Snail diversity, abundance and distribution in Arabuko Sokoke forest, Kenya
Abstract
enThis study sought to investigate land-snail diversity, abundance and distribution in Arabuko Sokoke forest. Sampling was done using standard timed direct search and litter sample methods. In total, 25 species and 1263 specimens were recorded. The 25 species rank Arabuko Sokoke forest as the second richest coastal forest in land snails in East Africa. The 25 species, however, are not uniformly distributed across the forest. The majority of the snails were localized with low levels of abundance. Eight species were restricted to one forest type, whereas those shared among forest types tended to concentrate in one forest type. Only five species, Gulella radius, Gulella foveolata, Gonaxis kibweziensis, Pseudoglessula biovini and Opeas gracilis, were widespread. The highest snail diversity was recorded in the mixed forest, implying that special conservation of this habitat is necessary for retention of the broadest molluscan diversity. The observation that some species rare in the mixed forest were recorded in excess from Brachystegia and Cynometra forest, further suggests that conservation of the entire A. Sokoke forest is critical for efficient molluscan conservation. Species recorded in low levels of abundance are probably declining and investigation on the snails' ecology to understand factors that influence the snails' diversity is recommended.
Résumé
frCette étude s'est penchée sur la diversité des escargots terrestres, leur abondance et leur distribution dans la forêt d'Arabuko Sokoke, au Kenya. On a récolté des échantillons en utilisant la recherche standard chronométrée directement et l'échantillonnage de la litière. On a rapporté un total de 25 espèces et de 1263 spécimens. Les 25 espèces classent Arabuko Sokoke comme la deuxième forêt la plus riche d'Afrique de l'Est en escargots terrestres. Toutefois, les 25 espèces ne sont pas uniformément distribuées dans la forêt. La majorité des escargots étaient localisés et peu abondants. Huit espèces étaient limitées à un type forestier alors que celles qui se distribuaient dans plusieurs types forestiers avaient tendance à se concentrer dans un seul. Cinq espèces seulement, Gulella radius, Gulella foveolata, Gonaxis kibweziensis, Pseudoglessula biovini et Opeas gracilis, étaient largement réparties. La plus haute diversité en espèces a été rapportée dans la forêt mixte, ce qui implique que cet habitat doit bénéficier d'efforts de conservation spéciaux puisqu'il abrite la plus grande diversité de ces mollusques. Le fait qu'on ait remarqué que certaines espèces, rares dans la forêt mixte, se trouvent en abondance dans la forêt de Brachystegia et de Cynometra suggère si besoin était qu'il est absolument nécessaire de conserver toute la forêt d'A. Sokoke pour garantir la conservation efficace des mollusques. Les espèces dont on a rapporté la faible abondance sont probablement en diminution, et on recommande d'entreprendre une étude de l'écologie des escargots pour comprendre les facteurs qui influencent leur diversité.
Introduction
Molluscs (snails and slugs) comprise the second most diverse animal phyla after arthropoda with an estimated 80,000 species world-wide (Solem, 1984; Emberton et al., 1997). The majority of the species are aquatic dwellers (marine and freshwater) with the terrestrial species comprising about 25% of the total fauna (Emberton et al., 1997). From a biodiversity conservation perspective, the terrestrial species are of regional and global concern, mainly because the majority of the species are forest dwellers and sensitive to habitat disturbance (Verdcourt, 1972; Emberton, 1995; Tattersfield, Seddon & Lange, 2001). Secondly the forests are rapidly declining (Sayer, Harcourt & Collins, 1992) leading to a malacological conservation crisis, especially in tropical Africa.
Despite the conservation challenge facing the terrestrial malacofauna, information on their biodiversity patterns from many parts of the world remain poorly understood. In East Africa, only the Congo-Guineo rainforests (Kakamega) in Western Kenya (Tattersfield, 1996; Lange & Maes, 2001), sections of Kenya's central highlands (Warui, 1998) and Tanzanian Eastern Arc mountains have been properly investigated (Emberton et al., 1997; Tattersfield, 1998). The biogeographically prominent East African coastal forests have only been examined at a superficial level (Emberton et al., 1997), yet they have remained in isolation for the last 30 million years (Dickinson, Burgess & Clarke, 1992) and are currently threatened by human exploitation (Burgess & Clarke, 1999). Past biodiversity studies have singled out these coastal forests as critical hotspots and centres of endemism for birds, mammals and arthropods (Larsen, 1991; Fitzgibbon, 1994; Matiku, Bennun & Odanga, 1998; Burgess, Ponder & Goddard, 1999).
The high biological significance of these forests, the escalating human exploitation and the negative impact that human disturbance has on the snails suggests that there is need for a detailed malacological investigation to document most of these forests snails before some disappear. Such information is vital for formulation of the malacofauna conservation priorities. In the present study, we report on the diversity, abundance and distribution of land snails in Arabuko Sokoke forest of coastal Kenya, which is the largest and biologically most important coastal forest in East Africa (Burgess & Clarke, 1999).
Materials and methods
Study site
Arabuko Sokoke forest (3°29′S, 39°48′E) is the largest block of native lowland forest on the East African coast, with an approximate area of about 372 km2 (Kelsey & Langton, 1984). The forest is generally a flat coastal plain (Matiku et al., 1998) with elevation barely rising above 60 m (Munir, 1994). There are three soil types in the forest, the typically red, well-drained and infertile soil referred to as magarini, forming a shelf 60 m above sea level (Matiku et al., 1998), the loose and compact white sands, and finally coral rag bordering the coastal belt (Kifcon, 1995). The mean rainfall ranges from below 600 mm to over 1000 mm per annum, with the driest area of the forest being the north-west, which is dominated by Cynometra webberi thicket, and the wettest comprising of the mixed forest (Kelsey & Langton, 1984). Mean temperatures range between 27 and 30 °C, although it may rise further during the driest part of the year in February (Kifcon, 1995).
The vegetation of Arabuko Sokoke is varied with distinct habitat types related to factors such as soil types and rainfall (Briton & Zimmerman, 1979). Generally, the forest can be classified into three habitat types (Matiku et al., 1998). The mixed forest comprised of mixed tree species, Cynometra forest dominated by Cynometra webberi and more open Brachystegia forest dominated by Brachystegia speciformis. The forest fauna is amazingly rich. Over 230 species of birds have been recorded from the forest (Kifcon, 1995) with six species being listed as globally threatened (Bennun, 1995). The reserve is also the home of numerous mammals of which three are very rare world-wide. These are the Adder's duiker (Cephalopus adersi) found only here and Zanzibar, Sokoke bushy tailed Mongoose (Bdeogale crassicauda omnivora) and the charming long snouted golden-rumped elephant shrew (Rhynchocyon chrysopygus) almost entirely confined to this forest (Bennun, 1995). Many rare and unusual species also occur in other animal groups such as arthropods, reptiles, amphibians and molluscs (Larsen, 1991; Bagine, Muhangani & Ruthuri, 1992; Chira, 1993; Drewes, 1997).
The forest is an important source of forest based products particularly to the coastal communities. The majority of households living adjacent to the reserve depend on the forest as a source of fuelwood, polewood and medicinal plants as well as honey (Blackett, 1994). Illegal commercial and semicommercial timber plus carving wood harvesting are also popular in the forest (Wass, 1994; Matiku et al., 1998). A survey carried out in 1983 indicated that over the years, these activities had drastically altered this reserve, leaving only 28% of the primary forest (Kelsey & Langton, 1984).
Site selection
Preliminary surveys were carried out in June 1999 to locate the prime forest habitats. Ten sampling plots of 40 × 40 m were marked at each of the three forest types. Plot selection was done after a thorough survey to ensure that they were distributed over much of each forest and were true representatives of the forest type.
Sampling molluscs
Sampling was undertaken by a combination of standardized timed direct search and litter sieving methods (De Winter, 1995; Tattersfield, 1996). In each plot, live snails and dead shells were searched and collected for 1 h by two individuals. During the search, all potential snail microhabitats (tree trunk base, under dead fallen wood, under leaf surfaces, on the forest floor litter) were searched and all snails found collected. Live snails collected during the search were first separated from the dry shells on returning to the campsite. They were then drowned and treated with 70% alcohol. Dry shells were kept in specimen vials.
At the end of each search period, each individual collected 4 L of forest floor litter from each plot. The litter was later thoroughly air-dried under hot sun and sieved using stacked sieves (mesh size of top sieve = 2 mm and bottom 0.5 mm). Litter retained in the top sieve was searched for any snail specimen then discarded. The fraction retained in the bottom sieve was examined for small snails (overlooked during timed direct search) under strong illumination.
Snail identification
All snails from the direct search and litter sieving were identified using relevant keys and reference materials available at the National Museum of Kenya (NMK). The majority of the snails were determined to species level using the NMK reference collection and literature. Several others, however, could not be identified because of lack of good matching material or the necessary literature. They were thus morph-typed as sp. A, B or C based on differences of characteristic morphological features.
Data analysis
Three methods of determining species richness and diversity were used. These have been calculated using the full data set of 30 replicate plots. These are S – the total number of species recorded, α– mean number of species per plot, and H′– the Shannon Weaver index. The total number of individuals (both living and dead specimens) were used as a measure of relative abundance, because the sampling effort was the same on each plot. Chi-squared tests were used to examine the association of species with different forest types. One-way analysis of variance with replicate was used to examine differences between the different forest types. One-way anova was applied to square root transformed species number (S), and total number of individuals (snail abundance). Multiple comparison using Tukey test was used to identify forest types supporting comparable and different levels of snail diversity and abundance. The conservation status of the snails in Arabuko Sokoke was explored by determining the frequency of species encountered among the sampled plots and relative abundance.
Results
Species diversity and abundance
A total of 25 species and 1263 specimens were recorded during the entire study. Snail diversity was highest in the mixed forest where higher values of species richness and diversity indices were recorded compared to both Brachystegia woodland and Cynometra thickets (Table 1). An anova test to infer whether there were differences in species richness, indicated significant differences (F2,27 = 1.921, P < 0.25). Multiple comparison using Tukey test detected the differences between mixed forest and Cynometra forest only.
| Habitat/forest types | |||
|---|---|---|---|
| Mixed Forest | Brachystegia Forest | Cynometra Forest | |
| Replicates/plots | 10 | 10 | 10 |
| Species richness(S) | 21 | 16 | 15 |
| Mean Spp. per plot | 8.00 ± 0.86 | 7.4 ± 0.62 | 6.3 ± 0.15 |
| Sp. range in plots | 6–13 | 5–11 | 6–7 |
| Shannon Index | 2.29 | 2.07 | 1.54 |
| Mean snail per plot | 37.0 ± 6.31 | 58.8 ± 11.48 | 30.5 ± 5.48 |
| Specimen total | 370 | 588 | 305 |
The snail abundance was highest in the Brachystegia woodland, where 46.6% of the total snail specimens were collected compared to 29.3% from mixed forest and 24.1% from Cynometra thickets (Table 1). An anova test of the mean number of specimens from the three forest types indicated a significant difference (F2,27 = 3.27; = P = 0.05), which was only detected between Brachystegia and Cynometra forests (Tukey test, q27,3 = 3.45; P < 0.1).
Faunal composition
There were major differences in species composition and relative abundance of species in the three forest types. A number of species were only found in one forest type, although all were recorded at very low frequencies. For instance, six species were only recorded in the mixed forest. These include Pupisoma orcula, Nesopupa sp. A, Rhachistia picturata, Tropidophora letourneuxi, Gulella gwendolinae and Edouardia sp. A. One species (Tropidophora anceps) was confined to Brachystegia woodland while Edentulina obesa was only recorded in Cynometra thickets.
The overall relative species abundance varied from 0.08 to 25.3%, with three species (Pseudoglessula boivini, Opeas gracilis and Gulella denticulatus) being the most abundant (Table 2). They contributed between 15.5 and 25.3% of the total sample, whereas the rest was in the range 0.08–7.36%. In the mixed forest, relative species abundance varied from less than 0.1 to 34.05%, while in Brachystegia forest it was in the range < 0.1–32.99%. The Cynometra forest presented the most varied species relative abundance (< 0.1–56.39%). Generally, the majority of the species recorded low abundance.
| Species | MF % abundance | BF % abundance | CF % abundance | Total % abundance |
|---|---|---|---|---|
| Opeas gracilis | 34.05 | 3.57 | 56.39 | 25.26 |
| Pseudoglessula boivini | 15.14 | 22.11 | 7.87 | 16.63 |
| Maizania wahlbergi | 8.92 | 1.02 | 0.00 | 3.09 |
| Guppya quadrisculpta | 5.95 | 0.34 | 0.00 | 1.90 |
| Gonaxis kibweziensis | 5.68 | 8.33 | 7.54 | 7.36 |
| Gulella foveolata | 4.59 | 3.91 | 11.15 | 5.86 |
| Achatina sp. | 4.05 | 5.61 | 0.66 | 3.96 |
| Rachidina braunsii | 3.78 | 0.34 | 0.00 | 1.27 |
| Afrodonta sp. A | 3.78 | 0.00 | 5.25 | 2.38 |
| Gulalla radius | 3.51 | 6.12 | 6.89 | 5.54 |
| Rhachistia picturata | 2.70 | 0.00 | 0.00 | 0.79 |
| Opeas lamoense | 2.43 | 4.25 | 0.33 | 2.77 |
| Nesopupa sp. A | 1.08 | 0.00 | 0.00 | 0.32 |
| Pseudoglessula sp. B(cf ingloria ) | 0.81 | 1.70 | 0.33 | 1.11 |
| Tropidophora letourneuxi | 0.81 | 0.00 | 0.00 | 0.24 |
| Pupisoma orcula | 0.81 | 0.00 | 0.00 | 0.24 |
| Gulella gwendolinae | 0.54 | 0.00 | 0.00 | 0.16 |
| Curvella sp. A | 0.54 | 6.29 | 0.33 | 3.17 |
| Edouardia sp. A | 0.27 | 0.00 | 0.00 | 0.08 |
| Rhachidina chiradzuluensis | 0.27 | 0.34 | 0.33 | 0.32 |
| Subulina sp. A (cf mrimensis ) | 0.27 | 0.00 | 0.66 | 0.24 |
| Gonaxis denticulatus | 0.00 | 32.99 | 0.33 | 15.44 |
| Tropidophora anceps | 0.00 | 0.85 | 0.00 | 0.40 |
| Trochonanina sp. A | 0.00 | 0.85 | 1.31 | 0.71 |
| Edentulina obesa | 0.00 | 0.00 | 0.33 | 0.08 |
| Unidentified | 0.00 | 1.36 | 0.33 | 0.71 |
| Total no. of specimens | 370 | 588 | 305 | 1263 |
The frequency of species encountered in the sampling sites/plots showed that only five snail species (Gonaxis kibweziensis, Gulella radius, Gulella foveolata, Pseudoglessula boivini and Opeas gracilis) were common or widespread in Arabuko Sokoke forest. They were recorded in more than half of each forest type total sampling plots and comprised over 70% of the entire sample size. In contrast, thirteen species were only found in about 17% of the total sampling plots and in not more than three of the ten sites investigated from each forest type.
A chi-squared test performed on fourteen species, recorded in sufficient numbers to test whether the distribution of snails was independent of forest type, showed that all the species differed significantly between the three forest types (Table 3). The association of each species with the forest types was explored by calculating the percentage deviation from the expected numbers, assuming equal abundance across the forest types. Table 4 identifies those species whose abundance deviated from the expected number by over 5%, 20% and 30%. Twelve species were recorded in greater abundance than expected in the mixed forest, eleven in Brachystegia forest and six species in Cynometra thickets (Table 4).
| Species | Forest types | χ2 | Prob. | ||
|---|---|---|---|---|---|
| CT | MF | BW | |||
| Maizania wahlbergi | 0 | 33 | 6 | 47.538 | 0.0000*** |
| Gonaxis denticulatus | 1 | 0 | 194 | 84.031 | 0.0000*** |
| Gonaxis kibweziensis | 23 | 21 | 49 | 15.742 | 0.0004*** |
| Gulalla radius | 21 | 13 | 36 | 11.686 | 0.0029** |
| Gulella foveolata | 34 | 17 | 23 | 6.027 | 0.0491* |
| Pseudoglessula boivini | 24 | 56 | 130 | 84.457 | 0.0000*** |
| Opeas gracilis | 172 | 126 | 21 | 112.671 | 0.0000*** |
| Opeas lamoense | 1 | 9 | 25 | 25.513 | 0.0000*** |
| Curvella sp. A | 1 | 2 | 37 | 63.050 | 0.0000*** |
| Achatina sp. | 2 | 15 | 33 | 28.229 | 0.0000*** |
| Rachidina braunsii | 0 | 14 | 2 | 21.490 | 0.0000*** |
| Afrodonta sp. A | 16 | 14 | 0 | 15.199 | 0.0005*** |
| Guppya quadrisculpta | 0 | 22 | 2 | 37.000 | 0.0000*** |
- *** P < 0.001;
- ** P < 0.01;
- * P < 0.05.
| Mixed Forest | Cynometra | Brachystegia | |
|---|---|---|---|
| Species found in excess in one forest type only | |||
| Guppya quadrisculpta | *** | ††† | †† |
| Rachidina braunsii | *** | ††† | †† |
| Maizania walbergi | *** | ††† | †† |
| Gulella foveolata | † | * | |
| Pseudoglessula boivini | † | †† | ** |
| Rachidina chiradzuluensis | † | † | ** |
| Achatina sp. | ††† | *** | |
| Curvella sp. A | †† | ††† | *** |
| Pseudoglessula sp. B (cf ingloria ) | † | †† | *** |
| Opeas lamoense | † | †† | *** |
| Gonaxis kibweziensis | † | † | ** |
| Gonaxis denticulatus | ††† | ††† | *** |
| Species found in excess in more than one forest type | |||
| Afrodonta sp. A | * | ** | ††† |
| Gulella radius | * | ††† | ** |
| Trochonanina sp. A | ††† | * | ** |
| Opeas gracilis | * | ** | ††† |
| Species recorded in low frequencies but confined to one forest type | |||
| Edouardia sp. A | ** | ††† | ††† |
| Gulella gwendolinae | *** | ††† | ††† |
| Tropidophora letournexi | *** | ††† | ††† |
| Tropidophora anceps | ††† | ††† | *** |
| Rhachistia picturata | *** | ††† | ††† |
| Subulina sp. A (cf mrimensis ) | ††† | *** | ††† |
| Edentulina obesa | ††† | ** | ††† |
| Nesopupa sp. A | *** | ††† | ††† |
| Pupisoma orcula | *** | ††† | ††† |
- Deviation from expected number, expressed as a percentage of total specimens.
- *** ≥ 30% excess;
- ** ≥ 20% excess;
- * ≥ 5% excess.
- †††≤ 30% deficit;
- ††≤ 20% deficit;
- †≤ 5% deficit.
Discussion
The 25 snail species recorded in the present study represent a high level of molluscan diversity from a single East African coastal forest. In comparison, studies elsewhere at the coastal forests of Tanzania documented 29 species from Amboni caves, six species from Miono forest, 22 species from Pindiro forest and eleven species from Nanganga forest (Tattersfield, 1998). Similarly, Emberton et al. (1997) previously reported seven species from Kazimzumbwi forest, seven species from Masasi forest and 21 species from Ngarama forest in coastal Tanzania. This information ranks Arabuko Sokoke forest as the second richest coastal forest in terrestrial snails after Amboni caves of Tanzania. Additionally, it is currently the known richest Kenyan coastal forest in land-snails.
The above malacological conclusion is a contribution to further evidence of the critical role of Arabuko Sokoke forest to conservation of terrestrial fauna. Ornithological studies have singled out the forest as the second most important bird conservation centre in Africa after Itombwe forest of the Democratic Republic of Congo (Collar & Stuart, 1988). Other surveys also suggest that Arabuko Sokoke forest is a stronghold for numerous coastal mammals, butterflies, reptiles and amphibians, particularly those endemic to the East African coast (Larsen, 1991; Chira, 1993; Fitzgibbon, 1994; Reid, 1996; Drewes, 1997).
Despite the reported high regional molluscan biodiversity, the majority of the species were recorded in low levels of abundance. In view of past human activity resulting in forest fragmentation (Kifcon, 1995; Matiku et al., 1998) and the fact that most land-snails are sensitive to environmental disturbances (Berry, 1973; Emberton, 1995; Tattersfield et al., 2001), it probable that these species occurring in low abundance may have declined owing to forest disturbance. Thus, they probably reflect changes in environmental quality and hence their potential as indicator species needs investigation. Secondly, a further survey on the snails ecology to investigate critical environmental variables that influence the snail diversity patterns is necessary for the implementation of appropriate conservation strategies that safeguard this healthy molluscan biodiversity. In this context, it is noteworthy that five of the species that are widespread and highly abundant are more resilient to habitat disturbances.
The distribution of snails revealed that the fauna is not uniformly distributed across the three prime habitat types, notably mixed forest, Brachystegia forest and Cynometra forest. The diversity results show that the greatest snail diversity was concentrated within the mixed forest, with the least diversity being registered at Cynometra forest. Faunal composition analysis further indicated that eight species were restricted to one forest type, whereas those shared among the forest types tended to cluster in one forest type. The comparatively high molluscan diversity reported in mixed forest is in line with other studies, which show that many terrestrial species of Arabuko Sokoke forest are concentrated within the mixed forest. The forest biodiversity inventory report by Blackett (1994), indicated that although the mixed forest is of limited extent, overall it shows the greatest diversity of terrestrial species. Similarly, ornithological survey by Bennun & Waiyaki (1991), recorded most of the forest birds species from the mixed forest.
The high malacological significance of the mixed forest and vulnerability owing to its characteristic fertile soil for agriculture and commercially-valuable trees like Afzelia quanzensis for wood-carving, implies that special conservation of this habitat is a priority. Apparently, the entire mixed forest zone should be declared a nature reserve so that all sorts of human impacts likely to alter the size and nature of the forest are controlled. The comparatively high snail diversity recorded in the mixed forest does not, however, mean that the other two forest regions are malacologically useless. Analysis of the species affinities among the three forest types identified several species occurring in greater abundance within Brachystegia and Cynometra forests than in the mixed forest. Thus, efficient and effective malacological conservation requires conservation of most of the existing regions of Arabuko Sokoke forest.
Acknowledgements
This project was supported by partial grants from Wildlife conservation Society, WWF-Prince Bernard scholarship and Darwin Molluscan Biodiversity Project to whom we are most grateful. We also wish to thank all the staff of A. Sokoke forest station, KEFRI and KWS for the assistance they accorded the first author during fieldwork.
