Spatial and seasonal variation of the gonad index of Diadema antillarum ( Echinodermata : Echinoidea ) in the Canary Islands

The sea urchin Diadema antillarum Philippi has a tropical-subtropical amphiatlantic distribution. According to Lessios et al. (2001), there are two genetically separate forms, one from the western and central Atlantic (D. antillarum-a) and one from the eastern Atlantic (D. antillarum-b). The latter form is found from the Madeira Archipelago to the Gulf of Guinea, including the Selvages, Canary and Cape Verde Islands. SCIENTIA MARINA 70(4) December 2006, 689-698, Barcelona (Spain) ISSN: 0214-8358


INTRODUCTION
The sea urchin Diadema antillarum Philippi has a tropical-subtropical amphiatlantic distribution.According to Lessios et al. (2001), there are two genet-ically separate forms, one from the western and central Atlantic (D. antillarum-a) and one from the eastern Atlantic (D. antillarum-b).The latter form is found from the Madeira Archipelago to the Gulf of Guinea, including the Selvages, Canary and Cape Verde Islands.
High densities of the sea urchin Diadema antillarum occur at the Canary Islands and Madeira (Alves et al., 2001;Brito et al., 2002;Tuya et al., 2004a, b).These urchins have produced extensive barren grounds on rocky substrata between 0.5 and 50 m depths (unpublished observations).The intensity of the effect of the sea urchins is inversely proportional to water movement (Alves et al., 2001).Barren grounds of D. antillarum are shallower in sheltered areas than in exposed areas (unpub.obs.).
A peak in the gonad index of Diadema antillarum was observed by Lewis (1966) in Barbados during low temperature periods (May 1963 andApril 1964).Bauer (1976) also found a maximum gonad index related to low temperature periods at the Florida Keys.Iliffe and Pearse (1982) suggested that high temperature periods inhibit gonad growth.Capo et al. (2001) noted that Diadema antillarum in the laboratory had more spawning episodes at 24ºC than at 26ºC.Other factors such as benthic food availability, photoperiod and population characteristics (densities) may also influence gonad growth (Lessios, 1981;Iliffe and Pearse, 1982).
In sea urchins, gametogenesis and intragonad reserves of nutrients in nutritive phagocytes (NP) are linked.Nutritive phagocytes are renewed during pregametogenesis and used in gametogenesis.In both sexes, the gonad grows based on the increase in nutrient reserves in the NP (Walker et al., 2001).The nutrient status of the sea urchin has been shown to influence gonad production in both field (Meidel and Scheibling, 1998) and laboratory studies (Garrido and Barber 2001).
Negative growth in D. antillarum is densitydependent thus allowing increased survivorship and reproductive output under resource constraints.This ability can be important for maximizing the energy available for reproduction (Levitan, 1989).
Only two studies on the reproductive periodicity of the eastern Atlantic populations of D. antillarumb have been conducted (Bacallado et al., 1987;Garrido et al., 2000).Both studies were carried out in the Canary Islands and showed a gonad index peak during the end of spring that was related to an increase in temperature.
In the Canary Islands the temperature reaches a distinct maximum in late summer and a minimum in winter in shallow waters.In winter, the upward movement of nutrients maximizes phytoplankton production (Braun et al., 1980).Whether benthic algal production is seasonal in the Canary Islands is often debated, but it seems that there is one clear peak of algal cover and algal richness during spring (M.C.Gil-Rodríguez and J. Afonso Carrillo, unpublished observations).This seasonal algal availability suggests a possible spatial and seasonal connection between sea urchin gonad production and benthic food availability in the Canary Islands.
In this study, we examine the spatial and seasonal variation of the gonad indices of the sea urchin Diadema antillarum at two contrasting sites off the southeast coast of Tenerife, that differ in sea urchin density and food availability.The study of these parameters could clarify the role of food availability (algal cover and number of species) in determining gonad periodicity.

Study sites
Two study sites were established at 4-6 m on the south-eastern coast of Tenerife Island, Abades (28º08'26''N, 16º26'04''W) and Boca Cangrejo (28º24'22'' N, 16º18'52''W) (Fig. 1).Current and wind regimes and the seasonal fluctuations of other oceanographic parameters are similar in both localities (J.G.Braun personal communication;Brito et al., 2002).Nevertheless, Abades has high densities of small individuals of D. antillarum and low macro-algal cover, usually called "urchin barren grounds", and Boca Cangrejo has low densities of large individuals and high macro-algal cover.
Therefore, there are two well defined situations: Abades with high densities and low food availability and Boca Cangrejo with low densities and high food availability (see results).

Characterization of the sea urchin populations
Density and individual test diameter were determined using the same method as Harmelin et al. (1980), Turón et al. (1995) and Casañas et al. (1998).Densities were estimated by counting all individuals in 10 × 2 m linear transects.Ten replicate transects placed at random were made at each locality every season during 2002 (March, June, September and December).Thirty individuals were collected at each date to measure test diameter (TD) to the nearest 0.1 mm.The total number of individuals to be measured was obtained for each locality following Kingsford (1998) sampling effort indications.

Algal cover measurement and species richness
Quadrats (0.25 × 0.25 m) were used for visual sampling of algal cover (Dawes, 1998).Sampling was done every season (March, June, September and December, 2002) at Boca Cangrejo and Abades.The percentage of algal cover of dominant species in each quadrat was noted; species that filled less than 1% of the quadrat were considered as 1% of the cover.Twenty replicates placed at random were made at each sampling date.For the statistical analysis, four "life form" groups were considered:   identified species; Articulated algae (AR) with 3 identified species; and Encrusting algae (CR) with 2 identified species (Table 1).These algal groups facilitate the ecological interpretation of the algae community's influence on the gonad periodicity of Diadema antillarum.Algal species were identified using the Afonso and Sanson (1999) identification key and the field guide by Haroun et al. (2003).The mean number of species observed in each season was termed benthic algal species richness.We used algal cover and benthic algal species richness (number of species) as an indirect method for measuring the spatial and seasonal changes in food availability.

Analysis of gut contents
In both localities we seasonally (March, June, September and December 2002) obtained guts from thirty sea urchins (the same urchins used for obtaining gonad material) and preserved them in 70% alcohol for posterior algae identification.To avoid daily rhythm variation in feeding (Lawrence and Sammarco, 1982;Carpenter, 1984) all collections were made in the morning.We counted the number of species of algae present in the gut of each individual and calculated the average number of species for each season.For the statistical analysis, the algal species were grouped according to their life form as in Table 1.

Analysis of gonad index
Thirty individuals were collected monthly (February 2002 to January 2003) at the same phase of the moon to avoid complications from a lunar spawning cycle (Pearse, 1975;Lessios, 1981;Illife and Pearse, 1982), in both localities.The horizontal diameter and oral-aboral axis of the tests were meas-ured.The sea urchins were dissected and the total volume of the gonad was measured to the nearest 0.1 ml by displacement in sea water.To obtain the dry body weight, the test, gut and lantern were dried for 24 hours at 110ºC and weighed to the nearest 0.01 g (Garrido et al., 2000).

Data analysis
Seasonal and spatial variations in density, test diameter, algal cover, species richness in gut content and gonad index were analyzed using a two-way ANOVA.We performed an ANOVA for each lifeform group.The two orthogonal factors, locality (L) and season (S), were used as fixed factors.
Cochran's test was used to evaluate homoscedasticity.Data that did not meet this requirement were log transformed (log (x + 1)).We lowered the significance level to 0.01 for transformed data that still had a heterogeneous variance (Underwood, 1981 and1997).

Sea urchin populations
The sea urchin populations were different in the two localities and the interaction locality x season had a significant effect, but no seasonal effect was detected (Table 2).Populations at Abades had a higher density (annual mean: 9.44 ind./m 2 ± standard deviation: 0.43) than those at Boca Cangrejo (annual mean: 3.62 ind./m 2 ± 1.52) (Fig. 2).
Test size showed differences between localities and the interaction locality x season had a significant effect (Table 2).The sea urchins from Abades were smaller (TD annual mean: 36.88 mm ± 11.05) than those from Boca Cangrejo (TD annual mean: 63.83 mm ± 11.98) (Fig. 2).
The interaction locality x season had a significant effect on species richness.Figure 3 shows a species richness peak during spring in Abades but not in Boca Cangrejo.This locality showed higher species richness than Abades throughout the entire year (Fig. 3).
These results indicate two clearly differentiated situations: Abades, with high densities of small individuals of D. antillarum, low algal cover and algal species richness, usually called "urchin barren grounds"; and Boca Cangrejo with low densities of large individuals, high macro-algal cover and higher species richness (Fig. 3).

Spatial and seasonal variation of algal life-form groups
Filamentous and fleshy (FF) algae increased during spring (31% of total cover) in Abades but were present during all seasons without conspicuous peaks (between 5 and 18 % of total cover) in Boca Cangrejo (Fig. 4).
Like FF, LC had maximum cover during spring in Abades (10% of total cover).In Boca Cangrejo, maximum cover was in summer (20% of total cover), followed by spring (10% of total cover) (Fig. 4).
Cover of AR in Boca Cangrejo was consistently higher than in Abades (Fig. 4).Abades did not have conspicuous cover in any season (Fig 4).In Boca Cangrejo this algal group (AR) was present in all seasons, varying from 3% cover in winter to 20% cover in summer (Fig. 4).

Spatial and seasonal variations of Diadema antillarum gut contents
In Boca Cangrejo, even though other algal groups had greater cover on rocky bottoms (Fig. 4), filamentous and fleshy (FF) group was more abundant in the digestive tracts of Diadema antillarum (Fig. 4).The same algal group was the most abundant in the guts of D. antillarum at Abades in all seasons with a clear peak in spring (Fig. 4).
The number of algal species in the digestive tracts of D. antillarum differed with locality (Table 4, note that the significant level was lowered to p<0.01).The gut contents of the sea urchins of Boca Cangrejo had a higher number of algal species (mean of 4.17 species) than those of Abades (mean of 1.53 species) (Fig. 4).
FF was significantly affected by the main factors locality and season (Table 4).Boca Cangrejo had a higher number of FF species (Fig. 4) than Abades.In Abades, a small peak in FF richness occurred in spring, in Boca Cangrejo the peak was in autumn.In Abades, this peak coincided with the increase in species richness but in Boca Cangrejo the species richness was always high (Fig. 4).
LC was significantly affected by the locality and the interaction locality x season.This shows a seasonal variation in gut richness that depends on the location's characteristics (Table 4).In Abades during the same season, a peak in LC species richness coincided with a high number of species of LC being found in the gut contents (Fig. 4).
AR showed differences in the interaction locality x season (Table 4).Some species of AR appeared in the guts during spring, summer and autumn.In Boca Cangrejo, this group of algae always appeared in the guts.
CR did not show any significant variations between locality, season or interaction.

Spatial and seasonal variation of gonad index
Seasonal changes in the gonad index (GI) differed significantly with locality (Table 5).Monthly mean Gonad Indices in Boca Cangrejo were higher than in Abades (Fig. 5).For example, the maximum value of GI in the former was 47.38 and 12.04 in the latter.
The GI increased from April to June in Boca Cangrejo but no clear peak was detected in Abades (Fig. 5).After the maximum peak in spring, the mean GI decreased slowly to a minimum in late summer / early autumn in Boca Cangrejo (Fig. 5).DISCUSSION Levitan (1988;1989) indicates that D. antillarum is able to change its body size in response to changes in food availability.The reduction in body size is related to energy conservation, since a smaller animal invests less energy in growth, which maximizes survivorship of individuals in areas with low food availability (Levitan, 1989).Thus, there may be a relative allocation of resources to gonad growth.Our results show that sea urchins in Abades are present in high densities and have small tests, which is indicative of low food availability.In contrast, sea urchins in Boca Cangrejo are present in lower densities and have larger tests than those at Abades, which is associated with greater food availability.
The monthly values of GI are different in the two localities (Abades-Boca Cangrejo) and only seasonal variation is clearly distinguished in Boca Cangrejo.In spite of the seasonal changes in the GI, it is probable that the sea urchins of Boca Cangrejo have a higher annual gamete output compared with urchins at Abades.In a high algal cover situation, such as at Boca Cangrejo, each sea urchin may make a larger contribution per capita to the larval pool.Lamare et al. (2002) noted that in favourable nutritive conditions the sea urchin Evechinus chloroticus (Val.) had a higher annual gamete output than urchins in areas with less conspicuous macrophytes.However, reproductive effort in Diadema antillarum is density-dependent (Levitan, 1989;1991), thus at high population densities, increased fertilization success can compensate the decreased gamete production.Thus there is decreased production per individual, but there are more individuals.Therefore, production may be similar in both Abades and Boca Cangrejo.A seasonal cycle in the gonad index of Diadema antillarum similar to that found here was reported by Bacallado et al. (1987) in Tenerife (Las Caletillas), and Garrido et al. (2000) for two localities in Gran Canaria (Puerto Rico and Sardina del Norte).However, there are timing differences between the GI peaks in their results and ours.Garrido et al. (2000) reported a gonad index peak in Sardina del Norte that was 1-2 months before (April) the peak in our data from Boca Cangrejo.This suggests inter-annual and inter-insular variations.Individuals at Abades have a gonad cycle without clear peaks, while those of Boca Cangrejo have a clear peak in spring.Sea urchin populations of Boca Cangrejo have a clearer periodicity and higher gonad indices than those at Abades.
Higher availability of algal cover and a greater variety of algal species provide more food and more preferred species (FF) for Diadema antillarum in Boca Cangrejo than in Abades.John et al. (1992) showed that Diadema antillarum of the tropical African coast preferred filamentous and fleshy algae (or FF in our groups).Sammarco (1977;1982) noted that high densities of D. antillarum in the Caribbean result in a change in algal composition, favouring crustose algae and cyanobacteria assemblages that may be less palatable.Vadas (1977) showed that Strongylocentrotus droebachiensis (Muller) and S. franciscanus (A.Agassiz) exhibit strong algal preferences.Algae with opportunistic strategies (r-selection) (FF in our algal group) were greatly preferred by these urchins.Moreover, the greater variety of species FF, LC and AR seems to favour gonad development and seasonality in the individuals of Boca Cangrejo.This result is supported by the higher algal specific richness found in the gut contents of the sea urchins of Boca Cangrejo during all seasons.Tuya et al. (2001) demonstrated that at Gran Canaria Diadema antillarum in the laboratory preferred algae such as Dictyota (included in our list as FF), which supports the interpretation of the gut contents found here.Other algae such as Lobophora (included in our list as LC), Padina (included in our list as LC) and Cystoseira (LC) were the least preferred.This demonstrates that Diadema antillarum has food preferences and selects algae such as FF, which could be due to the greater palatability of these algae.Spatial variability in the reproduction of Diadema antillarum seems to be attributed, at least in part, to different nutritive regimes between localities.The availability of its preferred algal group (FF) seems to play an important role.Sea urchins eat a range of algal species, but which species are more nutritive?Future isotopic studies may clarify this aspect.
The existence of a clear seasonality in the GI for Diadema antillarum in the Canary Islands does not agree with data from several authors who have studied Caribbean populations of D. antillarum-b.Seasonality is less conspicuous in tropical areas such as Fort Randolph (Panama), or nearly nonexistent in Maria Chiquita (Panama) (Lessios, 1981), Virgin Islands (Randall et al., 1964), Barbados (Lewis, 1966) and Bermuda (Illife and Pearse, 1982).Only Bauer (Bauer, 1976) found strong seasonality in the northern Caribbean area (Florida).In oceanographic terms, Florida is more comparable to Tenerife than any other locality in the Caribbean.However, Diadema mexicanum (Agassiz) of the American tropical Pacific (Culebra Island and Wraba) has strong seasonality in the GI.It has been suggested that these differences are due to differences in environmental factors in each area (Lessios, 1981).Our results supported the hypothesis that the seasonal GI variation may be linked to food availability.
Diverse reproductive strategies might take advantage of the most favourable times of the year.In this sense, gonad maturation peaks are inconspicuous and extend for several months in areas where annual variations of the environmental parameters are nearly non-existent (tropics).The waters of the Canary Islands show a peak in planktonic primary production in February-March (end of winter) (Braun, 1980), simultaneously with minimum water temperature and with important effects on the trophic network.The peaks in algal cover and species richness (food availability) that we observed during spring are related to this nutrient concentration in the water column.This peak in algal cover is more dramatic in Abades than Boca Cangrejo, because when there are few urchins the standing crop is limited by space and increasing productivity cannot increase algae standing crop (Lessios, personal communication).In the Canary Islands, a spring peak in food availability may be related to the clearly defined GI peaks of D. antillarum that occur during spring in Canary Islands waters (Bacallado et al., 1987;Garrido et al., 2000 and this paper) (by gonad nutrient storage).
The increase in algal cover and species richness is linked to higher species richness found in the sea urchin gut contents and to higher gonad indices.Moreover, the seasonal changes in algal cover and gut content richness are due mainly to the most preferred algae (FF) that may be related to gonad periodicity of the Canary Islands populations of Diadema antillarum.
FIG. 1. -The Abades and Boca Cangrejo locations where algal cover was measured and sea urchins were collected.

TABLE 2 .
-Two-way ANOVA on the effects of locality and season on density and test diameter of Diadema antillarum.

TABLE 3
. -Two-way ANOVA on the effects of locality and season on total algal cover and algal species richness.FIG. 3. -Mean of algal cover / 0.0625 m 2 and algal richness (nº algal species) for each locality.

TABLE 5 .
-Two-way ANOVA on the effects of locality and season on gonad index of Diadema antillarum.
FIG. 5. -Annual variation in the monthly mean (± SE) values of the gonad index (30 replicates) from February 2002 to January 2003 in both localities and seawater surface temperature.

TABLE 4 .
-Two-way ANOVA on the effects of locality and season on total algal richness and for the algal groups in the gut contents (log-transformed).Locality: Abades and Boca Cangrejo.[ª heterogeneity variance in spite of transformation of data, the significant level was lowered to p<0.01].