Contributions to Zoology, 80 (1) – 2011
Community ecology of tropical forest snails: 30 years after Solem
Keywords: Mollusca, Gastropoda, rain forests, species-richness, biodiversity, species-abundance-distributions
Since Solem’s provocative claim in the early 1980s that land snails in tropical forests are neither abundant nor diverse, at least 30 quantitative-ecological papers on tropical land snail communities have appeared. Jointly, these papers have shown that site diversity is, in fact, high in tropical forests; often more than 100 species have been recorded per site, which is somewhat more than normally found at sites in higher latitudes. At the same time, however, point diversities (which usually range between 10 and 30 species per quadrat) appear to be no different from the ones recorded for temperate localities, which suggests that the number of ways in which syntopic resource space can be subdivided among different land snail species has an upper limit that is no higher under tropical conditions. The available data do not allow much analysis of the ecological structuring processes of communities besides very coarse ones, e.g. the proportions of carnivores versus herbivores and Pulmonata versus non-pulmonates. Also, these first 30 years of research have shown that a number of serious methodological and conceptual issues need to be resolved for the field to move ahead; in particular whether empty shells from the forest floor may be used as a proxy for the contemporaneous communities. I make a number of suggestions for ways in which these obstacles may be removed. First, studies should be preceded by exploratory nested sampling in contiguous quadrats of increasing size, spanning several orders of magnitude. The shape of the triphasic species-area curve and nonlinear regression of the small-area end of the curve will help identify the quadrat and site areas that allow ecologically more meaningful studies. Second, researchers should be more aware of the trophic levels of species and restrict their analyses within guilds and within body size classes as much as possible. Testing species abundance distributions against ecologically explicit theoretical models may be a fruitful avenue for research. Finally, I argue that studies of this nature require species abundances that may only be found in tropical land snail communities that live on calcareous substrate, and therefore I suggest that malacologists aiming to understand community structure focus on limestone sites initially.
Almost all papers on tropical land snail diversity include a reference to Solem’s seminal paper of 1984 (Solem, 1984), in which he surmised that land snails in tropical rainforests are ‘neither diverse nor abundant’. Solem’s statement was based on very little evidence and was mostly conjectured from ecological generalizations about rainforests such as their sparse leaf litter, high incidence of fires, and high predation rates. As such, it was deliberately provocative, aimed at spurring tropical malacologists into action to prove him right or wrong. After all, whereas in the early 1980s studies of tropical rainforest insect diversity were in their heyday (Sutton, 2001), similar studies of Mollusca were still virtually non-existent, and Solem was right in calling attention to this lag. Not only are molluscs on land generally considered the next most-diverse phylum after the Arthropoda (Chapman, 2009), they have four ecological characteristics that allow their students to answer questions about tropical forest ecology that complement those addressed by entomologists. First, whereas the greatest diversity in tropical insects is in the monophagous and oligophagous herbivores, niche differentiation among land snails tends to be related to soil type, soil structure and bedrock type, as well as microclimatic aspects of the habitat. Second, snails’ extremely low vagility would result in lower point diversity but greater species turnover than for insects. Third, the fact that almost the entire malacofauna accumulates as empty shells on the forest floor may allow for more exhaustive and unbiased sampling methods than in other groups of invertebrates. And fourth, the presence of a single land snail individual in a site will almost always signify the presence of a population, and rarely represent an incidental migrant, as might be the case in more mobile taxa (Cameron, 1998).
The almost 30 years that have passed since Solem appealed to the malacological community (his 1984 paper was presented to the 8th International Malacological Congress in Budapest, 1983), have seen a steady increase in quantitative ecological studies of forest malacofaunas in tropical Africa, Southeast Asia, and, to a lesser extent, Central and South America. Besides proving Solem wrong, these studies have produced a first spectrum of insights into point diversity, species turnover, community structure, determinants for species abundance distributions, altitudinal and edaphic gradients in species diversity, and have made inroads into using land snails as model systems for biodiversity monitoring and as targets for conservation. In addition, the diversity of sampling methods used has highlighted certain methodological issues. In this paper, I present a synthetic review of the literature (specifically: quantitative studies of more or less complete terrestrial snail and slug communities in tropical forests, inclusive of tree plantations and timber production forests, but excluding mangroves), point out important gaps in our knowledge, and give recommendations for future work. Where relevant, I also refer to the considerably larger body of work done on terrestrial gastropods outside of the tropics. Although I do discuss changes in community structure over short distances (<10 km) and along local elevational and land-use gradients, I will not deal with the longer-distance beta-diversity (‘allopatric diversity’ sensu Solem, 1984).
Like all descriptive/analytical studies of communities of species, a manner of controlled sampling of these communities is called for. Students of tropical forest snail communities have tended to use a nested sampling design, usually consisting of replicate small units (regardless of their square or circular shape (which may influence sampling - see Brummer et al., 1994), they are here termed ‘quadrats’, after Cameron and Pokryszko, 2005) that together represent a sample of the land snail community within a larger unit or ‘site’, which, in turn, could be grouped into ‘areas’ (Cameron and Pokryszko, 2005). Typically, quadrats comprise a more or less homogeneous macrohabitat and have a surface area that ranges from the 8 m2 that Reuben Clements used to sample limestone outcrops in Peninsular Malaysia (Clements et al., 2008; Liew et al., 2008) to the 400 m2 plots introduced by Ken Emberton (Emberton et al., 1996; Emberton, 1997) in Madagascar and used since by various other researchers in Africa (De Winter and Gittenberger, 1998; Oke and Alohan, 2006; Oke et al., 2008) and Asia (Schilthuizen and Rutjes, 2001; Liew et al., 2010). A number of quadrats are arranged throughout a site either randomly (e.g. Fontaine et al., 2007a) or systematically along transects (Emberton, 1997; Emberton et al., 1999; Schilthuizen et al., 2002; Clements et al., 2008; Raheem et al., 2009). Although Cameron and Pokryszko (2005) consider sites to be small (< 1 km2) I here include in the ‘site’ category entire forests (e.g. Fontaine et al. (2007a) used 116 randomly placed quadrats to sample Lopé National Park in Gabon), forest fragments (Sri Lankan forest fragments were sampled with transects of twenty quadrats each by Raheem et al. (2009)), or predefined plots within forests, such as the 1 km2 block in Danum Valley Conservation Area in Borneo that Schilthuizen and Rutjes (2001) sampled with 36 randomly placed quadrats. Areas, finally, could be any grouping of sites to a higher hierarchical level, which is a spatial scale beyond the scope of this paper. Although most studies sample quadrats only once, some workers have maintained quadrats for longer periods for resampling in consecutive seasons (De Winter and Gittenberger, 1998; Bloch et al., 2007) or even decades (Vermeulen, 2003; Bloch et al., 2007).
Sampling within a quadrat usually takes place by manual search (by day and/or night) of suitable microhabitats, usually standardized by searching effort, expressed as person hours. This is then complemented with either or both of the following bulk methods: beating and shaking of (a standardized number of) branches over an inverted umbrella, and (as advocated by Emberton et al., 1996) collecting, sieving and sorting (with or without flotation) a standardized amount of leaf litter and/or soil (either sampled randomly across the quadrat or extracted from predefined blocks—the ‘Oekland sampling’ of Cameron and Pokryszko, 2005). The following are examples of different sampling regimes within quadrats. Bloch et al. (2007) looked for live snails and slugs at night for 0.5 person-hour in circular 3-m-radius plots in Puerto Rico, did not disturb the litter, and returned the animals after identification. De Winter and Gittenberger (1998) spent two person-hours searching in 20 × 20 m quadrats in Cameroon; this searching time included beating of the vegetation, 0.5 h searching of tree trunks, and collection of four litres of litter, which was subsequently sieved, dried, and searched manually in the lab. Clements et al. (2008), finally, sampled West-Malaysian karst forests by collecting four litres of litter and topsoil from 2 × 4 m quadrats at the bases of limestone cliffs. This was then enriched by coarse sieving, flotation (the collection of flotsam after submergence in a bucket of water), drying, and further sieving in a sieve cascade.
While some authors have adopted near-identical sampling techniques to improve comparison across studies (e.g. De Winter and Gittenberger, 1998; Schilthuizen and Rutjes, 2001; Cameron et al., 2003; De Chavez and De Lara, 2011), others have pointed out that sampling methods are best optimized for completeness under the prevailing conditions (Cameron and Pokryszko, 2005). Liew et al. (2008) computed incidence-based completeness estimators (ICE; Chao and Lee, 1992; Chazdon et al., 1998) for hexaplets of quadrats in limestone and non-limestone forests across Malaysia and found that the ones from limestone had high (c. 90%) completeness, regardless of quadrat sampling intensity, but the ones away from limestone were significantly incomplete. For most community-ecological purposes, therefore, sampling regimes are best optimized for the expected local species diversity and abundance (since completeness is generally a prerequisite for any kind of analysis), rather than for comparison with other studies.
Having said that, some further consideration of the concept of ‘completeness’ may be warranted. As Cameron and Pokryszko (2005) imply, an inventory of a quadrat is complete if all species actually present are found. However, in species with such limited active dispersal capacities as land snails, species ‘presence’ is an elusive concept. Even in medium and large-bodied (sub)tropical land snails, active dispersal rates are often in the order of just 1-5 m gen-2, leading to extremely narrow demes of tens of metres across or less (Schilthuizen and Lombaerts, 1994; Parmakelis and Mylonas, 2004; Giokas and Mylonas, 2004; Schilthuizen et al., 2005b). In the small-bodied microsnails in complex three-dimensional microhabitats that often dominate tropical snail communities, the scale of well-mixed population units is likely to be much smaller, and substantial heterogeneity may already be present below the scale of standard quadrats. In other words, a 20 × 20 m quadrat may contain 35 species, but only subsets of these may actually be coexisting in the sense that they share the same resources and potentially engage in ecological interactions with one another (the ‘point diversity’ of MacArthur, 1965). It is important to keep in mind that such clumping in land snails is fundamentally different in character from similar patterns of micro-scale clumping in organisms with mobile individuals, gametes, or seeds; in the latter cases, micro-scale clumping of adults affects demographic and genetic population structure much less than in land snails, and some of the unevenness seen in adjacent quadrats along transects (e.g. Schilthuizen et al., 2002; Dinarzarde Raheem, unpublished data) may thus be real rather than sampling artefacts. It also means that, in order to decide on the quadrat size adequate for capturing true alpha diversity in a certain habitat (i.e., the remaining diversity after area has been factored out), we may need to do a pilot study on a series of contiguous, nested quadrats, and determine the y-axis intercept q in the fitted Arrhenius equation S = q + cAz (Rosenzweig, 1995); see Fig. 1. The largest quadrat size that has an S not significantly different from q could be taken as the preferred quadrat size.
Fig. 1. The y-axis intercepts q for the fitted Arrhenius equation S = q + cAz represents point diversity and may be estimated from a non-linear regression of data at several non-zero area sizes.
Sampling previous generations
Another methodological issue that repeatedly comes up, at least in studies on non-acidic soils, is the inclusion of empty shells in the sample. While some authors have deliberately limited themselves to the contemporary populations by sampling living snails only (e.g. Alvarez and Willig, 1993; Bloch et al., 2007), most have included empty shells in their sampling either by necessity (in many tropical forests, live snails and slugs are extremely hard to find) or by design. Clements et al. (2008) and Schilthuizen et al. (2002), for example, sampled only empty shells, and most other studies mix living snails and empty shells. Often, the empty shells form the majority of specimens: the pristine and disturbed Bornean karst forest samples in Schilthuizen et al. (2005a), for example, contained less than 1% living animals.
Sampling empty shells has both advantages and disadvantages, depending on the study objectives. Among the advantages, the possibility to sample almost the entire malacofauna in a more or less unbiased fashion stands out (more or less unbiased, because there will be a bias towards the larger, thick-shelled species to survive weathering for much longer); this has proven useful for biogeographical questions addressing species-area relationships (Clements et al., 2008). All shelled snails and semi-slugs, living at all forest strata will eventually end up as empty shells on the forest floor – provided they do not live in enclosed spaces like phytotelmata (Kitching, 2000), nest ferns (Ellwood and Foster, 2004), or social insect nests (Eguchi et al., 2005). Sampling shells from the forest floor is thus comparable in efficiency to insecticide fogging for canopy insects (Coddington et al., 1991). Also, in theory, comparisons of previous diversities with present-day diversities could reveal changes in species composition, provided the empty shells date back long enough and can be dated (see below). In most studies, however, the empty shells are assumed to represent the contemporary malacofauna, which is only correct if the shells are younger than an average snail generation (Cernohorksy et al., 2010). The truth is that we still know very little about the rate by which shells on the forest floor degrade and dissolve. While Solem (1984; on authority of P. Mayhill) states that in New Zealand, ‘the rate of shell decay is a fast three weeks during dryer periods and less than ten days when the litter is wet’, Cameron and Pokryszko (2005), on the other hand, make the point that in calcareous areas, empty shells may actually be ancient or subfossil. The presence of old specimens in the sample is particularly problematic if sites are compared where major land use changes have taken place. For example, Schilthuizen et al. (2005a), Oke et al. (2008), and Tattersfield et al. (2001b) all included empty shells in their sampling, which was intended to study the impact of logging, quarrying, and forest-to-plantation conversion on the snail communities. These impacts may have been underestimated if their samples included remnants of the pre-disturbance faunas. Similar problems exist with other natural skeletal death assemblages such as vertebrate bones (Terry, 2010a, 2010b).
There are several ways in which these potential problems may be assessed, but few have been explored properly. An indirect method would be to compare the contemporary species diversity with the one apparent in the empty shells. If there are no obvious discrepancies, the samples may be pooled (Rundell and Cowie, 2004; Cameron and Pokryszko, 2005) and safely assumed to represent the present-day populations. Direct measurements of shell degradation are, however, very rare, especially for tropical settings. Lee (2009) reviews several published (e.g. Barrientos, 2000; Pearce, 2008); and unpublished experiments which appear to indicate that shell dissolution takes between several months and several years, depending on size and environment, and that it is bioerosion (calcium harvesting by other organisms), not chemical dissolution, that is the main factor responsible for shell disappearance. However, the admirably tenacious multi-species, long-term experiments that Pearce (2008) did in Delaware (and similar experiments are underway in the Czech Republic; Řihová et al., 2010) are yet to be replicated in tropical settings. Relatively small soil samples taken from the foot of limestone cliffs can sometimes yield enormous numbers of shells (Fig. 2; see below) and it is hard to imagine that these represent the ‘graveyard’ of the local fauna for just the most recent years. Intercepting falling shells with trays placed at the foot of the hill or using radiometric or amino-acid racemization dating of shells in the soil (Pigati et al., 2004; Hearty and Schellenberg, 2008) might be suitable methods to answer this important question.
Fig. 2. A large number of microsnails from a limestone hill in Kelantan, Malaysia, collected as empty shells. Photo: Reuben Clements.
Local species richness and diversity
If the past 30 years have shown anything, it is that Solem’s (1984) prediction that land snails in tropical forests are species-poor was incorrect. The highest land snail species diversities on record now all come from tropical sites, and no longer from wet-temperate locations (e.g. New Zealand’s South Island), as was the case when Solem wrote his review. In Borneo, Vermeulen (2003) recorded 108 species in three days’ collecting in the 31 km2 Batu Niah National Park, whereas Liew et al. (2010) collected 109 species from 142 quadrats in the 754 km2 Kinabalu Park. Emberton (1995) reports on ‘a small patch of lowland rainforest’ adjacent to the village of Manombo in Madagascar where 52 species live sympatrically. Rosenberg and Muratov (2006) found 91 species on a single one-hectare site in Jamaica. Schilthuizen and Rutjes (2001) found 61 species in a 1 km2 block of lowland rainforest in Borneo. De Winter and Gittenberger (1998), finally, recorded 97 species from a square kilometre of Cameroonian forest. Also in individual quadrats, high species richness has been found. The richest ≤400 m2 quadrats have yielded 30 species in Western Kenya (Tattersfield et al., 2001b), 48 species in West Malaysia (Liew et al., 2008; Clements et al., 2008), 44 species in Uganda (Wronski and Hausdorf, 2008, 2009), 45 species in Cameroon (De Winter and Gittenberger, 1998), and 42 species in Jamaica (Simpson (1894: 112) reports ‘in a narrow limestone gorge of the Rio Cobre near Bogwalk in the talus under a ledge some two rods long we found no less than forty-five species, all living....’).
Solem did appear to be correct, though, about the low abundances in tropical forests. The high species numbers from acidic rainforests mostly were painstakingly accumulated from small numbers of individuals from many quadrats, representing a great collecting effort and a relatively meagre output. The 142 Kinabalu quadrats of Liew et al. (2010) yielded only 2,832 specimens; the large study in Cameroon by De Winter and Gittenberger (1998) resulted in 2,654 individuals; and a similar collecting effort in Danum Valley (Schilthuizen and Rutjes, 2001) produced just 546 snails. Given that a large portion of the specimens in these studies were empty shells, living populations of tropical land snails may be considered extremely sparse. However, this may just reflect their cryptic mode of life. Many land snails and slugs are nocturnal and are simply not seen during the day. Others are soil-dwelling or canopy-dwelling and also not easily encountered, unless exceptional conditions reveal their true abundance. A few examples from Borneo (Schilthuizen, unpublished) may illustrate this. A single pitfall trap baited with lamb meat and placed in secondary forest gathered (besides the carrion beetles it was intended to catch) almost 400 individuals of Subulina octona (Bruguière, 1789), and similar traps in montane forest attracted several dozens of the otherwise rare ariophantid Kalamantania whiteheadi (Godwin-Austen, 1891); a heavy rain shower in lowland primary forest washed down many live specimens of Amphidromus martensi Boettger, 1894, a tree canopy snail, of which only the empty shells are normally found at ground level.
Still, although true densities may be higher than they appear to be, most tropical rainforests on acidic soil support demonstrably sparser land snail populations than those on alkaline, calcareous soils. Schilthuizen et al. (2003a) found that abundances in quadrats on limestone soil in Borneo were up to an order of magnitude higher than in those placed in surrounding non-limestone soils and similar observations were reported for Gabon by Fontaine et al. (2007a). Liew et al. (2008), in their compilation of a large number of Malaysian limestone and non-limestone plots, showed that the abundances on limestone are on average more than an order of magnitude greater than on other soil types.
It may be worth exploring what was Solem’s ecological reasoning for expecting land snails to be species poor in tropical forests. Among the chief factors he invoked were high predation, negligible leaf litter, and frequent fires. However, these ecological rationales may not apply. Among land snails’ chief predators, especially in Africa but also in other tropical regions (Schilthuizen et al., 2006; Rowson, 2010), are other land snails, which would add to, rather that detract from, land snail diversity. The widely held belief that tropical rainforests have a negligible litter layer is certainly not generally true (see also Corlett, 2009). As pointed out by De Winter and Gittenberger (1998), ‘thickness, distribution and composition of litter layers are dynamic in both space and time’. And: ‘great diversity in size, shape and firmness of the leaves, combined with differential decomposition rates by spot differences in microclimate, moisture and soil conditions, potentially provides a wide array of microhabitats supporting a high land snail diversity of vegetable and fungi consumers, and associated predator species’. Fire, finally, indeed affects tropical forests periodically, but probably not as disastrously as in many subtropical and temperate vegetations; moreover, since intermediate disturbance can delay competitive exclusion, fire may affect diversity positively rather than negatively (Connell, 1978).
Species abundance distributions
As soon as their curiosity about the actual species richness of their quadrats, sites, and areas is satisfied, community ecologists working on any taxon tend to turn to the next feature of interest, the species abundance distribution (‘SAD’). Tropical land snail researchers are no different: many papers explicitly display the pattern of commonness and rarity among the species found (e.g. De Winter and Gittenberger, 1998; Schilthuizen and Rutjes, 2001; Schilthuizen et al., 2002, 2005a; Oke and Alohan, 2006). Although this is usually presented as an untransformed bar graph, I will here use the now standard ‘Whittaker plots’ (Whittaker, 1965; Magurran, 2004; Cameron and Pokryszko, 2005; Nekola et al., 2008), which display rank (from commonest on the left to rarest on the right) on the x-axis, and log relative abundance on the y-axis (Fig. 3). SADs are interesting for a variety of reasons. First, they describe the community in a more informative way than by simply enumerating species, as they incorporate heterogeneity in abundance and hence are the basis for calculations of diversity (such as the Simpson Index; Magurran, 2004). Second, the rare-species tail of the SADs allows estimation of the numbers of species missed altogether and, therefore, of the true species richness. Chao (1984) developed a simple, non-parametric estimator for this, based on the numbers of singletons and doubletons in the sample. In many studies, percentages of singletons are high (11%, 12%, and 23%, respectively, in De Winter and Gittenberger, 1998; Fontaine et al., 2007b; Liew et al., 2010), leading to relatively high richness estimations. Third, SADs allow seasonal (De Winter and Gittenberger, 1998) or disturbance-induced (Schilthuizen et al., 2005a) changes in species dominance to be visualized. However, the property of SADs that has most captured the imagination in recent years is the possibility to deduce from them the ecological processes underlying the community structure. For example, George Sugihara (Sugihara, 1980; Sugihara et al., 2003) developed a model for sequential subdivision of niche space which results in SADs that approach the lognormal distribution seen in many natural communities, including those of gastropods. (In a Whittaker plot, a lognormal distribution would give a sigmoid curve that first descends fast, then levels off, then descends again.) However, a distinct right-skew – an overrepresentation of rare species – is often observed (Tokeshi, 1999), and Hubbell (2001) showed that a niche-free ‘neutral’ model in which slots in the community are filled randomly (the zero-sum multinomial) is better at predicting this skew. Testing goodness of fit between the predictions of particular models and an observed SAD should lead to an insight into the potential ecological processes structuring the community, as Cameron and Pokryszko (2005) did for a few tropical and non-tropical land snail communities (although many SADs available in the literature may suffer from the fact that they have been obtained with sampling methods that are not particularly unbiased; Cameron pers. comm.). However, a caveat is in order as indications exist that the properties of SADs are shared by many highly unrelated patterns both in nature and in culture and may reflect general statistical properties of all complex dynamical systems. If true, then this would limit the possibility to extract the ecological structuring processes from the shape of SADs alone (Nekola and Brown, 2007).
It should be stressed, moreover, that most biological SAD models only apply to a single guild in a single community. ‘Single guild’ and ‘single community’ are criteria that, in land snails, require a little elaboration. First, we have, above, already dealt with the concept of a single community, and I argue that this concept applies only at the spatial scale where true point diversity is in effect (a scale possibly smaller than most quadrats). If SADs are determined at a substantially larger spatial scale, the chances of finding patterns that derive from ecological rather than statistical causes decline rapidly (Šizling et al., 2009). Second, a guild is a community of species ‘with a similar ecological function’ (Rosenzweig, 1995), competing for roughly the same resource (Hubbell, 2001). To consider communities of tropical land snails as guilds would probably be wrong. Even with our stunted knowledge of tropical land snails’ autecology, we are certain that, for example, many African faunas are composed to a considerable fraction of predators (De Winter and Gittenberger, 1998; Wronski and Hausdorf, 2008), and even among non-molluscivores, sizeable but unknown proportions of fungivores, folivores, and detritivores could probably be distinguished. Yet another factor to be reckoned with is body size: throughout its life, a snail’s body mass passes through several orders of magnitude, and since it is likely that ecological interactions are stronger among individuals of similar size, we should really consider different size classes as different communities. This would imply that the juveniles of a particular species would sit in a different community than the adults. (As an aside, it may even be incorrect to consider such a community to be composed of land snails only, as some snail species may compete with other low-mobility invertebrates - Isopoda, Diplopoda, certain Coleoptera - more strongly than with other snails, as do marine invertebrates in a similar community, the rocky intertidal (Connell, 1961; Wootton, 1994).)
Although large, small-scale samples are required to distill structuring processes from SADs, the community structure may also be assessed in a more qualitative manner, by distinguishing functional groups within the malacofauna (Emberton, 1995). Unfortunately, as mentioned above, our knowledge of the niches and trophic levels of tropical forest snails is still extremely limited. Anecdotal information exists on carnivory, necrophagy, fungivory, frugivory, and folivory in some species (see above), but these data are without exception too few and scattered to compare entire communities. Until more detailed data are available, only more coarse patterns in composition can be studied.
One of these is the proportion of ‘non-pulmonates’ (mostly Caenogastropoda and Cycloneritimorpha). In the Cameroonian fauna sampled by De Winter and Gittenberger (1998), only three out of the 97 species are non-pulmonates (one cyclophorid and two maizaniids), all of moderate rarity. Conversely, Bornean faunas have roughly 50 : 50 proportions of Pulmonata and non-pulmonates (Schilthuizen et al., 2006a), whereas Raheem et al. (2008, 2009) recorded about one-third non-pulmonates in Sri-Lankan faunas. These differences are quite mysterious and may require composite explanations from both evolutionary ecology and historical contingencies. Schilthuizen et al. (2005a) showed that in Borneo, where pulmonates and non-pulmonates are roughly equally speciose and abundant, the former dominate in limestone habitats that are pristine, whereas Pulmonata do in ones disturbed by fire and/or logging (Fig. 4). This underscores the observation (Solem, 1974) that non-pulmonates are less resistant to drought and less able to survive under extreme conditions – which may also explain why most, if not all invasive snails are Pulmonata. If the proportion of non-pulmonates would generally be determined by the degree of seasonality, then this may be behind global and regional differences as well. A systematic study of drought-related climate parameters in relation to non-pulmonate species richness and commonness may reveal such patterns.
Fig. 4. Species abundance distributions for a Bornean limestone hill (Mawas). Non-pulmonates (a) and Pulmonata (b) are shown separately, as are the patterns for disturbed (open bars) and pristine (filled bars) quadrats. The patterns show that in disturbed habitats, Pulmonata increase in dominance at the expense of non-pulmonates (after Schilthuizen et al., 2005a).
Another conspicuous large-scale difference among tropical mollusc faunas is the proportion of carnivores. Workers studying the Afrotropical and Madagascan faunas, have repeatedly highlighted the large proportions of the exclusively molluscivorous family Streptaxidae in their samples. In many sites in tropical Africa, about a quarter of the species are streptaxids (Emberton et al., 1997; Tattersfield et al., 2001a; Fontaine et al., 2007b; Wronski and Hausdorf, 2008), and in the study by De Winter and Gittenberger (1998) in Cameroon, this proportion was even as high as 34%. In other faunal regions, the proportion of carnivores (as far as is known) is generally much lower: in most limestone sites in Sabah, only two known carnivore species (one diapherid and one rathouissid) occur among up to 100 non-carnivorous species; Schilthuizen, unpublished data), and in the Manakau Peninsula of New Zealand, only four (Rhytididae) out of 72 species are carnivores (Solem et al., 1981). The causes for these differences are totally unclear. De Winter and Gittenberger (1998) have speculated that the African streptaxids occupy niches filled by non-Mollusca in rainforests in other tropical regions, but even within East Africa (especially along elevational gradients) and in Southeast Asia (regionally) strong differences exist in the proportions of the carnivorous Streptaxidae and Diapheridae, respectively (Tattersfield, 1996; Emberton et al., 1997; Tattersfield et al., 2001a; Clements et al., 2008). Stable-isotope studies (Michener and Lajtha, 2007) in combination with considerations based on evolutionary rates of ecological traits and historical biogeography may be required to understand better the trophic levels occupied by members of different tropical snail communities.
In the absence of detailed information on niches and trophic levels of the constituent species, morphospace patterns may serve as a proxy. It is well-known that land snail size and shape are adapted to abiotic and, to some extent, also to biotic aspects of their niche and habitat (Cain, 1977; Goodfriend, 1986). In principle, thus, differences in community structure should to some degree be reflected in differences in morphospace occupied. Several studies (e.g. Peake, 1968; Emberton, 1995; De Winter and Gittenberger, 1998; Liew et al., 2010) have looked at this. Part of these treatments consider patterns in shell size or shell diameter, which I shall not discuss here because these will be related to species-biomass distributions and thus largely a different manifestation of the same dynamic ecological processes that generate the SAD (Henderson and Magurran, 2010). As for patterns in shape, several ways of distinguishing shape categories have been developed. The most applied is Cain’s (1977) shell height/width index, which in many faunas has a bimodal frequency distribution and tropical faunas seem to be no exception (Fontaine et al., 2007b), although the presumed relationship of these modes with vertical and horizontal substrates is not confirmed (Emberton, 1995; De Winter and Gittenberger, 1998). Emberton (1995) performed an admirable analysis on snail assemblies from North America, New Zealand, and Madagascar, in which he used forest structure data to predict the shell shape proportions from the available ‘flat’, ‘globose’, and ‘tall’ spaces in the habitat and combined these with phylogenetic constraints in shell shape within the dominant families. For the Madagascan fauna, this approach suggested that phylogenetic constraints were more important for structuring the morphospace than ecological niches. Yet, Emberton advocates developing more information-rich methods for evaluating morphospace, e.g. by applying Raupian coiling parameters rather than height/width measurements. In an attempt to enlarge the set of data used for assessing morphospace, Liew et al. (2010) added aperture height and width to shell height and width and reduced these four parameters to two principal components to show that occupied morphospace (and presumably also niche space) declined with increasing elevation, which they attribute to the reduced primary productivity. Further basic studies exploring the potential for morphospace to reflect more or less accurately the underlying ecological diversity are certainly needed, and measures such as the disparity index developed by McClain (2005) for deep sea gastropods may be useful.
Species diversity gradients and turnover
Although I do not discuss diversity patterns such as ‘allopatric diversity’ as defined by Solem (1984), which take place at the large scales that arguably fall outside of the realm of community ecology, I will here briefly discuss diversity patterns on the medium spatial scale, that is, within single sites. Such diversity patterns fall into two main categories: those taking place along gradients (e.g. elevational, land-use, or humidity gradients), and those represented by quadrat-to-quadrat spatial species turnover. The latter type is the hardest to interpret, and several authors have attempted to distinguish sampling error from ‘real’ species turnover. Cameron and Pokryszko (2005) have pointed out that frequently used indices for unevenness, such as Whittaker’s I (the total number of species divided by the mean number of species per quadrat), are prone to sampling error, as they tend to be negatively correlated with sample size: the smaller the quadrat samples (and in tropical land snail studies they often are small), the higher the apparent degree of between-quadrat variation. However, as we have above argued that many tropical forest snail populations will be very small and highly scattered, it is unlikely that each quadrat contains demes of all the species present in the site. Therefore, some degree of spatial species turnover is expected, even on small spatial scales. Fontaine et al. (2007a, 2007b), for example, in Lopé National Park, Gabon, sampled 116 quadrats differing strongly in forest type, habitat, and degree of anthropogenic disturbance, and found great unevenness; in fact, 22 of the 71 recorded species were found only at single stations. Schilthuizen et al. (2002) showed that Whittaker’s I within 20 × 200 m sites in Sabah was twice as great for non-pulmonates as for Pulmonata, which might at least in part reflect a greater degree of species clustering in the former taxon, possibly related to its narrower ecological tolerances (see above).
In sites where quadrat completeness is high (e.g. in limestone areas) and/or where the number of quadrats is very high, and where ecological parameters are measured for each quadrat, multivariate analyses of species-by-ecology may allow sampling error to be distinguished from the habitat diversity responsible for some of the spatial species turnover, although the truly relevant parameters may often be missed. However, if the changes take place along obvious gradients in one or several parameters, it is easier to distinguish sampling from ecological effects. Alvarez and Willig (1993) sampled across treefall-gaps in Puerto-Rican forest and recorded the appearance, disappearance and indifference of macrosnail species with regards to the gap habitat. Liew et al. (2010) and Tattersfield et al. (2001) sampled long elevational transects on Mount Kinabalu and Mount Kenya, respectively, and were able to map the community change over elevational gradients (Fig. 5).
Fig. 5. Pairwise community similarities among 142 quadrats taken along elevational transects on Mount Kinabalu, Borneo. The x-axis shows Jaccard’s index for pairs of plots; the y-axis shows the corresponding difference in elevation for each pair (this figure is based on the work published in Liew et al. (2010); the diagram itself has not been published before).
Conclusion and outlook
The latitudinal gradient in species richness is probably the single most obvious macro-ecological pattern, with species richness on both hemispheres declining with increasing latitude; this pattern is seen in the vast majority of taxa (Brown and Maurer, 1989; Rosenzweig, 1995; Gaston, 2000). In land snails, however, the trend is less than obvious. Although recorded quadrat richness in tropical localities can sometimes be as high as 48 species, and is thereby higher than previously expected (Solem, 1984), this is, in fact, not substantially greater than quadrat richness in subtropical (Stanisic et al., 2007; Cameron et al., 2003) and temperate (Cameron and Pokryszko, 2004; Nekola and Smith, 1999) localities. At a larger spatial scale, tropical land snail species richness is more conspicuously higher than those at higher latitudes due to greater rates of faunal turnover (Nekola, 2005; Stanisic et al., 2007). However, the limited degree of elevated point diversity in the tropics and the much smaller scale at which the first phase transition occurs in Rosenzweig’s (1995) triphasic species-area curve are perhaps the more interesting phenomena, because in these respects snails differ most from other taxa (notably insects (Morse et al., 1988) and plants). The fact that tropical forest snail communities appear to reach almost the same maximum point diversity as temperate ones may suggest that the number of ways in which resource space can be subdivided is quickly exhausted and probably not directly related to the surrounding biotic diversity, but more likely has to do with the available microclimatic and microchemical gradients.
For this reason, perhaps, now that land snail point diversity in the tropics appears to be higher than expected, our next challenge is understanding why it is not even higher. For this, a new style of study will be required. First, more than has been the case until now, malacologists will need to ‘think like a snail’. That is, design their sampling and experiments in such a way that they take place at the scale relevant for the questions to be answered. In many ways, our methods are still burdened by the fact that they have been adopted from methods developed for other organisms, with diversity patterns at scales many orders of magnitude greater. Hence, studies should be preceded by an exploratory nested sampling at ever increasing spatial scales to map the shape of the species-area relationship. Although limited in the largest scales by the fact that complete surveying is needed, this allows an estimation of the appropriate quadrat size. A second exploratory sampling at that quadrat size may then permit an assessment of the correct sampling effort to reach acceptable completeness and abundance for species-abundance distributions. Finally, samples should be segregated into classes of equal body size and separate guilds.
Such a series of conditions, however, poses practical constrictions: low abundances will usually prevent a single size class for a single trophic level in a quadrat of just a few square m to be sampled with any acceptable degree of completeness. An exception, however, are the communities of microsnails that feed on the microvegetated surfaces of limestone rocks (Clements et al., 2006). In these communities, population densities are generally very high, e.g. using 25 × 25 cm census quadrats, Berry (1961) and Schilthuizen et al. (2003b) measured average adult densities of 129-724 per m2 for Opisthostoma species in Malaysia. Another advantage of sampling such limestone rock faces (provided that the potential problem of time-averaging of death assemblages is resolved; see above) is that a steady accumulation of dead representatives collects at the foot of a vertical limestone wall, amounting to thousands of individuals, which can be obtained and separated into size classes by flotation and sieving with differing mesh sizes. Finally, Schilthuizen (2000), Schilthuizen et al. (2003a), and Liew et al. (2008) showed that besides their higher abundances, limestone communities are not particularly exceptional: their apparently higher richness is an illusion caused by the higher abundances (Fig. 6). Also, very few species are obligate limestone-dwelling: the vast majority are also found in adjacent non-calcareous habitats. Similar results were obtained by Nekola (2011) for temperate snail faunas.
Fig. 6. Richness vs. abundance log-log plot for quadrats sampled in forest on limestone (closed symbols) and away from limestone (open symbols), showing that perceived species richness of limestone habitats is an illusion created by higher abundances (modified after Liew et al., 2008).
As an example of the potential of studying SADs in these communities, I sampled quantities of c. 10 i of litter and top soil from underneath flat, near-vertical microvegetated limestone rock faces of 5 × 10 m on several limestone outcrops in Sabah, Malaysian Borneo, and collected the shells by flotation, drying, and sieving. Preliminary data on two such quadrats, ‘Simbaluyon’ and ‘Tinahas’, taken from separate limestone hills c. 10 km apart show that for just one size category (2.0 - 3.5 mm shell diameter), the former yielded 1,590 individuals and 46 species, and the latter gave 10,301 individuals and 41 species (Fig. 7). The large sizes of these small-scale samples are more than sufficient to test SAD models within ‘true’ communities (Fig. 3). Pooling of SADs from different size classes or adjoining quadrats would then allow to test at what spatial scale or body-size scale these models break down to give, e.g. canonical lognormal patterns. I would like to encourage students of tropical forest snail communities to focus on such high-density localities and study SADs in nested series of samples, nested in terms of sampling quadrats, size classes, and, ideally, also trophic levels. Only under such ideal conditions can many of the issues raised in this paper be properly addressed, allowing tropical forest snail community ecology to move from the current, descriptive stage into a higher level of analysis and understanding.
Fig. 7. A portion of a representative collection of same-size microsnails from a Bornean limestone hill, showing the numbers achieved for evaluating the species-abundance distribution.
I would like to thank Peter Tattersfield and Dinarzarde Raheem for inviting me to present a keynote lecture on ‘Tropical forest snail community ecology’ at the World Congress of Malacology, Phuket, Thailand, in July 2010. This paper is an expanded and slightly updated version of that lecture. I also thank various colleagues and friends, who have helped me shape my views on the community ecology of tropical snails, especially, Reuben Clements, Liew Thor-Seng, Wim Maassen, Jaap Vermeulen, and Ton de Winter. A draft of this paper was improved by the comments of Reuben Clements, Liew Thor-Seng, and Ton de Winter, by the reviews by Robert Cameron, Jeff Nekola, and Gary Rosenberg, and by editorial comments by Geerat Vermeij. I thank Liew Thor-Seng for preparing Fig. 6.
Received: 24 July 2010
Revised and accepted: 5 November 2010
Published online: 18 January 2011
Editor: G.J. Vermeij
Alvarez J, Willig MR. 1993. Effects of treefall gaps on the density of land snails in the Luquillo experimental forest of Puerto Rico. Biotropica 25: 100-110.
Barrientos Z. 2000. Population dynamics and spatial distribution of the terrestrial snail Ovachlamys fulgens (Stylommatophora: Helicarionidae) in a tropical environment. Revista de Biología Tropical 48: 71-87.
Berry AJ. 1961. The habitats of some minute Cyclophorids, Hydrocenids, and Vertiginids on a Malayan limestone hill. Bulletin of the National Museum of Singapore 30: 101-105.
Bloch CP, Higgins CL, Willig MR. 2007. Effects of large-scale disturbance on metacommunity structure of terrestrial gastropods: temporal trends in nestedness. Oikos 116: 395-406.
Boettger O. 1894. Ein neuer Amphidromus aus Borneo. Nachrichtsblatt der Deutschen Malakozoologischen Gesellschaft 26: 66-67.
Brown JH, Maurer BA. 1989. Macroecology: the division of food and space among species on continents. Science 243: 1145-1150.
Bruguière JH. 1789. Encyclopédie Méthodique. Histoire Naturelle des Vers. Paris: Panckoucke.
Brummer JE, Nichols JT, Engel RK, Eskridge KM. 1994. Efficiency of different quadrat sizes and shapes for sampling standing crop. Journal of Range Management 47: 84-89.
Cain AJ. 1977. Variation in the spire index of some gastropod shells, and its evolutionary significance. Philosophical Transactions of the Royal Society of London Series B 277: 377-428.
Cameron RAD. 1998. Dilemmas of rarity: biogeographical insights and conservation priorities for land Mollusca. Journal of Conchology Special Publications 2: 51-60.
Cameron RAD, Pokryszko BM. 2005. Estimating the species richness and composition of land mollusc communities: problems, consequences and practical advice. Journal of Conchology 38: 529-548.
Cameron RAD, Mylonas M, Triantis K, Parmakelis A, Vardinoyannis K. 2003. Land-snail diversity in a square kilometre of Cretan maquis: modest species richness, high density and local homogeneity. Journal of Molluscan Studies 69: 93-99.
Cernohorsky N, Horsak M, Cameron RAD. 2010. Land snail species richness and abundance at small scales: the effects of distinguishing between live individuals and empty shells. Journal of Conchology 40: 233-241.
Chao A. 1984. Nonparametric estimation of the number of classes in a population. Scandinavian Journal of Statistics 11: 265-270.
Chao A, Lee SM. 1992. Estimating the number of classes via sample coverage. Journal of the American Statistical Association 87: 210-217
Chapman AD. 2009. Numbers of living species in Australia and the world (2nd Edition). Canberra: Department of the Environment, Heritage, and Arts.
Chazdon RL, Colwell RK, Denslow JS, Guariguata MR. 1998. Statistical methods for estimating species richness of woody regeneration in primary and secondary rain forests of NE Costa Rica. Pp 285-309 in: Dallmeier F, Comiskey JA, eds, Forest biodiversity research, monitoring and modeling: Conceptual background and Old World case studies. Paris: Parthenon Publishing.
Clements R, Sodhi NS, Schilthuizen M, Ng P. 2006. Karsts of Southeast Asia: neglected and imperiled arks of biodiversity. BioScience 56: 733-742.
Clements R, Lu XX, Ambu S, Schilthuizen M, Bradshaw C. 2008. Using biogeographical patterns of endemic land snails to improve conservation planning for limestone karsts. Biological Conservation 141: 2751-2764.
Coddington JA, Griswold CE, Davila DS, Penaranda E, Larcher SF. 1991. Designing and testing sampling protocols to estimate biodiversity in tropical ecosystems. Pp. 44-60 in: Dudley EC, ed., The Unity of Evolutionary Biology: Proceedings of the 4th International Congress of Systematics and Evolutionary Biology. Portland: Dioscorides Press.
Connell JH. 1961. The influence of interspecific competition and other factors on the distribution of the barnacle, Chthamalus stellatus. Ecology 42: 710-723.
Connell JH. 1978. Diversity in tropical rain forests and coral reefs. Science 199: 1302-1310.
Corlett RT. 2009. The Ecology of Tropical East Asia. Oxford: Oxford University Press.
De Chavez ERC, de Lara AV. 2011. Diversity, abundance and spatial distribution of macro land snails in Mount Makiling, Philippines. Asia Life Sciences 20: 185-201.
De Winter AJ, Gittenberger E. 1998. The land snail fauna of a square kilometer patch of rainforest in southwestern Cameroon: high species richness, low abundance and seasonal fluctuations. Malacologia 40: 231-250.
Eguchi K, Bui TV, Janssen R. 2005. Gastropod guests (Prosobranchia: Pupinidae, and Pulmonata: Subulinidae) associated with the ponerine ant Diacamma sculpturatum complex (Insecta: Hymenoptera: Formicidae). Sociobiology 45: 307-315.
Ellwood MDF, Foster WA. 2004. Doubling the estimate of invertebrate biomass in a rainforest canopy. Nature 429: 549-551.
Emberton KC. 1995. Land-snail community morphologies of the highest-diversity sites of Madagascar, North America, and New Zealand, with recommended alternatives to height-diameter plots. Malacologia 36: 43-66.
Emberton KC. 1997. Diversities and distributions of 80 land-snail species in southeastern-most Madagascan rainforests, with a report that lowlands are richer than highlands in endemic and rare species. Biodiversity and Conservation 6: 1137-1154.
Emberton KC, Pearce TA, Randalana R. 1996. Quantitatively sampling land-snail species richness in Madagascan rainforests. Malacologia 38: 203-212.
Emberton KC, Pearce TA, Kasigwa PF, Tattersfield P, Habibu Z. 1997. High diversity and regional endemism in land snails of eastern Tanzania. Biodiversity and Conservation 6: 1123-1136.
Emberton KC, Pearce TA, Randalana R. 1999. Molluscan diversity in the unconserved Vohimena and the conserved Anosy mountain chains, southeast Madagascar. Biological Conservation 89: 183-188.
Fontaine B, Gargominy O, Neubert E. 2007a. Priority sites for conservation of land snails in Gabon: testing the umbrella species concept. Diversity and Distributions 13: 725-734.
Fontaine B, Gargominy O, Neubert E. 2007b. Land snail diversity of the savanna/forest mosaic in Lopé National Park, Gabon. Malacologia 49: 313-338.
Gaston K. 2000. Global patterns in biodiversity. Nature 405: 220-227.
Giokas S, Mylonas M. 2004. Dispersal patterns and population structure of the land snail Albinaria coerulea (Pulmonata: Clausiliidae). Journal of Molluscan Studies 70: 107-116.
Godwin-Austen HH. 1891. On a collection of land-shells made by Mr. A. Everett, with descriptions of supposed new species—Part II. Zonitidae and Helicidae. Proceedings of the Zoological Society of London 22-47.
Goodfriend GA. 1986. Variation in land snail shell form and size and its causes: a review. Systematic Zoology 35: 204-223.
Hearty PJ, Schellenberg SA. 2008. Integrated Late Quaternary chronostratigraphy for San Salvador Island, Bahamas: Patterns and trends of morphological change in the land snail Cerion. Palaeogeography, Palaeoclimatology, Palaeoecology 267: 41-58.
Henderson PA, Magurran AE. 2010. Linking species abundance distributions in numerical abundance and biomass through simple assumptions about community structure. Proceedings of the Royal Society Series B 277: 1561-1570.
Hubbell SP. 2001. The Unified Neutral Theory of Biodiversity and Biogeography. Princeton: Princeton University Press.
Kitching RL. 2000. Food Webs and Container Habitats: The Natural History and Ecology of Phytotelmata. Cambridge: Cambridge University Press.
Lee HG. 2009. Interpreting empty land snail shells. The Shell-O-Gram 50: 3-6.
Liew TS, Clements R, Schilthuizen M. 2008. Sampling micromolluscs in tropical forests: one size does not fit all. Zoosymposia 1: 271-280.
Liew TS, Schilthuizen M, Lakim M. 2010. The determinants of land snail diversity along a tropical altitudinal gradient: insularity, geometry, and niches. Journal of Biogeography 37: 1071-1078
MacArthur RH. 1965. Patterns of species diversity. Biological Reviews 40: 510-533.
Magurran AE. 2004. Measuring Biological Diversity. Oxford: Blackwell Science.
McClain CR. 2005. Bathymetric patterns of morphological disparity in deep-sea gastropods from the western North Atlantic Basin. Evolution 59: 1492-1499.
Michener R, Lajtha K. 2007. Stable Isotopes in Ecology and Environmental Science. Oxford: Blackwell Science.
Morse DR, Stork NE, Lawton JH. 1988. Species number, species abundance and body length relationships of arboreal beetles in Bornean lowland rainforest. Ecological Entomology 13: 25-37.
Nekola JC. 2005. Geographic variation in richness and size of eastern North American land snail communities. Records of the Western Australian Museum 68: 39-51.
Nekola JC. 2011. Acidophilic terrestrial gastropod communities in North America. Journal of Molluscan Studies (in press).
Nekola JC, Brown JH. 2007. The wealth of species: ecological communities, complex systems, and the legacy of Frank Preston. Ecology Letters 10: 188-196.
Nekola JC, Smith TM. 1999. Terrestrial gastropod richness patterns in Wisconsin carbonate cliff communities. Malacologia 41: 253-269.
Nekola JC, Šizling AL, Boyer AG, Storch D. 2008. Artifactions in the log-transformation of species abundance distributions. Folia Geobotanica 43: 259-268.
Oke OC, Alohan FI. 2006. The land snail diversity in a square kilometre of tropical rainforest in Okomu Natioal Park, Edo State, Nigeria. African Scientist 7: 135-142.
Oke OC, Alohan FI, Uzibor MO, Chokor JU. 2008. Land snail diversity and species richness in an oil palm agroforest in Egbeta, Edo State, Nigeria. Bioscience Research Communications 20: 249-256.
Parmakelis A, Mylonas M. 2004. Dispersal and population structure of two sympatric species of the Mediterranean land snail genus Mastus (Gastropoda, Pulmonata, Enidae). Biological Journal of the Linnean Society 83: 131-144.
Peake JF. 1968. Habitat distribution of Solomon Island land Mollusca. Symposia of the Zoological Society of London 22: 319-346.
Pearce TA. 2008. When a snail dies in the forest, how long will the shell persist? Effect of dissolution and micro-bioerosion. American Malacological Bulletin 26: 111-117.
Pigati JS, Quade J, Shahanan TM, Haynes CV. 2004. Radiocarbon dating of minute gastropods and new constraints on the timing of late Quaternary spring-discharge deposits in Southern Arizona, USA. Palaeogeography, Palaeoclimatology, Palaeoecology 204: 33-45.
Raheem DC, Naggs F, Preece RC, Mapatuna Y, Kariyawasan L, Eggleton P. 2008. Structure and conservation of Sri Lankan land-snail assemblages in fragmented lowland rainforest and village home gardens. Journal of Applied Ecology 45: 1019-1028.
Raheem DC, Naggs F, Chimonides PDJ, Preece RC, Eggleton P. 2009. Fragmentation and pre-existing species turnover determine land-snail assemblages of tropical rain forest. Journal of Biogeography 36: 1923-1938.
Řihová D, Janovský Z, Juřičková L. 2010. Land snail shell degradation in temperate forest. Tropical Natural History Supplement 3: 265.
Rosenberg G, Muratov IV. 2006. Status report on the terrestrial Mollusca of Jamaica. Proceedings of the Academy of Natural Sciences of Philadelphia 155: 117-161.
Rosenzweig ML. 1995. Species Diversity in Space and Time. Cambridge: Cambridge University Press.
Rowson B. 2010. Systematics and Diversity of the Streptaxidae with Particular Reference to the East African Region. Unpublished PhD thesis, University of Cardiff, Wales.
Rundell RJ, Cowie RH. 2004. Preservation of species diversity and abundances in Pacific island land snail death assemblages. Journal of Conchology 38: 155-170.
Schilthuizen M. 2000. The evolution and conservation of limestone biotas in Malaysia, with special reference to land snails. Pp. 358-362 in: Leigh M, ed., Borneo 2000; Environment, Conservation, and Land; Proceedings of the Sixth Biennal Borneo Research Conference, Kuching, Sarawak, July 10-14, 2000. Kuching: Borneo Research Council.
Schilthuizen M, Rutjes HA. 2001. Land snail diversity in a square kilometre of tropical rainforest in Sabah, Malaysian Borneo. Journal of Molluscan Studies 67: 417-423.
Schilthuizen M, Lombaerts M. 1994. Population structure and levels of gene flow in the Mediterranean land snail Albinaria corrugata (Pulmonata: Clausiliidae). Evolution 48: 577‑586.
Schilthuizen M, Teräväinen MIF, Tawith NFK, Ibrahim H, Chea SM, Chuan CP, Daim LJ, Jubaidi A, Madjapuni MJ, Sabeki M, Mokhtar A. 2002. Microsnails at microscales in Borneo: distributions of Prosobranchia versus Pulmonata. Journal of Molluscan Studies 68: 259-262.
Schilthuizen M, Chai HN, Kimsin TE, Vermeulen JJ. 2003a. Abundance and diversity of land snails on limestone in Borneo. Raffles Bulletin of Zoology 51: 35-42.
Schilthuizen M, Rosli RB, Ali AMBM, Salverda M, van Oosten H, Bernard H, Ancrenaz M, Lackman-Ancrenaz I. 2003b. The ecology and demography of Opisthostoma (Plectostoma) concinnum s.l. (Gastropoda: Diplommatinidae) on limestone outcrops along the Kinabatangan River. Pp 55-71 in: Mohamed M, Takano A, Goossens B, Indran R, eds, Lower Kinabatangan Scientific Expedition 2002. Kota Kinabalu: Universiti Malaysia Sabah.
Schilthuizen M, Liew TS, Elahan B, Lackman-Ancrenaz I. 2005a. Effects of karst forest degradation on Pulmonate and Prosobranch land snail communities in Sabah, Malaysian Borneo. Conservation Biology 19: 949-954.
Schilthuizen M, Scott BJ, Cabanban AS, Craze PG. 2005b. Population structure and coil dimorphism in a tropical land snail. Heredity 95: 216-220.
Schilthuizen M, van Til A, Salverda M, Liew TS, James SS, Elahan B, Vermeulen JJ. 2006. Micro-allopatric divergence in a snail associated with behavioural differences in its predator. Evolution 60: 1851-1858.
Simpson CT. 1894. Notes on collecting shells in Jamaica. Nautilus 7: 110-113.
Šizling AL, Storch D, Šizlingová E, Reif J, Gaston KJ. 2009. Species abundance distribution results from a spatial analogy of central limit theorem. Proceedings of the National Academy of Science USA 106: 6691-6695.
Solem A. 1974. The Shell Makers: Introducing Mollusks. New York: Wiley.
Solem A. 1984. A world model of land snail diversity and abundance. Pp. 6-22 in: Solem A, van Bruggen AC, eds, World-Wide Snails; Biogeographical Studies on Non-Marine Mollusca. Leiden: Brill.
Solem A, Climo FM, Roscoe DJ, 1981. Sympatric diversity of New Zealand land snails. New Zealand Journal of Zoology 8: 453-485.
Stanisic J, Cameron RAD, Pokryszko BM, Nekola JC. 2007. Forest snail faunas from S.E. Queensland and N.E. New South Wales (Australia): patterns of local and regional richness and differentiaton. Malacologia 9: 445-462.
Sugihara G. 1980. Minimal community structure: an explanation of species abundance patterns. American Naturalist 116: 770-787.
Sugihara G, Bersier LF, Southwood TR, Pimm SL, May RM. 2003. Predicted correspondence between species abundances and dendrograms of niche similarities. Proceedings of the National Academy of Science USA 100: 5246-5251.
Sutton SL. 2001. Alice grows up: canopy science in transition from Wonderland to Reality. Plant Ecology 153: 13-21.
Tattersfield P. 1996. Local patterns of land snail diversity in a Kenyan rain forest. Malacologia 38: 161-180Tattersfield P, Warui CM, Seddon MB, Kiringe.
JW. 2001a. Land-snail faunas of afromontane forests of Mount Kenya, Kenya: ecology, diversity and distribution patterns. Journal of Biogeography 28: 843-861.
Tattersfield P, Seddon MB, Lange CN. 2001b. Land-snail faunas in indigenous rainforest and commercial forestry plantations in Kakamega Forest, western Kenya. Biodiversity and Conservation 10: 1809-1829.
Terry RC. 2010a. The dead do not lie: using skeletal remains for rapid assessment of historical small mammal community baselines. Proceedings of the Royal Society Series B 277: 1193-1201.
Terry RC. 2010b. On raptors and rodents: testing the ecological fidelity and spatiotemporal resolution of cave death assemblages. Paleobiology 36: 137-160.
Tokeshi M. 1999. Species Coexistence: Ecological and Evolutionary Perspectives. Oxford: Oxford University Press.
Vermeulen JJ. 2003. The terrestrial mollusk fauna of Batu Niah N.P. and Gunung Mulu N.P., Sarawak, Malaysia. Copenhagen: DANIDA.
Whittaker RH. 1965. Dominance and diversity in land plant communities. Science 147: 250-260.
Wootton JT. 1994. Predicting direct and indirect effects: an integrated approach using experiments and path analysis. Ecology 75: 151-165.
Wronski T, Hausdorf B. 2008. Distribution patterns of land snails in Ugandan rain forests support the existence of Pleistocene forest refugia. Journal of Biogeography 35: 1759-1768.
Wronski T, Hausdorf B. 2009. Diversity and body-size patterns of land snails in rain forests in Uganda. Journal of Molluscan Studies 76: 87-100.