Indiscriminate males: Betta splendens males do not prefer larger females
The basis for mate choice lies in parental investment theory (Trivers 1972), which assumes that the sex that invests less in parental care will compete for access to the sex that invests more. The existence of anisogamy means that females invest heavily in egg production, and are thus limited by resources, while males produce excess sperm and are thus limited by access to females. This energy allocation pattern results in the male-male competition and female choice that we so often see in nature. This pattern may be altered by male parental investment, however, a pattern that is seen in a large number of fish species.
Mate choice arises from the assumption that the ability to assess mate quality will result in higher reproductive success. The sex that invests more in parental care is thus expected to be selected for the ability to assess quality in the opposite sex (Costello et al. 2009). In species that display male parental investment, mating with a lower quality female and investing in her offspring could result in a "missed opportunity" to mate with a higher quality female (Costello et al. 2009). In these species reversal of the sex roles can be seen, with male mate choice and female competition occurring (Mattle and Wilson 2009). For this switch to occur, however, there must be a cost to mating with a low quality female.
Betta splendens is a freshwater fish well known for its spectacularly coloured males, and it is no surprise that mate choice studies in this species have concentrated on female choice. However, males of this species make a significant parental investment by building and maintaining bubble nests for their young, which are said to protect eggs from anoxia during development and are strong predictors of offspring survival in other fish species (Clotfelter et al. 2006). Males are territorial and defend these nests from other males (Bronstein 1982); they also chase away the female after spawning to prevent egg cannibalism, leaving the male to care for the eggs and fry themselves (Jaroensutasinee and Jaroensutasinee 2001). Given the paternal care seen in B. splendens, we propose to examine male mate choice in this species.
It is suggested that the cost of mating with low quality females to male B. splendens is that of nest-building. As the time and energy spent in the building, maintenance and defence of these nests is considerable, it is not unreasonable to assume that this behaviour may be costly to males. Paternal care in other species of fish has been shown to be costly in terms of reductions in feeding and body condition, and increases in predation risk (Kvarnemo et al. 1998). Furthermore, investigations into nest-building behaviour in fish have demonstrated a relationship between male condition and nest-building. Male reproductive behaviour in sand gobies (Pomatoschistus minutus) is similar to that of Betta splendens: males build and defend nests, and display to females during courting; males fed ad libitum were found to have bigger and better quality nests than food deprived males (Olsson et al. 2009). This relationship has been also been shown in other species: in the common goby (Pomatoschistus microps) starved males were found to have poorly constructed nests (Kvarnemo et al. 1998), and Rushbrook et al. (2008) found nest-building to be linked to male condition in three-spined sticklebacks (Gasterosteus aculeatus). These studies support the hypothesis that nest-building may be costly to males.
This study aims to look for the existence of male mate choice in B. splendens by looking at nest-building behaviour in males. We propose to alter female quality by altering food availability. Food availability is strongly linked to growth and size in most animals, and as female body size is positively associated with fecundity in fish (Clotfelter et al. 2006), we will use female treatment and size as a proxy for fecundity. Male size will be used as a proxy for male condition. If nest building is costly we predict males exhibiting mate choice will invest more in nest-building in the presence of high-quality females, and that this effect will be mediated by male condition. As nest building may be a sporadic activity, we propose to also examine male display behaviour in response to female quality. Male courtship behaviour in B. splendens can include zigzagging, fin spreading, and operculum gill flaring (Dzieweczynski 2009), all of which might be assumed to be metabolically expensive. Studies into courtship behaviour have long suggested costs to investment in male display (e,g. Moller and de Lope 1994; Kotiaho 2000)Kotiaho, J. S. 2000. Testing the assumptions of conditional handicap theory: costs and condition dependence of a sexually selected trait. Behav. Ecol. Sociobiol. 48:188-194. CrossRef See All Re, and investment in courtship display might thus be expected to mirror male parental investment.
Trials were run in November 2009 using 20 male and ten female adult Bettas. Prior to trials the mass of each fish was determined by weighing each fish in a container with a known weight of water; females were weighed a second time once trials were complete. To control for possible colour effects only blue males were used in this experiment; females were of varying colours and so were paired according to colour and approximate size. Each individual within a female pair was randomly assigned to either the high or low food treatment. Female food treatments began ten days before trials were conducted and continued for the duration of the experiment (approximately three weeks). Fish were fed daily Monday to Friday; males were fed three pellets of Wardleys Essentials Tropical Fish Food, while females were fed either one (low) or four (high) pellets daily. Trials were conducted in rectangular glass tanks; each fish was in 2340 cm2 (15 x 13 x12 cm) of dechlorinated water for the duration of a trial. In between trials all Bettas were housed individually in large jars (vol?) filled with dechlorinated water at room temperature. All fish (except trial pairs) were kept visually isolated from one another during and between trials.
A total of four trials were conducted in two blocks; each block consisted of ten new males encountering the low- and high-condition female of a pair in a random order. Each male was assigned a random female at the beginning of the first trial of a block; the matching female of the pair was assigned to that same male for the second trial of the block. At the beginning of each trial a male and female were placed in tank separated by an opaque partition. Fish were given six hours to acclimatize to the new environment; following this the opaque partition was removed, leaving a transparent partition behind. This partition was lifted temporarily to allow chemical cues to pass into each half of the tank. Male operculum flaring duration and frequency was observed for a ten minute period. Operculum gill flaring was used as it is easy to identify and is a well-recognized courtship behaviour (Dzieweczynski 2009). Only full operculum flaring events (gill flared at a 90 angle to the head, from Clotfelter et al. 2006) were recorded. Fish were left in visual contact for 48 hours. On the second and third morning following introduction tanks were checked for the presence of a bubble nest; bubble nest area was measured using a transparent 1x1mm grid. Following nest measurement on the third morning males and females were returned to their home tanks and given at least a day between trials to recover. Observers were alternated for fish pairs, and the order (HL or LH) in which a male encountered the females was randomized.
As no significant difference was found between the weights (unpaired two-sided t test; t = -1.00, df = 18, P = 0.330) or display behaviours of Block 1 and Block 2 males (Table 1), the data were pooled across the two blocks for analysis. Male behaviours were averaged across each individual female and unpaired two sided t tests were then conducted to test the effect of female food treatment on the time to first display, percent time spent displaying, total number and mean duration of operculum displays performed by males. Nests generally did not appear until the third day and only those data are used. Only one male built a nest in Block 1; given this only nests from Block 2 were included in the analysis. As no effect of female treatment on male behaviour was found (Figures ), low and high females were pooled for regressions of male and female weights against male behaviours. Female receptivity was noted sporadically and could not be included in the analysis. All statistical analyses were run in Microsoft Excel (MINITAB for regressions) at a statistical significance level of P = 0.05.
There appeared to be no effect of treatment on female weight as no significant difference was found between the two groups in their final weights (unpaired two-sided t test; t = -0.65, df = 8, P = 0.533). Female treatment was found to have no significant effect on the total number of displays, mean duration of display, percent time spent displaying or nest area (Figures ). However a nearly significant difference was found for time to first display, with males showing a longer time to first display for high food females (Figure ). No significant correlation was found between male or female weight and any measure of male opercular display (Figures ) or nest area (Figures ).
Previous studies into mate choice in B. splendens have all concentrated on female choice. However, given the considerable male parental investment seen in this species, we proposed to look for evidence of male mate choice by looking at the effect of female condition on operculaum display behaviour and nest size. To this end twenty males were exposed to one of five pairs of high and low food treatment females. Neither female weight nor female food treatment were found to have an effect on any measures of display behaviour or on nest area (Figures ).
There could be several reasons no effect of female treatment was found for display behaviour or nest area in this experiment. Firstly we were limited in the number of females available, and male response was effectively reduced to five data points per treatment. As each male only encountered one female pair, a paired t test could have been used for individual male responses to high and low food females. However as only five female pairs were used for 20 males, a paired analysis of all 20 males for each treatment would have constituted pseudoreplication. This small sample size could have made it very hard to find an effect of treatment were one present.
Another key issue is the lack of an effect of food treatment on female size.If males are exhibiting mate choice according to female size one might expect a relationship between female size and male display behaviour and/or nest area regardless of treatment: this effect was not found (Figure _). This is in contrast to results found in other animal species showing male parental care, where males have been found to prefer larger females. In the pot-bellied seahorse H. Abdominalis, males show parental care by carrying eggs in a brood pouch. In this species a strong male preference for larger females has been demonstrated (Mattle and Wilson 2009); this same result has been shown in another species with paternal care, the pipefish Syngnathus typhle (Berglund et al. 2006).These results suggest that either female size is not an indicator of fecundity or desirability in this species, or that male mate choice is not occurring. In a study conducted by Clotfelter et al. (2006), female size was not found to be an indicator of fecundity in B. splendens, and males did not show a preference for larger females. In light of these results it is still possible that male choice is occurring if males are choosing according to some other marker of female health or fecundity. In this case one would expect an effect of female treatment on male display behaviour or nest area, however the lack of an effect of treatment on female size means that treatment may also not have had an effect on any other female characteristics.
An important factor that was not accounted for when looking at male response was female behaviour. Clotfelter et al. (2006) found males displayed more to females that displayed more. As male display behaviour has found to be condition-dependent in _ ( ), it is equally possible that female display behaviour is also condition-dependent. Thus males may still be exhibiting mate choice according to female quality when choosing females that display more. If female display levels were unrelated to food treatment (as is likely given the lack of an effect of treatment on size) this could have created noise around our results. Future studies might look at the relationship of female display behaviour to fecundity and health. Clotfelter did... In addition, this study was limited by observing only two male response types. Although opercular gill flaring is the main component of male courtship display in Bettas (Clotfelter 2006), other male display behaviours such as zig sagging and fin flaring could also have been looked at. In this study it was not possible to record more than one behaviour in detail, however future studies might look at several male and female behaviours by use of a video camera.
In this study male weight was used as a proxy for male condition, and was not found to correlate with either display behaviour or nest area (Figures ). This suggests there may not be a cost to display or nest building, however male size is does not necessarily indicate condition as it is also genetically based. Given the small sample size it was not possible to alter male condition, making it hard to prove the assumption that nest building and display behaviours are costly to males. In addition only 9 nests were built out of 40 opportunities to do so, and this small sample size could result in a non-robust correlation.
It is interesting to note that there was a nearly significant effect of female treatment on time to first operculum display (Figure ). This result is in a direction opposite to what might be expected if male mate choice were occurring: males were found to have a longer time to first display for high food females. This may be an artefact however, as males were sometimes facing the wrong direction at the beginning of a trial, and would not see the female for a while. This could be controlled for in the future by only beginning trials once the male observed the female.
Our results suggest that male mate choice is not occurring in response to female size, but do not conclusively disprove the possibility of male mate choice occurring in response to other female characteristics. This study could be improved by the use of more females, and by manipulating male and female condition by starting food manipulations during the juvenile stage. Future studies might allow actual mating to occur, and could look at additional male and female courtship and mating behaviours. Paternal investment in offspring of different females could be observed by looking at the speed in building and maintenance of bubble nests. A standard choice assay test could be conducted for display behaviours or other indicators of preference; this was not possible in this study, as we would not have been able to determine which female a nest was built for.