69 Ectoparasites of the whitespotted rabbitfish,Siganus sutor (Valenciennes, 1835) off the Kenyan Coast: distribution within the host population and site selection on the gills A. GEETS*, II. COENE and F. OLLE VIER Laboratory for Ecology and Aquaculture, Catholic University Leuven, Naamsestraat 59, 3000 Leuven, Belgium (Received 21 November 1996; revised 11 January 1997; accepted 11 January 1997) SUMMARY Different populations of the whitespotted rabbitfish,Siganus sutor, were examined for ectoparasites: adults from the Mombasa area (sampled in December 1990) and different age classes (adult, subadult and juveniles) from Gazi Bay (sampled in December 1992 and August 1993). The most common gili parasites were: the monogeneansPseudohaliotrema sp., Tetrancistrum sigani and Microcotyle mouwoi, the copepods Hatschekia sp., Pseudolepeophtheirus sp. and juvenile Caligidae, and prazina larvae of the isopodG nathia sp. Adult siganids had a higher parasite load than subadults. Juvenile rabbitfish did not harbour any gili parasites. Temporal differences in the parasite load of subadult rabbitfish were observed for M . mouwoi (highest in the December samples) and for juvenile Caligidae (highest in August). The microhabitat of the 5 most common gili parasites was species specific. M ost parasite species showed distinct site preferences with respect to both gili arches and gili sectors, within the gili arches. Niche breadth of the different gili parasite species was independent of the abundance of any of the other species present. However, niche breadths Mof . mouwoi, Tetrancistrum sp. and H atschekia sp. increased with their own abundance. This suggests that interspecific competition for space is low and that intraspecific factors could play an important role in the microhabitat choice of these gili parasites. The hypothesis that niche restriction leads to higher intraspecific contact and an enhancement of chances to mate was tested on 2 monogenean species, Pseudohaliotrema sp. and Tetrancistrum sigani. Their highly aggregated distribution over the gili filaments, leading to increased intraspecific contact, is consistent with the hypothesis. Key words: gili parasites, Siganus sutor, Indian Ocean, microhabitat use, niche. interactions and niche availability (Kennedy & INTRODUCTION Guégan, 1996). Rohde (1976, 1977, 1978, 1991), The whitespotted rabbitfish, Siganus sutor studying gili parasites of fish of the Indo-Pacific and (Valenciennes, 1835), is one of the most important the Australian coastal waters, suggested that most commercial fish species in artisanal marine fisheries gili parasite species live in low-density populations along the East African coast (Ntiba & Jaccarini, 1988). in resource-rich habitats. Whereas resource par­ Rabbitfish are herbivorous (Lam, 1974), live in titioning seems common and would reduce inter­ schools on reef flats and seagrass beds and harbour specific interactions, microhabitat restriction would several ecto- and endoparasite species (Diamant & lead to increased intraspecific competition as well as Paperna, 1986; Jones & Hine, 1983; Martens & enhance the chance to find a mate. Koskivaara, Moons, 1995; Toman, 1977, 1989). Geets& Ollevier Valtonen & Vuori (1992) and Bagge & Valtonen (1996) described the distribution of the endoparasitic (1996) equally found no limitations of resources and helminths within an adult host population of £.sutor an apparent lack of interspecific competition when off the Kenyan coast. The microhabitat preferences studying 10 dactylogyrid species inhabitating the of the intestinal helminth communities indicated a gills of roach in Finnish lakes. selective site segregation of the parasite species The present study focuses on the ectoparasite present. For testing niche theories, inter- and fauna of S. sutor and describes the characteristics of intraspecific interactions, and their impact on micro­ the component communities as well as the micro­ habitat use, parasites on the gills of marine teleost are habitat choice of the infracommunities of the gili very useful models (Rohde, 1991). The species parasites. The specific aim of our study was to assess richness of infracommunities of parasites will be whether the microhabitat use of gili parasites of a influenced by bottom up processes, i.e. interspecific common marine fish species of the Indian Ocean,S. sutor, fits the prediction of the hypotheses of Rohde (1991) on the relative importance of intra- and * Corresponding author. Tel: +32 16 323966. Fax: +32 interspecific competition and the reinforcement of 16 324575. E-mail: [email protected] reproductive barriers. Parasitology (1997), 115, 69-79 Printed in the United Kingdom. © 1997 Cambridge University Press subito e.V. licensed customer copy supplied and printed for Flanders Marine Institute Library (SLI05X00225E) A . Geets, H. Coene and F . Ollevier 70 set on 0 05. The critical value for comparison of a MATERIALS AND METHODS group of 33 (MADD) and 20 (GADD, GSUBD, Five populations of whitespotted rabbitfishSiganus G SUBA) observations was 20-17 and when 2 groups sutor (Valenciennes, 1835) were sampled. Two of 20 fish were involved it was 22-5. groups of adult rabbitfish were caught in December Data on infection levels of the 4 gili arches were 1990 and 1992: 1 sample from the seagrass beds off matched (measured on the same fish) within a host the Mombasa coast (MADD) and 1 from Gazi Bay group. Therefore Friedman ANOVA by ranks was (60 km south of Mombasa) (GADD) respectively. used when testing the overall differences of infection Two samples of subadults were collected in Gazi of the gili arches within a host group. We followed Bay: 1 in D ecem ber 1992 (G SU B D ) and 1 in August Siegel & Castellan (1988) for post hoc m ultiple 1993 (GSUBA). All fish were obtained from local comparisons. fishermen who used cabbage-baited dematraps (tra­ In order to get a better insight into the micro­ ditional bottom traps). In August 1993 we also habitat preferences of the gili parasites of S. sutor, caught small juvenile siganids from Gazi Bay by the gills of 10 adult fish (MADD) were drawn with a beach seining (GJUVA). Table 1 gives an overview drawing mirror (external and internal hemibranches of the samples and abbreviations used further in the drawn separately) and the exact position and number text. Fish were weighed, measured and fixed in 7 % of parasites was noted. Each hemibranch was divided formaldehyde solution immediately upon landing of in 8 equal sectors (Fig. 1) and the number of the fish (which was at the most 1 h after catching the parasites per sector was determined. Site specificity fish). The MADD sample contained 33 animals, all of the parasites on these 10 fish was tested on 3 other samples consisted of 20 individuals. Some levels: between the 4 gili arches (Friedman ANOVA additional fish were examined to collect living by ranks and multiple comparison), between external parasites for identification purposes. The parasites and internal hemibranchs (Wilcoxon’s matched pair were fixed using the methods described by Berland test), and between the 8 sectors (Friedman ANOVA (1984). The terms prevalence, intensity, mean by ranks and multiple comparisons). When a intensity and abundance are used according to difference of abundance on one of the first two Margolis et al. (1982). We used data on the levels (gills or hemibranchs) was noticed, the infracommunity level (defined as the numbers of a Friedm an ANOVA fo r the sectors was carried out on certain parasite species on 1 individual fish) for the the numbers of parasites per separate gili arches or calculations. The pseudobranch and the 4 gills of hemibranch, as well as on the summed data for each both left and right side of the fish were dissected and sector (all gills and hemibranchs together). analysed for parasites in all fish of samples MADD, To define spatial niche breadth, the 8 gili sectors GSUBD and GJUVA. The gills were screened for were seen as different resources. Levins’ niche parasites under a stereo-microscope (magnification breadth, B = 1 /£ p j w here p¡ is the proportion of range x6-x50). Since there was no significant specimens of a species found on sectorj, was difference in parasite num bers between left and right calculated for each infrapopulation and standardized gills in any of these samples (Wilcoxon’s matched between 0 and 1 (J30) (Krebs, 1989). For niche pair test, P > 0-05), we screened only the left gills of overlap, Renkonen index ~1 \Z \p ia— pjc\, was used, the fish of GADD and GSUBA. To calculate mean where p (a is the proportion of specimens of speciesi intensities and abundances for these last two on sectora and pja the proportion of speciesj on samples, the number of parasites found was doubled, sectora. Differences in both measures were tested by while for the other samples, intensities on left and Wilcoxon’s matched pair test. Spearman Rank right gills were summed. correlation was used for testing correlations between No parasites were found on the pseudobranch (P), niche breadth and number of parasites. except in 1 fish from GSUBA harbouring 2 In view of the hypothesis of Rohde (1991) that specimens Gnathia sp. Therefore the data on the niche restriction leads to higher intraspecific contact pseudobranch were not included in the statistical and enhancement of mating chance by the parasites, analyses. Gili arches were numbered I to IV in the distribution of the 2 most abundant monogenean anterior-posterior succession. Gili arch references species Pseudohaliotrema sp. and