The way organisms allocate their resources to growth and reproduction are key attributes differentiating life histories. Many organisms spawn multiple times in a breeding season, but few studies have investigated the impact of serial spawning on reproductive allocation. This study investigated whether resource allocation was influenced by parental characteristics and prior spawning history in a serial spawning tropical damselfish (Pomacentrus amboinensis). The offspring attributes of isolated parents of known characteristics were monitored over a 6-week breeding period in the field. Smaller females produced larvae of longer length and larger energy reserves at hatching. This finding is contrary to several other studies that found larger females produce offspring of greater quality. We found that resource allocation in the form of reproductive output was not influenced by the number of spawning events within the breeding season, with larger females producing the greatest number of offspring. Larval characteristics changed as spawning progressed. There was a general decline in length of larvae produced, with an increase in the size of the larval yolk-sac, for all females regardless of size as spawning progressed. This trend was accentuated by the smallest females. This change in larval characteristics may reflect a parental ability to forecast unfavourable conditions as the season progresses or a mechanism to ensure that some will survive no matter what conditions they encounter. This study highlights the importance of accounting for temporal changes in reproductive allocation in studies of reproductive trade-offs and investigations into the importance of parental effects.
Citation: Maddams JC, McCormick MI (2012) Not All Offspring Are Created Equal: Variation in Larval Characteristics in a Serially Spawning Damselfish. PLoS ONE 7(11): e48525. doi:10.1371/journal.pone.0048525
Editor: Howard Browman, Institute of Marine Research, Norway
Received: September 20, 2011; Accepted: October 1, 2012; Published: November 14, 2012
Copyright: © 2012 Maddams, McCormick. 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 credited.
Funding: This study was funded through the Australian Research Council Centre of Excellence for Coral Reef Studies (http://www.coralcoe.org.au/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
A central question in life history theory is how individuals allocate resources between current reproduction and future survival , . One of the basic principles of life-history theory is that reproduction carries a cost for subsequent survival and therefore for individual fitness , . This cost may lead to a trade-off between reproductive effort at any given time, and the anticipation of future reproduction , . Given that reproduction is a costly process where resources have been temporarily sacrificed or traded-off from other life functions , , energy invested into reproduction should have a fitness benefit.
Reproduction in the simplest sense can be measured by the number of offspring produced, and the characteristics of those resulting offspring. Both of these measures of reproductive allocation have been reported to vary seasonally in a diverse group of organisms (e.g. insects, fish, amphibians and reptiles). In general, variations in offspring number or characteristics have been attributed to genetic or non-genetic parental effects –. Of these effects, maternal influences are regarded as of particular importance due to the females' large investment of energy into eggs . Maternal effects may influence the developmental schedules , , growth – and performance ,  of offspring, and inevitably the numbers that survive to contribute to the next generation. Maternal investment to reproduction may change in response to her size , , , age – or body condition –, and may also be influenced by the environmental conditions she experiences –. Ultimately, her reproductive fitness results from a balance between the cost to herself and the gain to her offspring. While it is known how resource allocations to offspring change with age, size and social status of parents for a diverse array of taxa (e.g. butterflies, sheep mites, fishes) , , little is known of whether or how allocations change among successive reproductive events within a breeding season.
Early stages (i.e. eggs and larvae) of teleostean fishes are often subject to high mortality –, such that very large investments in reproductive material are necessary to ensure the persistence of populations . Producing very high numbers of offspring is a common reproductive strategy among a variety of reef fishes with pelagic eggs or larvae , as it is with other organisms with complex life cycles (e.g. many insects and amphibians). Most coral reef fishes spawn more than once over a breeding season. This is a strategy common to other taxa, for example 64 species of Neo-tropical frogs  and at least 10 lizard families produce multiple clutches in a breeding season . Spawning in batches over a period of several weeks is considered to be one factor that may actually increase the probability of larval survival by enhancing the chance to encounter optimal feeding conditions , . Currently, our understanding of energy allocation to reproduction over successively produced clutches is limited. It has been stated that the difficulty associated with accurately recording fecundity estimates for sequential spawning species has excluded this reproductive method from being appropriately modelled in current life history theory .
Evidence suggests that parents have the ability to adjust their resource allocations within a breeding season –. It has been suggested that resource competition between offspring can cause selective variation in offspring characteristics. For example, if resource competition between offspring is high, mothers may increase their per offspring investment as the season progresses to ensure that offspring spawned later in the season will not be at a competitive disadvantage , . This indicates that mothers have the ability to recognise prior energy investment to offspring and to alter their offspring accordingly. Alternatively, it has been suggested that variation is a form of risk spreading ensuring that no matter what conditions the offspring encounter some will survive . Whether or not the parents actually make a tactical decision to adjust resource allocation or if the response is purely a product of environmental conditions (and the fish's physiological reaction to these), such as an increase in food supply ,  is difficult to determine.
Without an understanding of how an individual's history of prior spawning, either within or between breeding seasons, has influence on subsequent spawning events it is impossible to realistically parameterise an essential component of life history theory. In an energy limited environment , we may predict that investment into egg production in serially spawning organisms may diminish with each successive spawning bout within a breeding episode.
The objective of the present study was to examine the influence of parental characteristics on reproductive output and larval characteristics of a serially spawning marine teleost, the benthic spawning damselfish, Pomacentrus amboinensis (Pomacentridae, Bleeker 1868). As such, this is the first field-based study to assess how the energy allocated to reproduction changes throughout a breeding season in a serial spawning marine teleost.
Pomacentrus amboinensis was selected as an example of a serially spawning marine teleost with a complex life cycle for a number of reasons: (1) it has a small home range, is site attached and can be transplanted onto patch reefs made of coral and rubble; (2) the female lays a monolayer of eggs on a nesting surface guarded by the male prior to hatching, so that the total number eggs and their development can be readily quantified; (3) the fertilized eggs hatch after about 5d (at 28°C) at ~20 min after sunset allowing for easy access to newly hatched larvae; (4) the species is widespread throughout the Indo-Pacific and has life history features that are commonly displayed by many other fishes. Over the breeding season this species can spawn several times. The present study encompassed the majority of the spawning season at the study site. Field observations of spawning before and after the six week study found no evidence of spawning prior to the start of the project, and very little after its completion. This species is protogynous hermaphrodite living in groups typically containing one male, which guards a benthic nest site, and one to seven females . Tagging studies using visible tags and passive integrated transponder tags have found that females do not travel more the 12m along a continuous reef to visit male nest sites , . To be able to accurately determine the parentage of the eggs it was necessary to isolate breeding pairs by transplanting them to patch reefs established on a sand-flat. Patch reefs are a part of the habitat they normally occupy . Previous studies that have individually tagged pairs have shown that the fish do not move the 20 m distance between patches or the reef edge , . A study on the demography of the species at the study reef found that females become reproductively mature at ~2 year old, and reach a maximum age of 6 years on the fringing reef around Lizard Island (McCormick unpublished data). Once the larvae hatch from the benthic nests they spend between 18–23 days as pelagic larvae before settling back to a reef habitat .
The study was conducted on Lizard Island (14°40′449S, 145°27′400E) in the northern Great Barrier Reef, Australia, over a six week period (October to December 2006). A matrix of 40 patch reefs was constructed on sand at a depth of 2–3 m within the blue lagoon. Each patch reef was composed of live and dead coral (Pocilliopora damicornis) with a basal area of approximately 2 m2 and extended approximately 40 cm above the substrata. Adjacent patch reefs were separated by 20 m of sand and were a minimum of 20 m away from contiguous reef. A 250 mm length of PVC pipe (170 mm diameter) was positioned within each patch reef to provide an artificial nesting surface. The artificial nests provided a uniform concave nesting surface of similar dimensions and defensibility as natural nests observed in the field and were readily adopted by P. amboinensis.
An initial assessment of reproductive activity was conducted over a 2-day period on the contiguous reef, to establish if the breeding season had started. No reproductive activity was recorded, and no active nests were found. Behavioural interactions between males and females were recorded giving some indication that reproduction was imminent. Forty females of differing size were captured from the contiguous main reef using hand nets and a clove oil/ethanol solution. The male most closely associated with the female was also captured – forming the breeding pairs. In situ (underwater) the fish were sexed by visual inspection of their genital papilla and standard length (SL) was measured to the nearest millimetre using callipers through a plastic bag to prevent damage and reduce stress. Pairs were randomly chosen under the constraint that their overall size distribution should span the full range of female sizes present in the local population. The standard length of females in the population at this location ranges from 35 to 65 cm (McCormick unpublished data). To allow individual identification, all fish were tagged with a non-toxic subcutaneous fluorescent elastomer injected just beneath the epidermis . Each tagged pair was haphazardly allocated and released onto an isolated experimental patch reef. Spawning activity of the fish was monitored daily within two hours of dawn, the presence of the breeding pairs (individually identifiable), and the presence of and developmental stage of eggs were recorded. Each clutch of eggs was digitally photographed in situ the morning of spawning. A transparent 1cm2 grid was placed over the clutch to facilitate measurement calibration. The area occupied by the egg mass and the density of eggs was estimated from the calibrated digital photographs using the image analysis software Optimas 6.5 (Media Cybernetics, 1999). The density was quantified by counting the number of eggs within five 10 mm2 areas. The number of eggs per clutch was a product of the total area occupied by the egg mass and the density of eggs. Due to the resolution of the in situ digital photographs, and the poor relationship between egg size and larval characteristics in this species , , this study did not measure egg size.
On the 5th day following egg deposition at about 2–3 h before hatching of embryos, the artificial nests were collected from the patch reefs and moved to an aerated flow-through seawater system. Each nest was placed in individual aquarium maintained with aerated sea water with a unidirectional flow over the clutches. Water temperature within the aquaria was equivalent to that in the field (24–26°C). All collected nests were replaced with another nesting surface at this time.
At hatching a sample of ~150–200 larvae was collected and placed in solution of 2.5% glutaraldehyde in seawater . The larvae were fixed for two hours at room temperature (26°C), then rinsed and transferred to fresh filtered seawater and stored at 4°C. Thirty larvae from each of the 90 clutches (n = 2700 larvae) were individually photographed using a digital camera attached to a stereomicroscope. A scale bar was placed adjacent to the larvae to facilitate calibration of size measurements. The image analysis program Optimas 6.5 (Media Cybernetics, 1999) was used to measure three larvae attributes: total larval length (mm), yolk sac area (mm2), oil globule area (mm2). Each attribute was measured blind on three occasions to identify any discrepancies.
At the conclusion of the experimental period all successful breeding pairs (n = 25) were collected from the reefs using hand nets and euthanised by cold shock. All fish were measured (SL ±0.01 mm) and weighed (± 0.001 g). The sagittal otoliths and the gonads of the maternal fish were removed. The gonads were blotted dry and weighed (±0.001 g). The sagittal otoliths were prepared following Fowler  to determine the age of the maternal fish. Five replicate counts on the number of opaque bands were made for each fish, and any discrepancies between the measurements were identified. A linear regression model was used to determine if size and age could be considered independent variables for the female population at the study location (Table S12). The relationship between female size and age for this population is weak (Figure S1, R2 = 0.08), allowing us to test size and age as independent variables.
To assess the effect of both female attributes and sequential spawning on reproductive output and larval characteristics at hatching, only data collected from breeding pairs that commenced spawning within 5 days of each other, spawned at least six times over the duration of the experiment (mean = 6.4 clutches, range = 2–10 clutches) and had similar inter-brood times were used in the analysis (n = 15). This choice was made so that a valid comparison could be carried out between parents that had undertaken the same number of reproductive events over the experimental period. Examining the influence of spawning frequency (i.e. the individually variable number of clutches spawned) within the spawning period on the link between parental attributes and offspring was beyond the scope of the present study.
Parental influence on reproductive output.
A number of measures were calculated as indices of body condition. The residuals of a standard length by wet weight plot were used as an index of relative condition  for both the female and male adults. Since adults are from the same population this is generally believed to be an effective way of standardizing weight for variable body size. The further individuals differ from the general size/weight relationship, the greater or lower their body condition (i.e. they either have higher or lower than average standardized weight). Female gonad somatic index (GSI) was calculated as the ratio of ovary weight (g) to body wet weight (g), multiplied by 100. GSI gives an indication of the reproductive state at the time of sampling, and how spawning history may have impacted energy stores. All statistical analyses were conducted using Statistica version 9 (Stat Soft, Tulsa, USA).
To investigate the influence of parental characteristics on the total number of embryos produced over the six reproductive events (dependent variable) a best-subset multiple regression model using a Mallows CP selection criteria was preformed. The relationship between female size and maturity for Pomacentrus amboinensis at this location is weak (Figure S1), therefore we can explore the influence of both size and age on production and larval characteristics as independent variables. The parental characteristics included in the analysis were maternal size (SL), maternal age, maternal GSI, maternal condition, paternal size (SL), and paternal condition. The assumptions of multiple regression, linearity and normality, were investigated by examining bivariate scatterplot of the variables of interest and residual analysis.
The influence of maternal size, age and condition on the total number of embryos produced from each successive spawn was explored using a repeated measures ANOVA. Maternal fish were grouped into three size classes (small, 47–52 mm SL, n = 5; medium, 55–60 mm SL, n = 5; and large, 63–68 mm SL, n = 5), three levels of condition (lower than average, average and better than average) and three age classes (2–3 years old n = 5, 4–5 years old n = 5, ≥6 years old n = 5) (Bonferroni adjustment was applied for repeated tests, corrected alpha = 0.016). The assumptions of ANOVA (homogeneity of variance and normality) were examined with residual analysis. The assumption of compound symmetry was tested using Mauchley's test of sphericity .
A Friedman ANOVA with Kendall Coefficient of Concordance was used to test if there was a difference in inter-brood time among females (spawning frequency).
To explore the relationship between larval attributes (dependent variable: larval length, and energy reserves (yolk sac area and oil globule area)) and parental characteristics (independent variable) and at a particular point in the spawning sequence for P. amboinensis three best subset multiple regressions were conducted. To be conservative, Bonferroni adjustment was applied for repeated tests since the analysis was conducted on three potentially co-varying dependent variables (larval length and energy reserves (2)) corrected alpha = 0.016). The parental characteristics included in the regression models were maternal size (SL), maternal age, maternal GSI, maternal condition, paternal size (SL), and paternal condition. To standardise for prior spawning history and to reduce the chance of over estimating variability due to the variable nature of the first batch of eggs  the analysis was conducted on the second clutch of all spawning pairs. Multiple regression models were also conducted on the fourth and sixth clutch of all spawning pairs to test if the same adult characteristics were influential throughout the spawning season. The selection criterion was based on the model with the smallest Mallows Cp and the largest adjusted R2 value. The assumptions of multiple regression, linearity and normality, were investigated by examining bivariate scatterplot of the variables of interest and residual analysis.
To investigate the influence of maternal size on the three larval characteristics (i.e. larval length, yolk sac area, and oil globule area) across six successive spawning events a series of repeated measures ANOVA were conducted (Bonferroni adjusted for co-linearity between larval traits alpha = 0.016). Maternal fish were again grouped in three size classes. The assumptions of ANOVA were examined with Levene's tests, and residual analysis. The assumption of compound symmetry was tested using Mauchley's test of sphericity .
Of the six parental characteristics examined, only maternal size was found to have an effect on the total number of eggs produced over the six spawning events (Table S1. adjusted R2 = 0.729, p = 0.0024). Total egg number increased with maternal size (Fig. 1). Large females (63–68mm SL) on average produced 51% more embryos than smaller females (47–52 mm SL) over the six spawning events.
Figure 1. Relationship between female standard length (mm) and total reproductive output (from six clutches) over the duration of the experiment.
Filled triangle represents the small females (47–52 mm; n = 5), filled square represents the medium females (55–60 mm; n = 5), and filled circle represents the large females (63–68 mm; n = 5).doi:10.1371/journal.pone.0048525.g001
Examining the relationship between maternal characteristics and egg production for each sequential spawning event revealed an influence of maternal size but no influence of clutch number in determining the number of eggs produced (Table S2; F10,60 = 2.058, p = 0.006). Larger females produced significantly greater number of eggs than the smaller females across the six sequential spawning events (Fig. 2). Within each maternal size class there was no trend in the number of eggs produced across the six spawning events. There was no effect of maternal age or body condition on the number of eggs produced across the sequential spawning events (Table S2).
Figure 2. Comparison of the mean (± SE) number eggs produced per clutch between females of different size for six successive clutches.
Female size had a significant influence on the number of offspring produced. C1-C6 represents six successive clutches from the same females. Dark grey bar represents the small sized females (47–52 mm; n = 5), light grey bar represents the medium sized females (55–60 mm; n = 5), and white bar represents the large sized females (63–68 mm; n = 5).doi:10.1371/journal.pone.0048525.g002
A Friedman ANOVA with Kendall Coefficient of Concordance was used to test if there was a difference in inter-brood time among the 15 females. No difference of inter-brood time was found (X2(5, 14) = 16.0393, p = 0.3109).
The influence of parental characteristics (maternal size, condition, GSI, and age, and paternal size and condition) on larval attributes was examined using multiple regression models. As female size was the only parental characteristic found to have influence on larval length (Table S3, adjusted R2 = 0.633, p = 0.0014) and larval energy reserves (Table S4, adjusted R2 = 0.788, p = 0.0002) throughout the spawning season (Table S3, S4, S5, S6, S7, and S8), female size is the characteristic that we have explored further in relation to larval characteristics over sequential spawns.
The total length of larvae was influenced by an interaction between maternal size and clutch number (Table S9, F10, 435 = 32.061, p = 0.0001). In the first clutch the smaller females produced larvae that were longer (3.54±0.02 mm, n = 5) than larvae from medium (3.42±0.02 mm) or large females (3.35±0.03 mm). This pattern was maintained across five of the six spawning events (Fig. 3). In general, there was a decline in the size of the larvae produced, irrespective of female size, as the breeding season progressed (Fig. 3). The greatest decline in larval length is seen in the small and medium sized females, declines of 4.7% and 3.6% respectively. This trend suggests that there may be a negative influence of successive spawning on larval size at hatching for the two smaller size class females.
Figure 3. Mean (± SE) larval total length (mm) at hatching from different size females through six successive spawns.
Both female size and clutch number had a significant influence on larval length (RM ANOVA; F10, 435 = 32.061, p = 0.0001). Filled triangle represent the small sized females (47–52 mm n = 5), filled square represents the medium sized female (55–60 mm n = 5), and filled circle represents the large sized females (63–68mm n = 5).doi:10.1371/journal.pone.0048525.g003
Larval energy reserves in the form of oil globule size were significantly influenced by an interaction between female size and clutch number (Table S10. F10, 435 = 14.710, p = 0.0001). Fluctuations within a general trend for increasing oil globule size over the successive spawns lead to an interaction between female size and clutch (Table S10). There was a general increase in oil globule area over the successive spawns for all larvae irrespective of the maternal size (Fig. 4); however, the degree of increase differed. Oil globule area of larvae from the smallest and largest females increased by approximately 35% from clutches one to six, while the oil globule area of larvae from the medium females increased by only 13% from clutches one to six (Fig. 4.). Similarly, larval energy reserves in the form of yolk sac area were also significantly influenced by an interaction between female size and clutch number (Table S11. F10, 435 = 17.610, p = 0.0001). The size of the yolk sac generally increased as the spawning season progressed for larvae produced by the small females.). Greater variability in yolk sac area was seen for both the medium and large females between the spawning events (Fig. 5). Larvae produced by the smallest females had larger yolk sac area than medium or large females across all spawning events (Fig. 5). In the first spawning event small females produced larvae with a yolk sac area 29% larger than that of larvae from the large females and 8% larger than larvae from the medium sized females (Fig. 5).
Figure 4. Mean (± SE) oil globule area (mm2) at hatching from different size females through six successive spawns.
Both female size and clutch number had a significant influence on larval oil globule area (RM ANOVA; F10, 435 = 14.710, p = 0.0001). Filled triangle represent the small sized females (47–52 mm n = 5), filled square represents the medium sized female (55–60 mm n = 5), and filled circle represents the large sized females (63–68 mm n = 5).doi:10.1371/journal.pone.0048525.g004
Figure 5. Mean (± SE) yolk sac area (mm2) at hatching from different size females through six successive spawns.
Both female size and clutch number had a significant influence on larval yolk sac area (RM ANOVA; F10, 435 = 17.610, p = 0.0001). Filled triangle represent the small sized females (47–52 mm n = 5), filled square represents the medium sized female (55–60 mm n = 5), and filled circle represents the large sized females (63-68 mm n = 5).doi:10.1371/journal.pone.0048525.g005
To date the importance of between season variability in resource allocations has been demonstrated in many species (e.g. red deer, mites, moose, willow tit and gulls) with several broad ecological implications, including dramatic cohort effects on condition , , breeding success , , and survival . However, at present the way organisms that serially spawn partition their energy within a reproductive season is poorly understood, particularly in tropical systems where demographic rates are often high and breeding can occur for extended periods.
In the present study we found no predictable change in egg production across the six spawning events for P. amboinensis. This suggests current production is unaffected by prior energy investment into reproduction, for this tropical damselfish. This was surprising, given the costs associated with reproduction, and that coral reefs are generally energy limited systems . It was anticipated that the production of several large clutches (>3000 eggs per clutch) over a relatively short time period would incur an energetic cost, expressed as a decline in production over time .
While no influence of serial spawning on production was apparent in this study, the size of the mother influenced gross reproductive output. Larger females produced a greater number of embryos per clutch, resulting in a greater total reproductive output. The positive relationship between female size and reproductive output is well recognised and widespread in reptiles –, amphibians , , insects ,  and fishes , , . Fecundity increases with body size because the amount of energy available for egg production ,  and the body cavity accommodating the eggs increase with size , . As most of these factors are interrelated to age and growth it is difficult to disentangle the effects of size and maturity for breeding populations. Although the relationship between size and age for the female Pomacentrus amboinensis population at the breeding site is weak, it is possible that the significant difference in egg production between smaller and larger females is due to a covariance of size and maturity. We would like to acknowledge that it is possible that we are comparing first time spawners to those that have spawned in previous years, a variable we cannot determine or control for when sampling from a wild population. Fish life history models commonly contain direct correlations between fecundity and a female's age or size , . Yet there is a pronounced absence in fish life history theory of consideration for variation in offspring quality or viability with change in parent size or reproductive effort . Production is only one element of the complex of factors that influence the number of offspring that survive from a spawning. Exploring only the quantity of eggs a female produces provides limited power to predict individual fitness.
Investigating characteristics of the offspring at hatching provided us with greater information on reproductive allocation in this iteroparous species, and allowed us to hypothesize benefits of spawning multiple times within a reproductive season. We found that smaller females produced both the longest larvae and larvae with greater energy reserves. This result is curious and contrary to many studies that have found a positive relationship between female size and egg size –. One possible explanation for the pattern is that there is a morphological restriction to the number of offspring a small female can produce and therefore excess reserves are allocated to offspring quality. A study conducted on the bicolor damsel-fish (Stegastes partitus) found that when females were supplemented with additional resources there was no change in the number of offspring produced but the quality of the offspring changed . The author suggests that production is mediated by population demographics which change over longer time scales, whereas offspring quality may change in response to short term variation in the environment. A short term change in the condition of P. amboinensis offspring has been observed in response to an increase in natural food availability .
In our study the characteristics of the offspring changed significantly over the breeding season. Over the six spawns there was a general decline in larval length, which was most apparent in larvae from the small females. Little research has investigated the traits of the resulting larvae from successively produced clutches. Declines in the size of eggs produced successively have been detected in a few species of turtles: Kinosternon subrubrum, Sternotherus odoratus, and Pseudemys floridana  and fishes: Gadus morhua , and Rhinogobius sp. . Whether or not a decline in the size of the eggs produced has a fitness consequence for the resulting offspring is yet to be fully explored. Egg size may not always be a good proxy for egg quality and may not accurately reflect maternal investment in reproduction . Larvae of shorter length but with larger reserves could occupy the same size egg as larvae of longer length but with fewer reserves.
Exploring only one larval characteristic at hatching can lead to invalid conclusions of the workings of this complex system. For example, in this study we could have assumed an associated cost to the quality of offspring when produced successively, expressed as a decline in larval length. However, both the size of the oil globule within the yolk and the size of the yolk-sac increased over successive reproductive events. Lipids contained within the oil globule are believed to be a principal source of energy after hatching and important for post-hatching longevity .
In this study we have demonstrated that in a serially spawning fish the characteristics of offspring vary over relatively short time periods. Initially offspring were longer in length and had fewer energy reserves, but as the season progressed offspring were characterised by shorter length but with larger energy reserves. One explanation for the variation in offspring characteristic observed in this study is that by producing offspring that vary the parents ensure that some offspring will survive no matter what environmental conditions are encountered . Pomacentrus amboinensis offspring disperse away from the parent environment. Therefore, the parents cannot directly predict the environmental conditions where development will take place. Consequently, we may not expect to find a strong link between the offspring's provisioning and parents' environment at the time of gametogenesis. The production of variable offspring may be under parental control, it may not however be a direct response to the environmental conditions encountered by the parents during the reproductive process.
Alternatively, the change in offspring characteristics may reflect selection for larger reserves later in the season. Offspring spawned later in the season will recruit to a reef system where several other cohorts have already established residence. The more eggs the mothers lay, the greater the likelihood that the offspring will recruit to a population with higher density. Females that produce better provisioned offspring later in the season should have enhanced reproductive success, by making their offspring more competitive. Apportioning more energy to each larval offspring should enhance maternal fitness  as endogenous reserves influence initial larval growth , and fast-growing, early-metamorphosing damselfish show the highest larval survivorship , . Therefore, selection for larger reserves later in the season may not only enhance larval survivorship but may also create equality between all offspring produced over the season. By allocating resource differently over the reproductive season it is possible that the parents are mediating competition for resources. This type of resource allocation has been seen in several species where there is high competition for resources between offspring , .
Studies investigating the importance of larval characteristics at the time of recruitment in P.amboinensis provide evidence that mortality is a selective process and certain characteristics increase survival probability. Hoey and McCormick  found that larvae of lower total lipid content and low pre-settlement growth rates were selectively preyed upon. They also found that total lipid content was negatively related to standard length, suggesting that smaller recruits had a tendency to have a higher proportion of lipid than larger recruits . Similarly, Meekan et al.  found that larger, faster growing P.amboinensis larvae suffered high mortality rates at settlement. It appears that body condition, in the form of available or excess energy, is important at settlement and may determine which individuals survive this period of high selection , .Whether or not the phenotypic characteristics offspring have at hatching are maintained over the entire larval phase remains in doubt, as it is very difficult to determine .
The findings of this study highlight that examining a single reproductive event only provides a snapshot of the investment into reproduction for multiple spawning species. Caution is needed when extrapolating total successful reproductive output from one clutch. Our results also highlight the variability in early life history characteristics among sequential reproductive events. This stresses the importance of assessing the quality of larvae in addition to quantity when investigating the fitness consequences to successively spawned offspring. If fitness consequences of iteroparity are to be accurately evaluated, life history models should be appropriately revised to include information on the influence of successive spawns on production, larval quality and viability .
In summary, evidence suggests that small females produced larvae of higher quality. All females regardless of size show flexible resource allocation, producing larvae that vary in quality over the successively produced clutches. Variation in offspring characteristics may not only enhance chances of offspring survival in the patchy pelagic environment into which they are liberated, but may also mediate density dependent resource competition between offspring. These finding may have implication for all animals that reproduce multiple times within a breeding season such as insects, reptiles and amphibians.
Relationship between size and age (y = 2.3215× = 41.782, adjusted R2 = 0.083) for the female population of Pomacentrus amboinensis in the Blue Lagoon at Lizard Island, QLD Australia. The filled gray diamonds represent the female population at the study site and the filled black squares represent the females used in the study.
Relationship between total reproductive output with female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC) at the conclusion of the six week experiment.
Influence of successive spawns (clutch) of females of different (a) standard length (size), (b) age and (c) condition on the total number of embryos produced. Results are from a repeated measure analysis of variance involving repeated sampling of egg clutches from six successive spawns of individually identified females of differing sizes (corrected alpha = 0.016).
Relationship between larval length (dependent variable) from clutch 2 and female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC). Using a best sub set regression model.
Relationship between larval energy reserves (dependent variable) from clutch 2 and female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC). Using a best sub set regression model.
Relationship between larval length (dependent variable) from clutch 4 and female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC). Using a best sub set regression model.
Relationship between larval energy reserves (dependent variable) from clutch 4 and female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC). Using a best sub set regression model.
Relationship between larval length (dependent variable) from clutch 6 and female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC). Using a best sub set regression model.
Relationship between larval energy reserves (dependent variable) from clutch 6 and female standard length, age, GSI and body condition (BC), and male standard length and body condition (BC). Using a best sub set regression model.
Influence of successive spawns of females of different standard length (size), on the length of the larvae produced. Results are from a repeated measures analysis of variance involving repeated sampling of larvae from six successive spawns of individually identified females of differing size.
Influence of successive spawns of females of different standard length (size) on the oil globule area of the larvae produced. Results are from a repeated measures analysis of variance involving repeated sampling of larvae from six successive spawns of individually identified females of differing size.
Influence of successive spawns of females of different standard length (size) on the yolk sac area of the larvae produced. Results are from a repeated measures analysis of variance involving repeated sampling of larvae from six successive spawns of individually identified females of differing size.
Results from a linear regression analysis conducted on the relationship between size and age for the female population of P.amboinensis at the breeding site, Blue Lagoon Lizard Island, Australia (McCormick, unpublished data).
Many thanks to Raechel Littman for assistance with field and laboratory work, and to Arnold Mangott, Stefan Walker and Andrew Hoey who kindly provided useful comments on a draft of the manuscript. Research was conducted under the James Cook University animal ethics guidelines (A1152) and permitted by the Great Barrier Reef Marine Park Authority (G000593).
Conceived and designed the experiments: JM MIM. Performed the experiments: JM. Analyzed the data: JM. Contributed reagents/materials/analysis tools: MIM. Wrote the paper: JM MIM.
- 1. Roff DA (1992) The evolution of life histories: Theory and Analyses, New York: Chapman & Hall.
- 2. Stearns SC (1992) The Evolution of Life Histories., New York: Oxford University Press.
- 3. Fisher RA (1958) The genetical theory of natural selection. 2nd ed, New York: Dover Publications.
- 4. Williams GC (1966) Adaptation and natural selection: a critique of some current evolutionary thought, Princeton, N.J: Princeton University Press.
- 5. Reznick D (1985) Cost of reproduction: an evaluation of the empirical evidence. Oikos 44: 257–267.
- 6. Stearns SC (1989) Trade-offs in life-history evolution. Funct Ecol 3: 259–268.
- 7. Reznick D (1991) Maternal effects in fish life histories. In: Dudley E, editor. Evolutionary biology. Portland, Oregon: Discorides Press. 780–793.
- 8. Bernardo J (1996) Maternal effects in animal ecology. Amer Zool 36(2): 83–105.
- 9. Badyaev AV, Uller T (2009) Parental effects in ecology and evolution: mechanisms, processes and implications. Phil Trans R Soc B 364: 1169–1177.
- 10. Mousseau TA, Uller T, Wapstra E, Badyaev AV (2009) Evolution of maternal effects: past and present. Phil Trans R Soc B 364: 1035–1038.
- 11. Nakatsuru K, Kramer DL (1982) Is Sperm Cheap? Limited Male Fertility and Female Choice in the Lemon Tetra (Pisces, Characidae). Science 216(4547): 753–755.
- 12. Heath DD, Fox CW, Heath JW (1999) Maternal effects on offspring size: Variation through early development of chinook salmon. Evolution 53(5): 1605–1611.
- 13. McCormick MI, Nechaev IV (2002) Influence of cortisol on developmental rhythms during embryogenesis in a tropical damselfish. J Exp Zool 293(5): 456–466.
- 14. Parichy DM, Kaplan RH (1992) Maternal effects on offspring growth and development depend on environmental quality in the frog Bombina orientalis. Oecologia 91: 579–586.
- 15. Bowen WD, Ellis SL, Iverson SJ, Boness DJ (2001) Maternal effects on offspring growth rate and weaning mass in harbour seals. Can J Zool 79: 1088–1101.
- 16. Clemmesen C, Buhler V, Carvalho G, Case R, Evans G, et al. (2003) Variability in condition and growth of Atlantic cod larvae and juveniles reared in mesocosms: environmental and maternal effects. J Fish Biol 62: 706–723.
- 17. Sinervo B (1990) The Evolution of Maternal Investment in Lizards: An Experimental and Comparative Analysis of Egg Size and Its Effects on Offspring Performance. Evolution 44(2): 279–294.
- 18. Bogdanova MI, Nager RG, Monaghan P (2006) Does parental age affect offspring performance through differences in egg quality? Funct Ecol 20(1): 132–141.
- 19. Hendry AP, Day T (2003) Revisiting the positive correlation between female size and egg size. Evol Ecol Res 5(3): 421–429.
- 20. Marshall DJ, Keough MJ (2004) When the going gets rough: effect of maternal size manipulation on larval quality. Mar Ecol Prog Ser 272: 301–305.
- 21. Croxall JP, Rothery P, Crisp A (1992) The effect of maternal age and experience on egg-size and hatching success in wandering albatrosses Diomedea exulans. Ibis: The International Journal of Avian Science 134(3): 219–228.
- 22. Bowen WD, Oftedal OT, Boness DJ, Iverson SJ (1994) The effect of maternal age and other factors in birth mass in the harbour seal. Can J Zool 72: 8–14.
- 23. Berkeley SA, Chapman C, Sograd SM (2004) Maternal age as a determinant of larval growth and survival in a marine fish, Sebastes Melanops. Ecology 85(5): 1258–1264.
- 24. Schluter D, Gustafsson L (1993) Maternal inheritance of condition and clutch size in the collared flycatcher. Evolution 47: 658–667.
- 25. Keech MA, Bowyer RT, Ver Hoef JM, Boertje RD, Dale BW, et al. (2000) Life-history consequences of maternal condition in Alaskan moose. J Wildl Manage 64(2): 450–462.
- 26. McCormick MI (2003) Consumption of coral propagules after mass spawning enhances larvae quality of a damselfish through maternal effects. Oecologia 136: 37–45.
- 27. Einum S, Fleming IA (1999) Maternal effects of egg size in brown trout (Salmo trutta): norms of reaction to environmental quality. Proc R Soc Lond [Biol] 266(1433): 2095–2100.
- 28. Saino N, Dall'ara P, Martinelli R, Møller AP (2002) Early maternal effects and antibacterial immune factors in the eggs, nestlings and adults of the barn swallow. J Evol Biol 15(5): 735–743.
- 29. Gagliano M, McCormick MI, Meekan MG (2007) Temperature-induced shifts in selective pressure at a critical developmental transition. Oecologia 152(2): 219–225.
- 30. Hoey J, McCormick MI, Hoey AS (2007) Influence of depth on sex-specific energy allocation patterns in a tropical reef fish. Coral Reefs 26: 603–613.
- 31. Plaistow SJ, Benton TG (2009) The influence of context-dependent maternal effects on population dynamics: an experimental test. Phil Trans R Soc B 364: 1049–1058.
- 32. Crean AJ, Marshall DJ (2009) Coping with environmental uncertainty: dynamic bet hedging as a maternal effect. Phil Trans R Soc B 364: 1087–1096.
- 33. Duckworth RA (2009) Maternal effects and range expansion: a key factor in dynamic process? Phil Trans R Soc B 364: 1075–1086.
- 34. McCormick MI (2009) Indirect effects of heterospecific interactions on progeny quality through maternal stress. Oecologia 118: 744–752.
- 35. Bernardo J (1996) The particular maternal effect of propagule size, especially egg size: Patterns, models, quality of evidence and interpretations. Amer Zool 36(2): 216–236.
- 36. Fortier L, Leggett WC (1985) A drift study of larval fish survival. Mar Ecol Prog Ser 25: 245–257.
- 37. McGurk MD (1986) Natural mortality of marine pelagic fish eggs and larvae: role of spatial patchiness. Mar Ecol Prog Ser 34(227–242).
- 38. Bradford JH, Cabana G (1997) Interannual variability in stage-specific survival rates and the causes of recruitment variation. In: Chanbers RC, Trippel EA, editors. Early life history and recruitment in fish populaions. London: Chapman & Hall. 469–493.
- 39. Duarte CM, Alcaraz M (1989) To produce many small or few large eggs a size independent reproductive tactic of fish. Oecologia 80: 401–404.
- 40. Sale PF (1991) The Ecology of fishes on coral reefs, San Diego Academic Press.
- 41. Crump ML (1974) Reproductive strategies in a tropical anuran community. University of Kansas, Museum of Natural History, Miscellaneous Publications 61: 1–68.
- 42. Dunham AE, Miles DB, Reznick D (1988) Life history patterns in squamate reptiles, in In Biology of the Reptilia, C. Gans and R.B. Huey, Editors, Alan R. Liss Inc.: New York. p. 441–522.
- 43. Benton TG, St Clair JJH, Plaistow SJ (2008) Maternal effects mediated by maternal age: from life histories to population dynamics. Journal of Animal Ecology 77: 1038–1046.
- 44. Landa K (1992) Adaptive seasonal variation in grasshopper offspring size. Evolution 46(5): 1553–1558.
- 45. Trippel EA, Neil SRE (2004) Maternal and seasonal differences in egg sizes and spawning activity of northwest Atlantic haddock (Melanogrammus aeglefinus) in relation to body size and condition. Can J Fish Aqat Sci 61: 2097–2110.
- 46. Philippi T, Seger J (1989) Hedging one's evolutionary bets, revisited. Trends Ecol Evol 4(2): 41–44.
- 47. Samhouri JF (2009) Food supply influences offspring provisioning but not density-dependent fecundity in a marine fish. Ecology 90(12): 3478–3488.
- 48. Jones GP, McCormick MI (2002) Coral Reef Fishes: dynamics and diversity in a complex ecosystem. In: Sale PF, editor. Coral Reef Fishes: dynamics and diversity in a complex ecosystem. California: Academic Press, San Diego.
- 49. McCormick MI, Smith S (2004) Efficacy of passive integrated transponder tags to determine spawning-site visitations by a tropical fish. Coral Reefs 23(4): 570–577.
- 50. Kerrigan BA (1997) Variability in larval development of the tropical reef fish Pomacentrus amboinensis (Pomacentridae): The parental legacy. Mar Biol 127(3): 395–402.
- 51. McCormick MI (1998) Behaviorally Induced Maternal Stress in a Fish Influences Progeny Quality by a Hormonal Mechanism. Ecology 79(6): 1873–1883.
- 52. McCormick MI (2006) Mothers matter: crowding leads to stressed mothers and smaller offspring in marine fish. Ecology 87(5): 1104–1109.
- 53. Kerrigan BA (1996) Temporal patterns in size and condition at settlement in two tropical reef fishes (Pomacentridae: Pomacentrus amboinensis and P. nagasakiensis). Mar Ecol Prog Ser 135: 27–41.
- 54. Gagliano M, McCormick MI (2009) Hormonally mediated maternal effects shape offspring survival potential in stressful environments. Oecologia 160: 657–665.
- 55. McCormick MI, Gagliano M (2008) Carry-over effects - the importance of a good start. in Proceedings of the 11th International Coral Reef Symposium, 2008. Ft. Lauderdale, Florida.
- 56. Oozeki Y, Hirano R (1988) Effects of Glutaraldehyde Fixation on the Body Size of Red Sea Bream (Pagrus major) Larvae. Aquaculture 71: 265–269.
- 57. Fowler AJ (1990) Validation of annual growth increments in the otoliths of a small, tropical coral reef fish. Mar Ecol Prog Ser 64: 25–38.
- 58. Koops MA, Hutchings JA, McIntyre TM (2004) Testing hypotheses about fecundity, body size and maternal condition in fishes. Fish Fish 5: 120–130.
- 59. Zar JH (1999) Biostatistical Analysis. 4 ed, New Jersey: Prentice Hall.
- 60. Buckley LJ, Bradley TM, Allen-Gulimette J (2000) Production, quality, and low temperature incubation of eggs of Atlantic Cod Gadus morhua and Haddock Melanogrammus aeglefinus in captivity. Journal of the World Aquaculture Society 31(1): 22–9.
- 61. Gaillard JM, Delorme D, Jullien J (1993) Effects of cohort, sex, and birth date on body development of roe deer (Capreolus capreolus) fawns. Oecologia 94: 57–61.
- 62. Thessing A, Ekman J (1994) Selection of the genetical and environmental components of tarsal growth in juvenile willow tits (Parus montanus). J Evol Biol 7: 713–726.
- 63. Rose KE, Clutton-Brock TH, Guinness FE (1999) Cohort variation in male survival and lifetime breeding success in red deer. J Anim Ecol 676: 979–986.
- 64. Coltman DW, Smith JA, Bancroft DR, Pilkington J, MacColl ADC, et al. (1999) Density-Dependent Variation in Lifetime Breeding Success and Natural and Sexual Selection in Soay Rams. Amer Nat 154: 730–746.
- 65. Spear L, Nur N (1994) Brood Size, Hatching Order and Hatching Date: Effects on Four Life-History Stages from Hatching to Recruitment in Western Gulls. J Anim Ecol 63: 283–298.
- 66. Forsman A (2001) Clutch Size versus Clutch Interval: Life History Strategies in the Colour-Polymorphic Pygmy Grasshopper Tetrix subulata.. Oecologia 129(3): 357–366.
- 67. Iverson JB (1992) Correlates of reproductive output in turtles (Order Testudines). Herpetol Monogr 6: 25–42.
- 68. Shanbhag BA, Radder RS, Saidapur SK (2000) Maternal Size Determines Clutch Mass, whereas Breeding Timing Influences Clutch and Egg Sizes in the Tropical Lizard, Calotes versicolor (Agamidae). Copeia 4: 1062–1067.
- 69. Wilkinson LR, Gibbons JW (2005) Patterns of reproductive allocation: clutch and egg size variation in three freshwater turtles. Copeia (868–879).
- 70. Blouin-Demers G, Weatherhead PJ (2007) Allocation of offspring size and sex by female black ratsnakes. Oikos 116: 1759–1767.
- 71. Kuramoto M (1978) Correlations of Quantitative Parameters of Fecundity in Amphibians. Evolution 32(2): 287–296.
- 72. Dziminski MA, Roberts JD (2006) Fitness consequences of variable maternal provisioning in quacking frogs (Crinia georgiana). J Evol Biol 19: 144–155.
- 73. Larsson FK (1989) Female longevity and body size as predictors of fecundity and egg length in Graphosoma lineatum L. (Hemiptera, Pentatomidae). Deutsche Entomologische Zeitschrift 36(4): 329–334.
- 74. Van Dijk TS (1994) On the relationship between food, reproduction and survival of two carabid beetles: Calathus melanocephalus and Pterostichus versicolor. Ecol Entomol 19: 263–270.
- 75. Trippel EA, Kjesbu OS, Solemdal P (1997) Effect of adult age and size structure on reproductive output in marine fishes. In: Chambers CR, Trippel EA, editors. Early life history and recruitment in fish populations. London: Chapman & Hall. 63–102.
- 76. Kamler E (2005) Parent-egg-progeny relationships in teleost fishes: an energetics perspective. Rev Fish Biol Fish 15(4): 399–421.
- 77. Jonsson N, Jonsson B (1999) Trade-off between egg mass and egg number in brown trout. J Fish Biol 55: 767–783.
- 78. Lambert Y, Yaragina NA, Kraus G, Marteinsdottir G, Wright P (2003) Using environmental and biological indices as proxies of egg and larval production of marine fish. Journal of Northwest Atlantic Fishery Science 33: 115–159.
- 79. Stearns SC, Koella JC (1986) The evolution of phenotypic plasticity in life-history traits: predictions of reaction norms for age and size at maturity. Evolution 40: 893–913.
- 80. Kjesbu OS (1989) The spawning activity of cod, Gadus morhua.. J Fish Biol 34: 195–206.
- 81. Marteinsdottir G, Steinardsson A (1998) Maternal influence on the size and viability of Iceland cod Gadus morhua eggs and larvae. J Fish Biol 52: 1241–1258.
- 82. Beacham TD, Murray CB (1985) Effect of female size, egg size, and water temperature on development of chum salmon (Oncorhynchus keta) from the Nitinat River, British Columbia. Can J Fish Aqat Sci 42: 1755–1765.
- 83. Kjesbu OS, Solemdal P, Bratland P, Fonn M (1996) Variation in annual egg production in individual captive Atlantic cod (Gadus morhua). Can J Fish Aqat Sci 53: 610–620.
- 84. Tamada K, Iwata K (2005) Intra-specific variations of egg size, clutch size and larval survival related to maternal size in amphidrmous Rhinogobius goby. Environ Biol Fishes 73: 379–389.
- 85. Gagliano M, McCormick MI (2007) Maternal condition influences phenotypic selection on offspring. J Anim Ecol 76: 174–182.
- 86. Chambers RC, Leggett WC, Brown JA (1989) Egg size, female effects, and the correlation between early life history traits of capelin, Mallotus villosus: an appraisal at the individual level. Fish Bull 87: 515–523.
- 87. Ronnestad I, Fyhn HJ, Gravningen K (1992) The importance of free amino acids to the energy metabolism of eggs and larvae of turbot (Scophthalmus maximus). Mar Biol 114: 517–525.
- 88. Wilson DT, Meekan MG (2002) Growth-related advantages for survival to the point of replenishment in the coral reef fish Stegastes partitus (Pomacentridae). Mar Ecol Prog Ser 231: 247–260.
- 89. Nemeth RS (2005) Linking larval history to juvenile demography in the bicolor damselfish Stegastes partitus (Perciformes:Pomacentridae). Revista de Biología Tropical 53(1): 155–163.
- 90. Dziminski MA, Alford RA (2005) Patterns and fitness consequences of intra-clutch variation in egg provisioning in tropical Australian frogs. Oecologia 146: 98–109.
- 91. Hoey AS, McCormick MI (2004) Selective predation for low body condition at the larval-juvenile transition of a coral reef fish. Oecologia 139: 23–29.
- 92. Meekan MG, von Kuerthy C, McCormick MI, Radford B (2010) Behavioural mediation of the costs and benefits of fast growth in a marine fish Anim Behav. 79(4): 803–809.
- 93. McCormick MI, Ryen CA, Munday PL, Walker SPW (2010) Differing mechanisms underlie sexual size-dimorphism in two populations of a sex-changing fish. PLoS One 5(5).
- 94. Trippel EA (1998) Egg size and viability and seasonal offspring production of young Atlantic cod. Trans Amer Fish Soc 127(3): 339–359.