Ryen , Philip L. Munday , and Stefan P. Received Dec 7; Accepted Apr Copyright McCormick et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited. This article has been cited by other articles in PMC.
Abstract Variability in the density of groups within a patchy environment lead to differences in interaction rates, growth dynamics and social organization.
In protogynous hermaphrodites there are hypothesised trade-offs among sex-specific growth, reproductive output and mortality.
When differences in density lead to changes to social organization the link between growth and the timing of sex-change is predicted to change.
The present study explores this prediction by comparing the social organisation and sex-specific growth of two populations of a protogynous tropical wrasse, Halichoeres miniatus, which differ in density. At a low density population a strict harem structure was found, where males maintained a tight monopoly of access and spawning rights to females. In contrast, at a high density population a loosely organised system prevailed, where females could move throughout multiple male territories.
Otolith microstructure revealed the species to be annual and deposit an otolith check associated with sex-change. Growth trajectories suggested that individuals that later became males in both populations underwent a growth acceleration at sex-change. Moreover, in the high density population, individuals that later became males were those individuals that had the largest otolith size at hatching and consistently deposited larger increments throughout early larval, juvenile and female life.
This study demonstrates that previous growth history and growth rate changes associated with sex change can be responsible for the sexual dimorphism typically found in sex-changing species, and that the relative importance of these may be socially constrained.
Introduction Sex-allocation theory suggests that the timing of sex change in sequential hermaphrodites is dependent on the relationship among sex-specific growth, reproductive output and mortality  — . When individuals are brought together by a common requirement for limited resources, dominance hierarchies lead to the monopolisation of some resources, differential growth of individuals and the social control of sex-ratios  , . In the marine environment, resource availability can be unpredictable due to environmental patchiness and variability in population density .
The complex life-history of most marine organisms also means that juveniles enter social environments that may be very different from their natal state. This unpredictability has led to plasticity in the way individual fitness is maximised; individuals in different populations may change sex at different sizes and ages due to the different patterns of sex specific growth, fertility and mortality among populations .
In fishes there are strong links among growth, the sex of an individual and the mating system it operates within. In protogynous mating systems where males monopolise matings with many females, male reproductive success is strongly linked to size .
Males tend to be larger than similar aged females within the social group. This size difference can either be due to a history of faster growth in sex changing individuals  ,  , or a product of a growth spurt that occurs coincident with sexual transition  , . While it is commonplace for males to be larger than females in a protogynous mating system  ,  , the developmental aspects of sexual size dimorphism SSD have seldom been explored see for exceptions  ,  ,  , .
Indeed, this is the case not only for fishes, but for vertebrates in general . Recently, the microstructural increments within otoliths earstones have been used to clarify the link between sex-change and growth history.
Once the deposition of increments has been appropriately validated  , the width of increments can be used as a proxy for somatic growth. Abrupt changes in increment structure, or checks, associated with key life history transitions, such as settlement  and sex-change  ,  , allow a growth history to be interpreted with respect to key life events.
This powerful tool gives researchers the opportunity to explore the link between growth history, sex-change and their mating system in a detail not previously possible.
The mating system adopted can depend on the density of individuals that are potential members of one or other sex. Monopolisation of resources by a small number of males may be difficult at high densities since interactions may be too frequent to allow a stable social group to form  — . In contrast, at low densities, males may be able to visit females sufficiently often to reinforce a social hierarchy, suppress growth of females, and monopolise environmental resources and females  — .
Hence, it has been suggested that social system should strongly influence the temporal and ontogenetic relationships between sex-specific growth, sex change and SSD, and specifically, the way in which males achieve relatively larger body size . We predict that growth of subordinate females should be reduced with the strength of social control by males, such that individuals rely more on accelerated growth during sex change to achieve SSD.
Furthermore, at low densities the previous growth history of an individual should be less important in determining which females change sex, and its timing, because transition will be triggered by the relaxation of social control through the loss of a dominant male e.
The present study compares the social organisation and sex-specific growth of two populations of a protogynous tropical wrasse, Halichoeres miniatus, which differ in density. The social organisation of the populations is first described by examining the space use and interaction regime of individuals within the groups. Detailed examination of growth allowed the mechanisms underlying the sexual size dimorphism found in the two populations to be characterised.
The presence of otolith checks associated with sex-change in this species  enabled an investigation of sex-specific growth in a detail not previously possible. Materials and Methods Study species and habitats The small coral-reef wrasse H. Males of this short-lived protogynous hermaphrodite are larger than females and display brightly coloured markings. Otolith increment formation has been validated as daily, and females have been experimentally shown to alter otolith accretion during sex change to form a check, which is characterised by a change in optical density and increment width .
Similar sex-change associated checks have been observed in the sandperch, Parapercis cylindrica and P. At Lizard Island the study population inhabited isolated patches of rubble and algae in shallow water separated by open sand flats.
Halichoeres miniatus was common but not densely populated on the rubble patches, with an average density of 0. In contrast, at Orpheus Island, the study location was part of a continuous macroalgal zone on shallow reef flat and H. Both locations were situated at the leeward side of the fringing reefs, where H. Data presented here show that at the Lizard Island location males defended non-overlapping territories containing females, whilst at Orpheus Island males were resident in specific areas and had areas of regular use, but these were seldom defended from neighbouring or transitory males.
Demography and social organisation The size and age distributions of H. Fish were collected from five sites located haphazardly around the leeward side of each island 49 individuals from Lizard Island and 69 from Orpheus Island. Age was determined by counting the increments in the transverse sections through the nucleus of one sagittal otolith from each fish, prepared using the protocol of Wilson and McCormick . Sex for each individual was initially determined by the colour patterns terminal or initial phase and then by macroscopic examination of their gonads under a dissection microscope.
Testes were identified by their smooth surface and cream colouration while ovaries were identified by their yellow colouration and a rough surface texture, indicating the presence of developed eggs . To determine the social organisation of H. To facilitate recording the location and movement of individuals, areas were mapped with the aid of a reference grid of nylon string at both study locations 2.
To facilitate individual identification all the males and the largest females were tattooed subcutaneously near the dorsal fin with a fluorescent elastomer Northwest Marine Technologies using a 27 gauge hypodermic needle while restrained by the plastic bag. During the tagging process fish were partially sedated due to the anaesthetic clove oil used in capture. This method of tagging minimised stress and scale damage through handling and could be done underwater to minimise processing time .
Tagging left a 0. Upon release fish quickly returned to their areas of residence and males resumed territorial behaviour. Behavioural observations began the day after tagging and were made over three days for five hours per day. Each observation period followed one individual for 15 minutes and all interactions were recorded during that time. A scuba diver followed individuals at a distance of 2—3 m and the proximity of the diver did not appear to influence fish behaviour.
At the end of the study, tagged fish were recollected and euthanised using the previously mentioned protocol to allow age determination from increments in otolith cross sections.
The location and movement of tagged individuals was plotted on a scale map of the study areas. Home range sizes and the degree of overlap of home ranges for males and the largest females were measured and compared between the two locations using a one-way analysis of variance ANOVA. Residual analysis was used to test whether data conformed to the assumptions of homogeneity of variance and normality.
Ontogeny, growth and SSD Microstructural increments on the sagittal otoliths were used to describe the growth history of the tagged fish.
At the end of the observation period, fish were recaptured as above , and euthanised by cold shock in a slurry of seawater and crushed ice to minimise stress. Sagittal otoliths were processed to produce a transverse section of the distal-rostral plane, following the methods of Wilson and McCormick .
Multiple regression was used to confirm an age-independent, predictive relationship between otolith growth and somatic growth i. The body size- and otolith size- maximum otolith radius, MOR distributions of females and males were compared within each population using t-tests, and sex-specific otolith increment width profiles were used to infer the timing and shape of growth divergence. No attempt was made to compare the magnitude of otolith growth between populations since different relationships exist between somatic growth and otolith growth between populations.
Increment width profiles were compared between sexes male, female , locations Orpheus and Lizard Islands and between initial larval growth and juvenile growth using a three-factor repeated measures MANOVA. Two ten-day periods were chosen to typify early larval and juvenile growth day 1—10 and day — respectively.
Pillai's trace was used as the test statistic for within subject i. Significant terms were interpreted from increment graphs. To explore whether there was an increase in increment width as a proxy for somatic growth associated with sex change, the otoliths of males were re-examined and increment widths were re-plotted so that they were centred on the check in the otolith associated with sex-change .
This makes it easier to distinguish changes in otolith growth that occur at the time of sex-change and avoids the problem of masking through the averaging of increment widths of fish that undergo the transition at a variable age.
We explored variation in both the age at sex change and the size at sex change between populations using Kolmogorov-Smirnov two-sample K—S tests. The age at sex change for each individual was determined by counting the number of increments from the otolith nucleus to the sex change-associated check mark.
The size at sex change for each individual was back-calculated using the biological intercept method . The size at hatching was estimated as the mean larval size at hatching of a congeneric, Halichoeres poecilopterus . For both populations a linear model was found to best describe the otolith radius versus body size relationship.
Results Population demography A total of 69 fish were collected from Orpheus Island 51 female and 18 male , and 49 fish from Lizard Island 23 female and 26 male.
The size and age and size-at-age distributions for both locations were characteristic of a protogynous species Fig. Examination of the gonads also revealed no initial phase males, suggesting that H. Overall, there was more overlap in the frequency distributions of female and male age than size.
Males and females at Orpheus Island showed a greater overlap in size and age classes than at Lizard Island indicating greater variability in the size and age at sex change in the Orpheus Island population.
The average male size at Orpheus Island was Average female sizes were Several females were larger than the smallest males at Orpheus Island but males were always larger than females at Lizard Island. At both locations no individuals were over days indicating that, at these locations, H.