Foraging behaviour of harvesting ants determines seed removal and dispersal

Harvesting ants can affect the regeneration of plants through at least two different processes: seed removal and seed dispersal. We analyse the role of different foraging strategies of ants on patterns of seed removal and dispersal by three Messor species with considerable differences in their foraging systems. Messor capitatus workers rarely leave the nest in well-formed columns, while the other two species form foraging trails, with M. bouvieri forming temporary trails and M. barbarus foraging on a stable system of permanent foraging trails. Overall seed intake of M. capitatus colonies is considerably less than that of the two group-foraging species. There are also differences in the size of seeds collected: M. barbarus and M. capitatus harvest similar amounts of large and small seeds, while M. bouvieri harvests small seeds more intensely than large ones, due to the smaller size of the worker caste. The three Messor species differ in the percent of seed dropping of the different seed type and in the seed dispersal distance. Moreover, M. bouvieri and M. capitatus redistributed dropped seeds preferentially in bare soil and low sparse vegetation habitats, while M. barbarus redistributed seeds mainly in the high vegetation habitat. These results show that the foraging systems of these harvesting ants determine different patterns of seed removal and dispersal and, thus, affect the abundance and redistribution of seeds in the area.     

Foraging,food selection and worker size in the seed harvesting ant <Emphasis Type="Italic">Messor bouvieri</Emphasis>

Interactions among granivores and seeds depend on the foraging behaviour and morphology of the granivores and on the attributes and availability of seeds. We investigated seed selection by the seed harvesting ant Messor bouvieri in three adjacent plant communities in Spain by relating the harvested seeds to those in the seed rain. Preference was positively correlated with seed size and abundance which accounted for 43% and 20% of the variance respectively. Contrary to predictions of central place foraging theory, the size of seeds harvested did not increase with distance from the nest. Inclusion of a less-preferred item in the diet was more strongly related to the abundance of more-preferred items (60% of the variance) than the abundance of the less-preferred item (14% of the variance). Worker size accounted for 20–30% of the variance in the size of harvested seeds, although small workers did not appear to be constrained by load size for the range of seeds available. The body size of ants was significantly larger in the community with the greatest proportion of large seeds, although this was not due to their ability to carry larger loads or due to the greater force required to crush these seeds. The strong preference of M. bouvieri for large seeds may have important consequences for the plant communities in which they forage. Received: 30 November 1999 / Accepted: 14 April 2000     

Foraging and community consequences of seed size for coexisting Negev Desert granivores

We examined the effects of seed size on patch use and diet selection for three co-existing Negev Desert granivores: Allenby's gerbil ( Gerbillus allenbyi ), greater Egyptian sand gerbil ( Gerbillus pyramidum ), and crested lark ( Galerida cristata ). We manipulated size and spatial distribution of seeds in experimental food patches and quantified foraging behavior by measuring giving-up densities (GUDs: the amount of food remaining in a resource patch following exploitation by a forager). In one experiment, we presented small (<1.4 mm in diameter cracked wheat), medium (2.0–3.3 mm), and large (>3.4 mm) seeds in separate trays; in a second, we presented small and medium seeds separately and mixed together. Gerbils had a higher handling time efficiency on smaller seeds, but a much higher encounter probability on larger seeds (20 times higher on large than medium seeds, and 2–5 times higher on medium than small seeds). This led gerbils to have significantly lower GUDs on larger seeds than smaller seeds and to harvest a higher proportion of the larger seeds. When presented with rich and poor patches, G. allenbyi tended to equalize GUDs in both patches, indicating a quitting harvest rate rule for patch exploitation. In contrast, larks appeared to use a fixed time rule for patch exploitation. For larks, seed size did not influence encounter probabilities, and they showed no seed-size selectivity. Still, larks had higher handling efficiencies on smaller than larger seeds, and consequently had a significantly lower GUD on small than medium seeds. Despite large differences between the gerbils and larks in their foraging, our results do not support species coexistence via seed-size partitioning: the larks had much higher GUDs than the gerbils on all seed sizes. Nonetheless, seed size, seed abundance, seed distribution and the animal's patch use behavior all played major roles in determining gerbils' and larks' diet selectivities and GUDs.     

Seed harvesting by a generalist consumer is context‐dependent: Interactive effects across multiple spatial scales

Granivore foraging decisions affect consumer success and determine the quantity and spatial pattern of seed survival. These decisions are influenced by environmental variation at spatial scales ranging from landscapes to local foraging patches. In a field experiment, the effects of seed patch variation across three spatial scales on seed removal by western harvester ants Pogonomyrmex occidentalis were evaluated. At the largest scale we assessed harvesting in different plant communities, at the intermediate scale we assessed harvesting at different distances from ant mounds, and at the smallest scale we assessed the effects of interactions among seed species in local seed neighborhoods on seed harvesting (i.e. resource–consumer interface). Selected seed species were presented alone (monospecific treatment) and in mixture with Bromus tectorum (cheatgrass; mixture treatment) at four distances from P. occidentalis mounds in adjacent intact sagebrush and non‐native cheatgrass‐dominated communities in the Great Basin, Utah, USA. Seed species differed in harvest, with B. tectorum being least preferred. Large and intermediate scale variation influenced harvest. More seeds were harvested in sagebrush than in cheatgrass‐dominated communities (largest scale), and the quantity of seed harvested varied with distance from mounds (intermediate‐scale), although the form of the distance effect differed between plant communities. At the smallest scale, seed neighborhood affected harvest, but the patterns differed among seed species considered. Ants harvested fewer seeds from mixed‐seed neighborhoods than from monospecific neighborhoods, suggesting context dependence and potential associational resistance. Further, the effects of plant community and distance from mound on seed harvest in mixtures differed from their effects in monospecific treatments. Beyond the local seed neighborhood, selection of seed resources is better understood by simultaneously evaluating removal at multiple scales. Associational effects provide a useful theoretical basis for better understanding harvester ant foraging decisions. These results demonstrate the importance of ecological context for seed removal, which has implications for seed pools, plant populations and communities.     

Quantifying the effect of colony size and food distribution on harvester ant foraging

Desert seed-harvester ants, genus Pogonomyrmex, are central place foragers that search for resources collectively. We quantify how seed harvesters exploit the spatial distribution of seeds to improve their rate of seed collection. We find that foraging rates are significantly influenced by the clumpiness of experimental seed baits. Colonies collected seeds from larger piles faster than randomly distributed seeds. We developed a method to compare foraging rates on clumped versus random seeds across three Pogonomyrmex species that differ substantially in forager population size. The increase in foraging rate when food was clumped in larger piles was indistinguishable across the three species, suggesting that species with larger colonies are no better than species with smaller colonies at collecting clumped seeds. These findings contradict the theoretical expectation that larger groups are more efficient at exploiting clumped resources, thus contributing to our understanding of the importance of the spatial distribution of food sources and colony size for communication and organization in social insects.     

A mast‐seeding desert shrub regulates population dynamics and behavior of its heteromyid dispersers

Granivorous rodent populations in deserts are primarily regulated through precipitation‐driven resource pulses rather than pulses associated with mast‐seeding, a pattern more common in mesic habitats. We studied heteromyid responses to mast‐seeding in the desert shrub blackbrush (Coleogyne ramosissima), a regionally dominant species in the Mojave–Great Basin Desert transition zone. In a 5‐year study at Arches National Park, Utah, USA, we quantified spatiotemporal variation in seed resources in mast and intermast years in blackbrush‐dominated and mixed desert vegetation and measured responses of Dipodomys ordii (Ord's kangaroo rat) and Perognathus flavescens (plains pocket mouse). In blackbrush‐dominated vegetation, blackbrush seeds comprised >79% of seed production in a mast year, but 0% in the first postmast year. Kangaroo rat abundance in blackbrush‐dominated vegetation was highest in the mast year, declined sharply at the end of the first postmast summer, and then remained at low levels for 3 years. Pocket mouse abundance was not as strongly associated with blackbrush seed production. In mixed desert vegetation, kangaroo rat abundance was higher and more uniform through time. Kangaroo rats excluded the smaller pocket mice from resource‐rich patches including a pipeline disturbance and also moved their home range centers closer to this disturbance in a year of low blackbrush seed production. Home range size for kangaroo rats was unrelated to seed resource density in the mast year, but resource‐poor home ranges were larger (P < 0.001) in the first postmast year, when resources were limiting. Blackbrush seeds are higher in protein and fat but lower in carbohydrates than the more highly preferred seeds of Indian ricegrass (Achnatherum hymenoides) and have similar energy value per unit of handling time. Kangaroo rats cached seeds of these two species in similar spatial configurations, implying that they were equally valued as stored food resources. Blackbrush mast is a key resource regulating populations of kangaroo rats in this ecosystem.     

Subcaste‐specific evolution of head size in the ant genus Pheidole

An organism's morphology is constrained by its evolutionary history and the need to meet a variety of potentially competing functions. The ant genus Pheidole is the most species‐rich ant genus and almost every species has a dimorphic worker caste (a few are trimorphic). This separation of workers into two developmentally distinct subcastes (smaller minors and larger majors with distinctively large heads) may partially release individuals from functional constraints on morphology, making Pheidole an ideal genus for addressing questions on the evolution of morphology in relation to ecological specialization. Major workers can perform a variety of tasks, although they are usually specialized for defence, as well as food retrieval and processing. Pheidole species vary in their diet, although many species gather seeds. The major workers mill the seeds using large jaws powered by mandible closer muscles that occupy a large proportion of the head cavity. In the present study, we examined the relationship between seed‐harvesting and morphology in Pheidole, hypothesizing that majors of seed‐harvesting species would have larger heads relative to non‐seed‐harvesters to accommodate the powerful mandibular muscles needed to mill seeds. By taking a phylogenetically controlled comparative approach, we found that majors of seed‐associated Pheidole did not have larger heads (width and length) than majors of non‐seed‐harvesting species. However, the head length of minors (and to a lesser extent head width) was smaller in seed‐harvesters. Additionally, we found the difference in head size between majors and minors was greater in seed‐harvesting species. These morphological differences in diet, however, were not related to changes in the rate of evolution in either seed‐harvesting or non‐seed‐harvesting lineages. These findings suggest that the morphologies of worker subcastes can evolve independently of each other, allowing colonies with polymorphic workers to specialize on new resources or tasks in ways not possible in monomorphic species.     

Effects of body size,seed density,and soil characteristics on rates of seed harvest by heteromyid rodents

Summary Although it is well established that coexisting heteromyid rodent species forage in different microhabitats, we do not yet know the basis for divergent microhabitat choice. One possibility is that seed harvest rates differ among microhabitats, and each species forages where it can extract seeds most efficiently. Microhabitats vary in several factors that could affect heteromyid foraging efficiency, including seed density, soil organic content and particle size distribution. We have explored the effect of each of these variables on harvest rates of several species feeding from petri dishes containing known densities of millet seeds embedded in soil of known particle size and density. Results indicate that the number of seeds harvested per second increases uniformly with seed density and soil density and decreases with soil particle size. Body size affects these relationships: larger animals have higher harvest rates for a given set of conditions and experience a greater relative change in harvest rate for a give change in conditions. This implies that heteromyids can be expected to exhibit species-specific microhabitat preferences while foraging in nature.     

The effects of owl predation on the foraging behavior of heteromyid rodents

Summary Researchers have documented microhabitat partitioning among the heteromyid rodents of the deserts of North America that may result from microhabitat specific predation rates; large/bipedal species predominate in the open/risky microhabitat and small/quadrupedal species predominate in the bush/safer microhabitat. Here, we provide direct experimental evidence on the role of predatory risk in affecting the foraging behavior of three species of heteromyid rodents: Arizona pocket mouse (Perognathus amplus; small/quadrupedal), Bailey's pocket mouse (P. baileyi; large/quadrupedal), and Merriam's kangaroo rat (Dipodomys merriami; large/bipedal). Both kangaroo rats and pocket mice are behaviorally flexible and able to adjust their foraging behavior to nightly changes in predatory risk. Under low levels of perceived predatory risk the kangaroo rat foraged relatively more in the open microhabitat than the two pocket mouse species. In response to the presence of barn owls, however, all three species shifted their habitat use towards the bush microhabitat. In response to direct measures of predatory risk, i.e. the actual presence of owls, all three species reduced foraging and left resource patches at higher giving up densities of seeds. In response to indirect indicators of predatory risk, i.e. illumination, there was a tendency for all three species to reduce foraging. The differences in morphology between pocket mice and kangaroo rats do appear to influence their behavioral responses to predatory risk.     

Positive interactions between desert granivores: localized facilitation of harvester ants by kangaroo rats

Facilitation, when one species enhances the environment or performance of another species, can be highly localized in space. While facilitation in plant communities has been intensely studied, the role of facilitation in shaping animal communities is less well understood. In the Chihuahuan Desert, both kangaroo rats and harvester ants depend on the abundant seeds of annual plants. Kangaroo rats, however, are hypothesized to facilitate harvester ants through soil disturbance and selective seed predation rather than competing with them. I used a spatially explicit approach to examine whether a positive or negative interaction exists between banner-tailed kangaroo rat (Dipodomys spectabilis) mounds and rough harvester ant (Pogonomyrmex rugosus) colonies. The presence of a scale-dependent interaction between mounds and colonies was tested by comparing fitted spatial point process models with and without interspecific effects. Also, the effect of proximity to a mound on colony mortality and spatial patterns of surviving colonies was examined. The spatial pattern of kangaroo rat mounds and harvester ant colonies was consistent with a positive interspecific interaction at small scales (<10 m). Mortality risk of vulnerable, recently founded harvester ant colonies was lower when located close to a kangaroo rat mound and proximity to a mound partly predicted the spatial pattern of surviving colonies. My findings support localized facilitation of harvester ants by kangaroo rats, likely mediated through ecosystem engineering and foraging effects on plant cover and composition. The scale-dependent effect of kangaroo rats on abiotic and biotic factors appears to result in greater founding and survivorship of young colonies near mounds. These results suggest that soil disturbance and foraging by rodents can have subtle impacts on the distribution and demography of other species.     

The spatial scale of seed collection by harvester ants

Colonies of the seed-eating ant, Pogonomyrmex barbatus, compete with neighboring colonies for foraging areas. In a conflict over foraging area, what is at stake? This depends on how resources are distributed in time and space: if certain regions consistently provide particularly nutritious seed species, or especially abundant seeds, such regions will be of greater value to a colony. During the summer, seeds were taken from returning foragers in colonies located in 4 different vegetation types. There was no relation between the vegetation currently growing in the foraging area, and the species of seeds collected by ants. During the summer, ants collect mostly seeds produced in previous seasons and dispersed by wind and flooding. In 1991, colonies in all vegetation types collected mostly Bouteloua aristidoides; in 1992, Eriastrum diffusum and Plantago patagonica. There was no relation between colony density and numbers of seeds collected. Seed species collected by ants were compared in different colonies, and on different foraging trails within a colony. The results show that seed patches are distributed on the scale of distances between nests, not the smaller scale of different foraging trails of one colony. It appears that colonies are competing for any space in which to search for seeds, not competing for certain regions of consistently high value.     

Discrimination among pinyon pine trees by Clark's Nutcrackers: effects of cone crop size and cone characters

Summary The influences of Colorado pinyon pine (Pinus edulis) cone crop size, cone and seed weight, cone length, number of seeds per cone, number of viable seeds, and percent viable seeds on the foraging behavior of avian seed dispersal agents were examined in field and laboratory settings. In the field, there was a significant positive relationship between cone number per tree and both the absolute number of cones and the percentage of the cone crop from which seeds were harvested. Cone weight and the number of viable seeds were also significantly related to seed harvest intensity. Laboratory experiments examined the relationship between crop size and cone characters on seed harvest by 18 Clark's Nutcrackers (Nucifraga columbiana). Nutcrackers were offered a choice of two tree types: one with 20 cones attached, and another with 10 cones attached. Significantly more birds chose to remove seeds first from the tree with 20 cones than the tree with 10 cones. In timed trials, they also harvested seeds from significantly more cones on the tree with the higher cone density. In the laboratory, cones chosen for seed removal by the nutcrackers had significantly more viable seeds, more seeds, and were longer compared to cones that were not chosen. Such discriminatory foraging behavior may increase avian foraging efficiency and result in differential reproductive success of pinyon pines. This behavior may therefore influence the evolution of pinyon pine reproductive traits.     

Teasing Apart the Effects of Seed Size and Energy Content on Rodent Scatter-Hoarding Behavior

Scatter-hoarding rodents are known to play a crucial role in the seed dispersal of many plant species. Numerous studies have indicated that both seed size and the energy content of seeds can affect rodent foraging behavior. However, seed size is usually associated with energy content per seed, making it difficult to isolate how seed size and energy affect rodent foraging preferences. This study used 99 treatments of artificial seeds (11 seed sizes×9 levels of energy content) to tease apart the effect of seed size and energy content on rodent seed-caching behavior. Both seed traits showed significant effects, but their details depended on the stage of the rodent foraging process. Seeds with higher energy content were harvested more rapidly while seed size only had a modest effect on harvest rate. However, after harvesting, seed size showed a much stronger effect on rodent foraging behavior. Rodents’ choice of which seeds to remove and cache, as well as seed dispersal distance, seemed to reflect an optimal seed size. Our findings could be adapted in future studies to gain a better understanding of scatter-hoarding rodent foraging behavior, and the co-evolutionary dynamics between plant seed production and seed dispersers.     

Promotion of exotic weed establishment by endangered giant kangaroo rats (Dipodomys ingens) in a California grassland

Giant kangaroo rats (Dipodomys ingens) continually modify their burrow precincts by digging tunnels, clipping plants, and other activities. In the valley grasslands of the Carrizo Plain Natural Area (San Luis Obispo County, California), this chronic disturbance to soil and vegetation promoted the establishment of exotic ruderal and early successional plant species. Erodium cicutarium, Bromus madritensis ssp. rubens, and other Mediterranean annuals were found to constitute a very large proportion of the vegetation on giant kangaroo rat precincts. When vegetation on precincts was compared with the vegetation in less disturbed intermediate areas located between precincts, species richness, cover and frequency of exotic plants were significantly greater on precincts. The reverse was found for native species. In addition, exotic species encountered in this study had significantly larger seeds than did native species, suggesting that these granivorous kangaroo rats preferentially cache large weed seeds on their precincts. Since the kangaroo rats depend on exotic plants for food and the exotic plants depend upon the kangaroo rats to disturb their habitat continually, the weed-kangaroo rat relationship is mutualistic. This strong relationship may also inhibit population growth of native grassland plants which occupy disturbed habitats but have difficulty competing with exotic weeds for resources. From a conservation perspective, this mutualism presents an intractable management dilemma. Restoration of valley grasslands where endangered giant kangaroo rats occur, to conditions where native species dominate, may be impossible.     

ENHANCEMENT OF INBREEDING DEPRESSION BY DOMINANCE AND SUPPRESSION IN IMPATIENS CAPENSIS

We investigated the effect of intraspecific competition on the magnitude of inbreeding depression in Impatiens capensis by planting seeds from chasmogamous (CH) and cleistogamous (CL) flowers in three experimental greenhouse treatments: in individual pots, in flats in dense pure stands according to seed type, and in flats with the two seed types intermixed in a checkerboard array. The size distributions of plants grown in flats were significantly more hierarchical than those of plants grown individually, indicating that larger plants competitively suppressed smaller plants in the high-density treatments. The magnitude of inbreeding depression at high density depended upon the planting arrangement of CL and CH seeds. CH advantage was greatest when CH and CL seedlings were grown in competition with one another, suggesting that fitness differences between outcrossed and inbred individuals were intensified by dominance and suppression. For plants grown individually, the effects of maternal parent, seed weight, and emergence date on seedling size disappeared with plant age, whereas at high density these effects remained at the final harvest. Thus, plant density may influence patterns of natural selection both on mating system and on juvenile traits in natural Impatiens populations.     

Foraging ecology of the South Australian glossy black‐cockatoo (Calyptorhynchus lathami halmaturinus)

The endangered South Australian glossy black‐cockatoo (Calyptorhynchus lathami halmaturinus Mathews 1912) feeds almost exclusively on the seeds of the drooping sheoak (Allocasuarina verticillata), and shows marked preferences for individual trees. This field study investigated foraging ecology and tree selection through observations of foraging birds and measurements of trees and seed cones. The cockatoos spent the vast majority of their foraging time (94%) handling seed cones, and handling behaviour was highly stereotyped. Handling time per cone was correlated primarily with cone size, while seed intake rate was correlated primarily with seed mass per cone. The cockatoos fed mostly in trees with signs of previous feeding. They tended initially to sample trees with large seeds, and to stay for long feeding bouts in trees with high ratios of seed‐to‐cone mass. As a result of these biases, feeding was concentrated in trees with high seed mass per cone. Preferred trees were also larger, with higher ratios of seed‐to‐cone mass and larger seeds containing more lipid and protein. By feeding from selected trees the cockatoos increased both their seed intake rate and the nutritional quality of the seeds ingested, thereby increasing their energy intake rate by an estimated 28%. They did not discriminate against trees that had re‐grown from basal shoots after fires. Insect larvae were present in some seed cones but the cockatoos did not appear to actively seek them. Males foraged 19% more efficiently than females, resulting in greater daily food intake. The characteristics of individual A. verticillata trees that determined the cockatoos’ feeding rates were also correlated with their distribution on a regional scale. This suggests that the distribution of this endangered cockatoo depends not only on the presence of food trees, but also on their regionally varying feeding profitability.     

Density-dependent foraging in the harvester ant Messor ebeninus: two experiments

Harvester ants are important seed predators in many xeric environments, and their foraging choices can influence the composition of plant communities. Seed abundance has been cited as an important factor in determining such foraging preferences. Three seed types (sesame, millet, and flax) were experimentally introduced in differing proportions near nests of the ant Messor ebeninus near the Dead Sea, in territory administered by the state of Israel. Two experiments were designed to investigate the effects of this density conditioning on two stages of the ants' subsequent foraging behavior: recruitment to seed patches and selection of seeds from within a patch. When seeds were presented in small, single-species patches, experimentally common seeds were discovered and exploited significantly faster than rare seeds, especially among less preferred seed types. When seeds were presented in large, mixed patches, however, no consistent effect of density was observed.     

A field assessment of optimal foraging in ants: trail patterns and seed retrieval by the European harvester ant Messor barbarus

Summary: The ant Messor barbarus is a major seed predator on annual grasslands of the Mediterranean area. This paper is an attempt to relate the foraging ecology of this species to resource availability and to address several predictions of optimal foraging theory under natural conditions of seed harvesting.¶Spatial patterns of foraging trails tended to maximise acquisition of food resources, as trails led the ants to areas where seeds were more abundant locally. Moreover, harvesting activity concentrated on highly frequented trails, on which seeds were brought into the nest in larger numbers and more efficiently, at a higher mean rate per worker.¶The predictions of optimal foraging theory that ants should be more selective in both more resource-rich and more distant patches were tested in the native seed background. We confirm that selectivity of ants is positively related to trail length and thus to distance from the nest of foraged seeds. Conversely, we fail to find a consistent relationship between selectivity and density or species diversity of seed patches. We discuss how selectivity assessed at the colony level may depend on factors other than hitherto reported behavioural changes in seed choice by individual foragers.     

Seed size selection by olive baboons

Seed size is an important plant fitness trait that can influence several steps between fruiting and the establishment of a plant’s offspring. Seed size varies considerably within many plant species, yet the relevance of the trait for intra-specific fruit choice by primates has received little attention. Primates may select certain seed sizes within a species for a number of reasons, e.g. to decrease indigestible seed load or increase pulp intake per fruit. Olive baboons (Papio anubis, Cercopithecidae) are known to select seed size in unripe and mature pods of Parkia biglobosa (Mimosaceae) differentially, so that pods with small seeds, and an intermediate seed number, contribute most to dispersal by baboons. We tested whether olive baboons likewise select for smaller ripe seeds within each of nine additional fruit species whose fruit pulp baboons commonly consume, and for larger seeds in one species in which baboons feed on the seeds. Species differed in fruit type and seed number per fruit. For five of these species, baboons dispersed seeds that were significantly smaller than seeds extracted manually from randomly collected fresh fruits. In contrast, for three species, baboons swallowed seeds that were significantly longer and/or wider than seeds from fresh fruits. In two species, sizes of ingested seeds and seeds from fresh fruits did not differ significantly. Baboons frequently spat out seeds of Drypetes floribunda (Euphorbiaceae) but not those of other plant species having seeds of equal size. Oral processing of D. floribunda seeds depended on seed size: seeds that were spat out were significantly larger and swallowed seeds smaller, than seeds from randomly collected fresh fruits. We argue that seed size selection in baboons is influenced, among other traits, by the amount of pulp rewarded per fruit relative to seed load, which is likely to vary with fruit and seed shape.     

Effects of kangaroo rat exclusion on vegetation structure and plant species diversity in the Chihuahuan Desert

Long-term (1977–90) experimental exclusion of three species of kangaroo rats from study plots in the Chihuahuan Desert resulted in significant increases in abundance of a tall annual grass (Aristida adscensionis) and a perennial bunch grass (Eragrostis lehmanniana). This change in the vegetative cover affected use of these plots by several other rodent species and by foraging birds. The mechanism producing this change probably involves a combination of decreased soil disturbance and reduced predation on large-sized seeds when kangaroo rats are absent. Species diversity of summer annual dicots was greater on plots where kangaroo rats were present, as predicted by keystone predator models. However, it is not clear whether this was caused directly by activities of the kangaroo rats or indirectly as a consequence of the increase in grass cover. No experimental effect on species diversity of winter annual dicots was detected. Our study site was located in a natural transition between desert scrub and grassland, where abiotic conditions and the effects of organisms may be particularly influential in determining the structure and composition of vegetation. Under these conditions kangaroo rats have a dramatic effect on plant cover and species composition.