Chapter 3. Reproductive biology of the protogynous temperate wrasse Odax pullus (Labridae) in the Hauraki Gulf, New Zealand
The ability to change sex is common amongst teleost fishes (Sadovy and Liu, 2008). The ontogeny of sex change may vary across families and genera (reviewed in Sadovy and Liu, 2008), but also within species (Munday et al., 2006b). Such flexibility in sexual ontogeny is thought to optimise individual reproductive value within variable environmental conditions (Munday et al., 2006a; Angilletta, 2009), and significant advances have been made as to our understanding of the underlying mechanisms driving the processes and timing of sex reversal (Allsop and West, 2003b; Allsop and West, 2003a; Munoz and Warner, 2003; Allsop and West, 2004; Buston et al., 2004; Munday et al., 2006a; Munday et al., 2006b).
The Labridae, with a total of eighty-two genera and around six hundred species (Westneat and Alfaro, 2005), display a wide range of patterns in sexual ontogeny, from gonochorism (separate sexes) to sequential hermaphroditism (sex change) (Sadovy and Liu, 2008). Many labrid species exhibit protogynous (female-to-male) sex reversal, which has been confirmed in twenty-one genera to date. While functional protogyny has been shown in many labrid tribes (Reinboth, 1962; 1967; 1968; 1970; Choat and Robertson, 1975; Reinboth, 1975; Robertson and Warner, 1978; Warner and Robertson, 1978), patterns of sexual ontogeny of odacine labrids are poorly understood. As it has been suggested that temperate environments at higher latitudes may provide less favourable thermal conditions for the underlying physiological processes of sex change (Sadovy and Liu 2008), the odacines are a potentially interesting group in this context. This chapter examines the reproductive biology of the odacine labrid Odax pullus.
Diagnosis of the sexual ontogeny of labrid species has proven to be complex. This results primarily from a mismatch between reproductive function (gonochorism vs. sex change) and gonad microstructure. In particular, testes that possess the morphological characteristics of previous female identity (Sadovy and Shapiro, 1987) and that, as such, are considered to be the product of protogynous sex-reversal, may not have passed through a functional female stage (Liu and Sadovy, 2009). Males may be derived from immature bisexual individuals (Hamilton et al., 2008), or directly from an immature female phase (Denny and Schiel, 2002). This is further complicated by the presence within many labrid species of two male development pathways (Reinboth, 1962; 1967; Choat and Robertson, 1975). Depending on local population structure, some males develop directly from a juvenile gonad (primary males) while others are the result of sex reversal of functional females (secondary males) (Shapiro and Rasotto, 1993; Munday et al., 2006b). The complexity of diagnosis of sexual ontogeny in labrid species raises a number of critical points. Firstly, establishing the schedules of sexual maturation and of sex reversal (size- and age-at-maturity and -at-sex change) will be necessary for the diagnosis of reproductive function. This may be challenging in temperate environments however, as the diagnosis of female spawning history may be hindered in conditions where females undergo long inter-spawning intervals. Secondly, examination of the gonad structure of immature juvenile individuals will be critical for the assessment of male and female ontogeny (Liu and Sadovy, 2009). Thirdly, histological examination of gonad structure will need to i) cover all sexual development stages (i.e. across the size and age ranges), and ii) be associated with demographic information on the relationship between size, age and sexual identity.
Prior examination of sexual ontogeny and reproductive function in O. pullus has suggested the presence of protogyny (Ritchie, 1969; Crabb, 1993; Bader, 1998). Evidence supporting this is based on (i) the presence of a sex-specific size and age distribution, with males occurring in the larger size- and older age- classes (Crabb, 1993; Bader, 1998), and (ii) the presence of an ex-ovarian lumen, of primary oocytes in developing testicular tissue, and of peripheral sperm sinuses (Crabb, 1993; Bader, 1998). However, the reliability of these diagnostic features as evidence for functional protogyny has been questioned (Liu and Sadovy, 2004; Sadovy and Liu, 2008; Liu and Sadovy, 2009). In particular, no evidence of prior female function in testes displaying female histological features has been found to adequately support the hypothesis of protogynous sex reversal in this species. The general aim of this study is to examine the sexual ontogeny and development of O. pullus in the Hauraki Gulf, northern New Zealand. Specifically, I first examine the morphology and microstructure of male and female gonads across the species’ size and age range at the sampling location, including juvenile individuals. Secondly, I combine histological and demographic information in order to establish the relationship between gonad microstructure, size, and age. This allows establishing the patterns of male and female recruitment into the population and the schedules of reproductive development and sexual maturation. Thirdly, I examine the histology of testis microstructure for the presence of prior female reproductive function, in order to assess the ontogeny of male function and establish the reproductive biology of the study species.
Materials & Methods
A total of 319 individuals of O. pullus were sampled in the Hauraki Gulf of New Zealand (35.9 – 36.6oS; 174.7 – 175.9oE) (Fig. 3.1). Fish were collected monthly over two and a half consecutive years between August 2005 and January 2008 (Table 3.1).
Fish were sampled by spearing and processed within one hour of collection. Fork Length (FL), Gutted Weight (GW), and gonad weight were recorded for each fish to the nearest millimetre, ten grams and gram, respectively. The sagittal pair of otoliths was removed, rinsed in 70% ethanol, and stored dry. Whole gonads were removed and placed in FAAC, a formalin-based fixative (formaldehyde 4%, glacial acetic acid 5%, calcium chloride 1.3%; Pears et al. 2006). After a minimum period of two weeks, fixed gonads were transferred into 70% ethanol before histological processing.
Table 3.1: Monthly sampling effort of Odax pullus in the Hauraki Gulf from August 2005 to January 2008, showing the number of individuals collected. * shows samples that had been frozen prior to being fixed – histology of these samples allowed determination of sex but not assessment of sexual maturity.
Gonads were blotted dry and weighed whole (including both lobes) to the nearest 10-4 gram. One gonad lobe was selected randomly, and a transverse section through the medial region was taken for histological processing. All gonad samples were embedded in paraffin wax, sectioned at 7 µm, mounted on glass slides and stained with Gill’s Haematoxylin and Eosin stain (H&E stain).
In order to examine the consistency of gonad development across the length of the gonad lobe as well as across gonad lobes, two sub-sample procedures were followed. (i) A subset of 39 gonad samples (collected November to April) was serially sectioned in the anterior, medial and posterior regions of the tissue to examine the consistency of gonad development along the gonad length. (ii) In a sub-sample of 10 gonads (collected in March), both lobes were sectioned to examine the consistency of reproductive diagnosis across gonad lobes analysed. Analysis of the sub-sampled histological sections showed no variation in diagnosis of sexual identity or reproductive status among gonad regions sampled and between lobes sectioned. Consequently, transverse sectioning of the medial region of one randomly chosen gonad lobe was considered to provide representative information for diagnosis of reproductive developmental stage and sexual identity of each gonad.
All histological sections were examined under a high power microscope with transmitted light. For each section, sexual identity, reproductive activity and spawning history were assessed. Examination of the slides was blind with respect to fish identity, age, or size. Stages of oocyte development were classified by the latest non-atretic stage present in the tissue, as follows: stage I – chromatin nucleolar stage, stage II – peri-nucleolar stage, stage III – cortical alveoli stage (formation of yolk vesicle precursors), stage IV – vitellogenic (yolk) stage, and stage V – hydrated stage (West, 1990). Oocyte atresia was identified following criteria presented in Hunter and Macewicz (1985), and the presence of α-stage atretic vitellogenic oocytes and of brown bodies was recorded. The latest and most abundant spermatogenetic stage present was used to identify the developmental stage of each testis (Grier, 1981). Gonads were classified as one of the following four categories: (i) “female”, when oogenic tissue was present throughout the section, (ii) “male”, when spermatogenic tissue was observed throughout the section, (iii) “early-transitional”, when atretic vitellogenic (or later stages) oocytes and/or post-ovulatory follicles could be seen within developing testicular tissue, and (iv) “late-transitional”, when the simultaneous presence of primary oocytes (pre-vitellogenic stages chromatin nucleolar and peri-nucleolar) and spermatogenic tissue (spermatogonia to spermatozoa) was observed, and no signs of previous spawning as a female (post-ovulatory follicles and / or atretic vitellogenic oocytes) could be seen (Sadovy and Shapiro, 1987; Sadovy and Liu, 2008). The presence in ovaries of intra-lamellar muscle bundles, atretic vitellogenic oocytes, and post-ovulatory follicles was recorded, and their incidence across months was established to examine their relationship to the timing of postspawning events. When appropriate, these criteria were used in the diagnosis of female spawning history (Adams and Choat, 2002; Liu and Sadovy, 2004). Criteria used in the diagnosis of gonad sexual identity, reproductive activity, developmental stage, and spawning history are summarized in Table 3.2.
Size- and age-frequency distributions of males and females were established (frequency of males or females per size or age class as a proportion of the total number of individuals collected). Evidence for putative protogynous sex change was assessed using criteria developed in Sadovy and Shapiro (1987) and Sadovy and Liu (2008). A critical factor was to correctly identify transitional individuals, a key criteria of interest being the presence or absence of atretic vitellogenic (or later stage) oocytes (AVOs). A subset of sections identified as early-transitional, mature inactive (resting) female, spawning female (with stage V oocytes) and post-spawning female (with AVOs) was unstained for H&E, and re-stained for identification of glycoproteins, a primary constituent of yolk, using Periodic-Acid Schiff (PAS) and Haematoxilin (counter-stain) (Wallace and Selman, 1981). In order to establish the ontogeny of testis development, the organisational microstructure of juvenile gonads was examined (Liu and Sadovy, 2009), and testis microstructure was compared to that of ovaries (Shapiro and Rasotto, 1993). Male gonads were further examined to: (i) determine the position of sperm sinuses (central duct situated dorsally indicative of primary male development, versus peripheral ducts indicative of secondary male development), and (ii) to check for the presence of a membrane-lined lumen (indicative of ex-ovarian lumen) and lamellar structure.
Timing and duration of the spawning season was assessed based on the period of highest female reproductive activity. Reproductive activity was evaluated from monthly variation in the gonadosomatic index (GSI). GSI was calculated using weight of both ovaries (gonad weight) as a proportion of gutted body weight, as follows: % 100. Mean monthly GSI was estimated for mature and immature females and the GSI profile examined across months sampled. GSI was plotted with mean monthly sea surface temperature (SST) recorded in the Hauraki Gulf across both years sampled.
The timing of the formation of intra-lamellar muscle bundles (ILMB), post-ovulatory follicles (POF) and atretic vitellogenic oocytes (AVO) was examined by plotting the frequency of ovaries showing ILMBs, POFs, or AVOs, respectively, across months.
Size- and age-at-maturity
Estimation of size- and age-at-maturity was performed based on females collected during the spawning season (Pears et al., 2006). Females with inactive ovaries (stage I & II oocytes only) during the spawning months were immature, and females displaying stage III (cortical alveoli stage) to stage V (hydrated stage) oocytes during the spawning season were mature.
Size- and age-at-maturity were estimated by calculating the size and age at which 50% of females were mature (proportion of mature females relative to the total number of females within each size or age class). For each female, information used was size (Fork Length FL in mm), age (in years), and maturity (immature or mature). A logistic function was fitted to the non-linear relationship between maturity (dependent variable, y-axis) and size or age (independent variable, x-axis) respectively, following Moore et al. (2007) and Williams et al. (2008). The best-fit logistic model was estimated by minimising the negative log of the likelihood based on a probability density function with a binomial distribution (Haddon, 2001). The logistic function used is of the form: – , where is the proportion of mature females in size or age class a, is the size or age at which 50% of females are mature, and is the size or age at which 95% of females are mature. Parameters estimated were size-at-50% maturity , size-at-95% maturity , age-at-50% maturity , and age-at-95% maturity . Size classes were based on 50 mm increments, and age classes were based on one year increments. 95% confidence intervals (CI) were estimated for each parameter value of and using a bootstrapping procedure (Moore et al., 2007). The data were randomly re-sampled 1000 times with replication. For each re-sampled data set, the logistic function was fitted and the best-fit combination of parameters and was estimated as described above. 95% percentile confidence intervals were calculated using the as 2.5 and 97.5 percentiles of the bootstrap estimates (Haddon, 2001).
An additional 108 fish (total of 427 fish) were sampled for size-at-age information; these additional individuals were sexed macroscopically. Based on the histological diagnosis of sexual identity of reproductively active males and females processed in this study, the error associated with macroscopic identification of sexual identity was calculated as 0.8%, and considered negligible. Sagittal otoliths were processed, and age was estimated as described in Chapter 2 (section 2.2). Size-at-age data was modeled separately for males and females using the re-parameterized equation of the von Bertalanffy growth function (rVBGF) (Francis, 1988) as described in Chapter 2 (section 2.2, “Growth modelling”). Best-fit rVBGF models were estimated for each sex separately by minimizing the negative log of the likelihood assuming a normal probability density function (Kimura, 1980; Haddon, 2001; Trip et al., 2008). The rVBGF parameters yielded estimates of mean size-at-age one (1), five (5) and nine (9) for females, and of mean size-at-age five (5) and nine (9) for males (no males were present in year one age class). A Likelihood Ratio Test (LRT) was used to compare the growth of males and females (Cerrato, 1990). The LRT statistic D was estimated as follows: 2log Ω, where is the maximum likelihood of the data under the null hypothesis of no difference in growth between males and females, and Ω is the maximum likelihood of the data under the alternative hypothesis that growth of males and females is better represented by two separate models. The null hypothesis of no difference in growth between the sexes was rejected at α = 0.05 if D exceeded the value of χ2(q), with q being the difference in the number of parameters under the two hypotheses (i.e. q = 3 for coincident curves) (. 37.81).
Table of contents
Table of contents
List of figures
List of tables
Chapter 1. General Introduction
Chapter 2. Age estimation of Odax pullus (Labridae): otolith microstructure, validation of increment periodicity, and position of the first annual band
2.2 Materials & methods
Chapter 3. Reproductive biology of the protogynous temperate wrasse Odax pullus (Labridae) in the Hauraki Gulf, New Zealand
3.2 Materials & Methods
Chapter 4. Latitudinal effects on the demography and life history of a temperate marine herbivorous fish, Odax pullus (Labridae)
4.2 Materials & Methods
Chapter 5. Latitudinal variation in life span: temperature effects on the rate of ageing in a temperate marine fish, Odax pullus (Labridae)
5.2 Materials and Methods
Chapter 6. General Discussion
GET THE COMPLETE PROJECT
Latitudinal variation in the demography and life history of a temperate marine herbivorous fish Odax pullus (Labridae)