* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Habitat destruction and metacommunity size in pen
Introduced species wikipedia , lookup
Source–sink dynamics wikipedia , lookup
Molecular ecology wikipedia , lookup
Unified neutral theory of biodiversity wikipedia , lookup
Extinction debt wikipedia , lookup
Biogeography wikipedia , lookup
Ecological fitting wikipedia , lookup
Restoration ecology wikipedia , lookup
Theoretical ecology wikipedia , lookup
Island restoration wikipedia , lookup
Assisted colonization wikipedia , lookup
Mission blue butterfly habitat conservation wikipedia , lookup
Occupancy–abundance relationship wikipedia , lookup
Biodiversity action plan wikipedia , lookup
Latitudinal gradients in species diversity wikipedia , lookup
Biological Dynamics of Forest Fragments Project wikipedia , lookup
Reconciliation ecology wikipedia , lookup
Habitat destruction wikipedia , lookup
Journal of Animal Ecology 2008, 77, 1175–1182 doi: 10.1111/j.1365-2656.2008.01444.x Habitat destruction and metacommunity size in pen shell communities Blackwell Publishing Ltd Pablo Munguia*,‡ and Thomas E. Miller† Department of Biological Science, Florida State University, Tallahassee, FL 32306–1100, USA Summary 1. In spatially structured communities, habitat destruction can have two effects: first, a main effect that occurs because of the loss of habitat area within a larger region, and a secondary effect due to changes in the spatial arrangement of local communities. Changes to the spatial arrangement can, in turn, affect the migration and extinction rates within local communities. 2. Our study involved the experimental destruction of entire local communities within larger regions in natural marine microcosms. Large and small arrays of dead pen shells were created in a shallow bay in north Florida, and the colonization by both encrusting and motile species on this empty substrate were followed through time. After most species had become established, half of the large arrays were perturbed to create small arrays by removal of half the shells, simulating habitat destruction. 3. After 48 days of further community development, comparisons of the large arrays, reduced arrays and original small arrays suggested that the mechanisms by which habitat destruction affects diversity could depend upon the size of the region affected and the natural history of the species being studied. 4. Habitat destruction reduced the diversity of motile species to a level lower than that found in the undisturbed small arrays, suggesting that the species that assembled in the original large metacommunities negatively influenced the species that occurred ultimately in the converted small arrays. 5. With sessile species, habitat destruction created richness levels that were intermediate to those of small and large arrays. The initial predestruction richness appears to have had a positive effect; because sessile species cannot disperse as adults, they may not respond to significant shifts in metacommunity size later in succession. Initial metacommunity size may be important for allowing individuals to select appropriate habitats before they settle. Key-words: abundance and distribution patterns, disturbance, diversity, fragmentation, marine habitats, rare species Journal of Animal Ecology (2007) doi: 10.1111/j.1365-2656.2007.0@@@@.x Introduction Habitat destruction is the primary explanation for the rapid loss of biodiversity in many habitats over the last century – for example, coral reefs have declined by an estimated 27% because of pollution and human exploitation, and more than 78 million acres of tropical forest are estimated to be lost each year to deforestation (Stone 1995; Gardner et al. 2003), with accompanying losses to biodiversity (e.g. Roberts et al. 2002). Despite these widespread effects, we still do not know the mechanisms by which habitat destruction affects diversity *Correspondence author. E-mail: [email protected] †Present address: UMDI-Sisal, Fac. Ciencias, UNAM, Sisal-Hunucma, Yucatan, Mexico, 97355. ‡Marine Science Institute, The University of Texas at Austin, 750 Channel View Drive, Port Aransas, TX, 78373, USA. (Debinski & Holt 2000; Gonzalez 2005). It may act directly by reducing available area: the positive relationship between habitat area and the number of species in a habitat is well known (MacArthur & Wilson 1967; Brown & Lomolino 1998). However, destruction may also have significant secondary effects by changing the spatial structure of populations and communities (Tilman et al. 1994; Gonzalez et al. 1998; Gonzalez 2005). Disturbance can occur at different spatial and temporal scales, with consequences for the resulting community patterns (e.g. Miller 1982). A common ecological scenario is the effect of localized disturbances creating small patches of destruction within a larger matrix of an established community (e.g. by Connell 1978; Paine & Levin 1981). When disturbance occurs on a greater scale (i.e. larger or more frequent disturbances), it can result in a large community being broken down into smaller units (fragments; Collinge © 2008 The Authors. Journal compilation © 2008 British Ecological Society 1176 P. Munguia & T. E. Miller 2000; Gonzalez 2000; Fahrig 2003). If the habitat is already spatially structured (or has been fragmented), such that established smaller ‘local’ communities occur as patches, then large-scale destruction can eliminate entire local communities, affecting both the density and distance among local patches (Gonzalez 2005). Habitat destruction thus has two effects; first, a main effect that occurs because of the loss of area in a habitat, and a secondary effect due to the change in the spatial arrangement of communities. Changes to the spatial arrangement can, in turn, affect the migration rates among fragments and the extinction rates within individual fragments. This secondary effect will be most important in systems where migration strongly affects diversity (Callum 1997). In communities undergoing habitat fragmentation, the movement of individuals from one patch to another has been considered an important mechanism controlling populations and diversity (Debinski & Holt 2000). Metapopulation theory has been used to explain how sources and sinks allow species persistence in habitat fragments (e.g. Gonzalez et al. 1998; Mouquet & Loreau 2003). Similarly, habitat destruction involves the removal of a local community (or fragment, if the community is already fragmented) from the environment and may be particularly important when local communities are linked to one another through migration (i.e. form a metacommunity; Leibold & Miller 2004). One of the proposed outcomes of such habitat destruction is an extinction debt (Tilman et al. 1994), which suggests that asynchronous population dynamics in sources and sinks will cause a reduction in diversity some time after habitat destruction takes place. Other outcomes could include the modification of the abiotic habitat or the loss of habitat heterogeneity, both of which could have potentially negative effects on diversity. We can understand the effects of habitat destruction on diversity by considering how local diversity changes with succession, habitat area and fragmentation (Mouquet et al. 2003). New habitats are expected to gain species through time up to some asymptote, with larger metacommunities (i.e. those with more local communities) supporting more species than smaller ones, either because of the species–area relationship (Connor & McCoy 1979) or because there are more habitats that offer refugia and heterogeneity to a larger suite of species (Leibold et al. 2004; Mouquet et al. 2006). The reduction in the number of local habitats through habitat destruction might therefore have one of three effects on diversity. First, the null expectation is that diversity might simply decrease to a level appropriate to the new metacommunity size. Secondly, the effects of a small metacommunity and some effect of the destruction itself on the spatial or landscape pattern (e.g. extinction debt) might combine to produce a diversity lower than that expected from the new metacommunity size (we call this a negative residual effect). Thirdly, species from the original larger metacommunity might persist, resulting in a new diversity level somewhat higher than that expected from the new metacommunity size (a positive residual effect). This last response can occur due to historical processes during community formation; positive species interactions at the local scale may allow species persistence after destruction. Alternatively, high rates of dispersal may counter local extinctions and allow species persistence even after habitat destruction has taken place (e.g. rescue or mass effects, Leibold et al. 2004). Destruction of local communities can also affect the patterns of species’ relative abundance. Studies have demonstrated a positive relationship between local abundance and regional distribution (Brown 1984; Gaston, Blackburn & Lawton 1997), so most species are believed to be on a continuum between locally rare and having narrow distributions to locally abundant and having broad distributions. This abundance–distribution relationship may change, however, with successional patterns (Mouquet et al. 2003), as well as with mechanisms such as habitat destruction that affect whole communities. Following habitat destruction, some species may increase their distribution or local abundance, while others will decline, depending on species-specific responses to disturbance (Mouquet & Loreau 2003; Shurin et al. 2004). We used the natural communities of small marine invertebrates found living in empty pen shells as experimental microcosms (see, e.g. Srivastava et al. 2004) to test the effects of metacommunity size and habitat destruction on local diversity and species commonness and rarity. Pen shell communities are a useful system for testing scaling issues in diversity (Munguia 2004), because each shell can be considered as a discrete habitat among a seagrass matrix where many species experience these habitats as metapopulations (Munguia, Mackie & Levitan 2007). In particular, we test whether small metacommunities have similar diversity levels as large metacommunities, and test if a shift in community density (number of habitats per unit area) through destruction will change diversity in different types of pen shell species. Then, we test whether the effect of habitat destruction is simply a reduction of habitat area or if it is a reduction of the species pool in that particular region. We also compare changes in species’ habitat occupancy and local abundance through successional time to determine whether species responded similarly to habitat destruction. Methods Our study was carried out in the summer of 2003 in St Joe Bay, Florida, a shallow, well-protected bay on the northern Gulf of Mexico. The substrate is composed of patches of sea-grass beds intermixed with sandy areas; very few natural hard surfaces are available other than the empty shells of dead pen shells Atrina rigida (Lightfoot). Pen shells are relatively large bivalves (~19 cm length) that live embedded in the sand within sea-grass beds. The shells remain in the sand once the mollusc dies, providing habitat for a large number of invertebrates and fish that are generally not found in the areas surrounding pen shells (Munguia 2004, 2007). The approximately 70 species occurring on dead pen shells experience processes at three discrete spatial scales: within an individual shell, where species interactions such as competition and predation generally occur, among neighbouring shells where individuals may move during their lifetimes (which we will define operationally as within the pen-shell metacommunity) and a much larger spatial scale © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182 Habitat destruction in metacommunities Fig. 1. Experimental design of the habitat destruction study. Top: three treatments of different densities were used: a large array (L) consisting of 16 shells, a small array (S) of four shells and a habitat destruction array (D) comprised of 16 shells, of which 12 shells were removed during community formation (dark circles). Bottom: during each sampling period, shells had to be sampled destructively; therefore there were replicates for each sampling date. At the 21-day collection time the destruction treatment was still equal to the large treatment, so we used the shells that were ‘destroyed’ as samples to estimate diversity in the destruction arrays (hence the ‘large’ box surrounding the D). Four blocks contained one array for each collection time, and four blocks contained one array that was collected at the last sampling time. The design had more power at the last sampling date (eight blocks) than when we first sampled (four blocks). 1177 evenly spaced shells and one small square array consisting of four shells, keeping the distance between shells constant (40 cm) within each array (Fig. 1, top). All the shells that were anchored were empty and fouling-free and of relatively the same size (average of 20 cm in length), allowing us to follow natural succession patterns (see Munguia 2004). In order to quantify all the species occurring on pen shells on any given sample date, shells had to be sampled destructively. Sampled shells were brought to the surface using sealed bags preventing the loss of organisms, and then taken to the laboratory for identification and quantification of inhabitants larger than 1 mm2 in size. This procedure does not affect neighbouring shells, as disturbance to the surrounding habitat is minimal, and species tend to remain within the refuge of the pen shell. After 21 days, 12 shells were removed from one of the two large arrays on each plot, leaving an array of four neighbouring shells: two shells that were on the outside of the original array and two on the inside, in order to account for potential edge effects. Within each plot, the large array represented the ‘large metacommunity’ treatment, the small array the ‘small metacommunity’ treatment and the reduced array the ‘habitat destruction’ treatment (Fig. 1, bottom). We defined the species pool as the species that were found at the scale of the entire array of four or 16 shells (an array therefore becomes a metacommunity), and local diversity was defined at the scale of a single shell within an array. Habitat destruction occurred at mid-succession, when all species in the species pool were present on at least some shells in the array (Munguia 2006). This timing allowed the effects of both habitat destruction and community age (Mouquet et al. 2003) to occur in our experiment. We established these arrays in eight areas chosen randomly within St Joe Bay. Four of these areas contained three arrays of each treatment type, one for each collection time, and four of these areas contained only one array of each type, which was collected at the last sampling time (Fig. 1, bottom). At 21 and 63 days after deployment of the arrays, one array of each treatment was collected for censusing of species present (n = 4 for 21 days, n = 8 for 63 days). Previous experiments on succession in pen shell communities (Munguia 2004, 2006) have demonstrated that this time-period (63 days) was sufficient for empty shells to be colonized and approach some relatively stable near-equilibrium state. CHANGES IN LOCAL SPECIES RICHNESS at which reproductive propagules may disperse (including the entire bay and possibly parts of the Gulf of Mexico). For analyses, we divided species found on pen shells into two groups based on their ability to move among shells within a metacommunity. Motile species, such as crustaceans and fishes, can move among communities as juveniles and adults, whereas sessile species, such as barnacles and bryozoans, have limited motility and generally move among communities only as propagules (Munguia 2004, 2007). The majority of motile species are found only on pen shells, but show random distribution patterns. However, there are few motile species that aggregate within shells (i.e. very abundant when present), such as the shrimp Palaemon floridianus (Chace) and the amphipod Dulichiella appendiculata (Say); in contrast, most sessile species show clumped distributions (Munguia 2007). Therefore, the fate of the habitat has the potential to cause different responses by species with these distribution patterns, particularly when species can have different stage-dependent dispersal strategies (Munguia et al. 2007). Replicate plots of three arrays of anchored pen shells were established within eight 2·25 m2 areas: two large square arrays consisting of 16 To address the primary question of the effects of destruction on species richness, we compared local diversity among all three treatments. Four shells from each of the large arrays that had the same position within their array as the habitat destruction treatment were used for statistical comparisons with the other two treatments, so that sample sizes and positions would be as similar as possible. Species richness of the three treatments was compared at each collection time by analysis of variance, treating plots as a blocking effect to account for variance due to spatial location. REDUCTION IN HABITAT AREA EFFECTS After quantifying the effects of destruction, we then tested two potential sources through which habitat destruction might affect species richness: the reduction of habitat area and the reduction of the species pool in that particular region. If destruction is simply a change in habitat area through decreasing the number of local communities (i.e. pen shells), then the size of the available pool of species should also decrease. We compared the regional richness © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182 1178 P. Munguia & T. E. Miller from the large array treatment (all 16 shells within an array for all eight sites), a simulated small array treatment (using data drawn from four subsampled shells from the 16-shell array) and the destruction treatment (four shells). Because of differences in sample size, and because the values in the subsample were part of the large array treatment, we tested these differences by bootstrapping regional richness of the large arrays (2000 iterations of 16 shells per site per treatment); the regional diversity of the small arrays were also bootstrapped for reference. Using the 95% confidence intervals across sites, we tested for the significance of the difference between two treatments with Tukey’s procedure (Zar 1999). If the confidence intervals crossed zero (i.e. difference between means is not different from zero), then the two compared treatments were not different from one another. Our a priori hypothesis was that regional diversities would vary in the following fashion: first, if regional diversity in the large array differed from the simulated small array, and this one in turn was similar to the diversity found in the habitat destruction treatment, then we would assume that the effect of habitat destruction is due simply to the reduction in area associated with the removal of pen shells. Alternatively, if regional diversity in the large array was similar to the diversity in the simulated small arrays, and the diversity in the simulated small arrays differed from the diversity in the habitat destruction treatment, then we would assume that habitat destruction has an additional effect on the species pool over and above that due to the loss of area associated with the removal of pen shells. CHANGES IN COMMONNESS AND RARITY For each colonizing species, the mean number of shells (habitats) occupied in each plot for each treatment was compared with the maximum local abundance in a plot for that species. Similar comparisons have been conducted previously in other systems at a single sampling time (Brown 1984; Gaston et al. 1997). In our analysis, we plotted the abundance against the distribution for each species for the first and last sampling dates, creating a vector between the two dates. We then standardized all the species vectors by setting the first collection point to zero; the vectors therefore reflected the direction and magnitude of change in local abundance and the number of shells occupied. In order to avoid pseudoreplication, we averaged all the vectors for each species within a treatment across the replicates, and then averaged across species to make comparisons among treatments. Changes in both abundance and proportion of shells occupied were subjected to angular transformation, and we compared all possible pairs of the average vectors for the three treatments using a non-parametric two-sample second-order analysis of angles (Zar 1999). Fig. 2. Species richness pre- and post-habitat destruction was implemented for motile (a) and sessile (b) species. Open bars represent small arrays of local communities (four communities), grey bars represent large arrays (16 communities) and black bars represent arrays that underwent habitat destruction (a shift from 16 to four communities). Bars are treatment means with one standard error; different letters represent statistically different levels of species richness (P < 0·05; Tukey’s honestly significant difference test); there were no predestruction treatment differences. Similarly, sessile species had similar local richnesses at 21 days (F = 1·07, P > 0·31). Species diversity in all three treatments continued to increase from day 21 to day 63 after the implementation of habitat destruction (Fig. 2b), but they differed significantly in species richness at 63 days (F = 5·37, P = 0·008). At the end of the experiment, richness in the undisturbed large array treatment was greater than in the small arrays, with the habitat destruction treatment having an intermediate level between the two. Results REDUCTION IN HABITAT AREA EFFECTS CHANGES IN LOCAL SPECIES RICHNESS Motile species richness in local communities increased with time in both large and small metacommunities: at 21 days, richnesses were not significantly different among the large and small arrays and the habitat destruction treatments prior to destruction (F = 1·49, P > 0·23; Fig. 2a). By the final sample (day 63), the treatments diverged in species richness (F = 7·09, P < 0·0001), both the large and small arrays had similar levels of diversity, while the habitat destruction treatment had lower diversity. Changes in regional species richness are due to more than simply changes in the number of local communities, based upon an inspection of the 95% confidence intervals from the bootstrapped values across the large array (L; considering a 16-shell sample), simulated small array (Ls; four-shell sample of the 16 shells) and destruction (D; four-shell sample) treatments (Fig. 3). Regional diversity of pen shells in the habitat destruction treatment appears to drop a couple of species due to changes in species pool size and in the number of habitats. Further, the destruction treatment presents less © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182 Habitat destruction in metacommunities 1179 Fig. 3. Bootstrapped species richness values for both motile and sessile species for the last sampling date (63 days). Open circles represent large arrays, shaded triangles simulated arrays and closed circles arrays subjected to ‘habitat destruction’; open triangles represent small arrays and are shifted to the right for clarity. Error bars are 95% confidence intervals; lack of overlap in error bars suggests differences between groups. variation in the number of species than the simulated small array and the actual small array. CHANGES IN SPECIES COMMONNESS AND RARITY In large metacommunities over time, motile species tended to increase in proportion of habitats occupied as well as in local abundance (Fig. 4a). In small metacommunities, the proportion of habitats occupied by each motile species decreased over time, but local population abundances actually increased. Local diversity in small arrays can therefore reach levels expected for larger arrays. Habitat destruction generally caused populations to experience lower population growth than did those in the large and small metacommunities and therefore little change in the relationship between abundance and habitat occupancy (Fig. 4a). Each of the treatment vectors was significantly different from the others [two-sample analysis of vectors: large and small arrays, degrees of freedom (d.f.) = 43,46, U2 = 0·46, P < 0·01; large and habitat-destruction arrays, d.f. = 43,50, U2 = 0·75, P < 0·01; small and habitatdestruction arrays, d.f. = 46, 50, U2 = 0·60, P < 0·01]. Sessile species increased in both local abundance and number of habitats occupied over time in all three treatments (Fig. 4b), but the magnitudes of increase differed; increases were greatest in large arrays and smallest in small arrays (two-sample analysis of vectors: large and small arrays, d.f. = 19,20, U2 = 0·44, P < 0·01; large and habitat-destruction arrays, d.f. = 19,21, U2 = 0·78, P < 0·01; small and habitatdestruction arrays, d.f. = 20,21, U2 = 0·45, P < 0·01). As with motile species, the habitat destruction correlated with lower rates of growth and habitat spread (Fig. 4b). Fig. 4. Species abundance–distribution trajectories under succession from the field experiment for (a) motile and (b) sessile species. Open circles represent large arrays, triangles small arrays and closed circles arrays subjected to ‘habitat destruction’. Discussion Results from our study suggest that habitat destruction can have complex effects on local community structure. For some types of species, destruction reduces available area, leading to a decrease in richness as described by species–area relationships. However, there are also secondary effects of successional history or the deleterious effects of destruction that affect the response of a community to habitat destruction. The regional species pool for both motile and sessile species was reduced under habitat destruction by only a couple of species. This small reduction in species pool suggests that habitat destruction is affecting local diversity much more strongly than regional diversity in pen shell communities. Furthermore, the effects of area alone and the effects of secondary mechanisms on diversity will vary depending upon whether species involved are motile or sessile. © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182 1180 P. Munguia & T. E. Miller Local richness of motile species in large arrays increased as the community underwent succession (Fig. 2a). Contrary to our expectations, small arrays reached the same level of richness, suggesting that there was no significant increase in richness with metacommunity area for these species. In large arrays subjected to habitat destruction during succession, however, richness was lower than in either the unmanipulated large or small arrays. However, the species pool was not affected severely by the destruction treatment; rather, the distribution of individual motile species and local diversity was mainly affected. Therefore, habitat destruction had a secondary, negative, effect on the number of motile species, suggesting that the species that assemble in large arrays dominate or otherwise affect negatively the communities that occur in the small arrays created through habitat destruction. The number of sessile species in both small and large arrays also increased through the last sampling date, with a lower local richness in small arrays (Fig. 2b). This result is consistent with our expectation that the initial metacommunity size should influence species richness. Arrays subjected to habitat destruction achieved richness levels that were intermediate between those of the original small and large arrays. The initial predestruction richness appears to have had a positive secondary effect on the sessile species richness after destruction. Because sessile species cannot disperse as adults, they may not respond to significant shifts in metacommunity size, such as occur with destruction – and this is reflected in the weak ecological effects observed. Alternatively, sessile species may be subject to a mechanism similar to the extinction debt (Tilman et al. 1994) due to a slower response by sessile species’ populations. Initial metacommunity size may be important for allowing individuals to select appropriate habitats before they settle (Mouquet & Loreau 2003). However, once these individuals are established, their presence may suppress incoming recruits, suggesting that priority effects could be important for this species group (see, e.g. Tilman, Lehman & Yin 1997; Almany 2003; Fukami 2004; Munguia 2004). H A B I T A T D E S T R UCT ION AND REGIONAL SPECIES P OOLS Habitat destruction also affected the regional species pool of potential colonizers. Both motile and sessile species pools were significantly different between the destruction treatment and the simulated small array. The reduction of number of habitats available had direct effects on the number of potential colonizers, independent of the size of the metacommunity array. Motile species diversity was not affected by array size, but rather by the direct effect of destruction that, in turn, reduced the species pool. Motile species, such as the amphipods D. appendiculata and Bemlos unicornis (Bynum and Fox), may have tracked available habitats (Munguia 2007; Munguia et al. 2007); therefore, a reduction in habitats had the potential to affect source populations. However, not all species responded the same way to habitat destruction; for example, the hermit crab Pagurus sp. was affected mainly by the original array size, having larger populations in the destruction and large arrays than in the small arrays; while the isopod Paracerceis sp. maintained higher populations in the destruction and small arrays than in the large arrays. Finally, the amphipod Melita nitida S. I. Smith showed a much higher abundance in the destruction treatment than in the large and small arrays, which could be due to the nomadic nature of this species (Munguia et al. 2007). Sessile species diversity was affected by the initial array size; the reduction of the species pool is linked probably to the number of habitats available. This pattern could have occurred because species such as the ascidian Didemnum sp. and the sponge Haliclona sp. require hard substrate to settle and grow, and their distribution patterns are similar to the distribution of pen shells (Munguia 2007). ABUNDANCE–DISTRIBUTION RELATIONSHIP In our experiment, habitat destruction is related to lower pen shell occupancy, due perhaps to lower migration of motile species among shells relative to the undisturbed large and small metacommunities. Destruction may have caused a reduction of potential sources for motile species which may, in turn, reduce the likelihood that a local population would serve as a source and seed neighbouring localities. Changes in species distribution among local habitat (shells) differed among the three treatments (Fig. 4a). Smaller arrays had lower local diversity, perhaps because of increased species interactions or lower rates of migration from fewer sources. Motile species may be sorting themselves out among habitats, with lower richness on each local habitat (Sale 1977; Leibold & Miller 2004), and a reduction of habitats increases the likelihood of a species from being excluded from the group of habitats. Changes over time in sessile species distribution and abundance suggest that short-distance dispersal is important for some of these species (Olson 1985; Bingham & Young 1991) and that communities at high densities therefore showed high species richness (Fig. 2b). Sessile species were not affected by metacommunity size or habitat destruction because the adults could not disperse: they had similar population-growth patterns and distributions in all arrays (Fig. 4b). Successional changes in sessile species were independent of the size of the metacommunity once the initial colonization pattern had occurred. What is interesting to note is that the initial assembly or colonization pattern (Belyea & Lancaster 1999) can affect local diversity but not the increase in abundance or distribution of individual species. The metacommunity concept seems especially appropriate for the communities associated with pen shells, as the shells define discrete spatial scales at which different processes may occur (Munguia 2007). However, as with other studies that attempt to relate local processes to regional mechanisms in natural systems, delimiting relevant spatial scales is problematic (Srivastava 1999; Munguia 2004). The possibility of very high long-distance dispersal for some species found on pen shells means that the community does not have a regionally closed system, as is assumed in most metacommunity theory (Leibold et al. 2004). It is likely that most natural metacommunities are not enclosed completely, at least at the scale that © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182 Habitat destruction in metacommunities regional processes (such as dispersal and habitat heterogeneity) are thought to operate. In pen shells, species respond to our manipulations of metacommunity size, which suggests that migration is important from neighbouring shells up to metres away. However, there is probably a low propagule input from larger scales, which could be crucial for establishing the first colonizers in pen shell metacommunities. An important question that would need to be addressed with natural systems is how different spatial scales change the influences of processes as communities undergo succession. Our study supports recent theoretical studies showing that the interaction between dispersal limitation, species interactions and habitat structure can affect diversity (Tilman et al. 1997; Chesson 2000; Amarasekare et al. 2004). Clearly, habitat destruction can have both main (through species–area relationships) and secondary effects (positive or negative residual effects) on community diversity. Species have different dispersal rates, growth rates and competitive abilities, which contributed to differences in abundance across treatments. In general, species that are locally rare also occupy few habitats (Brown 1984; Gaston et al. 1997) and may therefore also be more susceptible to the effects of habitat destruction (Gonzalez et al. 1998). The observed community patterns result from the different responses of and interactions among component species. The response of natural communities to habitat destruction depends clearly upon the scales of the habitat, the destruction and the species involved. Investigating patterns of species’ commonness and rarity (Magurran & Henderson 2003) provides insight into changes in species abundances, suggesting in particular that they are (a) dispersal-limited, (b) resourcelimited or (c) limited by species interactions (e.g. competition and predation). Habitat destruction can affect any of these parameters, but their combined effects can be understood only in the larger, metacommunity context. Manipulating whole communities has allowed us to study the interaction between regional size, habitat destruction and their combined effects on diversity. Acknowledgements We would like to thank Don R. Levitan, Andy Gonzalez, Brian D. Inouye, Casey Terhorst and Dave Ferrell for comments to the manuscript, as well as Kim Young for field assistance. K. Moriuchi assisted with the bootstrap procedure. This research was funded by a COFRS grant from Florida State University to T. E. Miller. References Almany, G.R. (2003) Priority effects in coral reef fish communities. Ecology, 84, 1920 –1935. Amarasekare, P., Hoopes, M.F., Mouquet, N. & Holyoak, M. (2004) Mechanisms of coexistence in competitive metacommunities. American Naturalist, 164, 310 – 326. Belyea, L.R. & Lancaster, J. (1999) Assembly rules within a contingent ecology. Oikos, 86, 402–416. Bingham, B.L. & Young, C.M. (1991) Larval behavior of the ascidian Ecteinascidia turbinata Herdman – an in situ experimental study of the effects of swimming on dispersal. Journal of Experimental Marine Biology and Ecology, 145, 189–204. Brown, J.H. (1984) On the relationship between abundance and distribution of species. American Naturalist, 124, 255–279. Brown, J.H. & Lomolino, M.V. (1998) Biogeography. Sinauer Associates Inc., Sunderland, MA. 1181 Callum, M.R. (1997) Connectivity and management of Caribbean coral reefs. Science, 278, 1454–1457. Chesson, P. (2000) General theory of competitive coexistence in spatiallyvarying environments. Theoretical Population Biology, 58, 211–237. Collinge, S.K. (2000) Effects of habitat fragmentation on insect species loss, colonization, and movement patterns. Ecology, 81, 2211–2226. Connell, J.H. (1978) Diversity in tropical rain forests and coral reefs. Science, 199, 1302–1310. Connor, E.F. & McCoy, E.D. (1979) The statistics and biology of the species–area relationship. American Naturalist, 113, 791–833. Debinski, D.M. & Holt, R.D. (2000) A survey and overview of habitat fragmentation experiments. Conservation Biology, 14, 342–355. Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity. Annual Review of Ecology and Systematics, 34, 487–515. Fukami, T. (2004) Assembly history interacts with ecosystem size to influence species diversity. Ecology, 85, 3234–3242. Gardner, T.A., Cote, I.M., Gill, J.A., Grant, A. & Watkinson, A.R. (2003) Long-term region-wide declines in Caribbean corals. Science, 301, 958 – 960. Gaston, K.J., Blackburn, T.M. & Lawton, J.H. (1997) Interspecific abundance– range size relationships: an appraisal of mechanisms. Journal of Animal Ecology, 66, 579–601. Gonzalez, A. (2000) Community relaxation in fragmented landscapes: the relation between species richness, area and age. Ecology Letters, 3, 441– 448. Gonzalez, A. (2005) Local and regional community dynamics in fragmented landscapes: insights from a bryophyte-based microcosm. Metacommunities: Spatial Dynamics and Ecological Communities (eds M. Holyoak, M.A. Leibold & R.D. Holt). Chicago University Press, Chicago, IL. Gonzalez, A., Lawton, J.H., Gilbert, F.S., Blackburn, T.M. & Evans-Freke, I. (1998) Metapopulation dynamics, abundance, and distribution in a microecosystem. Science, 281, 2045–2047. Leibold, M.A. & Miller, T.E. (2004) From metapopulations to metacommunities. Ecology, Genetics, and Evolution of Metapopulations (eds I. Hanski & O.E. Gaggiotti), pp. 133–150. Elsevier, San Diego, CA. Leibold, M.A.M.M.H., Mouquet, N., Amarasekare, P., Chase, J.M., Hoopes, M.F., Holt, R.D., Shurin, J.B., Law, R., Tilman, D., Loreau, M. & Gonzalez, A. (2004) The metacommunity concept: a framework for multi-scale community ecology. Ecology Letters, 7, 601–613. MacArthur, R.H. & Wilson, E.O. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Magurran, A.E. & Henderson, P.A. (2003) Explaining the excess of rare species in natural species abundance distributions. Nature, 422, 714 – 716. Miller, T.E. (1982) Community diversity and interactions between the size and frequency of disturbance. American Naturalist, 120, 533–536. Mouquet, N. & Loreau, M. (2003) Community patterns in source–sink metacommunities. American Naturalist, 162, 544–557. Mouquet, N., Miller, T.E., Daufresne, T. & Kneitel, J.M. (2006) Consequences of varying regional heterogeneity in source–sink metacommunities. Oikos, 113, 481–488. Mouquet, N., Munguia, P., Kneitel, J.M. & Miller, T.E. (2003) Community assembly time and the relationship between local and regional species richness. Oikos, 103, 618–623. Munguia, P. (2004) Successional patterns on pen shell communities at local and regional scales. Journal of Animal Ecology, 73, 64–74. Munguia, P. (2006) Diversity patterns of pen shell (Atrina rigida) communities. PhD dissertation, Florida State University, Tallahassee, FL. Munguia, P. (2007) Spatial structure of pen shell (Atrina rigida) communities. Marine Biology, 152, 149–156. Munguia, P., Mackie, C. & Levitan, D.R. (2007) The influence of stage-dependent dispersal on the population dynamics of three amphipod species. Oecologia, 153, 533–541. Olson, R.R. (1985) The consequences of short-distance larval dispersal in a sessile marine invertebrate. Ecology, 66, 30–39. Paine, R.T. & Levin, S.L. (1981) Intertidal landscapes: disturbance and the dynamics of pattern. Ecological Monographs, 51, 145–178. Roberts, C.M., McClean, C.J., Veron, J.E.N., Hawkins, J.P., Allen, G.R., McAllister, D.E., Mittermeier, C.G., Schueler, F.W., Spalding, M., Wells, F., Vynne, C. & Werner, T.B. (2002) Marine Biodiversity Hotspots and Conservation Priorities for Tropical Reefs. Science, 295, 1280–1284. Sale, P.F. (1977) Maintenance of high diversity in coral reef fish communities. American Naturalist, 111, 337–359. Shurin, J.B., Amarasekare, P., Chase, J.M., Holt, R.D., Hoopes, M.F. & Leibold, M.A. (2004) Alternative stable states and regional community structure. Journal of Theoretical Biology, 227, 359–368. © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182 1182 P. Munguia & T. E. Miller Srivastava, D.S. (1999) Using local-regional richness plots to test for species saturation: pitfalls and potentials. Journal of Animal Ecology, 68, 1– 16. Srivastava, D.S., Kolasa, J., Bengtsson, J., Gonzalez, A., Lawler, S.P., Miller, T., Munguia, P., Schneider, D. & Trzcinski, M.K. (2004) Miniature worlds: are natural microcosms the new model systems for ecology? Trends in Ecology and Evolution, 19, 379–384. Stone, L. (1995) Biodiversity and habitat destruction: a comparative study of model forest and coral reef ecosystems. Proceedings of the Royal Society of London, Series B, 261, 381–388. Tilman, D., Lehman, C.L. & Yin, C. (1997) Habitat destruction, dispersal, and deterministic extinction in competitive communities. American Naturalist, 149, 407–435. Tilman, D., May, R.M., Lehman, C.L. & Nowak, M.A. (1994) Habitat destruction and the extinction debt. Nature, 371, 65–66. Zar, J.H. (1999) Biostatistical Analysis, 4th edn. Prentice Hall, Upper Saddle River, NJ. Received 23 April 2008; accepted 22 May 2008 Handling Editor: Andy Gonzalez © 2008 The Authors. Journal compilation © 2008 British Ecological Society, Journal of Animal Ecology, 77, 1175–1182