Received: 10 May 2016 | Accepted: 17 September 2016 DOI: 10.1111/jai.13273

ORIGINAL ARTICLE

Dynamics of sex reversal in the marbled ( marmoratus Bloch, 1795), a diandric hermaphrodite from Marechal Dutra Reservoir, northeastern Brazil

N. H. C. Barros1 | A. A. de Souza1 | E. B. Peebles2,* | S. Chellappa3,*

1Postgraduate Programe in Psychobiology, Universidade Federal do Rio Summary Grande do Norte, Natal, Rio Grande do Norte, This study characterizes the dynamics of sex reversal in the marbled swamp eel, Brazil (Osteichthyes: Synbranchidae), a diandric hermaphrodite, 2College of Marine Science, University of South Florida, St. Petersburg, FL, USA within the context of managing with complex sex allocations. Monthly sam- 3Department of Oceanography and pling in Marechal Dutra Reservoir, northeastern Brazil, was conducted using metal eel Limnology, Universidade Federal do Rio traps from July, 2013, to June, 2014, during which a total of 288 individuals were Grande do Norte, Natal, Rio Grande do Norte, Brazil captured. Morphological and histological comparisons of gonads identified four sex types: primary males (n = 18), females (n = 197), transitional individuals (n = 30), and Correspondence Ernst Bryan Peebles, College of Marine secondary males (n = 43). Primary males were smallest, ranging 18–32 cm total length. Science, University of South Florida, St. Females were numerically dominant throughout the 1-­year sampling period, and Petersburg, FL, USA. Email: [email protected] ranged 20–60 cm. Transitional individuals ranged 32–60 cm, and secondary males ranged 46–74 cm. The otolith-­based age of 52 specimens ranged 0.5 to 5+ year. Funding information National Council for Scientific and Primary males were only observed at ages 0.5 and 1, and transitional individuals were Technological Development (CNPq); Post only observed at ages 3 and 4 during the female-­to-­secondary-­male transition, sup- Graduate Federal Agency of the Ministry of Education, Brazil (CAPES/MEC) porting the existence of two types of individuals: gonochoristic males and protogy- nous hermaphrodites. This observation was further supported by histological observations of deteriorating ovarian tissue in transitional individuals. Given the length of time required for individuals to attain secondary male status, this species ap- pears to be particularly vulnerable to over-­exploitation. Comparisons with results from other studies suggest sex allocations and adult size distributions vary substantially within this species’ range, adding complexity to management efforts.

Shapiro, 1987). Sequential hermaphroditism occurs as protandry or 1 | INTRODUCTION protogyny. Protandrous species change sex from adult male to female, whereas protogynous species change sex from adult female to male. Among vertebrates, fishes have notably flexible sex allocations and Protogynous species can be diandric, wherein two types of males are often hermaphroditic. Both sexes are functional at the same occur (Barros, Souza, & Chellappa, 2013; Reinboth, 1983). time in simultaneous hermaphrodites, although this condition is not The marbled swamp eel, Synbranchus marmoratus Bloch, 1975 common among teleosts (Reinboth, 1983). Sequential (dioecious) (: Synbranchidae), is an exploited fish native to the hermaphrodites, in contrast, are relatively common among teleosts, inland waters of the semiarid region of Brazil (Britski, Silimon, & Lopes, and the modality of sex reversal may be either spontaneous or con- 1999; Nascimento et al., 2011; Nelson, 1994). It is a diandric species, trolled by genetic, physiological, or behavioral factors (Sadovy & having both primary and secondary males (Barros et al., 2013; Lo

*Shared Senior Co-authorship. Nostro & Guerrero, 1996). Although other species may be considered

J Appl Ichthyol. 2017;33:443–449. wileyonlinelibrary.com/journal/jai © 2017 Blackwell Verlag GmbH | 443 444 | BARROS et al. diandric because two sex reversals can occur during the life of a sin- alcohol. The ovaries were later embedded in paraffin wax, sectioned gle individual (i.e., male to female to male), the marbled swamp eel is at 3–5 μm thickness, and stained with Harris hematoxylin and eosin considered diandric because it is believed that two types of individu- (Smith, Sansom, & Repetski, 1996). als exist, representing two independent types of male, with one type In order to avoid possible variation in the developmental stage being a lifetime, non-­hermaphroditic (gonochoristic) male and the of oocytes due to their position within the ovaries, histological ex- other type being a protogynous hermaphrodite. aminations were carried out on sections from the anterior (cephalic), Sex-­allocation theory provides some explanation of sex reversal middle (central), and posterior (caudal) regions of ovaries from differ- in terms of recognizing inherent population characteristics that foster ent developmental stages. Comparison of these regions indicated the gender plasticity (Allsop & West, 2003; Charnov, 1982). The general mid-­section of the ovary was representative of oocyte maturation idea is that sex change is favoured when (i) the reproductive fitness of throughout the ovary (Yoshida, 1964). Subsequently, only a central an individual is closely related to its age or size, and (ii) this relationship transverse section was prepared for each gonad (n = 40). Gonad mat- is different between the sexes (e.g., differences in mate-­finding prob- uration was determined using a compound microscope coupled with a ability, social dominance, or considerations related to the volumetric video camera and computer monitor. Oogenesis staging terminology packing of eggs and sperm within a size-­structured spawning stock). followed Wallace and Selman (1981) and Brown-­Peterson, Wyanski, Regardless of cause, switching to a sex that has higher reproductive Saborido-­Rey, Macewicz, and Lowerre-­Barbieri (2011). fitness would increase the individual’s lifetime reproductive potential, The proportions of the different sex types were calculated using and thus its total potential fitness. Conversely, changing the size/age the total number of individuals captured. Variations in sex ratio were 2 structure through fishing may diminish any fitness advantages brought analyzed using the chi-­square (χ ) test at a significance level of p < .05. about by this capability, and may even alter fundamental life history parameters, introducing additional complexity into the management 2.3 | Length–weight relationship process (Grift, Rijnsdorp, Barot, Heino, & Dieckmann, 2003; Hamilton et al., 2007; Watson, Anderson, Kendrick, Nardi, & Harvey, 2009). The Total length (TL ± 1 cm) and total body weight (W ± 1 g) of fish were present study investigates the dynamics of sex reversal in the marbled recorded. The total length–weight relationships of females and sec- swamp eel, and relates this topic to the prospect of managing species ondary males were estimated from the equation log W = log a + b log with plastic sex-­allocation capabilities. L, where W is total body weight, L is total length, a is the intercept, and b is the slope of the linear regression (Jobling, 2002). Log–log plots were generated to remove outliers (Froese, 2006), and 95% confi- 2 | MATERIALS AND METHODS dence limits for anti-­log a and b were calculated for females and sec- ondary males. A t test was performed to confirm whether b departed 2.1 | Sampling from the isometric value 3. A single weight–length equation (W = aLb) Eel sampling was carried out on a monthly basis from July, 2013 to was fitted to estimate the value of coefficient b using data obtained July, 2014 in the Marechal Dutra Reservoir (6°26′24″S; 36°38′00″W), from all individuals collected (Froese, 2006). located in northeastern Brazil. Eels were captured with the help of local fishermen, who used artisanal traps, locally known as “covos”. 2.4 | Age and growth estimation through The eel traps were made from sheets of tin and were tube-­shaped, otolith analysis with the tube closed at one end and an inverted funnel serving as an entrance at the other end. The traps were baited with prawns and Initially, all pairs of otoliths (lapilli, sagittae and asterisci) were removed­ tied into groups of two. Traps were left overnight at the bottom of from fish. The sagittae were selected as most suitable for age analy- the reservoir. The following morning, captured eels were individually sis based on annulus resolution quality and the regularity of micro- numbered, weighed and measured to obtain morphometric informa- structural patterns. Only the right sagittae were used for age analysis tion that was used to confirm taxonomic identity (Britski et al., 1999). (Campana, 1990). A total of 52 right sagittal otoliths from primary males (n = 3), females (n = 33), transitional individuals (n = 8), and sec- ondary males (n = 8) were analyzed at the College of Marine Science, 2.2 | Morphological and histological changes during University of South Florida, USA, according to the methods of Chilton sex reversal and Beamish (1982). All otoliths were read at 40× magnification using Stages of gonad maturity were determined using a macroscopic stag- a stereomicroscope coupled with a video camera and computer moni- ing system. The macroscopic characteristics used for sex differentia- tor. For initial readings, the first opaque ring was identified and the tion were based on the external morphology of the gonads, such as completeness of the last opaque ring (the marginal increment) was the size, form, coloration, presence of blood vascularization, presence checked. Only fully formed rings were considered in the counts. Three of visible oocytes and the amount of coelomic space occupied by the readings of otoliths were performed by three different readers and an gonads. The gonads of mature females and transitional individuals se- average of the three readings was calculated. lected for histological procedures (n = 40) were fixed in Bouin solution Marginal increment ratio (MIR) analysis was used to determine for 12–24 hr, washed for 24 hr in water, and preserved in 70% ethyl the seasonal period of annulus formation. The MIR was estimated BARROS et al. | 445

3ULPDU\ PDOHV )HPDOHV 7UDQVLWLRQDO LQGLYLGXDOV 6HFRQGDU\ PDOHV

 Q 







 FIGURE 1 Absolute frequency of

Number occurrence and total-­length classes (14 mm  Q  bins) of the four different sex types of  Synbranchus marmoratus (n = 236) collected Q   during the period July, 2013, to June, Q  2014, from Marechal Dutra reservoir,  northeastern Brazil. Different sex types are ± ± ± ± indicated by different fill types Total length (cm)

using the following formula: MIR = (R − Rn)/(Rn − Rn−1), where R is after sectioning (400–1,200 μm in width), particularly in the area of otolith radius, Rn is the radius of the last complete zone, and Rn−1 the third to fifth rings. The first and second rings were very close is the radius of the penultimate complete zone (Haas & Recksiek, to the core and were more difficult to discriminate. Estimated ages 1995). The adjustment in the accuracy of readings was performed ranged 0.5 to 5+ year (Figure 4). according to the calculation of average percent error established by The aging procedure identified six age classes (Figure 5). The age Beamish and Fournier (1981). The growth curve and associated pa- of primary males ranged from 0.5 to 1 year, females ranged from 1 to rameters were determined by nonlinear fit to the von Bertalanffy 2 years, and transitional individuals ranged from 3 to 4 years (Table 1). −k(t−t ) growth function, Lt = L∞ 1−exp 0 , executed using a quasi-­ Sex reversal occurred at an age of 3 to 4 years within the length range

Newton algorithm in STATISTICA (StatSoft, Inc.), where Lt is length of 28–48 cm. Secondary males were older and larger, with age vary- at age t, L∞ is maximum length, k is the growth coefficient, and t0 is ing from 4 to more than 5 years, and were the only represented in time at hatching. the largest length class (58–68 cm, Figure 5 and Table 1). The von Bertalanffy model indicated a relatively slow growth rate for all sex types combined (Figure 6). 3 | RESULTS

3.1 | Sex-­type composition 3.3 | Morphological and histological changes during sex reversal The entire collection of Synbranchus marmoratus (n = 288 individu- als) was composed of primary males (born as males and functionally Females had an unpaired, single, long and cylindrical ovary with male; n = 18), functional females (n = 197), transitional individuals rounded edges. The unpaired ovaries of young females were translu- (both male and female gonadal tissues present; n = 30), and secondary cent, ribbon-­like, and occupied a space less than 1/3 of the coelomic males (former females; n = 43). Primary males ranged 18–32 cm TL cavity. In the maturing stage, the ovary was larger, well irrigated by (26.3 ± 3.95). There was a predominance of females during the en- blood vessels, and was reddish to deep-­pink in colour. The mature 2 tire, 1-­year sampling period (χ = 57.51; p = .0001). Females ranged ovary was large, occupying almost all the available space of the coe- 20–60 cm (43.9 ± 6.85), transitional individuals ranged 32–60 cm lomatic cavity, and was yellowish-­orange in colour, with large oocytes (49.3 ± 6.7), and secondary males ranged 46–74 cm (57 ± 4.9). visible throughout the entire ovary. Secondary males had the largest body lengths and primary males The ovary of mature females was devoid of the medular zone had the smallest. The highest proportion of females and transitional and was covered by túnica albuginea consisting of conjunctive tis- individuals was in the range 42–60 cm (Figure 1). The fitted length– sue and blood vessels. In the ovigerous lamellae, either germinative weight relationship is presented in Figure 2. cells or oocytes­ in different phases of development were present. The ovarian stroma partly consisted of a prominent extravascular space at the ­periphery of the developing oocytes. The vitellogenic 3.2 | Age and growth estimation through oocytes had large quantities of yolk globules, the nucleus was otolith analysis smaller than in ­earlier stages, the nucleolus was randomly distrib- Sagittae measured 6.0–10.0 mm along their longest axis. On the inner uted, and the zona pelucida was narrow. In the ovary of mature side, irregular calcium-­carbonate deposition obscured the view of functional females, atretic vitellogenic oocytes were observed with growth rings. The otoliths had a flat profile on the inner face with a a characteristic irregular outline and detached zona pelucida, and convex outer face. The inner face was rhomboidal with a crenulated the nucleus was disorganized and had begun migrating towards the border and a median rostrum (Figure 3). Clear rings became visible periphery. 446 | BARROS et al.



  :WW /



(g) 

eight  FIGURE 2 Regression of weight on

W total length for all sex types of Synbranchus  marmoratus collected during the period July, 2013, to June, 2014, from Marechal  Dutra reservoir, northeastern Brazil. Outlier values are indicated with diamond  symbols; outliers were not included in the            calculation of the LWR equation (n = 202, Total length(cm) R2 = .99, slope p < .05)

(a) (b)

FIGURE 3 Whole sagittal otoliths of Synbranchus marmoratus viewed as (a) internal face with cauda identified (black line and asterisks) and (b) external face

(a) (b)

FIGURE 4 Photomicrographs (40× magnification) of (a) transverse section of a Synbranchus marmoratus sagittal otolith and (b) the same section with annuli indicated (gray lines and white numbers)

3ULPDU\ PDOH )HPDOH 6HFRQGDU\ PDOHV 7UDQVLWLRQDO LQGLYLGXDOV 



 FIGURE 5 Age composition for the four sex types of Synbranchus marmoratus collected during the period July, 2013, 1XPEHU  to June, 2014, from Marechal Dutra reservoir, northeastern Brazil. A total of 52  individuals were sex-­typed and aged using sagittal otoliths (three primary males, 33  females, eight transitional individuals, and       eight secondary males). Different sex types $JH \HDUV are indicated by different fill types BARROS et al. | 447

TABLE 1 Age-­class composition of Number of fish by age-­class (years) primary males, females, secondary males, and transitional individuals of Synbranchus Sex 0.5 1 2 3 4 5 Total marmoratus collected from July 2013 to June 2014, Marechal Dutra reservoir, Primary males 2 1 — — — — 3 northeastern Brazil Females — 8 22 3 — — 33 Transitional individuals — — — 6 2 — 8 Secondary males — — — — 1 7 8 Total 52

 ± W ± /W >±H[S @    FIGURE 6 Von Bertalanffy growth  curve for Synbranchus marmoratus collected   during the period July, 2013, to June,  2014, from Marechal Dutra reservoir,  northeastern Brazil (same specimens as in Figure 5, n = 52). All sex types are  combined (three primary males, 33 females, 7RWDOOHQJWK FP eight transitional individuals, and eight secondary males). The regression (black  line) was fit to mean total lengths for  each of six age classes (age-­class means  are plotted as symbols, with values in  parentheses; n = 6, R2 = .94, slope p < .05) $JH \HDUV

Transition individuals originate from females when they enter the 4 | DISCUSSION sex-­reversal phase. In the unpaired cylindrical ovary, there were dense groups of cells, white to pale yellow in colour, indicating cellular disor- 4.1 | Sex transition and ovarian maturation ganization with the first appearance of testicular tissue. During the pro- cess of sex reversal, testicular tissue replaced the female germinative If interpreted in isolation, the data in Figure 1 could support the exist- tissue. This phase was characterized by the presence of primary oocytes ence of either (i) a combination of gonochoristic males and protog- with cortical alveoli, several vitellogenic oocytes and spermatogenic tis- ynous individuals (scenario 1), or (ii) individuals that transition from sue. There were a small number of oocytes undergoing absorption. The male to female to male (scenario 2). However, when interpreted with detachment of the zona pellucida was observed in the atretic oocytes, age data, as in Figure 5, support for the presence of two types of indi- forming an extravascular space. The gonad showed a large proliferation viduals (scenario 1) strengthens, suggesting there is only one sex tran- of Leydig cells and migration of myoid cells. There was an invasion of sition in this species (i.e., transition from protogynous female to male). male tissue at the periphery of the ovigerous lamellae between the fe- This supports the sex allocation pattern attributed to Synbranchus male germ cells. During this phase, melano-­macrophagic centres began marmoratus by Lo Nostro and Guerrero (1996). to appear. These centres were pigmented cell aggregates or black spots, The present study also confirms that during the process of sex re- which occurred next to germ cells and were mainly responsible for the versal, adult female S. marmoratus undergo several, specific internal removal or phagocytosis of cellular debris. changes where the ovarian tissue is completely replaced by male tis- The subsequent phase was characterized by having few oocytes, sue. In some labrids, the process of sex reversal occurs in individuals and practically all of these were in the process of vitellogenic atre- prior to sexual maturity (Carruth, 2000). The gonads of transitional sia and were being gradually absorbed. Male tissue occupied almost individuals are very characteristic because they have a relatively pre- 75% of the gonad. Melano-­macrophagic centres and atretic oocytes dictable disorganization in gonad architecture, mass degeneration of became more numerous than in the previous phase. The main distin- female germ cells, the presence of melano-­macrophagic centers and guishing feature was the presence of cysts with male cells in various invasion of interstitial tissue by male germ cells. In the later phases stages of development that were visible as clusters with spermato- of transition, cysts containing spermatogonia appear at the edges of cytes. In the final phase, all the space formerly occupied by oocytes the ovigerous lamellae, which give rise to spermatozoids. A similar his- was filled with male tissue. The presence of many spermatocysts and tological pattern was also found in the Asian swamp eel, Monopterus spermatocytes, as well as the presence of lobes devoid of sperm, was albus (Liem, 1963). The presence of melano-­macrophages observed characteristic of this phase. during the transition phase of S. marmoratus (related to the absorption 448 | BARROS et al. and removal of tissue and cellular debris) has also been observed by 1996). Such apparently large differences in sex allocation could be due others (Asoh, 2005; Sadovy & Shapiro, 1987). The transition phase to differences in growth rate, differences in size/age distribution (pos- is characterized by the presence of both oocytes and spermatogenic sibly through selective fishing), or differences in the social organization tissue as well as atresia of vitellogenic oocytes in protogynous her- of the two populations (Edwards, 1998; Vicentini & Araújo, 2003). In maphrodites (Brown-­Peterson et al., 2011). This was confirmed in the addition, there also appear to be location-­specific differences in the present study. In transitional individuals, the sex reversal is character- maximum body length for this species in different parts of this spe- ized by the gradual decrease in female activity and degeneration of cies’ range. The maximum recorded body length for S. marmoratus is ovarian germ cells. At the end of this process, there is an increase in 100 cm (Britski et al., 1999), although most studies describe popula- spermatogenic activity, and secondary males become the end prod- tions of smaller individuals. In the present study, the secondary males ucts of the sex reversal. Similar cellular structures were encountered in (the largest sex type) measured 60–68 cm. In the Alto Uruguai River, other protogynous, diandric species such as the damselfishes Dascyllus Brazil, S. marmoratus upper lengths were reported to be shorter, rang- trimaculatus (Asoh & Kasuya, 2002), D. carneus (Asoh & Yoshikawa, ing 33–49 cm (Oliveira & Zaniboni-­Filho, 2009). In the hydrographic 2003) and D. flavicaudus (Asoh, 2004). basin of Tibagi River, Brazil, individuals were described as having The morphological characteristics of S. marmoratus female germ a mean length of 59 cm (Shibatta, 2003), which would indicate the cells in various development stages were described in an earlier study Tibagi River fish attained relatively large sizes. based on measurements of ovarian follicle diameters (Spadella, de Management decisions based on length and age at maturity of Castilho-­Almeida, Quagio-­Grassiotto, & Cesario, 2005). Their observa- females should more appropriately focus on the size and age at sex tions showed that it was possible to group oocytes by stages accord- transition when dealing with species that have complex and poten- ing to histological characteristics but not according to morphometric tially plastic sex allocations, and all sex types should be included in diameter, as there was a wide variation in diameter in each stage and estimates of spawning stock biomass. Moreover, if size/age distribu- overlap between different maturation stages. tions and sex allocations vary among locations (or over time within locations), then the practice of protecting sequential hermaphrodites from overexploitation becomes an even more complex challenge that 4.2 | Age at sex reversal requires a substantially larger amount of reproductive information The present study suggests that sex transition of female S. marmoratus than required for managing non-­hermaphroditic species. There is cur- occurs at 3–4 years of age. Sex ratios shifted from exclusively female rently no general theory that can be used as guidance while devel- individuals at age-2­ to predominantly transitional individuals by age-3.­ oping a spatial management plan for hermaphroditic fishes, although As such, the sex transition in S. marmoratus is related to age, maturity, models that respond to size-­specific fishing mortality have been re- and body size of the females (Figures 2, 5 and 6). The transitional-­stage cently developed by Easter and White (2016). Through a combination individuals had gonads that contained both degenerating ovarian tis- of morphological, histological, and otolith-­based analyses, we have sue and developing testicular tissue, further undermining the idea that confirmed the presence of small, short-­lived males (primary males) in primary males may transition to females (scenario 2 above). The sec- S. marmoratus. The presence of small, primary males in diandric spe- ondary males were the largest and oldest individuals (age-5­ year class) cies, when coupled with variation in size and age at sex transition, fur- in the population. Petracchi and Guerrero (1999) reported finding an ther complicates management of hermaphroditic species, and renders 11-year-­ ­old specimen of S. marmoratus. This 11-year-­ ­old was likely ex- conventional fisheries regulation based on size at protogynous-fe- ceptional; in the present study, no individuals older than 5+ were found. male maturation ineffective (Grandcourt, Al Abdessalaam, Francis, Al Shamsi, & Hartmann, 2009).

4.3 | Implications for management ACKNOWLEDGEMENTS Protogynous and protandrous fishes are more susceptible to overfish- ing because they need to reach an appropriate size for sex reversal, The authors wish to thank the National Council for Scientific and and if fishing pressure does not permit individuals to reach this size, Technological Development (CNPq) and the Post Graduate Federal then the entire stock could be threatened by gamete limitation (Alonzo Agency of the Ministry of Education, Brazil (CAPES/MEC), for the & Mangel, 2004; Beets & Friedlander, 1999). Diandric species such as financial support awarded during the study period. The authors are Synbranchus marmoratus present an even more complex management indebted to the artisanal fishermen of Marechal Dutra reservoir, challenge, as it is known that fish will alter fundamental life history traits northeastern Brazil, who helped to collect the fish used in this study. in response to size-selective­ mortality, such as that brought about by fishing (Grift et al., 2003; Hamilton et al., 2007; Watson et al., 2009). REFERENCES There is evidence that size/age distributions and sex allocations dif- fer substantially among locations within S. marmoratus’ range. Females Allsop, D. J., & West, S. A. (2003). Changing sex at the same relative body size. Nature, 425, 783–784. constituted 68.4% of the total number of S. marmoratus collected from Alonzo, S. H., & Mangel, M. (2004). The effects of size-­selective fisheries on Marechal Dutra Reservoir, differing from a study in Argentina where the stock dynamics of and sperm limitation in sex-­changing fish. Fishery females constituted only 36.9% of the catch (Lo Nostro & Guerrero, Bulletin, 102, 1–13. BARROS et al. | 449

Asoh, K. (2004). Gonadal development in the coral-­reef damselfish harvesting alters life histories of a temperate sex-­changing fish. Dascyllus flavicaudus from Moorea, French Polynesia. Marine Biology, Ecological Applications, 17, 2268–2280. 146, 167–179. Jobling, M. (2002). Environmental factors and rates of development and Asoh, K. (2005). Gonadal development and diandric protogyny in two growth. In P. J. Hart & J. D. Reynolds (Eds.), Handbook of fish biology populations of Dascyllus reticulatus from Madang, Papua New Guinea. and fisheries, Volume 1: Fish biology (pp. 97–122, 424). Oxford, UK: Journal of Fish Biology, 66, 1127–1148. Blackwell Publishing Ltd. ISBN 0632054123. Asoh, K., & Kasuya, M. (2002). Gonadal development and mode of sexual- Liem, K. F. (1963). Sex reversal as a natural process in the synbranchiform ity in a coral reef damselfish, Dascyllus trimaculatus. Journal of Zoology fish Monopterus albus. Copeia, 2, 303–312. (London), 256, 301–310. Lo Nostro, F. L., & Guerrero, G. A. (1996). Presence of primary and second- Asoh, K., & Yoshikawa, T. (2003). Gonadal development and an indication ary males in a population of the protogynous fish Synbranchus marmor- of functional protogyny in the Indian damselfish, Dascyllus carneus. atus. Journal of Fish Biology, 49, 788–800. Journal of Zoology (London), 260, 23–39. Nascimento, W. S., Araujo, A. S., Gurgel, L. L., Yamamoto, M. E., Chellappa, Barros, N. H. C., Souza, A. A., & Chellappa, S. (2013). Seasonal changes in N. T., Rosa, R. S., & Chellappa, S. (2011). Endemic fish communities and condition factor, gonadosomatic and hepatosomatic indices of the pro- environmental variables of the Piranhas-­Assu hydrographic basin in the togynous marbled swamp eel, Synbranchus marmoratus. Biology Brazilian Caatinga Ecoregion. Animal Biology Journal, 2, 97–112. Journal, 4, 15–25. Nelson, J. (1994). Fishes of the World, 3rd edn (600 pp.). New York, NY: Beamish, R. J., & Fournier, D. A. (1981). A method for comparing the pre- Wiley. cision of a set of age determinations. Canadian Journal of Fisheries and Oliveira, A. P. N., & Zaniboni-Filho, E. (2009). Length-­weight relationships Aquatic Science, 38, 982–983. of fish species caught in the upper Uruguay River, Brazil. Journal of Beets, J., & Friedlander, A. (1999). Evaluation of a conservation strategy: A Applied Ichthyology, 25, 362–364. spawning aggregation closure for red hind, Epinephelus guttatus, in the Petracchi, C. A., & Guerrero, G. A. (1999). The sagitta of Synbranchus mar- US Virgin Islands. Environmental Biology of Fishes, 55, 91–98. moratus Bloch, 1795 (Synbranchiformes). Neotropica (La Plata), 45, Britski, H. A., Silimon, K. Z. S., & Lopes, B. S. (1999). Peixes do Pantanal. In 113–114. H. A. Britski (Ed.), Manual de identificação (184 pp.). Brasília: EMBRAPA Reinboth, R. (1983). The peculiarities of gonad transformation in teleosts. SPI. ISBN: 9788573830538. Differentiation, 23(Suppl.), 82–86. Brown-Peterson, N. J., Wyanski, D. M., Saborido-Rey, F., Macewicz, B. J., & Sadovy, Y., & Shapiro, D. Y. (1987). Criteria for the diagnosis of hermaphro- Lowerre-Barbieri, S. K. (2011). A standardized terminology for describ- ditism in fishes. Copeia, 1, 136–156. ing reproductive development in fishes. Marine and Coastal Fisheries, Shibatta, O. A. (2003). Family Pseudopimelodidae. In R. E. Reis, S. O. 3, 52–70. Kullander & C. J. Ferraris Jr (Eds.), Check list of the freshwater fishes of Campana, S. E. (1990). How reliable are growth back-­calculations based South and Central America (pp. 401–405, 729). Porto Alegre, Brazil: on otoliths? Canadian Journal of Fisheries and Aquatic Science, 47, EDIPUCRS. ISBN 8574303615. 2219–2227. Smith, M. P., Sansom, I. J., & Repetski, J. E. (1996). Histology of the first fish. Carruth, L. L. (2000). Freshwater cichlid Crenicara punctulata is a protogy- Nature, 380, 702–704. nous sequential hermaphrodite. Copeia, 2000, 71–82. Spadella, M. A., de Castilho-Almeida, R. B., Quagio-Grassiotto, I., & Cesario, Charnov, E. L. (1982). The theory of sex allocation (355 pp.). Princeton, NJ: M. D. (2005). Follicular diameter range based on morphological features Princeton University Press. ISBN 9780691083124. in Synbranchus marmoratus (Bloch, 1795) (Teleostei, Synbranchiformes, Chilton, D. E., & Beamish, R. J. (1982). Age determination methods for Synbranchidae) from the south-­central region of Brazil. Tissue and Cell, fishes studied by the groundfish program at the Pacific Biological 37, 91–100. Station. Canadian Special Publication of Fisheries and Aquatic Sciences, Vicentini, R. N., & Araújo, F. G. (2003). Sex ratio and size structure of 60, 102. Micropogonias furnieri (Desmarest, 1823) (Perciformes, Sciaenidae) Easter, E. E., & White, J. W. (2016). Spatial management for protogynous in Sepetiba Bay, Rio de Janeiro, Brazil. Brazilian Journal of Biology, 63, sex-­changing fishes: A general framework for coastal systems. Marine 559–566. Ecology Progress Series, 543, 223–240. Wallace, R. A., & Selman, K. (1981). Cellular and dynamic aspect of oocyte Edwards, A. W. F. (1998). Natural selection and sex ratio: Fisher’s sources. growth in teleosts. American Zoologist, 21, 325–343. American Naturalist, 151, 564–569. Watson, D. L., Anderson, M. J., Kendrick, G. A., Nardi, K., & Harvey, E. S. Froese, R. (2006). Cube law, condition factor and weight-­length relation- (2009). Effects of protection from fishing on the lengths of targeted and ships: History, meta-­analysis and recommendations. Journal of Applied non-­targeted fish species at the Houtman Abrolhos Islands, Western Ichthyology, 22, 241–253. Australia. Marine Ecology Progress Series, 384, 241–249. Grandcourt, E. M., Al Abdessalaam, T. Z., Francis, F., Al Shamsi, A. T., & Yoshida, H. O. (1964). Skipjack tuna spawning in the Marquesas Islands and Hartmann, S. A. (2009). Reproductive biology and implications for Tuamotu Archipelago. Fishery Bulletin, 65, 479–488. management of the orange-­spotted grouper Epinephelus coioides in the southern Arabian Gulf. Journal of Fish Biology, 74, 820–841. Grift, R. E., Rijnsdorp, A. D., Barot, S., Heino, M., & Dieckmann, U. (2003). Fisheries-­induced trends in reaction norms for maturation in North Sea How to cite this article: Barros NHC, de Souza AA, Peebles EB, and plaice. Marine Ecology Progress Series, 257, 247–257. Chellappa S. Dynamics of sex reversal in the marbled swamp eel Haas, R. E., & Recksiek, C. W. (1995). Age verification of winter flounder in (Synbranchus marmoratus Bloch, 1795), a diandric hermaphrodite Narragansett Bay. Transactions of the American Fisheries Society, 124, 103–111. from Marechal Dutra Reservoir, northeastern Brazil. J Appl Ichthyol. Hamilton, S. L., Caselle, J. E., Standish, J. D., Schroeder, D. M., Love, M. 2017;33:443–449. https://doi.org/10.1111/jai.13273 S., Rosales-Casian, J. A., & Sosa-Nishizaki, O. (2007). Size-­selective