INTRODUCTION
Starfish species belong to the phylum Echinodermata and class Asteroidea, and about 2100 extant species are currently known (Pawson, Reference Pawson2007), of which the majority is known to be dioecious or gonochorics, and they reproduce sexually with external fertilization (Hyman, Reference Hyman1955; Lawrence, Reference Lawrence1987; Chia & Walker, Reference Chia, Walker, Giese and Pearse1991; Hendler et al., Reference Hendler, Miller, Pawson and Kier1995; Lawrence & Herrera, Reference Lawrence and Herrera2000). Although there are still inquiries about the factors that control reproduction of starfish, in a number of studies some answers have been found, providing evidence on the close relationship between the reproductive strategy and environmental, physiological and phylogenetic factors. Farmanfarmaian et al. (Reference Farmanfarmaian, Giese, Boolootian and Bennett1958) described the reproductive cycles of four species of asteroids from the coast of California using the gonad and pyloric caecal indices, which showed an inverse relationship indicating the translocation of nutrients from pyloric caeca to the gonads. A number of studies have been carried out with asteroids in different parts of the world (see Chia & Walker, Reference Chia, Walker, Giese and Pearse1991; Georgiades et al., Reference Georgiades, Temara and Holdway2006 for most of them; also Benítez-Villalobos et al., Reference Benítez-Villalobos, Díaz-Martínez and Tyler2007; Pastor de Ward et al., Reference Pastor-de-Ward, Rubilar, Díaz-De-Vivar, González Pisani, Zarate, Kroeck and Morsan2007; Benítez-Villalobos & Diaz Martinez, Reference Benítez-Villalobos and Díaz-Martínez2009), revealing different patterns in gonad and pyloric caecum development, but most of them resembling the results obtained by Farmanfarmaian et al. (Reference Farmanfarmaian, Giese, Boolootian and Bennett1958).
Seasonality of ecosystems increases gradually from the tropics to highly seasonal regions of the North and South Poles. This pattern mainly follows the annual cycle of solar energy received, and it is exacerbated in the marine environment by the oceanic ice in winter (Priddle et al., Reference Priddle, Hawes, Ellis-Evans and Smith1986). This fact is important in regulating the reproductive activity of echinoderms, because some dissimilarity has been found in the patterns of reproductive periodicity, which reflect the seasonal differences of the environment they inhabit (Byrne et al., Reference Byrne, Morrice and Wolf1997; Lessios, Reference Lessios2003). It has been observed that species from the poles or high latitudes show a notable seasonality in reproduction (Stanwell-Smith & Clarke, Reference Stanwell-Smith and Clarke1998), which decreases in species from mid-latitudes (Carvalho & Ventura, Reference Carvalho and Ventura2002; Rubilar et al., Reference Rubilar, Pastor de Ward and Díaz de Vivar2005) and is not evident in tropical species (Pearse, Reference Pearse1968; Guzmán & Guevara, Reference Guzmán and Guevara2002).
The reproductive biology of asteroids has been widely documented for different locations around the world, however, publications on tropical shallow-water species are scarce, and they have been carried out mainly in the Indo-Pacific and Caribbean.
This paper presents an analysis of the reproductive biology of Pharia pyramidatus, a starfish belonging to the family Ophidiasteridae. It is a species from shallow tropical waters, it is distributed from California to Peru as well as the Galápagos Islands, inhabiting rocky and sandy bottoms and associated with coral reefs (Caso, Reference Caso1943).
Pharia pyramidatus is considered a generalist predator, which preys mainly on molluscs, crustaceans, algae (Palma-Ruíz, Reference Palma-Ruíz2003), sponges, hydroids and bryozoans (Glynn & Wellington, Reference Glynn and Wellington1983), and it also has been reported as corallivore, mostly feeding on species of the genus Pocillopora (Dana & Wolfson, Reference Dana and Wolfson1970). The ecological importance of P. pyramidatus is related to the role that it plays in the trophic network as a predator of other populations, contributing significantly to the flux of matter and energy of the ecosystems in which it inhabits, especially in those highly productive ecosystems like coral reefs.
Economically, P. pyramidatus has become relevant in recent years, as it is among the species of marine invertebrates with the most important demand in Mexico for use as ornaments. The commercial extraction of P. pyramidatus started in the mid-1960s, but according to the Mexican National Institute for Fisheries, an historical record of catches does not exist (Piña-Espallargas, Reference Piña-Espallargas2004).
In light of the growing rate of extraction of species for ornamental purposes, it is necessary to improve the knowledge about this potential resource. Also the analyses of the reproductive patterns of P. pyramidatus will contribute to knowledge about its reproductive strategy as well as the importance of the environmental factors that generate those patterns, which become very significant in the context of the relative rapid change that the ecosystems are currently undergoing, especially those of great ecological value like coral reefs, where P. pyramidatus stands as a key species for the control of other populations.
MATERIALS AND METHODS
Sampling
About twenty individuals of Pharia pyramidatus were haphazardly collected by SCUBA diving monthly between April 2008 and May 2009, from a depth interval of 3–15 m on the Isla Montosa reef community (15°45′48″N 96°04′57″W), on the Pacific Coast of southern Mexico. During each monthly collection, salinity and pH of the water were measured using a waterproof multiparameter meter, and temperature was recorded at one hour intervals using a pendant HOBO® data logger, permanently attached to the bottom. Data of pluvial precipitation and photoperiod (light hours) were obtained from the weather centre of the Water National Commission (CNA) located in Puerto Angel, Oaxaca. Monthly means were calculated for every parameter recorded.
Gonad general structure and gametogenic cycle
All individuals were narcotized by leaving them in seawater at 2°C for 24 hours before dissection. The whole animals were damp-dried and weighed to the nearest mg. The major radius (R: centre of disc to the tip of the largest arm) and minor radius (r: centre of disc to the interradius) were measured for every individual to the nearest mm. The gonads and gut caeca of the five arms were dissected out by opening the oral side of the disc along each radius. Each organ was damp-dried and weighed separately to the nearest mg, and subsequently fixed in Bouin's solution and stored in 70% ethanol.
The gonad index percentage (gonads weight/total body weight ×100) and pyloric caecum index percentage (caeca weight /total body weight ×100) were calculated for each individual. These data were summated for each sample, and the mean and 95% confidence limits calculated in order to detect seasonal reproduction.
The volumes of the gonads were measured to the nearest 0.001 ml by the displacement of fluid method, using a variation of the hydrostatic balance of Mohr–Westphal (Scherle, Reference Scherle1970). The gonads were processed in paraffin wax, sectioned at 7 µm and stained with haematoxylin and eosin.
Five stages in gonad development were described, based on the frequency and characteristics of gametes, presence of phagocytes and cell remains, as well as thickness and shape of the gonad wall. Gonad stages were based on the developmental scales used with other asteroid species (McClary & Mladenov, Reference McClary and Madlenov1989; Byrne et al., Reference Byrne, Morrice and Wolf1997; Carvalho & Ventura, Reference Carvalho and Ventura2002; Rubilar et al., Reference Rubilar, Pastor de Ward and Díaz de Vivar2005).
Oogenesis and spermatogenesis were described using the sections. At least 100 oocytes per gonad, sectioned through the nucleolus, were measured (Feret diameter) using the image analysis package Image J. Oocyte size–frequency diagrams were constructed to identify evidence of seasonal reproduction. To compare the oocyte size–frequency data from different samples, the data were plotted into 20 µm classes. Two slides per male were prepared to identify the developmental stage of sperm.
Fecundity was quantified as the number of vitellogenic oocytes per female (actual fecundity), from the mean volume of oocytes and the volume of the ovary in each female. Oocyte volume (OV) is calculated assuming a spherical shape [OV= (4· × π × ·R3)/3] and averaged. We estimated the volume of the gonadal fluid that solidifies during fixation from ten slides in ten females as it was proposed by Ramirez-Llodra et al. (Reference Ramirez-Llodra, Tyler and Billett2002) and concluded that it occupies approximately 15% of the ovary. This percentage was extracted from the total gonad volume.
Fecundity was estimated as follows:
where Vg = volume of the gonad, Vpvo = mean volume of a previtellogenic oocyte, Vvo = mean volume of a vitellogenic oocyte, P = ratio between previtellogenic and vitellogenic oocytes: P = Npvo/Nvo, Npvo= number of previtellogenic oocytes counted on a subsample of 100 oocytes per gonad, Nvo= number of vitellogenic oocytes counted on a subsample of 100 oocytes per gonad.
Sex-ratio was tested by the Chi-square (χ 2) test and statistical differences in gonad index (GI) and pyloric caecum index (PCI) were tested using two-way analysis of variance (ANOVA) (sex and month as factors), and differences in fecundity for females were tested also using ANOVA. In order to know the relationship between salinity, temperature, pluvial precipitation and photo-period (independent variables) and the reproductive cycle of P. pyramidatus represented by GI and PCI (dependent variables), we performed a step by step multiple lineal regression analysis.
RESULTS
Sex-ratio and gonad morphology
Pharia pyramidatus did not show sexual dimorphism and it was confirmed that it is a gonochoric asteroid as not one of the 225 examined organisms was hermaphrodite. The ratio of males to females did not differ significantly, showing a 1:1 proportion throughout the year. The total number of individuals sexed was 218, with 107 males and 111 females (χ 2= 0.073, P > 0.05). The gonads were located in the coelomic cavity at the actinolateral zone underneath the pyloric caeca, and showed a serial arrangement, forming two rows of individual clusters (acini) running along every arm and separated by the ambulacral groove. Every acinus stayed attached to the lateral plates of the body wall by means of individual short gonoducts that conducted to their respective gonopore at the lateral papular zone.
The number of acini varied depending on the length of the arm, the largest specimens, with major radius (R) about 16 cm long, presented up to 26 clusters, whereas the smallest (R ≈ 8 cm) showed between 10 and 12 clusters. All the acini were equally small in an individual at the growing stage, but at the mature stage, the size of the acini was largest at the proximal zone and decreased gradually towards the distal zone, filling the coelomic cavity that was not occupied by the pyloric caeca (about 60%).
The microscopic analysis allowed for differentiation of the structural components of the gonad wall. Similarly to the majority of asteroids, ovaries and testes of P. pyramidatus possessed three layers of tissue: the inner and outer sacs and the genital coelomic sinus (perihaemal sinus) that separate them (Figure 1).
Fig. 1. Transverse section of testis showing the general arrangement of the tissues in the gonad wall. is, inner sac; lu, lumen; os, outer sac; ps, perihaemal sinus. Scale bar = 30 µm.
The outer sac was constituted by two layers of epithelial tissue separated by a layer of connective tissue, which showed a variable width. Conformation of the external layer of epithelium changed according to the gonad developmental phase, being formed by flat, cubic or cylindrical cells. The inner sac contained the haemal sinus, which was clearly visible during the active phase of the germinal epithelium when gametogenesis was activated.
Ovary development
GROWING STAGE
There was a predominance of previtellogenic oocytes with basophilic cytoplasm that stained dark purple with haematoxylin. The outer sac appeared extended and thin; the inner sac looked very thin and had a reduced perihaemal sinus. As this phase advanced, vitellogenic oocytes predominated and they had an eccentric nucleus and acidophilic cytoplasm that stained light pink with eosin. The gonad wall appeared thin and there were visible epithelial cells, the perihaemal sinus was occasionally visible (Figure 2A).
Fig. 2. Light histology of ovaries. (A) Growing stage: ovary showing the distribution of oocytes, the small oocytes remain at the periphery and the vitellogenic oocytes are accumulated in the lumen; (B) mature stage: mature ovary filled with vitellogenic oocytes; (C) spawned stage: ovary showing empty spaces due to gamete release; (D) spent stage: ovary almost empty and contracted with relict oocytes; (E) recovery stage: ovary containing cellular debris and phagocytes; (F) detail of recovery ovary with oogonia proliferating in the germinal epithelium. e, empty space; is, inner sac; lu, lumen; o, oogonia; os, outer sac; p, phagocytes; po, previtellogenic oocyte; ro, relict oocytes; vo, vitellogenic oocytes. Scale bar = 100 µm
MATURE STAGE
Ovaries were filled with agglomerated vitellogenic oocytes (110–130 µm diameter) resulting in an oval shape of the oocytes located at the lumen, while the ones at the periphery showed an almost rectangular shape. A pale thin meshwork of tissue surrounded each oocyte. Very scarce previtellogenic oocytes and oogonia remained. The gonad wall showed the thinnest and smoothest appearance of the entire reproductive cycle (Figure 2B).
SPAWNED STAGE
This stage was characterized by the presence of empty spaces in the lumen of the gonads, which evidenced the liberation of full-grown gametes. Oocytes with maximum development predominated in the ovaries. There were a few early previtellogenic oocytes and some atresic oocytes. The gonad wall remained thin but it started to show slight undulations (Figure 2C).
SPENT STAGE
All ovaries appeared almost empty and contracted, containing relict full-grown oocytes, most of them atresic or not well developed. In the areas where most of the occytes were present, the gonad wall remained thin, whereas in those areas where the lumen was almost empty, the outer sac started widening and undulating, the perihaemal sinus was evident, and the first phagocytes were observed (Figure 2D).
RECOVERY STAGE
There still existed relict oocytes with a cracked appearance and diffuse nuclei in the gonads. As the amount of cell remains increased, a proliferation of phagocytes was observed forming irregular agglomerations in the lumen of gonads. The gonad wall showed the thickest appearance of the entire reproductive cycle, the outer sac showed many undulations and the inner sac had a smooth appearance. Oogonia started to proliferate from the germinal epithelium (Figure 2E, F).
Testis development
GROWING STAGE
Spermatogonia (~4.2 µm) proliferated and they were organized in columns and through mitosis produced spermatocytes (~3 µm), which extended to the lumen forming colonettes. Spermatids (~2 µm) differentiated from the apices of the colonettes and gave rise to spermatozoa (~1.4 µm), which concentrated in the lumen of the gonads. The gonad wall appeared thin, the inner sac was extended and the epithelial cells were visible (Figure 3A, B).
Fig. 3. Light histology of testes. (A) Growing stage: testis showing the invaginations of the germinal epithelium in the lumen; (B) detail of growing testis; (C) mature stage: testis filled with spermatozoa and thin spermatogenic columns; (D) spawned stage: testis showing empty spaces due to gamete release; (E) spent stage: contracted testis with residual spermatozoa; (F) recovery stage: proliferation of phagocytes in the lumen. c, colonette; e, empty space; I, invagination; is, inner sac; lu, lumen; os, outer sac; p, phagocytes; ps, perihaemal sinus; rz, residual spermatozoa; s, spermatids; sc, spermatogenic columns; se, spermatocytes; sg, spermatogonia; z, spermatozoa. Scale bars: A–B = 25 µm; C–F = 100 µm.
MATURE STAGE
Almost the entire lumen was occupied by gametes ready to be spawned. The different layers of the gonad wall were not distinguishable. Spermatozoa were highly abundant and compressed so that they had the appearance of a purple uniform tissue, surrounded by a thin layer of spermatogenic columns (Figure 3C).
SPAWNED STAGE
A number of empty spaces with the appearance of fine striations were visible among the dense mass of spermatozoa. The inner and outer sacs were visible again with slight undulations, as well as the perihaemal sinus (Figure 3D).
SPENT STAGE
The testes appeared contracted and the lumen was almost empty. There were a few relict spermatozoa and the first phagocytes appeared. The outer sac started to thicken and showed undulations; a wide perihaemal sinus was visible (Figure 3E).
RECOVERY STAGE
The outer sac formed ample undulations and the perihaemal sinus diminished noticeably. Phagocytes at the periphery of the lumen proliferated and dispersed centrally, and were progressively abundant until most of the cell remains were eliminated. At the end of this stage the outer sac diminished in undulations, the perihaemal sinus expanded notably and the first spermatogonia were visible (Figure 3F).
Reproductive cycle
In females, the growing stage was observed from January to June. In January, 73% of females were at this stage and the most dominant size interval of oocyte was 30–50 µm. The percentage of females in the growing stage grew progressively to 91% in April, which coincided with a high abundance of previtellogenic (30–70 µm) oocytes followed by early vitellogenic (70–110 µm) oocytes. During May and June, 100% of females were mature, and most oocytes were in early vitellogenic stages. In July–August, 30% of females were at the mature stage and the dominant size interval of oocytes was 110–130 µm, followed by the adjacent intervals. Also in this period, 50% of females started spawning and continued until September. During the spawning period, the dominant size intervals of oocytes were similar to those of the previous stage. The spent stage was observed first in July–August in 20% of females, during September, it increased to 38% and decreased in October to 25% of females. During October, a noticeable change was observed in oocyte size intervals because oogonia (10–30 µm), which had remained in low numbers increased visibly, followed by a few relict full-grown oocytes (90–110 µm). The recovery stage was recorded in October in 75% of females, during November and December all females were at this stage and in January the frequency decreased to 27%. During this period, the oocyte size distribution was very similar; oogonia were the most dominant frequency, followed by previtellogenic oocytes (30–50 µm) and practically no vitellogenic oocytes (Figures 4A & 5).
Fig. 4. Gametogenic cycle of Pharia pyramidatus from April 2008 to May 2009. (A) Females; (B) males.
Fig. 5. Oocyte diameter frequency distribution of Pharia pyramidatus (April 2008–May 2009). Grey bars are modal size.
In males, the growing stage also started in January with a frequency of 40%. During February and March the frequency of males in this stage increased, and during April and May it was observed in the totality of organisms. In June 75% of males were mature while the rest were still in the growing stage. The highest intensity of spawning was observed during July and August, when testes of 83% of males were at this stage. During September, 21% of individuals were still spawning and during October, 14% of males still presented a considerable portion of full-grown gametes ready to be liberated. During September 36% of males were at the spent stage, the same percentage was observed in October, and in November the highest frequency was recorded with 43% of males at this stage. The percentage of males at the recovery stage in September, increased slightly during October and November. During December the gonads of the totality of males were at this stage (Figure 4B).
Gonad and pyloric caecum indices
From January (ẋ = 1.6; SE = 1.1; N = 20), and through the following months, a constant increase was observed in GI values for females, indicating that ovaries were at the growing stage. The highest value was reached in July–August (ẋ = 10.4; SE = 4.5; N = 20), when organisms initiated spawning. In September there was a high variability in GI values (ẋ =7.7; SE = 4.5; N = 20) as a result of spawning, characterized by the presence of organisms both mature and totally spawned. At the end of the spent stage (October), GI reached the lowest values (ẋ = 0.31; SE = 0.1; N = 20). From October to December, values were low, coinciding with the recovery stage (Figure 6A).
Fig. 6. Monthly variation of gonad index (GI) and pyloric caecum index (PCI) of Pharia pyramidatus. (A) Females; (B) males. Vertical bars indicate standard error.
In males values of GI also started to increase from February (ẋ = 0.3; SE = 0.1; N = 20) until July–August when it reached a peak (ẋ = 7.1; SE = 2; N = 20). In September it dropped abruptly (ẋ = 1.6; SE = 0.1; N = 20) as a consequence of the spawning and followed this trend until December, when the lowest value was recorded (ẋ = 0.2; SE = 0.1; N = 20) (Figure 6B). There were significant differences between months (ANOVA, F(10,196) = 29.13, P < 0.001) and sexes (ANOVA, F(1,196) = 40.43, P < 0.001), with females possessing the highest and most variable values (Figure 6).
From April to May 2008, for both sexes, a slight increase in PCI values was recorded, reaching a monthly mean of around 8, but in June it decreased (ẋ = 5.0; SE = 2.4; N = 20 for females and ẋ = 4.8; SE = 3.0; N = 20 for males). This trend continued until October, when PCI reached the lowest values (ẋ = 3.8; SE = 1.9; N = 20 for females and ẋ = 4.0; SE = 1.2; N = 20 for males). From November and through the subsequent months, values increased gradually until April–May, when the maximum values were reached for females (ẋ = 12.6; SE = 3.1; N = 20) and males (ẋ = 15.0; SE = 3.5; N = 20). There were no significant differences between sexes (ANOVA, F(1,196) = 0.451, P > 0.05), but there were differences between months (ANOVA, F(10,196) = 34.1, P < 0.001). The general GI showed a negative relation (β = −0.25) with the general PCI (Figure 6).
Fecundity
Actual mean (± SE) fecundity of Pharia pyramidatus was 7.3 ± 1.2 million of eggs per female. Fluctuations in fecundity values behaved similarly to those of the GI values in females. From April to June, an increase was observed in fecundity, reaching a maximum of 18.5 ± 1.7 millions of eggs. Starting in August, the values decreased continuously until October (0.28 ± 0.1 million). During November and December, no vitellogenic oocytes were recorded; therefore actual fecundity was zero for both months. In January, fecundity was 1.5 ± 0.2 million, and the values increased until April–May (7.5 ± 0.9 million). There were significant differences in monthly mean values of fecundity (ANOVA, F(8,74) =10.95, P < 0.001).
Environmental variables
The multiple regression analysis of GI with respect to the environmental variables was performed in two steps, resulting in the following model:
The model showed that the total explained variance over the GI, in terms of the selected environmental variables was 86.95% (R2 = 0.8695, R = 0.9325, P multiple < 0.05), indicating that the variables photoperiod (P) and bottom water temperature (T) are correlated 93.25% with GI (Figure 7). The first variable that went into the model was P providing 77.36% of the total explained variance, whereas at the second step T contributed to 9.6%.
Fig. 7. Monthly reproductive changes compared to environmental variables. (A) Mean values of gonad index (GI) and pyloric caecum index (PCI) of total population of Pharia pyramidatus compared with photoperiod; (B) monthly variation of bottom seawater temperature and pluvial precipitation.
Regarding the PCI the analysis was performed in two steps, resulting in the following model:
This model showed that the total explained variance in terms of the selected environmental variables was 68.16% (R2 = 0.6816, R = 0.8256, P multiple < 0.05). The variables T and P showed a multiple correlation value of 82.56% with the PCI (Figure 7). At the first step, the total explained variance was 40.32% as a result of the variable T and at the final step the variable P contributed to 27.84%.
Although pluvial precipitation was not considered as a significant variable by the model, when a simple lineal regression analysis was performed, the result showed that this variable had a significant correlation (R =0.80, R 2 = 0.64, P < 0.01) with GI (Figure 7B).
DISCUSSION
As it has been reported for other species (Kim, Reference Kim1968; Rubilar et al., Reference Rubilar, Pastor de Ward and Díaz de Vivar2005), the maturity stage of the gametogenic cycle of P. pyramidatus was the shortest and the growing stage was the most extensive, because it included vitellogenesis, which is considered the longest and most energetically expensive process of ovogenesis (Ramirez-Llodra, Reference Ramirez-Llodra2002). The spawning period is highly variable and it can last two (Chen & Chen, Reference Chen and Chen1992; Rubilar et al., Reference Rubilar, Pastor de Ward and Díaz de Vivar2005), three (Kim, Reference Kim1968; Stanwell-Smith & Clarke, Reference Stanwell-Smith and Clarke1998), four (Scheibling, Reference Scheibling1981; Tyler & Pain, Reference Tyler and Pain1982; Byrne et al., Reference Byrne, Morrice and Wolf1997) or up to five months (Carvalho & Ventura, Reference Carvalho and Ventura2002), according to the degree of seasonality of the species and synchrony of the population.
The histological analysis showed that females of P. pyramidatus reached the maturity stage slightly later than the males, and remained mature for a longer period as it has been reported for other asteroid species (Byrne et al., Reference Byrne, Morrice and Wolf1997). Nevertheless, gametogenesis in males and females occurred simultaneously and both sexes underwent a massive spawning during approximately three months (July–September). Hence, individuals of the population showed a high synchronicity, which is essential for spawning, in order to guarantee reproductive success (Mercier & Hamel, Reference Mercier and Hamel2009). In contrast with P. pyramidatus, Pearse (Reference Pearse1968) found that gametogenesis of Linckia laevigata was very asynchronous among individuals, and suggested that some individuals were reproductively active all year, at least when the populations were near the equator. This fact is interesting because L. laevigata is another tropical starfish closely related to P. pyramidatus as it belongs to the family Ophidiasteridae; however, their reproductive patterns are dissimilar.
Production of eggs of P. pyramidatus is shown to be seasonal rather than continuous, evidence that the females must concentrate their energy in order to make use of it in a single annual generation of numerous eggs. Giese & Pearse (Reference Giese, Pearse, Giese and Pearse1974) mentioned that precise breeding seasons may be important to provide larvae or juveniles with favourable environmental conditions. Baker (Reference Baker and De Beer1938) recognized two causes of reproductive seasonality: (1) exogenous factors which serve as clues to synchronize the cycles; and (2) environmental conditions that exert selective (evolutionary) pressure, favouring for a species one season for reproduction over others. Baker (Reference Baker and De Beer1938) distinguished these as the proximate and ultimate causes, respectively.
Photoperiod and sea temperature were parameters of high importance for the dynamics of reproduction of P. pyramidatus and apparently they constitute the proximate causes that generate the seasonal pattern observed, synchronizing the reproductive cycles. The statistical analyses suggested that photoperiod was the most significant variable and the gonadic maturation coincided with the increasing of day length, as it was recorded with Oreaster reticulatus (Scheibling, Reference Scheibling1981). In our study the stage of gonad reabsorption coincided with the months when the days were shorter, whereas the period of spawning occurred during the longest days. The apparent photoperiodic regulation of gametogenesis and gonadal growth in the case of P. pyramidatus coincides with what was observed for other asteroids at different latitudes (Pearse & Eernisse, Reference Pearse and Eernisse1982; Pearse & Beauchamp, Reference Pearse and Beauchamp1986; Pearse & Walker, Reference Pearse and Walker1986; Pearse & Bosch, Reference Pearse and Bosch2002; Georgiades et al., Reference Georgiades, Temara and Holdway2006). Byrne et al. (Reference Byrne, Morrice and Wolf1997) provided evidence that photoperiod regulates the gametogenic cycle of Asterias amurensis. Furthermore the authors compared reproduction of A. amurensis in Japan and Tasmania at similar latitudes, showing that gametogenesis of the northern and southern populations was six months out of phase. For this reason, it is important to carry out similar studies with P. pyramidatus in order to discern if the photoperiod has the same effect on populations of this tropical species at different latitudes.
Fluctuations in sea temperature generally act as a clue for the spawning or particular reproductive behaviours such as aggregations. For P. pyramidatus, the maximum values of sea temperature coincided with the spawning, and it is probably the external signal for that process. In another study carried out in eastern Canada, the spawning of Asterias vulgaris corresponded with a noticeable increase in sea temperature (Raymond et al., Reference Raymond, Himmelman and Guderley2007). Also the gonad maturation and spawning of the tropical asteroid Oreaster reticulatus occurred during the season of highest sea temperatures (Scheibling, Reference Scheibling1981).
Ultimate causes usually have to do with survival of offspring, and little is known about them, making them difficult to demonstrate, but most cycles are timed so that the production of larvae or juveniles must be synchronized with periods favourable for feeding or survival (Giese & Pearse, Reference Giese, Pearse, Giese and Pearse1974). On the coast of Oaxaca, there is a remarkable difference between dry (mean pluvial precipitation 0 mm) and the rainy (mean pluvial precipitation 800–1200 mm) seasons. The rainy season includes the period from May to September and this highly seasonal rainfall produces an important increase in availability of nutrients for larval food, through extensive runoffs from the main streams in the zone (especially the Copalita River). Furthermore, during the rainy season, the movement of tropical storms and hurricanes generate pulses of winds that propagate along the coast (Lavín et al., Reference Lavín, Robles, Argote, Barton, Smith, Brow, Kosro, Trasviña, Velez and García1992; Barton et al., Reference Barton, Argote, Brown, Kosro, Lavín, Robles, Smith, Trasviña and Velez1993; Trasviña & Barton, Reference Trasviña, Barton and Lavín1997), and probably the resultant currents favour the retaining of larvae near shore. On the other hand, between November and April, strong wind jets called ‘Nortes’ blow perpendicularly to the coast and generate upwelling in the Gulf of Tehuantepec. This specific oceanographic dynamic probably affects the reproductive pattern of marine organisms, carrying larvae offshore, and it is likely a condition that has been selected against. Consequently, the strong seasonal cycles of heavy rain and dry season, as well as the strong winds and the resulting upwelling in autumn–winter, are likely to constitute the ultimate causes of the reproductive seasonality of P. pyramidatus.
In this work GI and PCI showed an inverse relationship from May to August (Figure 6), which indicates a transfer of nutrients from the pyloric caeca to the gonads, as it has been documented by a number of authors for other species (Mauzey, Reference Mauzey1966; Rao, Reference Rao1966; Kim, Reference Kim1968; Crump, Reference Crump1971; Jangoux & Vloebergh, Reference Jangoux and Vloebergh1973; Ferguson, Reference Ferguson1975a, Reference Fergusonb; Barker, Reference Barker1979; Harrold & Pearse, Reference Harrold and Pearse1980; Scheibling, Reference Scheibling1981; Chia & Walker, Reference Chia, Walker, Giese and Pearse1991; Carvalho & Ventura, Reference Carvalho and Ventura2002; Rubilar et al., Reference Rubilar, Pastor de Ward and Díaz de Vivar2005; Georgiades et al., Reference Georgiades, Temara and Holdway2006). However, during five months the two indices showed no inverse relationship, probably because the available energy was abundant enough in order to allow an increase in the PCI even though gametogenesis was occurring in the gonads.
This behaviour of the two indices coincided with the period of more intense effect of the ‘Nortes' which generate upwellings in the Gulf of Tehuantepec and cause an important increase in primary production in the entire region of the Mexican south Pacific (Lavín et al., Reference Lavín, Robles, Argote, Barton, Smith, Brow, Kosro, Trasviña, Velez and García1992), providing a high availability of energy for all biological processes, in this case storage of energy in the pyloric caeca and gametogenesis in the gonads of P. pyramidatus. The intensity of the event varies from year to year and consequently so does the increase in available energy, which probably caused the pattern observed in the values of the PCI, as the results showed that the PCI was nearly twice as large in April–March of 2009 than in the same period in 2008.
When the features of the study site were first analysed it was supposed that P. pyramidatus would exhibit patterns characteristic of a continuous reproductive cycle because the fluctuations of temperature and photoperiod are apparently minimal and the availability of food is theoretically constant. However, the results of this study showed the opposite, because even though the variation of temperature (5.7°C) and photoperiod (1.9 hours) at Isla Montosa are relatively narrow, the statistical analyses showed that those variables were highly correlated (93%) with the cycle of the GI and they probably constitute the proximate causes. Although the study site of this asteroid species corresponds to a tropical locality, the local oceanic processes such as upwellings in the Gulf of Tehuantepec between November and April, and a remarkable difference between the rainy and dry seasons determine seasonal oceanographic regimes. Those regimes act as ultimate (evolutionary) causes, which influence the biological processes of the communities, determining the seasonal reproductive patterns observed in P. pyramidatus.
According to the reproductive features of P. pyramidatus (high fecundity, small eggs, relatively large adults and seasonality of gametogenesis), it is evident that the reproductive strategy of this species includes a larva with planktotrophic development, as is reported in other asteroid species studied in other parts of the world (Tyler & Pain, Reference Tyler and Pain1982; Ramirez-Llodra et al., Reference Ramirez-Llodra, Tyler and Billett2002).
ACKNOWLEDGEMENTS
This manuscript has been improved considerably by comments from two anonymous referees and Professor John Pearse. Jennifer Brown and Lindsey Franger made important suggestions to improve the English. This work was completed as part of the project Patrones Reproductivos de Tres Especies de Equinodermos Asociados a Arrecife en la Costa de Oaxaca supported by PROMEP (2IR0707).
