* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Convergence in Morphological Patterns and Community
Survey
Document related concepts
Banksia brownii wikipedia , lookup
Habitat conservation wikipedia , lookup
Biological Dynamics of Forest Fragments Project wikipedia , lookup
Unified neutral theory of biodiversity wikipedia , lookup
Biodiversity action plan wikipedia , lookup
Island restoration wikipedia , lookup
Gartons Agricultural Plant Breeders wikipedia , lookup
Occupancy–abundance relationship wikipedia , lookup
Molecular ecology wikipedia , lookup
Latitudinal gradients in species diversity wikipedia , lookup
Reconciliation ecology wikipedia , lookup
Ecology of Banksia wikipedia , lookup
Transcript
vol. 158, no. 5 the american naturalist november 2001 Convergence in Morphological Patterns and Community Organization between Old and New World Rodent Guilds Ariela Ben-Moshe,1 Tamar Dayan,1,* and Daniel Simberloff 2,† 1. Department of Zoology, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel; 2. Department of Ecology and Evolutionary Biology, University of Tennessee, Knoxville, Tennessee 37996-1610 Submitted August 22, 2000; Accepted May 23, 2001 abstract: We studied morphological relationships within three guilds of gerbillid rodents in Israel. We found a nonrandom pattern of overdispersed means (community-wide character displacement) for upper incisor widths among the species in these three guilds. Upper tooth-row lengths, condylo-basal skull lengths, and tooth-row surfaces displayed similar patterns. We also studied seed-size selection by two well-studied gerbil species, which have previously been found to compete, in order to test whether specializing on husking seeds of different sizes as a mechanism of coexistence may underlie the morphological patterns. The seed-size selection experiments took place in two large aviaries with artificial lighting simulating full-moon nights, which is when predation risk is perceived as high. Seeds of different sizes (commercial seeds in one experiment and husked wheat particles in the other) mixed with sand were offered in trays. The larger Gerbillus pyramidum took significantly larger commercial seeds and marginally larger wheat particles than the smaller Gerbillus allenbyi. The patterns attest to ecomorphological convergence at the guild level; we previously demonstrated size structuring in several North American heteromyid rodent guilds, and we now report similar size structuring among Israeli gerbillid guilds. The occurrence of convergent community structure strongly indicates general rules governing ecological communities or guilds. Keywords: community-wide character displacement, community structure, ecomorphology, guild convergence, seed-size selection. A major goal of ecology is identifying and interpreting recurring patterns in ecological communities. Elucidating * Corresponding author; e-mail: [email protected]. † E-mail: [email protected]. Am. Nat. 2001. Vol. 158, pp. 484–495. 䉷 2001 by The University of Chicago. 0003-0147/2001/15805-0003$03.00. All rights reserved. such patterns affords the possibility to understand the forces that structure ecological communities, to fathom the rules that govern them, in fact, to determine whether such rules exist at all (cf. Lawton 1999), and, if they do, to assess their generality. The occurrence of convergent community structure would strongly indicate general governing rules. If similar selection regimes produce similarly structured ecological communities, then general structuring rules can be sought and analyzed. Since whole ecological communities are extremely complex entities, ecologists usually study parts of communities. Ecological guilds (sensu Root 1967) have often been taken as such manageable units that allow detailed study (Simberloff and Dayan 1991); therefore, the search for guild convergence may be more promising than the search for convergence between entire communities. If guilds are building blocks of communities, understanding their structure and the forces operating within them may provide insight into the structure and function of communities. A conspicuous case of apparent convergence is found among two of the most thoroughly studied mammalian ecological guilds: the North American desert rodents of the family Heteromyidae and members of the Old World family Gerbillidae (Abramsky 1983). Most convergences and parallels are recognized by striking visual resemblances (Begon et al. 1990); general similarities in size and color, inflated tympanic bullae, and a tendency toward bipedal locomotion (albeit to different degrees) occur in both families. Seeds of ephemeral desert plants are a major item in the diet of heteromyids and gerbillids, and this similarity in resource base plus the general morphological similarities among these two groups of taxonomically distantly related rodents are often taken as a case of evolutionary convergence (Brown et al. 1979; Abramsky et al. 1985). However, Brown et al. (1994) propose that this apparent morphological convergence may be superficial; they base this conjecture on dietary differences between potential analogous members of heteromyid and gerbillid guilds rather than Old and New World Rodent Guilds 485 on specific measures of morphology. In fact, they see an intriguing lack of morphological convergence between granivorous rodents of the Old World and New World and suggest that factors that shape the more conspicuous attributes of species may not be the same as those that promote or inhibit species coexistence. One manifestation of community or guild morphological structure is community-wide character displacement (sensu Strong et al. 1979), which is the occurrence of nonrandom size relationships among guild members (means for the different species overdispersed). This phenomenon has generally been interpreted as resulting from interspecific competition and resource partitioning by food size as a mechanism of coexistence (e.g., Holmes and Pitelka 1968; Dayan et al. 1989a, 1989b, 1990, 1992; Dayan and Simberloff 1994a). Such size-structured communities or guilds may arise from species sorting and/or from an actual coevolutionary change and may attest to the ecological and evolutionary importance of competition in structuring ecological communities (Case and Sidell 1983; Case et al. 1983; Sinclair et al. 1985a, 1985b; Dayan and Simberloff 1998). Community-wide character displacement has been described for sets of species in both Old and New World seed-eating rodent guilds (Yom-Tov 1991 and Dayan and Simberloff 1994b, respectively). Were such morphological patterns convergent, they could indicate shared, underlying community-wide selective forces. Curiously, however, the patterns have been found in different morphological characters and have been given different interpretations. Dayan and Simberloff (1994b) demonstrated regular morphological relationships among the upper incisors of heteromyid rodents. Incisors of heteromyids are used for husking, some of which takes place aboveground where predation risk is high (Rosenzweig and Sterner 1970; Lemen 1978). Therefore, husking speed may be critical (Rosenzweig and Sterner 1970). Dayan and Simberloff (1994b) hypothesized that for each species there may be an optimal seed size, too large to be efficiently stored in the heteromyids’ external cheek pouches without husking but that can be husked efficiently enough to outweigh the risk of predation, which merits being the target of specialization. The strong relationship between pouch volume and incisor width for these species supports this hypothesis. Specializing on different seed sizes may result from competition between coexisting heteromyids (Dayan and Simberloff 1994b). For tooth-row lengths and surfaces, Dayan and Simberloff (1994b) found no regular pattern for heteromyids and suggested that, since grinding takes place in the safety of the burrow where there is no time constraint, selection pressure on this trait should not be strong. On the other hand, Yom-Tov (1991), who studied morphological relationships among the gerbillines of Israel, measured two characters, skull length and upper tooth-row length, and demonstrated overdispersed means for the latter trait. He suggested that the longer the tooth row, the more grinding area is available for each jaw movement; species with longer tooth rows may feed on harder and larger seeds and on a higher proportion of leaves and stems, which are often hard and include silica (Yom-Tov 1991). How do we reconcile these differing patterns and hypotheses for tooth morphology between the gerbillines and heteromyids? Do they indicate different guild-level selective forces in the two groups? The apparent convergence noted by Abramsky (1983) may simply reflect a striking overall visual resemblance, while ecological detail as reflected by teeth defies this interpretation (see pp. 23–25 in Begon et al. 1990). We carried out a comparative morphological study of Israeli gerbillines plus cafeteria studies of seed-size selection, with the following specific goals: to test for convergent morphological patterns between Israeli and North American seed-eating rodents and to determine whether seed-size selection facilitates coexistence among Israeli gerbillids, as it appears to do for North American heteromyids. We sought insight into the possible relationship between convergence in morphological patterns and convergence in forces organizing communities in Old and New World granivorous desert rodents. Material and Methods Seed-Eating Rodent Guild Yom-Tov (1991) studied the gerbilline rodents of Israel, which he viewed as a single group, and we follow his guild assignation. The jerboa (Jaculus jaculus) also constitutes part of the same rodent community (Brown et al. 1994), but neither Yom-Tov (1991) nor we include it in the gerbilline guild, owing to its different limb morphology (hence, locomotor behavior; reviewed by Dayan and Simberloff 1998) and to its herbivorous diet (Osborn and Helmy 1980; Bar et al. 1984). Yom-Tov (1991) included the fat sand rat (Psammomys obesus) in the gerbilline rodent guild, and we do so tentatively. The fat sand rat subsists mainly on Atriplex halimus (Daly and Daly 1973), but no research has focused on its precise diet. Seeds have been found in fat sand rat burrows (Anderson and de Winton 1902 cited in Harrison and Bates 1991; M. Kam, personal observation), and in captivity, fat sand rats can subsist on a mixed diet of seeds and A. halimus (M. Avishai, personal observation). However, because the evidence is far from conclusive, we analyzed the rodent guild both including and excluding this species. 486 The American Naturalist We follow Yom-Tov (1991) in choice of geographical regions and species studied: sand dunes of the southern coastal areas (with four psammophile species: Meriones sacramenti, Meriones tristrami, Gerbillus pyramidum, and Gerbillus allenbyi), sand dunes of the northern Negev Desert (with seven psammophile species: P. obesus, M. sacramenti, Meriones crassus, G. pyramidum, G. allenbyi, Gerbillus gerbillus, and Gerbillus henleyi), and the sand dunes of the Arava Valley (with four psammophile species: P. obesus, M. crassus, G. gerbillus, and G. henleyi). Patterns of local coexistence of these species have not been studied, and our study is concerned with patterns where overlap is at a geographical scale (see Dayan and Simberloff 1994b). Morphological Study We measured rodent skulls in the Tel Aviv University Zoological Museum. We took measurements with digital calipers (Mahr) to 0.01 mm precision for four characters: condylo-basal length of the skull (CBL); upper incisor width (IW), which estimates the width of the cutting edge (Dayan and Simberloff 1994b); upper tooth-row length (UTL), the length of the grinding surface; and width of the upper tooth row—multiplying UTL by this measure provides an estimate of molar tooth-row surface, or grinding surface (UTRS). These measures were also taken in our heteromyid rodent study (Dayan and Simberloff 1994b), with the working hypothesis that seed-size selection would be best reflected by incisor width, whereas the tooth-row length and, even more so, surface would reflect the volume of food ingested, associated with body size and metabolic needs (Gould 1975), as well as the amount of plant material ingested (Yom-Tov 1991). We measured specimens of the relevant species in the Tel Aviv University Zoological Museum. Tooth wear was noted for every specimen measured, and the degree of tooth wear, measured by an index developed by Rahmut Ben-Moshe (1996), was correlated with all four measures. We found increasing growth during tooth-wear stages 1 and 2 that ceased at tooth-wear stage 3; therefore, in this study, we analyzed only specimens from stage 3 growth wear and older. Sample sizes (stage 3 and older) were 26 G. allenbyi, 13 G. pyramidum, six G. gerbillus, two G. henleyi, 46 M. tristrami, 22 M. sacramenti, 20 M. crassus, and 27 P. obesus. Gerbils exhibit very low sexual-size dimorphism, with a maximum of 4% difference in size between the sexes (Rahmut Ben-Moshe 1996); therefore, we used mixed-sex populations in our current study, as did Yom-Tov (1991) for the same species and Dayan and Simberloff (1994b) for heteromyid rodents. To assess whether size ratios were more equal than ex- pected by chance within a ranked sequence, we used Barton-David (B-D) statistics (Barton and David 1956; Simberloff and Boecklen 1981). For a line uniform-randomly broken into segments, Barton and David calculated the cumulative frequency distributions for the ratios of the ith smallest segment to the jth smallest segment for all (i,j) such that i ! j. If morphological sizes are based on a logarithmic line, lengths of segments correspond to size ratios, and so the B-D statistics test whether pairs of size ratios are more similar than would be expected if the species’ sizes were independently distributed according to a loguniform distribution. Boecklen and NeSmith (1985) showed that points distributed log-normally give very similar B-D statistics. Throughout, we use three B-D statistics: G1, n, the ratio of the smallest to the largest size ratio; G1, n⫺1, the ratio of the smallest to the second largest size ratio; and G 2, n, the ratio of the second smallest to the largest size ratio. For detailed discussion of statistical tests of size ratios, see Dayan and Simberloff (1998). To test for morphological differences between mean measurements in different geographic regions, we carried out paired t-tests for each species because no species occurred in more than two regions. Because slight morphological differences may exist even where differences do not appear significant, we carried out our B-D analyses on both pooled assemblages (using means for specimens from all geographic regions) and regional ones (means for specimens only from specific region included). After we obtained results of the B-D tests, we sought the probabilities of achieving these significance values in three different regions by chance. Because the sets of species in the different regions overlap, there is no apparent way to calculate a formal total probability. As noted above, patterns of local coexistence for these rodents are unknown; therefore, we cannot emulate the methods of Bowers and Brown (1982). Following a technique of Losos et al. (1989), we used Fisher’s test of combined probabilities (Sokal and Rohlf 1981) on the G1, n statistics (the most extreme comparison of size ratios), combining probabilities from all three regions. However, if the same size ratio was used more than once in a test, the estimate of combined probability would be too low; that is, the test would not be conservative and significance could be inflated. Thus, we checked to see if ratios between the same two species were being included more than once in each calculation. Seed-Size Selection Studies of seed-size selection were carried out using only two of the gerbilline species: G. allenbyi (24 g) and G. pyramidum (39 g), which have been found to compete (Abramsky et al. 1985, 1990, 1991). A problem in labo- Old and New World Rodent Guilds 487 Table 1: Lengths (mm) of seeds used in cafeteria experiments Species Mean SD Helianthus annuus Avena sativa Avena strigosa Phalaris canariensis Setaria italica 24.04 11.12 10.33 4.99 1.61 2.87 1.12 .93 .31 .05 to using these seeds in our experiment was that the seeds were not husked; seeds found in nature are unhusked. Therefore, our assumption was that gerbils confronted with different sizes of palatable seeds would make similar choices to those made under natural conditions. Note: Means and standard deviations are for 10 individuals of each species. ratory experiments is that the rodents may not perceive the same risks as do free-ranging gerbils, and this perception affects their natural foraging patterns (Abramsky et al. 1991; Ziv 1991). Therefore, we carried out our experiments in a large, 5-m-high aviary (18 # 23-m enclosure divided into two equal parts by a 1-m-high fence) at the Blaustein Institute for Desert Research. We used artificial lighting to simulate full-moon nights, which is when predation risk for these rodents is perceived as high (Kotler et al. 1991). The seeds provided (2 Ⳳ 0.02 g of each seed type [weighed by digital scales to 0.01 g precision]) were mixed with 5 L of sieved sand in trays measuring 45 # 60 # 2.5 cm in order to simulate natural foraging conditions and costs. We carried out two experiments with two types of seed. Commercial Seeds. We used seeds of sunflower (Helianthus annuus), cultivated oats (Avena sativa), wild oats (Avena strigosa), canary grass (Phalaris canariensis), and foxtail millet (Setaria italica). Sizes of these seeds are listed in table 1. These specific seeds were chosen because of their different sizes, following a preliminary cafeteria experiment to determine that the seeds were all palatable to gerbils, and taken in significant quantities. The advantage Different-Sized Particles of Husked Wheat. The wheat (cf. Abramsky 1983; Price 1983; Price and Brown 1983) was ground and then electrically sieved to three size groups: 14,000 m; 2,000–4,000 m; and 1,000–2,000 m. This range of seed sizes is smaller than that of the commercial seeds described above, and most of these particles are smaller than most of the commercial seeds. The advantage of using husked wheat is that the particles differ only in size. A disadvantage is that we had essentially a continuum of seed particle sizes rather than discrete size groups, as when using commercial seeds of different sizes. Since these seed particles required no husking and can be eaten immediately, our working hypothesis was that seed-size selection would be of less importance in this case because the seeds require no handling. Seed trays were placed in the aviary at sunset and were collected at sunrise. The seeds were sifted and weighed by seed type. During an acclimatization period of at least a week (to become familiar with the aviaries and trays), and between experiments, the gerbils were fed with large, commercial mouse pellets. Before each experiment, we accustomed the gerbils to foraging in seed trays by feeding them with mouse pellets mixed in the sand. Because seeds may absorb humidity, we also placed a control tray outside, to which gerbils had no access. In the morning, we weighed the seeds in the control trays and performed the following calculation: we divided the original mass of the seeds by the mass of the seeds in the morning and multiplied the seed weight that remained in Table 2: Sample means and standard deviations (in parentheses) for the different measurements of the different species Species Gerbillus allenbyi Gerbillus gerbillus Gerbillus henleyi Gerbillus pyramidum Meriones crassus Meriones sacramenti Meriones tristrami Psammomys obesus UTRS CBL 5.57 (.34) 5.33 (.08) 2.73 (.04) 7.7 (.55) 10.27 (.92) 13.58 (.71) 8.99 (.92) 15.01 (1.68) 25.31 (.81) 25.08a (.19)a, 24.5b (.03)b 19.83 (.11) 28.87 (.83) 35.61 (1.64) 39.25 (1.90) 32.86 (2.29) 39.23a (4.20)a, 42.51b (1.88)b IW .81 .68 .49 .91 1.00 1.38 1.02 1.48 (.05) (.04) (.00) (.05) (.04) (.15) (.09) (.15) UTL 3.94 3.89 2.89 4.61 5.57 6.48 5.40 7.05 (.12) (.08) (.02) (.17) (.26) (.24) (.36) (.45) Note: The data are for specimens pooled from all regions. In the few cases where significant differences (by paired t-tests; P ! .05) were found between specimens in different regions, we present the sample statistics for the specific region. UTRS p molar tooth-row surface; CBL p condylo-basal length of the skull; IW p upper incisor width; UTL p upper tooth-row length. a Negev region. b Arava region. 488 The American Naturalist Table 3: Barton-David test results for the different characters measured in three different regions Trait Negev: UTL IW CBL UTRS Arava: UTL IW CBL UTRS Coastal plain: UTL IW CBL UTRS G1, 6 G1, 3 P G1, 5 G1, 2 P G2, 6 G2, 3 P .0430 .2135 .0022 .0658 … … … … .566 .073 .971 .424 .0675 .2172 .0024 .1360 … … … … .596 .186 .981 .352 .2837 .2878 .0389 .1496 … … … … .171 .165 .912 .473 … … … … .6564 .8359 .4736 .5672 .038 .007 .115 .068 … … … … .7930 .8497 .8375 .5786 .148 .106 .115 .327 … … … … .8278 .9837 .5655 .9803 .153 .014 .425 .017 … … … … .8614 .3775 .7285 .3753 .005 .186 .022 .188 … … … … .9929 .9803 .9837 .4784 .005 .013 .011 .421 … … … … .8675 .3851 .7406 .7845 .116 .637 .238 .195 Note: The tests were calculated using means for specimens pooled from all three regions and including all species. Significant or marginally significant (up to P p .075 ) results are in bold. For abbreviations, see table 2. the experimental trays by the resultant number, thereby accounting for humidity-induced changes in seed mass. We carried out two replicate experiments per species. We trapped individuals of both species in the field and placed five individuals of G. allenbyi in one part of the aviary and five individuals of G. pyramidum in the other part. Later, the gerbils were removed and new groups were placed for the second experiment. No individual was used in more than one replicate. The two gerbil species differ in body mass, and individuals of the larger species are expected to consume greater quantities of seed. Therefore, we calculated the proportion of each seed type taken per night as part of the total mass of seed taken that night and arcsine transformed the results. We carried out two-way ANOVAs (using STATVIEW; SAS 2000) to test for differences between the species in seed choice and differences between replicate experiments. Results Paired t-tests revealed that, except for two cases, CBLs of Gerbillus gerbillus and Psammomys obesus (Negev Desert and Arava Valley), there were no significant morphological Table 4: Barton-David test results for the different characters measured in three different regions Trait Negev: UTL IW CBL UTRS Arava: UTL IW CBL UTRS Coastal plain: UTL IW CBL UTRS G1, 6 G1, 3 P G1, 5 G1, 2 P G2, 6 G2, 3 P .07255 .2219 .06122 .04392 … … … … .390 .066 .449 .559 .10641 .26957 .06947 .07341 … … … … .442 .122 .587 .569 .27785 .23378 .15666 .14894 … … … … .179 .254 .451 .476 … … … … .85731 .61851 .55750 .71320 .005 .050 .072 .025 … … … … .97937 .77348 .93274 .76902 .014 .163 .046 .167 … … … … .87536 .79964 .59771 .92741 .109 .180 .389 .062 … … … … .79294 .29333 .66753 .38605 .012 .275 .035 .178 … … … … .98729 .64471 .87986 .57764 .0085 .269 .083 .328 … … … … .80314 .45498 .75867 .66833 .176 .553 .220 .313 Note: The tests were calculated using means for specimens from each region separately, and including all species. Significant or marginally significant results are in bold. For abbreviations, see table 2. Old and New World Rodent Guilds 489 Table 5: Barton-David test results for the different characters Trait G1, 5 G1, 2 P G1, 4 P G2, 5 P .4298 .28781 .3887 .6583 … … … … .647 .065 .674 .517 .6752 .29282 .4351 .13601 .693 .192 .790 .475 .50925 .35525 .41437 .41756 .067 .194 .132 .129 … … … … .82776 .84967 .56552 .98031 .094 .081 .278 .010 … … … … … … … … … … … … … … … … .07255 .23378 .15666 .04397 … … … … .485 .108 .219 .641 .10641 .28490 .17779 .7341 .560 .201 .376 .671 .48466 .28863 .32358 .39611 .08 .292 .237 .149 … … … … .85731 .77348 .59771 .92741 .077 .128 .252 .038 … … … … … … … … … … … … … … … … a A. Specimens pooled: Negev: UTL IW CBL UTRS Arava: UTL IW CBL UTRS B. Specimens measured in two regions:b Negev: UTL IW CBL UTRS Arava: UTL IW CBL UTRS Note: Significant or marginally significant results are in bold. For abbreviations, see table 2. a Excluding Psammomys obesus. b Measured where Psammomys obesus occurs but excluding this species. differences between conspecific populations of the different regions. Sample statistics for the measurements taken for the different gerbil species are given in table 2. Results of the B-D tests for pooled data are given in table 3 and for the unpooled data in table 4. A comparison of tables 3 and 4 shows several significant results, and they tend to be for the same statistics for the same traits in the same locations. Thus, for example, for IW, size ratios are significantly similar, as evidenced by G1, n in the Negev and Arava, though not the Coastal Plain, whether or not the data are pooled. Second, for all four traits—UTL, IW, CBL, and UTRS—there are some significant results. For example, again focusing on G1, n, we see both mean IW and mean UTL significantly overdispersed (or very close to it) in two sites, while CBL and UTRS show significance in one site each. Results excluding P. obesus are given in table 5 (pooled and unpooled data, respectively). It is apparent if one compares tables 3 and 5A or tables 4 and 5B that removal of the fat sand rat changes the results moderately. There are still significant results but fewer of them. Thus, for example, for pooled data, a highly significant result for the Arava has much greater null probability with this species excluded. The combined probability test of Fisher rarely violated the independence assumption. For 14 of the 16 tests, no size ratio (either the smallest or the largest) consisted of the same species pair for more than one site. For the other two tests, one ratio (either smallest or largest) consisted of the same species pair for two of the three sites. These tests show significance for most traits most of the time whether or not the data are pooled and whether or not the fat sand rat is included (table 6). Results of the two-way ANOVA with trays as repeated measures show a significant difference in seed preference between the two species (F p 21.36, df p 4, 110, P ! .0001) in the experiment using commercial seeds of different sizes. The interaction between species and seeds was significant (F p 20.61, df p 4, 110, P ! .0001). Gerbillus pyramidum, the larger species, preferred larger seeds, whereas Gerbillus allenbyi preferred smaller ones (fig. 1). We performed two-way ANOVAs with nights as repeated measures to test for differences between our replicate experiments and for differences in seed preference for each seed type separately between the two gerbil species. Our results (table 7) show drastic differences between rodent species in seed-type choice, a clear difference between replicate experiments for one seed type (Setaria italica), and no significant interactions. 490 The American Naturalist Table 6: Fisher’s combined probability statistic (associated with x2 test with df p 6) for three different regions (see text) Fisher’s statistic Nominal probability UTL IW CBL UTRS 18.316 18.522 12.018 10.435 .005 ! P ! .01 .005 ! P ! .01 .05 ! P ! .1 .1 ! P ! .25a UTL IW CBL UTRS 21.326 14.010 13.568 11.993 .001 ! P ! .005 .025 ! P ! .05 .025 ! P ! .05 .05 ! P ! .1 UTL IW CBL UTRS 21.193 13.857 10.983 13.893 .001 ! P ! .005 .025 ! P ! .05 .05 ! P ! .1 .025 ! P ! .05a UTL IW CBL UTRS 15.421 11.145 12.499 10.898 .01 .05 .05 .05 Trait A: B: C: D: ! ! ! ! P P P P ! ! ! ! .025 .1 .1 .1 Note: A, Pooled data, including Psammomys obesus. B, Separate data, including P. obesus. C, Pooled data, excluding P. obesus. D, Separate data, excluding P. obesus. For abbreviations, see table 2. a Two of three regions had same pair of species for one ratio. from very distantly related families, occur on different continents. Upper tooth-row lengths, skull lengths, and tooth-row surfaces of gerbils in Israel display patterns similar to those for incisor widths. The overall similarity among the patterns depicted by these characters may result from natural selection acting on all these traits or from a passive, correlated response to selection directed at incisor size or other traits (see Dayan et al. 1989b). It may also reflect foraging behavior. Heteromyid rodents have large, external cheek pouches in which they collect seeds aboveground (Reichman 1981); mean pouch volumes of heteromyids (Morton et al. 1980) also show regular spacing (Dayan and Simberloff 1994b), much as incisor widths do, while other traits, including dental ones, show no such patterns. Gerbillids do not possess cheek pouches, and they seem to stuff as many seeds as they can into their mouths and make repeated trips to their burrows or caches (Abramsky 1983). Mandible length, CBL, and tooth-row length may all provide a rough estimate of the volume available for seed transport, much as pouch volume does in heteromyids. This possibility may explain the relatively regular pattern found in these characters and the overdispersed mean upper tooth-row lengths found by Yom-Tov (1991). Nonrandom morphological patterns such as these may arise from species sorting and/or from coevolutionary change (see Brown and Wilson 1956; Case et al. 1983; Sinclair et al. 1985a, 1985b; Roughgarden 1989; Dayan and Our qualitative observations revealed a great difference between G. pyramidum and G. allenbyi husking behaviors; large quantities of seed husks were found in the morning in G. pyramidum seed trays, whereas few seed husks remained in G. allenbyi seed trays. We performed two-way ANOVAs with nights as repeated measures for the wheat particle–size groups and found at most a marginally significant difference between seeds (F p 2.56, df p 2, 51, P p .088), which suggests little or no difference in particle-size choice between the species. In our samples, the larger species, G. pyramidum, preferred slightly larger particles than G. allenbyi. No significant difference was found between replicate experiments (F p 1.10, df p 2, 30, P p .345). Discussion We found a nonrandom pattern of overdispersed means for upper incisor widths in three Israeli gerbilline guilds. This pattern conforms to that found for several heteromyid rodent guilds of North America. Thus, convergent sizestructured granivorous rodent guilds, composed of species Figure 1: Seed choice in cafeteria experiment in which Gerbillus pyramidum (white) and Gerbillus allenbyi (black) ate different species of commercial seeds. Seed size declines from left to right. Old and New World Rodent Guilds 491 Table 7: Different seed preferences between Gerbillus allenbyi (a) and Gerbillus pyramidum (p) in cafeteria experiments Difference between replicates Seed type Helianthus annuus Avena sativa Avena strigosa Phalaris canariensis Setaria italica F P Species favoring this type 3.22 .05 2.15 .05 14.83 .088 .825 .158 .827 .001 p p a a a Species difference Interaction F P F P 28.33 11.13 18.41 10.80 11.18 !.0001 .06 1.20 .85 .08 1.86 .804 .286 .367 .776 .188 .003 .0004 .004 .003 Note: All F ratios have df p 1, 20. Simberloff 1998). For heteromyids, there is geographic variation in size, and using this variation at the broad regional scale, we have some indication that the mechanism in play is coevolution (Dayan and Simberloff 1994b). At the scale of local coexistence, Bowers and Brown (1982) provide evidence for species sorting by body size in heteromyids. There is little geographic variation in size of the gerbillids in the south of Israel. Therefore, we were unable to use such variation to understand the relative roles of species sorting and coevolution in producing the regular pattern for these gerbillid guilds at the regional scale. Neither are there data to enable a study of local coexistence analogous to that of Bowers and Brown (1982). Whatever the underlying mechanism, the morphological patterns found for North American granivorous rodent guilds and for Israeli gerbillids attest to ecomorphological convergence at the guild level. Both guilds are size structured for incisor widths, and this morphological pattern reflects guild organization. Moreover, in recent years, other studies of both fossil and recent rodent guilds have demonstrated similar size structuring on the basis of incisor size (Michaux et al. 1999; Parra et al. 1999). Incisors of heteromyids are used for husking, some of which occurs aboveground where predation risk is high (Rosenzweig and Sterner 1970; Lemen 1978). Therefore, husking speed may be critical (Rosenzweig and Sterner 1970). Different species exhibit different husking behaviors related to their body sizes (Lemen 1978), resulting in apparent seed-size selection that is related to body size (Brown and Lieberman 1973; see also Dayan and Simberloff 1994b). It has also been demonstrated that very small and very large particles are rejected by heteromyid rodents (Lawhon and Hafner 1981). Lemen (1978) suggested that interspecific differences in husking behavior may be explained by an interaction between cheek pouch size, husking time (which decreases per seed with increasing rodent size; Rosenzweig and Sterner 1970), and exposure to predation. Abramsky (1983) compared seed-size preference of some Israeli gerbilline rodents (Gerbillus pyramidum, Gerbillus allenbyi, Gerbillus henleyi, and Meriones crassus) with that of ants and concluded that both taxa show a preference for large seeds, but he did not study seed-size preferences with a variety of seed sizes among the rodents. Bar et al. (1984) studied the food habits of gerbilline rodents, but this study was limited to gross dietary categories. Thus, seed-size preference of the different species has not been directly tested in Israel. Seed-size selection has been studied for North American heteromyid rodents by cheek pouch contents (Brown and Lieberman 1973; M’Closkey 1978, 1980) or by seed-choice preference (cafeteria) experiments in laboratory conditions. In these experiments, researchers used either commercial seeds (Mares and Williams 1977) or hulled, hardwinter red wheat ground to different-sized particles (Price 1983). Because gerbillids do not possess cheek pouches, we could not examine pouch contents, so we carried out cafeteria experiments with commercial seeds and with hulled wheat particles. We hypothesized that, if husking behavior is key, then seed-size selection would be a factor in the commercial seed experiment but not in the hulled wheat experiment. Results of our experiments with commercial seeds of different sizes (tables 1, 7) offered in sand trays to G. pyramidum and G. allenbyi support a hypothesis of seedsize selection in nature: the larger G. pyramidum took significantly larger seeds than the smaller G. allenbyi. Although the distribution of seed sizes in these sites is unknown, it is likely that most available seeds are smaller than the commercial seeds we offered. However, the experimental animals readily ate all offered seeds. Research on heteromyids, summarized by Price and Reichman (1987), suggests that there are many more small than large seeds in nature but that several species of rodents select the large seeds, perhaps to the point of changing the seedsize distribution in the field. Thus, it is quite possible that partitioning of large seeds by size in an experiment reflects an important process occurring in nature. 492 The American Naturalist Figure 2: Choice of wheat particle size by Gerbillus pyramidum (white) and Gerbillus allenbyi (black) in a cafeteria experiment. Seed size declines from left to right. Mares and Williams (1977) believed that using commercial seeds would be preferable to field-collected seeds because all of the seeds offered were probably unfamiliar to the rodents; therefore, commercial seeds should minimize the confounding effects of familiarity with wild seeds that could have caused animals to favor one or another type of seed because of flavor, consistency, toxicity, or unknown factors. Further, commercial seeds are generally selected for little or no toxicity; therefore, this factor would not interfere with selection for size. The disadvantage of using commercial seeds is that the seeds differ in taste and nutritional attributes as well as size (Price 1983). We used seeds that were totally unfamiliar to the rodents prior to the experiments and that, as previous experiments indicated (Rahmut Ben-Moshe 1996), were palatable to both species. Since these closely related gerbil species are very similar in their biology, their seed choices are expected to be similar, and seed size should be the major factor. No significant difference in seed-size choice was found in the hulled-wheat experiment, in which the food particles required no treatment. Abramsky et al. (1991) concluded by direct observations that G. allenbyi exhibits different foraging behaviors under laboratory and field conditions. In the laboratory, this gerbil ate the seeds on the seed trays and did not collect and cache them, whereas in the field, it collected seeds in its mouth without eating them to transport them frequently to burrows. In our experimental setup, under simulated natural conditions in which perceived risk of predation is considerable, gerbils did husk some of the seeds in the seed trays. This was particularly true for the larger G. pyramidum, which were also taking larger seeds. Thus, the rationale for selection on incisor width of gerbils parallels our rationale for heteromyid rodents (Dayan and Simberloff 1994b). Dayan and Simberloff (1994b) suggested that, if husking large seeds by small species necessitates staying exposed for a relatively long time at a particular place, the energetic advantage of taking this large seed may be outweighed by excessive risk of predation. In other words, there may be a trade-off between husking large seeds (which takes time aboveground), so as to carry more of them per foraging trip, and not husking them, thereby requiring more trips. Aboveground husking speed may be critical to these rodents. This hypothesis implies there may be considerable overlap in the intake of smaller seeds between the different species. In fact, Brown and Lieberman (1973) and Mares and Williams (1977) emphasized for heteromyids that it is the increased tendency of large species to select large seeds (in addition to many small seeds) that gives rise to their high average seed sizes. They suggested that this tendency gives the larger species a competitive advantage. In our experiment, combining data from table 1 and figure 1, we found the larger G. pyramidum to take larger seeds on average than G. allenbyi (length: 11.87 vs. 8.10 mm); similarly, using data from figure 2, we found G. pyramidum to take slightly larger ground seed particles than G. allenbyi (2,470 vs. 2,408 m). Larger species are frequently found to have a larger food-range size among particulate feeders (Schoener 1969; Wilson 1975). In our experiment with commercial seeds of five species, G. pyramidum took similar amounts of large and small seeds, while the smaller G. allenbyi took a much smaller amount of large than of small seeds (fig. 1). In the ground wheat experiment, with much smaller particles overall, the two species took similar proportions of the different-sized particles, although the larger G. pyramidum took slightly larger particles on average (fig. 2). In heteromyid rodents, the hypothesis of seed-size selection, at first supported by cheek pouch content sieving and feeding experiments (Brown and Lieberman 1973; Mares and Williams 1977), was later challenged by other studies (e.g., Smigel and Rosenzweig 1974; Lemen 1978; M’Closkey 1978, 1980; Price 1983; but see Dayan and Simberloff 1994b). Increasingly, other mechanisms facilitating coexistence, such as differing microhabitat preferences, seed dispersion preferences, trade-offs in the foraging efficiencies of different granivorous species, and interference competition (reviewed by Brown [1987], Kot- Old and New World Rodent Guilds 493 ler and Brown [1988], and Brown et al. [1994]), have been emphasized. Microhabitat selection has also been considered to be key to coexistence among Israeli gerbils (Rosenzweig et al. 1984; Abramsky et al. 1985, 1990, 1991). Another mechanism recently emphasized for gerbils is temporal partitioning (Kotler et al. 1993; Ziv et al. 1993), emphasizing that coexistence may be facilitated by more than one axis of heterogeneity. While the occurrence of interspecific competition has been demonstrated unequivocally in both heteromyids and gerbillids, it does not necessarily follow that competition will affect community structure (cf. Stone et al. 1996) or that it will do so in the same way in both rodent communities. Historical factors may strongly influence community structure, so that, although common processes of community assembly may operate, they may do so to very different degrees (Kelt et al. 1996). Our results suggest that seed-size selection is a further mechanism facilitating coexistence among Israeli gerbillines and North American heteromyids. We found size-structured seed-eating rodent guilds among gerbillines, as among heteromyids, attesting to convergence in mechanisms structuring community morphology between Old and New World granivorous desert rodents. Acknowledgments We thank Z. Abramsky, B. Kotler, J. Losos, and Y. YomTov for constructive suggestions on earlier drafts of this manuscript. A. Landsman provided technical assistance, while B. Kotler allowed us to use his aviary and L. OlsvigWhittaker let us use her sieving machine. Literature Cited Abramsky, Z. 1983. Experiments on seed predation by rodents and ants in the Israeli desert. Oecologia (Berlin) 57:328–332. Abramsky, Z., M. L. Rosenzweig, and S. Brand. 1985. Habitat selection in Israeli desert rodents: comparison of a traditional and a new method of analysis. Oikos 56: 45–53. Abramsky, Z., M. L. Rosenzweig, B. Pinshow, J. S. Brown, B. P. Kotler, and W. A. Mitchell. 1990. Habitat selection: an experimental field test with two gerbil species. Ecology 71:2358–2369. Abramsky, Z., M. L. Rosenzweig, and B. Pinshow. 1991. The shape of a gerbil isocline measured using principles of optimal habitat selection. Ecology 72:329–340. Bar, Y., Z. Abramsky, and Y. Gutterman. 1984. Diet of gerbilline rodents in the Israeli desert. Journal of Arid Environments 7:371–376. Barton, D. E., and F. N. David. 1956. Some notes on or- dered random intervals. Journal of the Royal Statistical Society B 18:79–94. Begon, M., J. L. Harper, and C. R. Townsend. 1990. Individuals, populations and communities. 2d ed. Blackwell Scientific, Boston. Boecklen, W. J., and C. NeSmith. 1985. Hutchinson’s ratios and log-normal distributions. Evolution 39:695–698. Bowers, M. A., and J. H. Brown. 1982. Body size and coexistence in desert rodents: chance or community structure? Ecology 63:361–367. Brown, J. H. 1987. Variation in desert rodent guilds: patterns, processes, and scales. Pages 185–203 in J. H. R. Gee and P. S. Giller, eds. Organization of communities: past and present. Blackwell Scientific, Oxford. Brown, J. H., and G. A. Lieberman. 1973. Resource utilization and coexistence of seed-eating desert rodents in sand-dune habitat. Ecology 54:788–797. Brown, J. H., O. J. Reichman, and D. W. Davidson. 1979. Granivory in desert ecosystems. Annual Review of Ecology and Systematics 10:201–227. Brown, J. S., B. P. Kotler, and W. A. Mitchell. 1994. Foraging theory, patch use, and the structure of a Negev desert granivore community. Ecology 75:2286–2300. Brown, W. L., and E. O. Wilson 1956. Character displacement. Systematic Zoology 5:49–64. Case, T. J., and R. Sidell. 1983. Pattern and chance in the structure of model and natural communities. Evolution 37:832–849. Case, T. J., J. Faaborg, and R. Sidell. 1983. The role of body size in the assembly of West Indian bird communities. Evolution 37:1062–1074. Daly, M., and S. Daly. 1973. On the feeding ecology of Psammomys obesus (Rodentia, Gerbillidae) in the Wadi Saoura, Algeria. Mammalia 37:545–561. Dayan, T., and D. Simberloff. 1994a. Character displacement, sexual size dimorphism, and morphological variation among the mustelids of the British Isles. Ecology 75:1063–1073. ———. 1994b. Morphological relationships among coexisting heteromyids: an incisive dental character. American Naturalist 143:462–477. ———. 1998. Size patterns among competitors: ecological character displacement and character release in mammals, with special reference to island populations. Mammal Review 28:99–124. Dayan, T., E. Tchernov, Y. Yom-Tov, and D. Simberloff. 1989a. Ecological character displacement in SaharoArabian Vulpes: outfoxing Bergmann’s rule. Oikos 55: 263–272. Dayan, T., D. Simberloff, E. Tchernov, and Y. Yom-Tov. 1989b. Inter- and intraspecific character displacement in mustelids. Ecology 70:1526–1539. ———. 1990. Feline canines: community-wide character 494 The American Naturalist displacement in the small cats of Israel. American Naturalist 136:39–60. ———. 1992. Canine carnassials: character displacement among the wolves, jackals, and foxes of Israel. Biological Journal of the Linnean Society 45:315–331. Gould, S. J. 1975. On the scaling of tooth size in mammals. American Zoologist 15:315–362. Harrison, D. L., and P. J. Bates. 1991. The mammals of Arabia. 2d ed. Harrison Museum, Kent. Holmes, R. T., and F. A. Pitelka. 1968. Food overlap among coexisting sandpipers on northern Alaska tundra. Systematic Zoology 17:305–318. Kelt, D. A., J. H. Brown, E. J. Heske, P. A. Marquet, S. R. Morton, J. R. W. Reid, K. A. Rogovin, and G. Shenbrot. 1996. Community structure of desert small mammals: comparisons across four continents. Ecology 77: 746–761. Kotler, B. P., and J. S. Brown. 1988. Environmental heterogeneity and the coexistence of desert rodents. Annual Review of Ecology and Systematics 19:281–307. Kotler, B. P., J. S. Brown, and O. Hasson. 1991. Owl predation on gerbils: the role of body size, illumination, and habitat structure on rates of predation. Ecology 72: 2249–2260. Kotler, B. P., J. S. Brown, and A. Subach. 1993. Mechanisms of coexistence of optimal foragers: temporal partitioning by two species of sand dune gerbils. Oikos 67:548–556. Lawhon, D. K., and M. S. Hafner. 1981. Tactile discriminatory ability and foraging strategies in kangaroo rats and pocket mice (Rodentia: Heteromyidae). Oecologia (Berlin) 50:303–309. Lawton, J. H. 1999. Are there general laws in ecology? Oikos 84:177–192. Lemen, C. A. 1978. Seed size selection in heteromyids: a second look. Oecologia (Berlin) 35:13–19. Losos, J. B., S. Naeem, and R. K. Colwell. 1989. Hutchinsonian ratios and statistical power. Evolution 43: 1820–1826. Mares, M. A., and D. F. Williams. 1977. Experimental support for food particle size resource allocation in heteromyid rodents. Ecology 58:1186–1190. M’Closky, R. T. 1978. Niche separation and assembly in four species of Sonoran desert rodents. American Naturalist 112:683–694. ———. 1980. Spatial pattern in sizes of seeds collected by four species of heteromyid rodents. Ecology 61: 486–489. Michaux, R. S., J. Mein, J. P. Aguilar, and J. C. Auffray. 1999. Patterns of size and shape differentiation during the evolutionary radiation of the European Miocene murine rodents. Lethaia 32:61–71. Morton, S. R., D. S. Hinds, and R. E. MacMillen. 1980. Cheek pouch capacity in heteromyid rodents. Oecologia (Berlin) 46:143–146. Osborn, D. J., and I. Helmy. 1980. The contemporary land mammals of Egypt including Sinai. Fieldiana Zoology 5:1–579. Parra, V., M. Loreau, and J. J. Jaeger. 1999. Incisor size and community structure in rodents: two tests of the role of competition. Acta Oecologica 20:93–101. Price, M. V. 1983. Laboratory studies of seed size and seed species selection by heteromyid rodents. Oecologia (Berlin) 60:259–263. Price, M. V., and J. H. Brown. 1983. Patterns of morphology and resource use in North American desert rodent communities. Great Basin Naturalist Memoirs 7:117–134. Price, M. V., and O. J. Reichman. 1987. Distribution of seeds in Sonoran Desert soils: implications for heteromyid rodent foraging. Ecology 68:1797–1811. Rahmut Ben-Moshe, A. 1996. Seed size selection and morphological relationships among the Gerbillidae of Israel. M.Sc. thesis. Tel Aviv University. [In Hebrew; English abstract.] Reichman, O. J. 1981. Factors influencing foraging in desert rodents. Pages 195–213 in A. C. Kamil and T. D. Sargent, eds. Foraging behavior: ecological, ethological, and psychological approaches. New York: Garland, STPM. Root, R. B. 1967. The niche exploitation pattern of the blue-gray gnatcatcher. Ecological Monographs 37: 317–350. Rosenzweig, M. L., and P. W. Sterner. 1970. Population ecology of desert rodent communities: body size and seed-husking as bases for heteromyid coexistence. Ecology 51:217–224. Rosenzweig, M. L., Z. Abramsky, and S. Brand. 1984. Estimating species interactions in heterogenous environments. Oikos 43:329–340. Roughgarden, J. 1989. The structure and assembly of communities. Pages 203–226 in J. Roughgarden, R. M. May, and S. A. Levin, eds. Perspectives in ecological theory. Princeton University Press, Princeton, N.J. SAS Institute. 2000. STATVIEW. SAS Institute, Cary, N.C. Schoener, T. W. 1969. Models of optimal size for solitary predators. American Naturalist 103:277–313. Simberloff, D., and W. Boecklen. 1981. Santa Rosalia reconsidered: size ratios and competition. Evolution 35: 1206–1228. Simberloff, D., and T. Dayan. 1991. Guilds and the structure of ecological communities. Annual Review of Ecology and Systematics 22:115–143. Sinclair, D. F., J. E. Mosimann, and D. A. Meeter. 1985a. Tests for character displacement. Biometrics 41: 835–846. Old and New World Rodent Guilds 495 ———. 1985b. Tests for deletion. Biometrics 41:847–857. Smigel, B. W., and M. L. Rosenzweig. 1974. Seed selection in Dipodomys merriami and Perognathus penicillatus. Ecology 55:329–339. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. 2d ed. W. H. Freeman, San Francisco. Stone, L., T. Dayan, and D. Simberloff. 1996. Communitywide assembly patterns unmasked: the importance of species’ differing geographic ranges. American Naturalist 148:997–1015. Strong, D. R., L. A. Szyska, and D. Simberloff. 1979. Tests of community-wide character displacement against null hypotheses. Evolution 33:897–913. Wilson, D. S. 1975. The adequacy of body size as a niche difference. American Naturalist 109:769–784. Yom-Tov, Y. 1991. Character displacement in the psammophile Gerbillidae of Israel. Oikos 60:173–179. Ziv, Y. 1991. Mechanisms of coexistence, based on theories of optimal foraging and habitat selection, between Gerbillus allenbyi and G. pyramidum in the sandy habitat of the Western Negev. M.Sc. thesis. Ben Gurion University. [In Hebrew; English abstract.] Ziv, Y., Z. Abramsky, B. P. Kotler, and A. Subach. 1993. Interference competition and temporal and habitat partitioning in two gerbil species. Oikos 66:237–246. Associate Editor: Jonathan B. Losos