Reproductive biology and development of sexually dimorphic structures in Aphyocharax anisitsi ( Ostariophysi : Characidae )

The reproductive period of Aphyocharax anisitsi is described through the analysis of specimens collected monthly from April 2001 to March 2002 in the arroio do Salso (30°22’27"S, 55°02’06"W), Rosário do Sul, Rio Grande do Sul, Brazil. The monthly variation of the Gonadosomatic Index mean (GSI) of both males and females allowed the recognition of a defined seasonal reproductive period occurring between September and February, showing a positive correlation with temperature and daylength. Absolute fecundity was 344.8 oocytes and relative fecundity 0.68 oocytes per mg of body weight. Absolute fecundity was correlated with standard length and total body weight. It was possible to identify two categories of males through the analysis of the number of rays bearing hooks on the anal fin: those with few or no rays bearing hooks identified as young males hatched in the last reproductive season, and those with a higher number of anal-fin rays bearing hooks identified as males that had already reproduced once or that are maturing and participating in their first reproductive period. Once developed, anal-fin rays are maintained permanently. A positive correlation was found between the GSI of maturing and mature males and the number of anal-fin rays bearing hooks from August to January. All specimens bearing gill glands were mature or undergoing maturation, and the frequency of occurrence of the gill glands increased with a rise of GSI.


Introduction
Characiformes is a group of Neotropical and African freshwater fishes that exhibits a wide range of behaviors and comprises the most morphologically diverse assemblage among ostariophysan fishes (Vari, 1998).The family Characidae, with over 1352 species (Reis et al., 2003), is the largest within the order and one of the most diverse families of Neotropical fishes.Aphyocharax is the single genus currently in the apparently monophyletic characid subfamily Aphyocharacinae,

P R O O F S
which includes a total of 10 species (Lima, 2003).
Aphyocharax anisitsi is found in the rios Paraná, Paraguay and Uruguay and the laguna dos Patos drainages in southern South America.It is a small species that reproduces by means of external fertilization.Mature males have small bony hooks on several anal and pelvic-fin rays, a putative apomorphic character shared by most characids (Malabarba & Weitzman, 2003).Mature males of this species possess gill glands, which have thus far been described in a few characid species (Burns & Weitzman, 1996;Bushmann et al., 2002).
In spite of reproduction being one of the most studied aspects of fishes, the available information is still very limited given the large number of fish species.Vazzoler & Menezes (1992) estimated that information on reproductive patterns is available for only about 10% of Neotropical characiforms.
In the current report, the reproductive cycle of Aphyocharax anisitsi is described from the analysis of the monthly variations of the gonadosomatic index (GSI) and stages of gonadal maturation in both sexes.We also determine female absolute and relative fecundity, as well as the influence of water temperature, day-length, rainfall, and stomach repletion index (SRI) on the reproductive cycle.Male secondary sex characters (gill glands and anal-fin hooks), which develop during sexual maturation, are also described.

Material and Methods
Specimens of both sexes were collected monthly from April 2001 to March 2002 in the arroio do Salso (30 o 22'27" S, 55 o 02'06" W), a stream of the rio Uruguay drainage, Rosário do Sul, Rio Grande do Sul, Brazil.This stream has a sandy bottom and slow current flow.Marginal vegetation is sparse and the abundance of floating macrophytes moderate.
Specimens were obtained with a seine net (5 mm mesh), and preserved in 10% formalin just after capture.Water temperature, oxygen saturation, pH and conductivity were measured in the field during each collection.Specimens were left in fixative a minimum of 10 days, and later transferred to 70% ethanol.Standard length (SL) and total body weight (Wt) were measured for each specimen, as well as gonad weight (Wg) and stomach weight (Ws) after dissection.All weights were calculated to the nearest 0.0001 g.Gonadosomatic index (GSI) and stomach repletion index (SRI) were calculated according to Vazzoler (1996): GSI = (Wg x 100)/ Wt and SRI = (Ws x 100)/Wt.
Total fecundity was estimated by counting the number of yolky oocytes in both ovaries of 10 mature females, and relative fecundity by dividing the total number of oocytes by total body weight (Adebisi, 1987).Stages of gonadal maturation were determined through macroscopic analysis following Azevedo et al. (2000) and confirmed through histological analysis, when necessary.For histology, formalin fixed gonads were dehydrated in an ethanol series, embedded in paraffin, sectioned at 0.5 µm and stained with hematoxylin-eosin.The fusion of gill filaments during gill gland formation was determined using scanning electron microscopy (SEM).For SEM analysis, the first gill arch of formalin fixed and alcohol preserved specimens was extracted, dehydrated with critical point dryer and covered with coat and gold.A stereomicroscope was used to analyze the presence and number of rays bearing hooks on the anal and pelvic fins.
Correlations were calculated between GSI and abiotic factors (water temperature, day-length and rainfall) and one biotic factor (SRI) using the non-parametric test of Spearman.Correlations were also calculated between absolute fecundity and the biometric factors (SL and Wt) using the correlation test of Pearson.Voucher specimens are deposited at Departamento de Zoologia of the Universidade Federal do Rio Grande do Sul fish collection, and numbered as UFRGS 6485.

Results
A total of 285 specimens were analyzed, comprising 130 males .Male GSI means ranged from 0.31 (March) to 1.09 (December) and female GSI means ranged from 0.45 (April) to 6.14 (January).The monthly variation of GSI of both sexes (Fig. 1) indicated a well defined reproductive period extending from September through February, which corresponds to spring and summer in the southern hemisphere.The monthly variation of mean GSI of both sexes was significantly correlated with the monthly variation of water temperature (r = 0.6200, p = 0.0347 for males and r = 0.5884, p = 0.0489 for females) and day-length (r = 0.7692, p = 0.0034 for males and r = 0.7622, p = 0.0055 for females) (Fig. 1).A significant negative correlation with rainfall was found for females only (r= -0.7343, p= 0.0065).No significant correlation was found between GSI and SRI for either sex (Table 1).Absolute fecundity ranged from 203 to 491 oocytes (mean = 344.8± 116.4), and the mean relative fecundity was 0.68 ± 0.14 oocytes/mg Wt.Absolute fecundity was significantly correlated with SL (r= 0.7693, p= 0.0093) and Wt (r = 0.8337, p = 0.0027).

P R O O F S
The number of anal-fin rays in males ranged from iv (unbranched) + 15 to 20 (branched).Maximal development of bony hooks on this fin was represented by the presence of hooks in the last unbranched ray and along all branched rays.All mature males examined from October through February, when the highest GSI means were seen, had a high number (11-20) of anal-fin rays with bony hooks (Fig. 2).From April through September 2001 and March 2002, however, it was possible to discern two categories of males: those with few to no rays (0-8) bearing hooks and those with a high number (11-20) of rays with hooks (Fig. 2).The range of body size of males with no hooks or fewer rays (0-8) bearing hooks is similar to and largely overlaps the range of body with a greater number of anal-fin rays (11-20) bearing hooks (19.6-28.2mm SL, mean 24.7 mm, vs. 21.5-31.6mm SL, mean 27.8 mm, respectively).Body size of males with few or no hooks, however, was significantly smaller than those with a greater number of rays bearing hooks when submitted to a Mann-Whitney U test (U = 456, p = 0.001).A significant correlation between GSI and number of anal-fin rays with hooks (r = 0.7409, p < 0.001) was found for males from August 2001 through January 2002, when we observed a progressive increase in the number of anal-fin rays bearing hooks.
Males with few to no hooks on the anal-fin rays from April through September 2001 and March 2002 were identified as young males that were probably hatched during the previous reproductive season.The presence of hooks on at least some fin rays suggest that some males had begun to mature, but once environment conditions (day-length, temperature) become unfavorable, further maturation was inhibited.Males with a high number of anal-fin rays with hooks captured outside of the reproductive period, i.e., from April to August 2001 and March 2002, have apparently already reproduced during the previous season.It appears that, once developed, the hooks remain on the anal-fin rays permanently.
Gill glands were either observed or not in males of A. anisitsi.All specimens that possessed distinct gill glands were males that were already mature or undergoing sexual maturation.When present, gill glands were formed from the fusion of two to five gill filaments on the first gill arch on either side of the body (Fig. 3).Higher mean GSIs were associated with a greater number of males possessing gill glands (Fig. 4).However, immature, maturing and even some mature males lacked discernible gill glands.

Discussion
According to Vazzoler and Menezes (1992), most characiforms, including several characid species, spawn during spring and summer.The peak reproductive period of Aphyocharax anisitsi also occurred during spring and summer when temperatures and day-lengths were both maximal.Indeed, in both sexes, mean monthly GSIs were significantly correlated with both water temperature and day-length, which shows an annual variation of approximately four hours.An-  nual day-length variations as small as 1.6 hours have been shown to significantly influence reproduction in low-latitude fishes (Burns, 1985;Burns & Flores, 1981).In temperate southern Brazil, the site of our collections, environmental conditions during spring and summer appear to be most favorable for the survival of the fry, when food availability is greatest.
In tropical regions, where the annual variations of water temperature and day-length may be minimal, rainfall appears to be a major environmental factor influencing reproductive cycling (Vazzoler & Menezes, 1992).In these areas, the increase in water levels provides better shelter for the fry, as well as a more abundant food supply.On the other hand, in southern Brazil, where no definitive pattern in monthly precipitation is evident, fish reproductive cycles are generally not correlated with rainfall (Vazzoler & Menezes, 1992).This is the case with several characid species from southern Brazil (Azevedo, 2000;Azevedo et al., 2000;Oliveira et al., 2002).However, in the present study, the mean monthly GSIs of female Aphyocharax anisitsi did show a significant negative correlation with monthly rainfall levels, although male GSIs showed no such pattern.

Fig. 1 .
Fig. 1.Monthly variation of the mean values of GSI (± SD) in males (a) and females (b) of Aphyocharax anisitsi collected from April 2001 to March 2002 in the arroio do Salso, Rosário do Sul, RS, Brazil, and monthly variation of the day-length (a) and water temperature (b) mean values.(GSI = Gonadosomatic Index; SD = Standard Deviation represented by vertical bars).

Fig. 2 .
Fig. 2. Monthly variation of the number of anal fin rays bearing hooks in males of Aphyocharax anisitsi between April 2001 and March 2002 in the arroio do Salso, Rosário do Sul, RS, Brazil (each dot may represent more than one specimen).At the bottom, males with few or no hooks identified as young males that were probably hatched during the previous reproductive season.At the top, males that matured during previous reproductive season or are maturing for the first time.

Fig. 3 .
Fig. 3. Lateral view of the ventralmost portion of the first gill arch of two males of Aphyocharax anisitsi collected in February 2001 in the arroio do Salso, Rosário do Sul, RS. (A) specimen with the initial development of the gill gland seen through the fusion of two gill filaments.(B) specimen with a fully developed gill gland seen through the fusion of five gill filaments (SEM).

Fig. 4 .
Fig. 4. Monthly variation of the GSI mean values and frequency of occurrence (%) of males of Aphyocharax anisitsi bearing gill glands, collected from April 2001 to March 2002 in the arroio do Salso, Rosário do Sul, RS, Brazil.

Table 1 .
Monthly mean values of GSI (± SD) and RSI (± SD)for males and females of A. anisitsi collected from April 2001 to March 2002 in the arroio do Salso, Rosário do Sul, RS, Brazil.GSI = Gonadosomatic Index; RSI = Stomach Repletion Index; SD = Standard Deviation.

Table 2 .
Absolute and relative mean fecundity values (± Standard Deviation) of small size characids.Relative fecundity values = Number of oocytes / mg of total body weight.