ineffective and that predator-prey interactions in turbid water will be primarily characterised by the direct effects of predator consumption of prey, rather than behavioural modification reducing the growth rates of prey. This hypothesis was tested in two laboratory experiments. The first experiment investigated how water turbidity, risk of predation, and their interaction affect habitat choice decisions by fathead minnows (Pimephales promelas). These data demonstrate that fathead minnows reduce their use of dangerous habitats, but that this effect is diminished in turbid water. A second experiment examined mortality patterns when these fish were preyed upon by yellow perch (Perca flavescens) in clear and turbid water. Absolute mortality rates were unaffected by visibility, but patterns of mortality were random in turbid water and skewed towards smaller individuals in clear water. Combined, these results support our hypothesis and suggest that the impact of predation risk will be reduced in turbid aquatic ecosystems. Received: 23 May 1996 / Accepted after revision: 18 November 1996     

Effects of density and simulated recruitment and mortality on sex change in a protandric simultaneous hermaphroditic shrimp, Lysmata wurdemanni

Sex change in many hermaphrodite animals has been suggested to be environmentally determined, especially socially. To investigate whether sex change in the protandric simultaneous hermaphrodite shrimp Lysmata wurdemanni (Gibbes 1850) is socially mediated, two experiments were conducted in the laboratory between September 2002 and April 2004 using laboratory-cultured shrimp that originated from Port Aransas, Texas, USA. The size at sex change (from male to simultaneous hermaphrodite) in this shrimp is variable, with the minimum around 2.4 cm in total length (TL). Large shrimp (2.4–4.5 cm TL) still in the male-phase (MP) have been found in the wild and laboratory environments. This study tested the hypothesis that large MP shrimp delay changing to the euhermaphrodite-phase (EP) due to social control. In the first experiment, ten shrimp were raised in large (110-l) and small (20-l) containers to test the effect of habitat size/density on sex change. The percentage of shrimp changing to EP was significantly higher in the large container (low density) than in the small container after 60 and 120 days. But after 570 days sex ratios were the same, 2 MP:8 EP. In the second experiment, group composition was changed over time to simulate population recruitment and mortality. MP shrimp delayed sex change when EP shrimp were present. However, if group structure is stable, some MP shrimp may not change sex during their lifetime. Under certain demographic conditions, such as when postlarvae (PL) were added to (simulating recruitment) or EP shrimp were removed from (simulating mortality) a group, all old MP (from original PL) shrimp changed to EP. The response of old MP shrimp to simulated recruitment is faster than to simulated mortality. The present study confirms that social control affects the size and timing of sex change in L. wurdemanni. However, some MP shrimp never change sex suggesting that genetics might also play a role in the sex ratios of L. wurdemanni populations.     

An ultraviolet signal generates a conflict between sexual selection and species recognition in a newt

Whether sexual selection and species recognition involve distinct preferences and signals is still debated. Earlier work showed that traits under sexual selection can reduce the efficiency of species recognition but remains uncertain on how frequently such a conflict occurs. We can, however, hypothesise that overlapping distributions of sexual signals may enhance the hybridization risk in many species. We tested this hypothesis in a newt, Lissotriton vulgaris, which hybridises with Lissotriton helveticus. The two species also share an ultraviolet (UV) colour trait, which influences male attractiveness in L. vulgaris, though this trait is probably not functional for sexual communication in L. helveticus. We predicted that the shared trait would affect species recognition when UV radiation is present in the environment. We staged binary choice preference tests under UV+ and UV? conditions. In the UV+ treatment, female preference depended on the values of the shared UV trait and total brightness, regardless of male species identity. Thus, species recognition was enhanced or reversed depending on the difference in the male trait. Females preferred no male type in the UV? treatment, likely explained by our design, which alternated different sensory environments, and the limited prior exposure of subjects to the other species’ morph. We conclude that the presence of this shared trait used in sexual communication contributes to the production of hybrids in syntopic ponds. To our knowledge, this is the first experimental evidence of the influence of an UV sexual signal modulating species recognition.     

Alternative strategies for a father with a small brood: mate, cannibalise or care

Brood cycling, the alternation of periods of courtship and mating activity ("mating phase") with days of full brood care ("parental phase"), and total filial cannibalism, the consumption of all current offspring by a parent, have been interpreted as a response to the conflicting demands faced by parental males during the breeding season. Under the assumption that these two behaviours result from trade-offs among mating, caring and feeding, the strategy adopted by the male should depend on egg number and age. This hypothesis was tested in the scissors-tail sergeant (Abudefduf sexfasciatus), a common tropical damselfish. Observations of undisturbed males and a brood reduction experiment confirmed all theoretical predictions. The shift from the mating to parental phase occurred when an appropriate number of eggs had been obtained. When the brood was reduced on the first day of the parental phase, males often reverted to the mating phase, but never did so if the manipulation occurred on the second or third day. Only broods that were smaller than average were consumed by the parental male. Artificial reduction of the brood increased the incidence of total filial cannibalism, but the probability of cannibalism decreased with brood age. Electronic Publication