Perognathus Longimembris Research Articles

A highly contiguous genome assembly for the pocket mouse Perognathus longimembris longimembris.

The little pocket mouse, Perognathus longimembris, and its nine congeners are small heteromyid rodents found in arid and seasonally arid regions of Western North America. The genus is characterized by behavioral and physiological adaptations to dry and often harsh environments, including nocturnality, seasonal torpor, food caching, enhanced osmoregulation, and a well-developed sense of hearing. Here we present a genome assembly of Perognathus longimembris longimembris generated from PacBio HiFi long read and Omni-C chromatin-proximity sequencing as part of the California Conservation Genomics Project. The assembly has a length of 2.35 Gb, contig N50 of 11.6 Mb, scaffold N50 of 73.2 Mb, and includes 93.8% of the BUSCO Glires genes. Interspersed repetitive elements constitute 41.2% of the genome. A comparison with the highly endangered Pacific pocket mouse, P. l. pacificus, reveals broad synteny. These new resources will enable studies of local adaptation, genetic diversity, and conservation of threatened taxa.

Interactions among rainfall, fire, forbs and non-native grasses predict occupancy dynamics for the endangered Pacific pocket mouse (Perognathus longimembris pacificus) in a Mediterranean-type ecosystem

It is important to understand species-habitat relationships to implement effective adaptive management for rare species. However, it can be challenging to assess habitat associations and their relationships to abiotic stressors in dynamic habitats without the insights that can be gained from long-term monitoring. We report results from the first six years of extensive track tube monitoring of the largest two of three remaining extant populations of federally endangered Pacific pocket mouse (Perognathus longimembris pacificus) in a coastal Mediterranean-type ecosystem on Marine Corps Base, Camp Pendleton in southern California, USA. We used dynamic occupancy and structural equation modeling to assess potential drivers of population trends that included habitat, fire history, rainfall, disturbance, and the presence of other small mammals. We found that the variables that best predicted mouse occupancy were moderate to high forb and perennial herb cover (40–80%), and moderate to high open ground (20–70%) and low non-native grass cover (<20%), Non-native grass cover (>20%) was also a strong predictor of lower PPM colonization and increased extinction probabilities, with the extent of non-native grass cover being strongly influenced by annual rainfall and recency of fire. Our study adds to the growing literature on effects of invasive annual grasses on native species in Mediterranean-type ecosystems. We suggest that habitat management could be based upon promotion of open forb and perennial herb dominated habitats with reduction of non-native grasses by prescribed fire and other methods. These types of spatial and temporal monitoring programs can support land managers by creating a monitoring and management feedback loop. They can reveal landscape and environmental variables associated with species persistence, inform habitat management goals, and help managers to assess the success of management actions on populations of conservation concern.

Taxonomic reassessment of the Little pocket mouse, Perognathus longimembris (Rodentia, Heteromyidae) of southern California and northern Baja California

The Little pocket mouse (Perognathus longimembris) encompasses 15 to 16 currently recognized subspecies, six of which are restricted to southern California and adjacent northern Baja California. Using cranial geomorphometric shape parameters and dorsal color variables we delineate six regional groups of populations from this area that we recognize as valid, but these differ in name combination and geographic range from the current taxonomy. We resurrect two names from their current placement in synonymies, synonymize two currently recognized subspecies, and we reassign a third. Importantly, we restrict the U. S. Federally endangered Pacific pocket mouse (P. l. pacificus Mearns) to the vicinity of its type locality at the mouth of the Tijuana River in the southwestern corner of San Diego County and resurrect P. l. cantwelli von Bloeker for the other two population segments along the coast, those that span the northwestern corner of San Diego County and adjacent Orange County and that in coastal Los Angeles County. The name cantwelli would now apply to the only extant populations of the Pacific pocket mouse, a reassignment with obvious management implications. Our taxonomic decisions also reconfigure the ranges of other subspecies of conservation concern, notably P. l. bangsi Mearns and P. l. brevinasus Osgood.

Quantitative SWOT analysis: A structured and collaborative approach to reintroduction site selection for the endangered Pacific pocket mouse

Species extinction and loss of biodiversity are major crises in the Anthropocene. Translocations of threatened and endangered species, the movement of individuals to augment existing or establish new populations, are increasingly important conservation tools, but have historically had limited success. Selection of a suitable receiver site is essential to translocation success, with poor site suitability cited as one of the most common reasons for relocation failure. We utilized a quantitative SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis to evaluate and prioritize potential receiver sites for the Pacific pocket mouse (Perognathus longimembris pacificus), an endangered subspecies of heteromyid rodent endemic to coastal southern California. With only three remaining extant populations, a conservation breeding and reintroduction program is underway with the goal of creating additional wild populations in new or historic locations throughout its indigenous range. Here we describe our use of SWOT analysis and discuss the strengths of this approach as well as improvements that could be made to the evaluation process for other species. Overall, we found that using a structured, transparent, and collaborative process was a valuable tool for prioritizing receiver sites. SWOT analysis is a flexible, repeatable, and proactive approach for identifying receiver sites and the preparations necessary to improve species-specific suitability. This approach has the potential to result in successful relocation compared to less structured site selection processes where poor site suitability is ultimately identified as a major factor in failure to establish wild populations.

Fitness costs associated with ancestry to isolated populations of an endangered species

Habitat fragmentation from urban development leaves species vulnerable to inbreeding depression and genomic erosion. Restoring gene flow can reduce inbreeding and preserve genetic diversity, but a common concern is that genomic incompatibilities may lead to outbreeding depression. The introduction of deleterious genetic load is less commonly considered. The endangered Pacific pocket mouse (Perognathus longimembris pacificus) persists in three isolated populations in southern California. Mitochondrial and microsatellite data indicated that effective population sizes were extremely small (Ne< 50), and continued declines prompted a conservation breeding program founded by individuals from each population. We tracked genetic diversity and individual fitness (measured by reproductive success) in a captive setting over six generations of admixture. Although we observed an increase in fitness in the F1 and F2 generations relative to the founding populations, inbreeding depression alone did not explain the fitness patterns, and outbreeding depression was difficult to rule out as reproductive success waned after F2. However, reproductive success was consistently higher in admixed individuals than founders from Dana Point, the smallest population with the lowest heterozygosity. Across generations, we saw a strong negative correlation between individual reproductive success and ancestry to Dana Point, leading to a rapid decrease of Dana Point alleles. Although the genomic underpinnings remain to be determined, reduced fitness associated with Dana Point ancestry is consistent with high deleterious genetic load in this population, and thus any facilitated migration should be unidirectional. Our findings highlight that, even in the absence of outbreeding depression, there may be a cost to restoration of gene flow if populations harbor high genetic load.

Caching Propensities and Effectiveness of Five Coexisting Heteromyid Rodent Species as Dispersers of Indian Ricegrass (Achnatherum hymenoides) Seeds

Recruitment of Indian ricegrass (Achnatherum hymenoides) seedlings occurs primarily from surface seed caches made by scatter-hoarding desert rodents of the family Heteromyidae. We used radiolabeled Indian ricegrass seeds in field seed-caching experiments with 5 coexisting heteromyid species (Dipodomys deserti, D. merriami, D. microps, Microdipodops pallidus, and Perognathus longimembris) to compare their potential effectiveness as dispersers. Dipodomys microps individuals deposited seeds almost exclusively in larders rather than in scatter hoards, implying that this species is unlikely to be an important seed disperser. Among the other species, individual caches made by larger species had more seeds than those made by smaller species, but smaller species made more caches. At the level of individual animals, M. pallidus was the most effective disperser in a quantitative sense; they made more caches than other species tested and placed fewer excess seeds in caches relative to optimal cache sizes for Indian ricegrass seedling establishment. However, because D. merriami individuals were considerably more abundant at the study site than other species and were also avid scatter hoarders, D. merriami is likely to be the most quantitatively effective disperser of Indian ricegrass seeds at the species level. Ranking species according to qualitative effectiveness (e.g., by considering effects such as the caching microsite on seedling establishment) was more ambiguous. For example, P. longimembris made relatively shallow caches that most closely match optimal planting depths for Indian ricegrass seedling emergence, but such shallow caches are probably more likely than deeper caches to be discovered by foraging rodents before they can germinate. With the possible exception of D. microps, any of the species we tested may be effective dispersers of Indian ricegrass seeds.

Cache pilfering in a granivore guild: Implications for reintroduction management

ABSTRACTReintroduction programs that release endangered species back into areas from which they have been extirpated rarely take competitive interactions between species into account. The endangered Pacific pocket mouse (Perognathus longimembris pacificus) is being reintroduced to parts of its former range where multiple species of native rodents have overlapping diets. The granivorous species in this foraging guild compete for seeds exploitatively and through direct interference interactions, and pocket mice are the smallest and least dominant species in the guild. Repeated aggressive interactions from resident heterospecifics could lower the chances of reintroduced pocket mice establishing burrows during the post‐release settlement period. As such, temporarily reducing the density of competing species through exclusionary fencing might be an advisable reintroduction strategy, in combination with other interventions, such as predator exclusion. The presence of other members of the seed‐foraging guild, however, could have a net benefit for pocket mice, if the pocket mice pilfer from the other species’ seed caches more than the other species pilfer from their caches. To test the frequency of cache pilfering between species, we conducted a field experiment with fluorescent dyed seeds. Two of 10 pocket mice pilfered from heterospecifics, but only 1 of 33 heterospecifics pilfered from pocket mice. In a field‐enclosure experiment, we could not conclude that any of the 4 species tested used heterospecific scent to find (or avoid) seed caches, and pocket mice were less efficient in pilfering from artificial caches, recovering fewer seeds than the larger species. We did not find strong evidence that Pacific pocket mice benefit from living in sympatry with larger, dominant species. Although further research is needed to elucidate the relationship between heterospecific density and the prevalence of cache pilfering, a conservative reintroduction approach would be to select receiver sites with low densities of known competitors to benefit pocket mice during the critical post‐release establishment phase. © 2019 The Wildlife Society.

Severe thermoregulatory deficiencies in mice with a deletion in the titin gene TTN.

Muscular dystrophy with myositis (mdm) mice carry a deletion in the N2A region of the gene for the muscle protein titin (TTN), shiver at low frequency, fail to maintain body temperatures (Tb) at ambient temperatures (Ta) <34°C, and have reduced body mass and active muscle stiffness in vivo compared with wild-type (WT) siblings. Impaired shivering thermogenesis (ST) could be due to the mutated titin protein causing more compliant muscles. We hypothesized that non-shivering thermogenesis (NST) is impaired. To characterize the response to cold exposure, we measured Tb and metabolic rate (MR) of WT and mdm mice at four nominal temperatures: 20, 24, 29 and 34°C. Subsequently, we stimulated NST with noradrenaline. Manipulation of Ta revealed an interaction between genotype and MR: mdm mice had higher MRs at 29°C and lower MRs at 24°C compared with WT mice. NST capacity was lower in mdm mice than in WT mice. Using MR data from a previous study, we compared MR of mdm mice with MR of Perognathus longimembris, a mouse species of similar body mass. Our results indicated low MR and reduced NST of mdm mice. These were more pronounced than differences between mdm and WT mice owing to body mass effects on MR and capacity for NST. Correcting MR using Q10 showed that mdm mice had lower MRs than size-matched P. longimembris, indicating that mutated N2A titin causes severe thermoregulatory defects at all levels. Direct effects of the titin mutation lead to lower shivering frequency. Indirect effects likely lead to a lower capacity for NST and increased thermal conductance through decreased body size.

Isotopic niche variation from the Holocene to today reveals minimal partitioning and individualistic dynamics among four sympatric desert mice

Species interact with each other and their environment over a range of temporal scales, yet our understanding of resource partitioning and the mechanisms of species coexistence is largely restricted to modern time-scales of years to decades. Furthermore, the relative magnitudes of inter- vs. intraspecific variation in resource use are rarely considered, despite the potential for the latter to influence a species' ability to cope with changing environmental conditions. Modern desert rodent communities are thought to be strongly structured by competitive interactions, with niche partitioning of food resources hypothesized to explain the coexistence of multiple sympatric granivores. Yet the stability of niche dynamics over extended temporal scales within desert rodent communities is unknown. I examined the isotopic niche dynamics of four common sympatric desert mice (three granivores: Chaetodipus formosus, Perognathus longimembris and Reithrodontomys megalotis, and one omnivore: Peromyscus maniculatus) in the Smoke Creek Desert of northwestern Nevada using 13 C and 15 N isotopes obtained from "Modern" (2008-2013 CE), "Historical" (1989-2005 CE) and Holocene fossil specimens spanning the last c. 7,500years. I found significant variation in niche position, niche breadth and interspecific niche overlap of these species through time. The niche breadth dynamics of the cricetids (P.maniculatus and R.megalotis) were positively correlated with one another, while the niche breadth dynamics of the heteromyid C.formosus were negatively correlated with those of all other species. Body size, dietary functional group, palaeoenvironmental trends and time-averaging provided little explanatory power. Importantly, Modern and Historical patterns of resource use and partitioning differed from Holocene baselines in terms of decreased niche overlap and in the absolute and relative position of each species' niche in at least one isotopic axis. These observations suggest that each species' resource use changed individualistically over the Holocene, hence niche dynamics are poorly explained by the hypothesis of temporally stable species interactions at millennial time-scales. Furthermore, changes to the resource base over the last century (likely due to the spread of invasive cheatgrass) may be increasing resource partitioning in the Modern, pushing species past their baseline ranges of resource use variation.

Hibernating Little Pocket Mice Show Few Seasonal Changes in Bone Properties

Periods of disuse or physical inactivity increases bone porosity and decreases bone mineral density, resulting in a loss of bone mechanical competence in many animals. Although large hibernators like bears and marmots prevent bone loss during hibernation, despite long periods of physical inactivity, some small hibernators do lose bone during hibernation. Little pocket mice (Perognathus longimembris) remain underground during winter hibernation and undergo bouts of torpor and interbout arousals, but the torpor bout duration is shorter than other rodent hibernators. Additionally, little pocket mice may enter torpor during summer estivation. In this study, cortical and trabecular bone architectural, mineral, and mechanical properties were analyzed for femurs from little pocket mice captured during 8 different months (March to October) to determine seasonal effects on bone. There were no differences in any bone properties between the pre-hibernation month of October and the post-hibernation month of March, suggesting winter hibernation did not adversely affect bone properties. However, cortical area was higher in March than April, May, and June. Bone mechanical and osteocyte lacunar properties were not different between any months. Trabecular bone in the distal femoral epiphysis showed no changes between months. The distal femoral metaphyseal region showed higher trabecular spacing and lower trabecular number in May than August, otherwise, there were no differences in trabecular parameters. The few monthly differences in bone properties may be due to physical inactivity from periodic summer estivation or from the timing of birth and growth in spring and summer months. Anat Rec, 300:2175-2183, 2017. © 2017 Wiley Periodicals, Inc.

Rodent Removal of Fallen Joshua Tree (Yucca brevifolia) Fruits

Abstract Joshua trees (Yucca brevifolia) produce large, indehiscent fruits that contain numerous large seeds. Seed dispersal in this species depends on rodents to dismantle fruits and extract the seeds which they disperse tens of meters from the source. Using camera trapping and fruits tied to bobbins, I show that white-tailed antelope squirrels (Ammospermophilus leucurus) and kangaroo rats (Dipodomys spp.) moved intact, fallen fruits 6 to 7 m from trees before opening them. Pocket mice (Chaetodipus fallax and Perognathus longimembris) and pinyon mice (Peromyscus trueii) dismantled fruits and harvested loose seeds but did not appear to move them although they readily harvested loose seeds. Mobilizing fruits may be an important, overlooked step in the seed dispersal process, especially if the fruits are indehiscent. Fruit-carrying behavior of rodents described in this study adds to the dispersal distance of Joshua tree seeds.

Coexisting desert rodents differ in selection of microhabitats for cache placement and pilferage

Seed caching by desert rodents in the family Heteromyidae is an important behavioral adaptation for animals living in environments with limited and unpredictable food resources. Heteromyids cache seeds throughout their home ranges, either concentrated in 1 location (larder hoard) or in multiple, small seed piles (scatter hoards). To minimize cache pilferage by other rodents, coexisting species may scatter hoard seeds in distinct microhabitats. We examined interspecific differences in caching and pilfering behaviors of 3 coexisting heteromyid rodents, Merriam's kangaroo rat (Dipodomys merriami), the pale kangaroo mouse (Microdipodops pallidus), and the little pocket mouse (Perognathus longimembris), to determine if caching microhabitat affects likelihood of pilferage. In outdoor enclosures we tracked cache placement and measured pilferage of artificial caches using radiolabeled Indian ricegrass (Achnatherum hymenoides) seeds. M. pallidus and P. longimembris placed seed caches mostly under shrubs, whereas D. merriami placed caches predominately in open microhabitat. However, D. merriami showed a significant preference for pilfering caches under shrubs, whereas P. longimembris did not show a significant preference for pilfering caches in either open or undershrub microhabitats. The 2 species pilfered similar numbers of caches. Coexisting heteroymid rodent species may contribute to spatial heterogeneity of available resources by caching in different microhabitats, thereby reducing but not eliminating cache pilferage by other species.

MARK–RECAPTURE ESTIMATES OF POPULATION PARAMETERS FOR SELECTED SPECIES OF SMALL MAMMALS

We used program CAPTURE to analyze 1,535 capture–recapture data sets from 33 species of small mammals for sources of variation in capture probabilities and to characterize species-specific responses to mark–recapture trapping. Program CAPTURE contains 8 models to account for all possible combinations of heterogeneity (h), behavioral response (b), and time (t) effects on capture probabilities. Model Mo (equal capture probabilities) was selected most often by CAPTURE as the most appropriate model for small sample sizes, reflecting the difficulty of detecting unequal probabilities of capture in small data sets. Heterogeneity in capture probabilities among individuals was evident in larger data sets for all species of small mammals examined. Strong behavioral responses were detected in several species. Sylvilagus, Tamias amoenus, Glaucomys sabrinus, Perognathus longimembris, and Microtus ochrogaster were trap shy, whereas Tamias striatus, Tamias townsendii, Spermophilus richardsonii, Perognathus parvus, Reithrodontomys megalotis, Peromyscus maniculatus, Peromyscus truei, Sigmodon hispidus, Microtus californicus, and Microtus pennsylvanicus were trap prone. Time effects were not apparent for most species except where they were accompanied with heterogeneity. Murid species had the highest estimated probabilities of capture, followed by heteromyids, sciurids, and leporids in decreasing order. Estimates of abundance from model Mh were significantly greater than those from models Mb, Mo, and Mt, which likely reflects the negative bias of these latter estimators.

PHYLOGENY OF THE PEROGNATHUS LONGIMEMBRIS SPECIES GROUP BASED ON MITOCHONDRIAL CYTOCHROME-b: HOW MANY SPECIES?

The Perognathus longimembris species group currently consists of 3 species and approximately 23 subspecies. Members of this species group have been of considerable interest to biologists in both physiological and ecological studies, and several members are threatened with extinction. No one has yet attempted a phylogenetic study including several populations from each of these 3 species to test the monophyly of these taxa and to elucidate the relationships among the species and subspecies. Mitochondrial DNA sequences from the cytochrome-b gene were used to assess the genetic variation among these species and to make a 1st approximation of the number of species that should be recognized taxonomically in this species group. Results indicate that the Arizona pocket mouse, P. amplus, and the San Joaquin pocket mouse, P. inornatus, are monophyletic entities, and should remain as single polytypic species, whereas the little pocket mouse, P. longimembris, is paraphyletic in the maximum-likelihood and maximum-parsimony phylogenies, suggesting that it might best be split into 2 polytypic species.

Morphological, chromosomal, and molecular evolution are uncoupled in pocket mice

C-, and G-banded chromosomes are presented for Perognathus amplus and Perognathus longimembris from Arizona, USA and Chaetodipus nelsoni from Coahuila, Mexico. The two species of Perognathus reveal similar C-band patterns, and extensive autosomal and X chromosome G-band identity with only pericentric inversions distinguishing pairs 4 and 6 and a difference in the morphology of pair 20. Three pairs of autosomal secondary constrictions were found in P. amplus and only one in P. longimembris. Only 50% of the amplus/longimembris G-banded karyotype could be aligned with that of C. nelsoni indicating extensive chromosomal restructuring has taken place since these genera last shared a common ancestor. A review of the literature suggests variable rates of morphological, chromosomal and molecular evolution in these animals.

Hierarchical genetic structure in fragmented populations of the Little Pocket Mouse (Perognathus longimembris) in Southern California

The geographic genetic structure, based onsequence variation of an 810 base pair fragmentof the mitochondrial cytochrome b gene,is described for populations of five subspeciesof the Little Pocket Mouse, Perognathuslongimembris, from Southern California. Oneof these, P. l. pacificus (Pacific PocketMouse), is listed as Endangered by the U.S.Federal Government. Sixty-two uniquehaplotypes were recovered from 99 individualssampled. Phylogenetic analyses of thesevariants do not identify regionallyreciprocally monophyletic lineages concordantwith the current subspecies designations, butmost haplotypes group by subspecies in networksgenerated by either statistical parsimony ormolecular variance parsimony. Moreover, asubstantial proportion of the total pool ofhaplotype variation is attributed to thesesubspecies, or to local populations withingeographic segments of each, indicating theirrelative evolutionary independence. The pooledextant populations of the endangered PacificPocket Mouse exhibit the same levels ofnucleotide and haplotype diversity as other,presumptively less-impacted populations ofadjacent subspecies, although the sample fromDana Point, Orange County, has markedly lowhaplotype diversity in comparison to allothers. These populations also show a geneticsignature of population expansion rather thanone of decline. Both pieces of evidence are atodds with current empirical populationestimates, which reinforces the fact thatpresent-day patterns of genetic diversity arethe product of coalescent history and will notnecessarily reflect recent anthropogenic, orother, perturbations. Comparison of haplotypevariation within and among extant populationsof the Pacific Pocket Mouse with those obtainedfrom museum samples collected more than 70years ago suggests that the pattern ofpopulation differentiation and diversity was inplace before the post-World War II exponentialurbanization of Southern California.

The foraging behavior of granivorous rodents and short-term apparent competition among seeds

The foraging behavior of a predator species is thought to be the cause of short-term apparent competition among those prey species that share the predator. Short-term apparent competition is the negative indirect effect that one prey species has on another prey species via its effects on predator foraging behavior. In theory, the density-dependent foraging behavior of granivorous rodents and their preference for certain seeds are capable of inducing short-term apparent competition among seed species. In this study, I examined the foraging behavior of two heteromyid rodent species (family Heteromyidae), Merriam’s kangaroo rats (Dipodomys merriami) and little pocket mice (Perognathus longimembris). In one experiment I tested the preferences of both rodent species for the seeds of eight plant species. Both rodent species exhibited distinct but variable preferences for some seeds and avoidance of others. However, the differences in preference appeared to have only an occasional effect on the strength of the short-term apparent competition detected in a field experiment. In another experiment, I found that captive individuals of both rodent species had approximately equal foraging effort (i.e., time spent foraging) in patches that contained a highly preferred seed type (Oryzopsis hymenoides) regardless of seed density and the presence of a less preferred seed type (Astragalus cicer) in the patches. The rodents also harvested a large proportion of O. hymenoides seeds regardless of initial seed density; this precluded a negative indirect effect of A. cicer on O. hymenoides. But there was a negative indirect effect of O. hymenoides on A. cicer caused by rodents having a lower foraging effort in patches that only contained A. cicer seeds than in patches that contained A. cicer and O. hymenoides seeds. The indirect interaction between O. hymenoides and A. cicer thus represented a case of short-term apparent competition that was non-reciprocal. Most importantly, it was caused by the foraging behavior of the rodents.

Effects of Cattle Grazing on Salt Desert Rodent Communities

Abstract Cattle grazing has been shown to alter various features of desert communities that may impact microhabitats required by various species of desert rodents, with unknown implications for desert rodent communities. We conducted a series of studies at heavily and lightly grazed sites to investigate effects of cattle grazing on desert rodent relative abundances, home range sizes and microhabitat use in salt desert shrub communities of the western Great Basin Desert. Monitoring of rodent populations with repeated live trapping showed that different levels of grazing were associated with differences in relative abundances of some species of rodents. Specifically, Dipodomys merriami was significantly more abundant in heavily grazed areas, and Perognathus longimembris was significantly more abundant in lightly grazed areas. Our studies showed that cattle, by preferentially feeding on certain plants, can create conditions that are more suitable for some species of rodents, while reducing important microhab...

Differences in Food Hoarding among Six Species of Heteromyid Rodents

Storing food is an important adaptation of heteromyid rodents to desert environments of southwestern North America. These rodents store seeds in burrows (larder hoarding) and in small, widely scattered caches buried at shallow depths in soil (scatter hoarding). Laboratory trials were done with five species from one site and two species from two other sites to test the hypothesis that larger species would larder hoard relatively more seeds than smaller species. Subjects were tested in arenas consisting of a sand-filled box for scatter hoarding and a nest box for larder hoarding. Chisel-toothed and Merriam's kangaroo rats ( Dipodomys microps and D. merriami ) from one site larder hoarded more than sympatric dark kangaroo mice ( Microdipodops megacephalus ) and long-tailed and little pocket mice ( Chaetodipus formosus and Perognathus longimembris ). Chisel-toothed kangaroo rats larder hoarded more than Merriam's kangaroo rats and little pocket mice larder hoarded more than dark kangaroo mice. These results were consistent with reported information on the relative aggressive behaviors of the species. For subjects from the other two sites, Ord's kangaroo rats ( D. ordii ) larder hoarded more than the smaller Merriam's kangaroo rats. Pocket mice made larger scatter hoards than kangaroo rats and placed these scatter hoards closer to edges and corners of arenas than kangaroo rats. Differences among species in food-hoarding behavior should be considered in future studies of the community ecology of desert rodents.

CAMBIOS EN LA COMPOSICION DE LAS COMUNIDADES DE ROEDORES EN RELACION A LOS TIPOS DE VEGETACIÓN Y GEOMORFOLOGÍA EN EL PINACATE, SONORA, MEXICO

Durante la primavera de 1981, se realizaron muestreos de roedores en las nueve comunidades vegetales principales del Gran Desierto de Sonora. Se registraron 16 especies de roedores: tres de sciúridos, una de geómido, siete de heterómidos y cinco de múridos. La subfamilia Perognathinae fue el grupo más diverso, con cinco especies. Se determinó su distribución respecto a las asociaciones vegetales y al sustrato. Se encontró que los suelos firmes fueron los más utilizados por los roedores. En ellos se localizó al 93.75% de las especies; el ardillón [SpermophHus variegatus) y la rata nopalera {Neotoma albigula) caracterizaron a los sustratos rocosos. Solamente la rata canguro (Dipodomys deserti) fue exclusiva de las dunas más arenosas. Esta última especie separó claramente dentro de la clasificación que se hizo de los sitios con base en las especies de roedores, a los sitios arenosos inestables de los arenosos consolidados. El ratón de bolsas {Perognathus longimembris) y las ratas {Neotoma albigula y N. lepida) separaron dentro de la clasificación a los habitats más diversos de los menos diversos. La vegetación riparia fue importante en cuanto a la diversidad y abundancia de los roedores