Chapter 15 in Causes of Evolution: a Paleontological Perspective, Ross, R. M., and Allmon, W. D. (eds.), Univ. Chicago Press, pp. 422-465 (1990).
QUATERNARY MAMMALS OF THE GREAT BASIN:
EXTINCT GIANTS, PLEISTOCENE RELICTS, AND RECENT IMMIGRANTS

Timothy H. Heaton

Department of Earth Sciences
University of South Dakota
Vermillion, SD 57069

The Great Basin desert comprises nearly all of the state of Nevada, the western half of Utah, and parts of California, Oregon, and Idaho (FIGURE 1). It is defined as that area between the Sierra Nevada and the Rocky Mountains where there is no external drainage. Geologically it is characterized by active regional extension which has produced north-south trending horsts and grabens. Because of the youthfulness of mountain building in the area, the mountain ranges tend to be very high, and 20 of these ranges have peaks over 3000 meters in elevation. In these localized high regions the climate is similar to the Sierra Nevada and Rocky Mountains. But these ranges are separated by desert valleys that are so dry that all the water they collect from the mountains evaporates within them. The topography is typical of areas that are dry and tectonically active. Massive alluvial fans extend from the mountain fronts down into the valleys, and many of the valleys contain pluvial lakes (or lakebeds) that expand and shrink as precipitation varies.

The low precipitation and temperature extremes of the Great Basin valleys permit the survival of only the hardiest xeric vegetation. The valleys contain mostly annual grasses, sagebrush, and other small desert plants. Trees occur only locally around water sources. The ground tends to be rocky near the mountain fronts and hard, dry, and alkaline in the valleys. Many mammalian species that are adapted to a desert habitat live in the interconnecting valleys of the Great Basin, and most of these species have ranges that extend southward into the Mojave Desert and other deserts of the American Southwest.

Table 1 lists a series of floral belts and life-zones as given by Hall (1946) for Nevada. The Upper Sonoran life-zone covers nearly all of the Great Basin and includes the pi¤on-juniper and sagebrush floral belts. The southernmost part of the Great Basin, the Mojave Desert, is in the Lower Sonoran life- zone. All the floral belts above the Sonoran are restricted to the high mountain ranges and occupy only a tiny fraction of the total area (~5%). Because of the steepness of the mountain ranges, these higher floral belts are reduced compared to other parts of the Western United States, and in many places one or more are entirely absent. The transition zone in particular is often absent, and the sharply contrasting Sonoran and Boreal floras are in abrupt contact.

In the boreal floral belts a host of mammals are found which also occur in the Sierra Nevada and/or Rocky Mountains, but none of these species are restricted to any one life-zone. In Table 1, mammalian species which occur no lower than the Transition life-zone are listed for the Canadian life-zone. The open woodlands of the pi¤on-juniper floral belt occupy the margins of the mountain ranges from about 2000 to 3000 m elevation, and several species are restricted to this upper section of the Upper Sonoran life-zone (Table 1). Juniper ranges slightly lower than pi¤on, and in the northwestern part of the Great Basin juniper occurs alone (Hall, 1946). By far the most extensive life-zone of the Great Basin is the main section of the Upper Sonoran covered primarily by sagebrush. A large suite of mammals is restricted to this zone (Table 1), but most of these species extend down into the Lower Sonoran and up into the Transition zone. Table 2 lists the mammal species currently found in the Great Basin, both living and late Pleistocene fossils, and pertinent information concerning these species.

THE MAMMALIAN BIOGEOGRAPHIC PUZZLE: BROWN'S NON-EQUILIBRIUM MODEL

The high-elevation mountain ranges of the Great Basin form montane islands with habitat similar to the more extensive high elevation regions of the Sierra Nevada and Rocky Mountains. Brown (1971, 1978) compared the distribution of birds and mammals on these montane islands to the distribution of species on oceanic islands studied by MacArthur and Wilson (1963, 1967) using the Sierra Nevada and Rocky Mountains as the equivalent of mainland. He found that island distribution of boreal birds and large mammals fit the colonization-extinction equilibrium model outlined by MacArthur and Wilson, but for small boreal mammals it does not. Because of this he challenged the idea that those species got to the montane islands by dispersal under present conditions.

MacArthur and Wilson (1963, 1967) showed that the probability of an oceanic island having a colony of mainland species was related to its distance from the mainland, distance from other islands having such colonies, and the size of the island. Island size not only affects the probability of colonization, but also the probability of extinction once colonization has occurred. Brown (1971, 1978) arbitrarily defined montane islands as ranges having peaks over 3000 m in elevation and being separated from their closest neighboring island or mainland by at least 8 km with elevation below 2300 m. The elevation of 2300 m corresponds to the lower limit of continuous pi¤on-juniper woodland in the Great Basin. Brown did not consider the northern part of the Great Basin in his study because records were lacking and because pi¤on-juniper woodlands are not well developed in that region. With these criteria Brown (1978) identified 19 montane islands in the Great Basin as shown in Figure 2, and data on these mountain ranges is given in Table 3.

The criteria Brown used to define boreal animals was somewhat arbitrary. He chose species that live at high elevations in the Sierra Nevada and/or Rocky Mountains and do not range lower than the pi¤on-juniper floral belt. The mountain ranges of the Great basin are truly islands to species thus defined. Table 4 shows which boreal mammals live on which islands. Mountain ranges containing the highest numbers of these species are not particularly close to a mainland or to each other, neither do they have a higher-elevation pass to a mainland or another island. The number of boreal species inhabiting an island is strongly related to its size, however, and this relationship is even more pronounced than on oceanic islands. This suggests a very low rate of colonization relative to extinction. Such a system is not in equilibrium and can only be explained by historical factors.

Hall (1946), in an early attempt to explain the discontinuous range of small boreal mammals in the Great Basin, proposed that these animals might disperse from one range to another during the winter, but all the available evidence contradicts this. Since they do not fit the colonization-extinction equilibrium model, Brown (1971, 1978) concluded that these small mammals reached the montane islands during the Pliestocene when boreal conditions were thought to have reached lower elevations in the Great Basin, and that there have been extinctions but no colonizations during the Holocene. Brown then considered factors that might affect the extinction rate. In addition to the correlation with island size, Brown showed that extinction is least likely in mammals with small body size, low trophic level, and high habitat diversity. The Great Basin islands show this very dramatically; species without these qualities occur only on the largest islands. The distribution of some species can be explained on more individual grounds. For example the Ruby Mountains, one of the larger ranges, does not support a pi¤on-juniper chipmunk (Eutamias dorsalis or E. panamintinus) probably because pi¤on-juniper woodland is poorly developed in that range (Brown, 1978).

Brown (1971, 1978) noted the conspicuous absence in the Great Basin of small mammals that are restricted to dense yellow pine forests of the surrounding mainland, in spite of the fact that suitable habitats are available. He listed 11 species in this category: Snowshoe Rabbit (Lepus americanus), Mountain Beaver (Aplodontia rufa), Alpine Chipmunk (Eutamias alpinus), Townsend's Chipmunk (E. townsendii), Lodgepole Chipmunk (E. speciosus), Red Squirrel (Tamiasciurus hudsonicus), Douglas' Squirrel (T. douglasi), Northern Flying Squirrel (Glaucomys sabrinus), Red-backed Mouse (Clethrionomys gapperi), Heather Vole (Phenacomys intermedius), and Pine Marten (Martes americana). Since this suite of species are not found living on any of the montane islands of the Great Basin, Brown concluded that their absence was due to never having colonized these islands rather than colonization and later extinction. He proposed that climatic conditions never allowed habitat bridges of this higher-elevation suite of plants to exist across the Great Basin. Although the mammals of this habitat are lacking, Brown noted that islands of yellow pine forest have many avian species characteristic of such habitats.

Brown did not consider the fossil record in any detail, but Grayson (1981, 1987) derived 3 implications from Brown's model that could be tested in the fossil record: 1) boreal mammals living on isolated montane islands must once have occupied the intervening lowlands, 2) boreal mammals found on only some of the mountains today should have been present on other ranges in the past, and 3) there may have been species of boreal mammals on isolated mountains in the past that are on none of the mountains today. As will be discussed subsequently, all of these implications have been confirmed by the fossil record. A fourth prediction is implied by Brown's observation that Great Basin ranges lack boreal species inhabiting only the highest elevations of the mainlands: 4) boreal mammals inhabiting only the yellow pine and higher life-zones in the Sierra Nevada and Rocky Mountains never inhabited the Great Basin. This prediction has been violated in several cases; Heather Vole, Pine Marten, and specimens thought to be Snowshoe Rabbit and Douglas' Squirrel have been recovered from Quaternary cave deposits in the Great Basin (Grayson, 1981; Heaton, 1985; Spiess, 1974).

Brown's explanation for the current distribution of boreal mammals in the Great Basin is very impressive, and the Quaternary fossil record has backed up almost all of its expectations. But the restricted subset of species and localities studied by Brown tells only a small part of a much greater story of the changes in mammals inhabiting the Great Basin during the Quaternary period, and it is this greater story that I want to present here. First, however, the changes in climate and resulting changes in vegetation in the Great Basin during the Quaternary period must be considered.

CLIMATE OF THE GREAT BASIN, PAST AND PRESENT

The great Wisconsin Ice Age, which ended about 10,000 years ago, has left all of the northern hemisphere in a state of transition. Geomorphic evidence is what first lead geologists to recognize the drastic changes that have taken place in the Great Basin. Numerous terraces attest to the former presence of sizable lakes in most of the intermontane valleys (Figure 3). The largest of these was Lake Bonneville which covered most of western Utah and adjacent parts of Nevada and Idaho and drained northward into the Snake River. Lake Lahontan and more than a hundred smaller lakes filled the closed basins of Nevada, Oregon, and California. Most of these lakes lacked outlets, but a series of rivers connected lakes in the southernmost Great Basin and carried water to Lake Manly in Death Valley (Smith and Street-Perrott, 1983). Glacial moraines in the mountain ranges show that valley glaciers existed on many of the highest peaks and came as low as 2800 m elevation (Porter et al., 1983), but they never covered enough area to have a strong direct impact on mammal ranges.

Climatic reconstruction of the late Pleistocene has been the subject of intense study. The mean annual temperature in the region decreases at about 6 C per 1000 m altitude and 0.8 C per degree of latitude (Porter et al., 1983). Pleistocene snow lines in the western United States reached about 1000 m lower than they do today, so this would suggest that the mean annual temperature during the late Pleistocene was about 6 C cooler than at present. But this assumes that the elevation-precipitation relation was also depressed 1000 m. At present mean annual precipitation increases 0.4 to 1.0 m per 1000 m increase in elevation, and depressing that gradient 1000 m would drastically increase total precipitation in the region. If precipitation has remained unchanged since the Pleistocene, then an increase in mean annual temperature of 10 to 15 C would be required to account for the 1000 m rise in snow lines (Porter et al., 1983). Changes in precipitation rates since the Pleistocene are very hard to calculate, and estimates vary widely. Smith and Street-Perrott (1983) studied the factors that lead to the development of pluvial lakes in the Great Basin during the Pleistocene, and they estimated that the streams that feed the basins must have carried 5 to 10 times more water than they do now. From this they estimated that Pleistocene precipitation reached 1-1/2 to 2 times its present value, and this translates to a depression in the elevation-precipitation relation of around 500 m. If this were the case, a temperature decrease of 8 to 10 C would account for the 1000 m depression of the snow line.

Plant fossils from late Pleistocene wood rat middens show enormous differences from the vegetation of today. The following patterns have been documented by Thompson and Mead (1982), Spaulding et al. (1983), and Wells (1983). The profusion of creosote-bush in the warm Mojave Desert of the southern Great Basin represents a Holocene expansion replacing pi¤on-juniper woodland in that region. The record also shows that pi¤on-juniper woodland was restricted to the Mojave Desert region during the late Wisconsin and has only recently expanded more than 500 km northward into the central and northern Great Basin where it now constitutes a major floral belt. Valleys of the central Great Basin, which are now occupied primarily by sagebrush and shadescale, were dominated by subalpine conifers. Bristlecone pine, Engelman spruce, and common juniper reached as low as 1660 m in the central Great Basin. Connecting divides higher than 1800m exist across the central Great Basin, so this allowed a continuous coniferous forest dominated by bristlecone pine. In southern Nevada bristlecone pine only reached as low as 1850 m, but the lower elevations were dominated by limber pine and Douglas- fir which expanded from the mountains of southern Utah. All of these species had ranges extending 500 to 1000 m lower than today. The presently dominant montane conifers of the middle-elevations (white fir, Douglas fir, and yellow pine) were limited to the southernmost part of the Great Basin during the last glacial. Brown's (1971, 1978) model for explaining the distribution of mammals in the Great Basin is complicated by the fact that neither of the two floral belts that he was contrasting (yellow pine and pi¤on-juniper) occurred in the main area he was considereng during the late Pleistocene.

Timing of the events outlined above has been worked out reasonably well by a large variety of methods. The Wisconsin Ice Age included a number of periods of glaciation interspersed with periods more like today, but earlier events have been obscured by later events. Based on studies of continental glaciation and sea level changes, the last pulse of glaciation (late Wisconsin) began at about 35,000 years ago and peaked at about 18,000 years ago (Bloom, 1983). The last rise and fall of Lake Bonneville followed a similar but slightly delayed time span. According to Scott et al. (1983) the lake started its last filling cycle prior to 26,000 years ago and was at its highest level from about 16,000 to 14,500 years ago at which time erosion at the outlet catastrophically lowered the lake level 100 m. About 13,000 years ago the lake evaporated below its outlet, and by 11,000 years ago a lake configuation similar to today was achieved (Currey, 1982; Scott et al., 1983). Lake Lahontan seems to have had a more complicated history. It was dry 36,000 years ago, then it reached its highest level at two different periods about 22,000 and 13,000 years ago (Benson, 1978; Davis, 1978). Plant fossil evidence indicates that the current desert conditions in the Great Basin were reached by 8000 years ago in the valleys (Brown, 1971; Wells, 1983).

Nearly all the Quaternary mammalian fossils from the Great Basin are younger than 30,000 years old (Webb, 1986) and are original bone preserved in dry caves, wood rat middens, and lake deposits. The mammalian record of the region prior to the late Wisconsin glacial surge is very poorly understood, so it will not be considered here. The floral and faunal records document the biotic changes that occurred as conditions became warmer and drier following the last glacial peak 18,000 years ago. In the rest of this paper I will consider the pathways by which this abiotic climatic shift altered the presence and distribution of mammals in the Great Basin.

CRYSTAL BALL CAVE: A CASE STUDY IN LATE QUATERNARY PALEOECOLOGY

My biggest contribution to the study of Quaternary mammals of the Great Basin was my studies at Crystal Ball Cave (Heaton, 1984, 1985) which lies in a hill in Snake Valley 10 km northeast of the Snake Range. This locality is well within the sagebrush floral belt although sagebrush is not a dominent plant on the hill itself, and there are no native trees in the area. Dry dust deposits deep within the cave contain abundant fossil bone, accumulated mainly by wood rats. Crystal Ball Cave has provided one of the largest late Pleistocene mammal faunas from the Great Basin, and the contrast between the fossil assemblage and the mammals now living in the area is very dramatic. The species recovered from Crystal Ball Cave can be arranged into four categories: 1) species that currently live in Snake Valley, 2) species that require boreal conditions, 3) species that require perennial water, and 4) extinct species. Each of these groups will be briefly discussed. Other Quaternary fossil assemblages in the Great Basin show the same general patterns as Crystal Ball Cave. But this site's location in a Great Basin valley, its proximity to Lake Bonneville, its singular mode of bone accumulation, and its abundant fossils (especially of small mammals) make it particularly good for documenting changes in mammal ranges over the last 20,000 years.

Extant species of Snake Valley

Crystal Ball Cave contains many species of mammals that are restricted to the Sonoran life-zone in which the cave is now located. Among these are Pygmy Rabbit, Black-tailed Jack Rabbit, Least Chipmunk, White-tailed Antelope Squirrel, Townsend's Ground Squirrel, Long-tailed Pocket Mouse, Dark Kangaroo Mouse, Canyon Mouse, Desert Wood Rat, Sagebrush Vole, Kit Fox, and Pronghorn. Most of these desert mammal species are outnumbered in the assemblage by more boreal ecological counterparts. Other local species of the Sonoran life-zone are conspicuous by their absence such as Desert Cottontail, Rock Squirrel, Great Basin Pocket Mouse, Little Pocket Mouse, Ord's Kangaroo Rat, Grasshopper Mouse, Gray Fox, and Spotted Skunk to name a few.

One drawback of Crystal Ball Cave is that its bone deposits are shallow and unstratified, and bone accumulation continues to occur. Fossil bone is therefore indistinguishable from recent bone. Four bones were C-14 dated at about 13,000, 18,600, 18,800, and over 23,000 years old, and these dates suggest that most of the accumulation occurred in the late Pleistocene. But the Sonoran life-zone species may represent recent additions to the cave deposits. The most notable thing about the assemblage is that most desert species are poorly represented or entirely absent. If bone deposition has occurred at a constant rate for the last 25,000 years and the cave has been in its present desert environment for the last 8,000 years, then species restricted to the Sonoran life-zone should make up a more sizable part of the assemblage than they do. A decrease in the rate of fossil deposition may have occurred at the end of the Pleistocene due to a decrease in local biomass and/or a change in the species of woodrat that was transporting bone into the cave. Another explanation is that boreal conditions remained in the area longer than currently believed. Wells (1983) reported that upward withdrawal of subalpine conifers was tardiest in the Snake Range, and Mead et al. (1982) reported bristlecone pine and common juniper associated with desert species in 7,000 year old middens from Streamview Rockshelter in the lower part of Smith Creek Canyon just 100 m higher than Crystal Ball Cave. Grayson (1985) found remains of mesic mammals at Hidden Cave (which no longer inhabit that area) in strata less than 1500 years old, but transport by birds and humans confuses the issue at that site.

A few desert species were too prominent in the Crystal Ball Cave fauna to be accounted for by recent invasion, and these need to be considered individually. By far the most abundant species from the cave was Townsend's Ground Squirrel (Spermophilus townsendii), and squirrels are not common in the area today. Of four species of ground squirrels currently living in the region (Hall, 1981) and possibly two others reported from the nearby Smith Creek Cave sediments (Mead et al., 1982), S. townsendii is the only species represented at Crystal Ball Cave. This high density is even more remarkable when one considers that ground squirrels are not cave- dwelling species. Yet there are twice as many bones of ground squirrels than of wood rats. When conditions are favorable S. townsendii can exist in incredibly high densities (Long, 1940; Smith and Johnson, 1985), and the area around Crystal Ball Cave must have presented such favorable circumstances at some time in the past. Durrant (1952) found that although S. townsendii is found throughout the Great Basin Desert, "it shows a marked increase in numbers in irrigated land and at desert springs." Snake Valley was much wetter in the past as will be discussed later. But this does not account for why a desert species like S. townsendii was favored over boreal ground squirrels that lived in the region. This preponderance of S. townsendii is not found at other fossil sites in the Great Basin, so it must represent a local anomaly.

The other well represented desert species in the Crystal Ball Cave assemblage is the Sagebrush Vole (Lagurus curtatus). According to Hall (1946) this species can be found at almost any elevation, but only in association with sagebrush (Artemisia and perhaps only A. tridentata). This rodent is about twice as abundant in the fauna as the similar but more montane species of voles (Microtus spp.), and this seems to violate the idea that the area was covered with subalpine conifers while fossils were accumulating. But the fact that Lagurus can survive even in isolated patches of sagebrush may provide an explanation. Gandy Mountain, the hill containing Crystal Ball Cave, is mostly rocky and has very limited and shallow soil deposits. Even with wetter conditions it is hard to imagine the growth of trees requiring deep roots. Perhaps this hill was therefore a refugium for sagebrush and other small woody plants during even the wettest periods of the Ice Age. This could account for the anomalous late Wisconsin abundance of certain desert species like Townsend's Ground Squirrel and Sagebrush Vole. It could also explain why no montane plant fragments were found in the cave. The plains surrounding Gandy Mountain have deep fertile soil which produces rich vegetation when irrigated, so subalpine conifers could easily have surrounded Gandy Mountain during the Ice Age and housed the rich boreal fauna that characterizes the Crystal Ball Cave assemblage. The scarcity of Townsend's Ground Squirrel and Sagebrush Vole in other late Pleistocene deposits supports this local explanation.

Boreal species in Crystal Ball Cave

Many fossils from Crystal Ball Cave are of extant species that are too boreal to live in Snake Valley at present. Included in this list of species are two small rodents that are restricted to the pi¤on-juniper floral belt: Cliff Chipmunk (Eutamias dorsalis) and Pi¤on Mouse (Peromyscus cf. truei). Although these are desert species, their absence from the sagebrush floral belt makes their presence in Snake Valley unexpected. Neither pi¤on nor juniper trees occur on or near Gandy Mountain today, and their presence in the lower elevations of the Snake Range is thought to be a recent invasion from southern Nevada (Wells, 1983). It is possible that pi¤on-juniper woodland extended closer to the cave at some time during the early Holocene, but the scarcity of these rodents in the cave fauna suggests that such woodland never dominated the area.

Many species in the Crystal Ball Cave assemblage occur no lower than the Transition Life-zone, so it seems very unlikely that they could have reached the cave in recent times. Some of these are still found at higher elevations in the nearby Snake Range: White-tailed Jack Rabbit (Lepus townsendii), Marmot (Marmota flaviventris), Bushy-tailed Wood Rat (Neotoma cinerea), Long-tailed Vole (Microtus cf. longicaudus), and Wapiti (cf. Cervus elaphus; extirpated in recent times). Others are extirpated from the Snake Range: Pika (Ochotona princeps), Snowshoe Rabbit (Lepus cf. americanus), Meadow Vole (Microtus cf. pennsylvanicus), Red Fox (Vulpes vulpes), and Pine Marten (Martes americana). Extinct species considered to be boreal will be discussed later.

Two main factors restrict boreal species from inhabiting deserts: 1) inability to survive under hot and/or dry conditions, and 2) lack of preferred food items. These limitations are most pronounced when there is direct competition for resources with species better adapted to desert conditions. Such competition seems to have been an important factor for three of the above species: White-tailed Jack Rabbit was replaced by Black-tailed Jack Rabbit, Bushy-tailed Wood Rat by Desert Wood Rat, and Red Fox by Gray Fox. Also Nuttall's Cottontail was replaced by Desert Cottontail as the dominant species in Snake Valley. Some of these replacements are very dramatic. Bushy-tailed Wood Rat is better represented in the assemblage than Desert Wood Rat by a ratio of at least 20 to 1, yet Desert Wood Rat is the only wood rat living in or near the cave today. Gray Fox was not found in the assemblage at all, yet it is now the dominent fox in the region. The same applies to Desert Cottontail. Other alpine species like Marmot and Pika are better represented as fossils than most of the desert mammals. All these replacements can be explained by the climatic and vegetational changes that have occurred in the Great Basin. When conditions were wetter and cooler and subalpine conifers covered much of the Great Basin valleys, boreal mammals extended their ranges and increased in numbers. When conditions became hotter and drier and conifers receded to higher elevations, desert mammals extended their ranges at the expense of boreal mammals. Crystal Ball Cave is located in a place where this replacement shows up very dramatically.

Species requiring perennial water

The Crystal Ball Cave assemblage includes bones of several kinds of fish and of two semiaquatic mammals: Muskrat (Ondatra zibethicus) and Mink (Mustela cf. vison). These animals live in a wide variety of life-zones, but only where there are perennial streams and/or lakes. There is a large warm spring at the south end of Gandy Mountain, but upon discovery it contained only small minnows. Muskrat and Mink do not presently occur in or near the Snake Range, but they do range into the northern Great Basin. Since present conditions cannot support these species, conditions must have been wetter in the past.

Lake Bonneville, when at its highest level, included a shallow arm that extended southward into Snake Valley (Figure 3), and during that time the lake was only 1.7 km east of and 195 m lower than Crystal Ball Cave. The lake was within 8 km of the cave for about half the time that fossils were accumulating. Washes near Gandy Mountain that drain the northern Snake Range may also have contained perennial streams during the Ice Age and provided suitable habitat for semiaquatic animals. These species are rare in Quaternary fossil assemblages of the Great Basin, so Crystal Ball Cave probably contains them only because of its close proximity to Lake Bonneville. The rarity of these fossils even at Crystal Ball Cave is probably because the cave is on a hill and therefore not in the immediate proximity of suitable habitat.

Extinct species from Crystal Ball Cave

Fossils of many extinct mammals were found in Crystal Ball Cave: Saber- toothed Cat (Smilodon), Horse or Zebra (Equus; at least two extinct species), Yesterday's Camel (Camelops), Large-headed Llama (Hemiauchenia), Woodland Muskox (cf. Symbos), and Short-faced Skunk (Brachyprotoma). These include the largest animals represented in the assemblage, and they are represented by only a few of their smallest bones since these are the only elements small enough for a wood rat to transport. Of these genera, Equus is the most abundant with bones of at least six individuals represented, and several new specimens have been recovered since my study. Camelops is the second most common. The two carnivores are the most poorly represented, probably because herbivores outnumber carnivores in life.

Crystal Ball Cave has a richer extinct fauna than most fossil assemblages of the Great Basin. This is probably due to both its age and size. Horse and camel fossils are quite common in cave deposits of the Great Basin, but the other genera are not. This was the first recovery of Short-faced Skunk from the western United States. It may turn out, however, that Short-faced Skunk material from other sites has been misidentified as the living Spotted Skunk.

LATE QUATERNARY EXTINCTIONS IN THE GREAT BASIN

The Quaternary paleontological record of the Great Basin has become quite extensive in recent years, and many important studies are currently in progress. Some of the more significant sites are shown in Figure 4 and listed in Table 5. Table 6 lists taxa found in sites where the most species were identified. The purpose of this paper is not to review all the work that has been done, but to discuss patterns of change and the factors that have caused those changes.

One important contribution of these late Pleistocene faunas has been the discovery of about 20 species of extinct mammals that formerly lived in the Great Basin (Table 7). Short-faced Bear (Arctodus simus), Mammoth (Mammuthus cf. columbi), Horse (Equus sp.), Camel (Camelops cf. hesternus), and Muskox (Symbos cavifrons, including Bootherium bombifrons as probable females) have been found in deposits of Lake Bonneville (Nelsen and Madsen, 1978, 1980, 1983). All these extinct species have been found in at least one cave fauna or at Tule Springs. These include the largest Pleistocene herbivores (mammoths, ground sloths, horses, peccaries, camels, and oxen) and carnivores (short-faced bears, saber tooth cats, lions, and cheetahs). The only smaller mammals to go extinct were Diminutive Pronghorn, Noble Marten, and Short-faced Skunk. None of these mammals (with the possible exception of Western Short-faced Skunk) were restricted to the Great Basin; they ranged over much of North America.

Since mainly large mammals were affected, it is interesting to note the status of extant large mammals. Of seven species extirpated from the Great Basin at the end of the Pleistocene, four were large: brown bears, wolverines, jaguars, and mountain goats. This is even more striking when one considers that small species far outnumber large species. Of the large mammals reported living in the Great Basin in recent times, most are very rare and possibly extirpated: beavers, wolves, black bears, otters, lynx, wapiti, and bison. Bighorn sheep have also become scarce. The only large mammals found today in any abundance in the Great Basin are coyotes, foxes, mountain lions, bobcats, and mule deer. Mountain lions feed primarily on deer while the other carnivores feed mainly on smaller mammals (Hall, 1946).

It is possible that the Great Basin is presently unable to support a community of very large mammals. The region is primarily a desert with only xeric vegetation. Mesic regions are limited to isolated mountain ranges that make up only about 5% of the land area, and these may not be sufficient to maintain many species of large mammals. This is in sharp contrast to the Late Pleistocene. Although the largest mammals such as mammoths, sloths, bison, and oxen never seem to have been abundant in the Great Basin, some large mammals are found in great numbers at most fossil sites, especially horses (several varieties), camels, llamas, and bighorn sheep. When the valleys, which make up 95% of the land area, had much more vegetation and water during the last glacial, the region could probably support a much greater mammalian biomass.

Extinctions due to competition

Some mammal extinctions can probably be attributed to direct competition. Short-faced Bear was the largest native bear of North America when Brown Bear immigrated from Asia during the last glacial. These two bears are of similar size and carnivorous habit, and they have been found associated at Labor-of- Love Cave in Nevada (Emslie and Czaplewski, 1985) and Little Box Elder Cave in Wyoming (Kurten and Anderson, 1974). This suggests the possibility that Brown Bear drove Short-faced Bear to extinction.

Noble Marten is the only marten found in fossil deposits throughout the Great Basin, although Pine Marten has been found in northern and eastern sites. During the late Pleistocene Pine Marten inhabited primarily the eastern part of the continent and Noble Marten the western part, and Pine Marten replaced Noble Marten in the west probably by competition (Kurten and Anderson, 1980). The only Holocene records of Noble Marten are from the Great Basin where it survived until about 3300 years ago, and it has been found both in boreal and desert associations (Grayson, 1987). The failure of Pine Marten to invade the central Great Basin where these late records are found may account for why Noble Marten survived longer there. Why Noble Marten survived the climatic changes of the Pleistocene-Holocene boundary only to go extinct in the late Holocene is still a mystery.

The extinct Short-faced Skunk is similar in size to the living Spotted Skunk, and competition between them may have promoted the extinction of Short-faced Skunk. Short-faced Skunk is only known from a few sites and is always associated with a boreal fauna. The Crystal Ball Cave skull is the only citing to date from the western United States, so the range of this rare skunk is uncertain. Kurten and Anderson (1980) discounted competition because the two species coexisted in the eastern United States. Short-faced Skunk was the only skunk found at Crystal Ball Cave, however, so competition may have been a factor in the Great Basin.

Competition may have been a factor in other extinctions as well. Bighorn Sheep may have promoted the extinction of Harrington's Mountain Goat as the amount of area above timberline decreased at the end of the Pleistocene. These two species coexisted for a long time at Smith Creek Cave, however, though Mountain Goat was much rarer at that low elevation. Because most of the large extinct mammals were not replaced by ecological equivalents, competition does not seem to have been an important factor in their extinctions.

Extinctions due to human overkill

One major difficulty with understanding extinct species is that we cannot study living populations to understand their exact dietary and ecological needs. These can only be inferred by their distributions and associations or by studying related extant species. Because of these difficulties, the cause of extinctions in the Great Basin is more poorly understood than the less- dramatic range shifts among extant species.

Martin (1984) has long proposed that the primary cause of the late Pleistocene megafauna extinctions was overkill by human hunters. He envisioned a sudden human invasion across the Bering Land Bridge about 12,000 years ago which encountered a megafauna that had no fear of Man. Mosimann and Martin (1975) postulated that the human population grew rapidly in this plentiful environment and swept across the continent driving the megafauna into a smaller and smaller area and finally to extinction. Evidence cited for this model includes the rapidity of the extinctions, the fact that mainly large species were affected, and the coincidence with abundant human archaeological sites.

While most workers allow human activities some role in the extinctions, Martin's theory that Man was the exclusive cause has not been well accepted. Evidence cited against the overkill hypothesis has included the similarity of its climatic backdrop to many pre-human extinctions (Webb, 1984) and the archaeological distinctness of long-term human cultures in different parts of North America (MacNeish, 1976). Space does not permit a further review of this controversy. My work in the Great Basin suggests that the extinctions can best be seen as one componant of a dramatic climate-induced restructuring of the mammalian community, so human intervention is an unnecessary explanation. Man was clearly contemporary with many now-extinct species of large mammals in the Great Basin, but no unequivocal evidence has been found to show any association between the two (Grayson, 1982). Because climatic changes were particularly severe in the Great Basin and because early Man does not seem to have been abundant there, I feel that any human role in the extinctions was very small.

EXTIRPATIONS AND RANGE REDUCTIONS

Unlike extinct species whose habitat and food preferences can only be inferred, the ecology and range of extant mammals are generally known. Range changes in these animals can therefore tell us much about changes in environment. Because current ranges and habits of living mammal species are well known, the limiting factor in making reconstructions is the fossil record because it consists of a limited number of sites with ambiguities in their age and mode of accumulation. Many mammalian species may have had more extensive ranges in the Great Basin during the Pleistocene than we are now aware of, and some fossil recoveries may not represent local occupation. In spite of these limitations, both current distributions and the fossil record conclusively document the extirpation of many Pleistocene mammals from the Great Basin and a significant reduction in the range of others.

Table 4 lists 16 species of small boreal mammals that are restricted to isolated mountaintops in the Great Basin but which have an extensive range in the Sierra Nevada and/or Rocky Mountains: Vagrant Shrew, Water Shrew, Pika, Nuttall's Cottontail, White-tailed Jack Rabbit, Yellow-pine Chipmunk, Cliff Chipmunk, Uinta Chipmunk, Marmot, Belding's Ground Squirrel, Golden-mantled Ground Squirrel, Bushy-tailed Wood Rat, Long-tailed Vole, Western Jumping Mouse, Ermine, and Wolverine. Heather Vole, Pine Martin, and probably Snowshoe Rabbit and Douglas' Squirrel can be added to this list based on fossil occurrences. Some mammalian species have a continuous range in the northern Great Basin but a discontinuous range farther south. Panamint Chipmunk, Northern Pocket Gopher, Montane Vole, Long-tailed Weasel, Mule Deer, and most of the species listed above have at least one isolated mountaintop population far south of their main ranges (Hall, 1946, 1981). Great Basin valleys become much hotter and drier from north to south, and the ranges of many mammals reflect this. Brown (1971, 1978) and Grayson (1981, 1987) argued that these discontinuous ranges could only be explained in terms of the reduction of a continuous late Pleistocene distribution over the Great Basin. Numerous extralimital recoveries from fossil deposits confirm this scenario, and many of these sites are located at low elevations and at latitudes far south of species' current ranges (Figure 4, Tables 5 and 6). In a few cases isolation has led to speciation, such as the evolution of Palmer's Chipmunk from Uinta Chipmunk discussed below, and this also lends strong support to the relectual interpretation.

The fossil record provides the most conclusive evidence that many species had much more extensive ranges during the late Pleistocene than at present. Some species now extirpated from the Great Basin have been found at fossil sites far from their current ranges: Snowshoe Rabbit (Lepus cf. americanus), Meadow Vole (Microtis cf. pennsylvanicus), and Pine Marten (Martes americana) from Crystal Ball Cave (Heaton, 1985); Douglas' Squirrel (Tamiasciurus ? douglasii) and Mountain Pocket Gopher (Thomomys monticola) from Bronco Charlie Cave (Spiess, 1974); Dusty-footed Wood Rat (Neotoma fuscipes) from Kokoweef Cave (R. E. Reynolds, unpub.); Heather Vole (Phenacomys cf. intermedius) from Gatecliff Shelter and Smith Creek Cave (Grayson, 1981; Thompson and Mead, 1982); Brown Bear (Ursus arctos) from Deer Creek, Hidden, and Labor-of-Love Caves (Ziegler, 1963; Grayson, 1985; Emslie and Czaplewski, 1985); Wolverine (Gulo luscus) from Snake Creek Burial Cave (Barker and Best, 1976); and Jaguar (Panthera onca) from Smith Creek Cave (Miller, 1979). Fossils also document the former presence of species that have been eliminated from the Great Basin in recent times: Black Bear (Ursus americanus; Labor-of-Love Cave; Emslie and Czaplewski, 1985), Gray Wolf (Canis lupus), Wapiti (Cervus elaphus), and Bison (Bison bison; Table 6). In addition to documenting range limits, the fossil record also shows the prehistoric relative abundances of species. Some species now rare and restricted to mountaintops were vastly more abundant during the late Pleistocene. This is particularly true of Pika and Marmot whose fossils are found in enormous numbers in many fossil deposits (Grayson, 1987; Mead, 1987; Zimina and Gerasimov, 1969).

Another group of mammals whose range has been greatly reduced is those species requiring perennial lakes and streams. In the Great Basin this group consists of Water Shrew, Beaver, Muskrat, Mink, and River Otter. All except river otter have been found in extralimital fossil deposits (Table 6). During the last glacial when the Great Basin region had more lakes and streams these species enjoyed a large range that included the Great Basin valleys. Now they are rare and limited to a northern distribution and local mountainous habitats where perennial water is available.

RANGE EXTENSIONS AND ECOLOGICAL REPLACEMENTS

While the extinct and extirpated large mammals were mere subtractions from the mammalian fauna of the Great Basin, some of the small mammals that were extirpated or had their ranges reduced were replaced by ecological counterparts that were better adapted to desert conditions. In some cases these xeric mammals probably promoted the range reduction of the more mesic species through competition. At Crystal Ball Cave I documented range extensions in four species, Desert Cottontail, Black-tailed Jack Rabbit, Desert Wood Rat, and Gray Fox, each at the expense of a closely-related boreal species with similar habits (as discussed above). Another possible example involves the two living species of Grasshopper Mouse (Onychomys). Interspecific competition is suggested by the fact that these two species have adjoining ranges with very little overlap (Hall, 1981). Carleton and Eshelman (1979) discovered that Great Basin subspecies of Northern Grasshopper Mouse (O. leucogaster) are very distinct from populations elsewhere. Riddle and Choate (1986) investigated this further and stated: "We hypothesize that O. leucogaster was a component of the fauna in a 'cold' desert refuge in the location of the present Mohave 'hot' desert during [full-glacial] time. This would place geographic isolation of populations of northern grasshopper mice in the Great Basin from the remainder of the species sometime prior to 20,000 years B.P., and would account for the existence at the close of the Pleistocene of size differences among populations of O. leucogaster comparable to those among extant populations." Southern Grasshopper Mouse (O. torridus) now inhabits the southern third of the Great Basin and may be extending its range northward. Northern Grasshopper Mouse has been recovered from Hidden Cave where the ranges of the two species now meet (Grayson, 1985), but there is presently no additional paleontological evidence to confirm this replacement.

Range extensions in many species can be attributed to an expansion of their preferred resources. Dry lakebeds in Great Basin valleys are only inhabited by small mammals that are well-adapted to hot, dry conditions. These include rodents such as antelope squirrels, pocket mice, kangaroo mice, kangaroo rats, and sagebrush voles. These have all been found as rare elements in Pleistocene faunas, but their relative abundances have increased as the habitats they prefer have enlarged. Some species have enlarged their ranges as the foods they rely on have become more abundant. Pygmy Rabbit and Sagebrush Vole eat only sagebrush and are only found living where this plant is present (Hall, 1946). Late Pleistocene fossils of these species have been found south of their present range where it is currently too hot for sagebrush (Tables 1 and 6), but these species are currently expanding their ranges eastward in the Great Basin coincident with the expansion of sagebrush (Green and Flinders, 1980). Some rodents require specific substrates. Long- tailed Pocket Mouse has a strong preference for stony ground, Dark Kangaroo Mouse prefers coarse soils, and Pale Kangaroo Mouse and Desert Kangaroo Rat are restricted to fine, wind-blown sand containing some plants (Hall, 1946). Long-tailed Pocket Mouse, Pale Kangaroo Mouse, Merriam's Kangaroo Rat, and Desert Kangaroo Rat all have similar northern extensions of their ranges into the alkali flats of the Great Basin that were occupied by lakes during the late Pleistocene (Hall, 1946), and these habitats represent new range that was not available to terrestrial mammals during the glacial epoch. Pronghorn is rare in fossil deposits in comparison with Bighorn Sheep and Mule Deer, but it is presently the dominent artiodactyl in Great Basin valleys due to its preference for open rangeland. This and other desert species must have been more restricted during the last glacial when lakes and coniferous forests covered the Great Basin valleys.

In spite of these examples, range extensions at the end of the Pleistocene are few and unimpressive in comparison with range reductions and extinctions. Clearly the Great Basin supports only a small fraction of the biomass that it supported during the late Pleistocene. This is documented by the reduction in the number of species and by the decrease in fossil deposition at sites such as Crystal Ball Cave.

EVOLUTIONARY TRENDS AND POSSIBILITIES

The Great Basin has been the site of much ecological reorganization and discontinuity during the Quaternary Period, and this has afforded many opportunities for evolutionary change. This paper covers a timeframe of only 30,000 years, however, and it is therefore not surprising that no major evolutionary transformations have taken place. Evolutionary modifications have been limited to changes in physiological tolerance, body size, and various minor characters. These short-term changes, which are well-documented because of their recency and because they involve living remnants, show patterns that may have been characteristic of such environments in past ages.

Mammalogists have documented several cases of physiological evolution in species with ecologic opportunity. Northern Pocket Gopher inhabits the northern latitudes and higher elevations of the Great Basin, and it competes with Southern Pocket Gopher which inhabits the southern latitudes and lower elevations. Only Southern Pocket Gopher was recovered from Crystal Ball Cave, and this is the only pocket gopher found living in the Snake Range, where it reaches exceptionally high elevations (Heaton, 1985). Hall (1946) proposed that only the Southern Pocket Gopher was in the area when the Snake Range became available for colonization, and in the absence of the northern species it adapted to the higher elevations (although not in as great of numbers as Northern Pocket Gopher has in other ranges). The same situation apparently exists in the Toquima Range (Hall, 1946; Grayson, 1987). In the Pilot Range, in the absence of a high elevation chipmunk such as Uinta Chipmunk, Cliff Chipmunk, usually restricted to the pi¤on-juniper floral belt, has extended its elevational range upward into the mixed conifers on the single peak (Brown 1978; Harper et al., 1978). Cases such as these clearly document that restrictions are normally imposed on species by congeneric competition. When these restrictions are lifted, physiological traits can evolve very rapidly to allow the invasion of new habitats.

Ecologic fluctuations have been mirrored by changes in body size in some species. Stokes and Condie (1961) described enormous individuals of Pleistocene Bighorn Sheep from lake deposits throughout the Great Basin. Because of their large size, great orbital width, wide rostrum, and massive horn cores, they were assigned to a distinct species from the living form (Ovis canadensis). Further investigation by Stock and Stokes (1969) and Harris and Mundel (1974) showed that this large form evolved over about 2000 years into the smaller modern form. Geist (1971, 1983) proposed that bighorn sheep and many other Pleistocene mammals, during the glacial period when there was an overabundance of forage, altered their morphology because of selection for males with generalized feeding structures but exceptional sexual structures (large body size and enlarged ornate characters) and high intelligence (larger brains). Geist sees Irish Elk, Moose, Muskoxen, Polar Bears, and Man as examples of this pattern. When the exceptional environmental conditions deteriorate, these species are forced to reduce their investment in ornate characters or be driven to extinction. There are other similar examples of size change documented outside the Great Basin. Wolverine underwent a gradual size increase during Rancholabrean times then got smaller again following the glacial epoch (Kurten and Anderson, 1980). The most famous case was the dwarfing (prior to extinction) of mammoths on Santa Rosa and San Miguel Islands off the California coast (Orr, 1956b; Kurten and Anderson, 1980). The numerous extinctions of giant mammals at the close of the Pleistocene can be seen as part of this same trend.

Speciation events

A number of new species appear to have arisen in the Great Basin during the Quaternary Period as a result of isolation. Little fossil data is currently available to document when these species originated (most of them are very localized), but their current distribution suggests their recency of origin. None of these evolutionary events represent major morphological change, but mammalogists consider them worthy of species recognition.

Mt. Lyell Shrew (Sorex lyelli) inhabits a small area in central California and westernmost Nevada, and the Great Basin separates it from its parent species, Masked Shrew (S. cinereus), which is found throughout northern North America (Hall, 1946, 1981). Boreal habitat bridges across the northwestern Great Basin during the Pleistocene may have allowed the invasion of this species into California. Inyo Shrew (S. tenellus) lives in a few isolated mountainous environments in the southwestern Great Basin, and it clearly evolved from Ornate Shrew (S. ornatus) of the Sierra Nevada and coastal ranges of California (Hall, 1946, 1981). The development then deterioration of boreal habitat bridges between mountain ranges in the Great Basin could easily have led to the dispersal then isolation of Ornate Shrew, and finally the origin of Inyo Shrew. Panamint Chipmunk (Eutamias panamintinus) has a similar distribution and probably developed in the same way from Yellow-pine Chipmunk (E. ameonus).

Palmer's Chipmunk (E. palmeri) is restricted to Charleston Peak in southern Nevada. Its parent species, Uinta Chipmunk (E. umbrinus), is found on mountain peaks across the central Great Basin with the closest population, a distinct subspecies, being restricted to the Sheep Mountains just 50 km northwest of Charleston Peak (Hall, 1946, 1981). The fact that isolated populations of this species complex become increasingly more distinct to the south can be attributed to habitat bridges being lost earlier in that region, and this fits the climatic model presented in this paper perfectly. The fact that a normal population of Uinta Chipmunk is lacking from Charleston Peak is evidence that Palmer's Chipmunk developed since the Wisconsin glacial. The development of Yellow-eared Pocket Mouse (Perognathus xanthonotus) and White-eared Pocket Mouse (P. alticola) from Great Basin Pocket Mouse (P. parvus) on mountain ranges in southern California follows an identical pattern (Hall, 1981). As thermal threshholds for boreal species migrated northward at the end of the Pleistocene, populations first became isolated on the most southern mountaintops and there have had the longest time period to diverge from their parent species.

Eldredge and Gould (1972) proposed that most speciation events occur rapidly in small, isolated, peripheral populations subjected to unusually high selection pressures. They contrasted this model with the prevailing view which holds that evolutionary change occurs by the transformation of entire species. They proposed this model as an alternative explanation to the characteristic lack of transitional fossil forms between species, formerly attributed to a highly imperfect fossil record. According to this model of "punctuated equilibria," large populations remain very stable (equilibria) and are replaced abruptly by new species (punctuations) which develop in small populations. The recent speciation events discussed above fit the punctuated equilibria model well. Evolutionary change has occurred rapidly in localized areas, and consequently none of these new species have yet been found in fossil or subfossil deposits. To fit the punctuated equilibria model completely these new species should become sympatric with their parent species and eventually replace them. This could potentially occur at a future date if habitat bridges are reestablished. Some widespread species may have originated in this manner during former glacial epochs. While the current scenario in the Great Basin provides a highly detailed configuration that fits the punctuated equilibria model, the timeframe is too short to follow the model through a complete cycle.

CAUSAL PATHWAYS OF CHANGE

All of the major climatic and ecological changes described in this paper can be attributed to temperature fluctuations caused by variations in the earth's orbit, the so-called Milankovitch Cycles (Hays et al., 1976; Imbrie and Imbrie, 1980; Kutzback, 1983). But while the ultimate cause of these changes seems to be simple, the pathways by which they occurred and interacted are not. Few mammals would have been affected appreciably by the temperature fluctuations alone, but the indirect effects were monumental. Some of the pathways of change were abiotic such as increased precipitation, the development of lakes and glaciers, and the consequential lowering of sea level. Other pathways were biotic such as mammalian exchanges between continents, changes in the abundance and types of vegetation, development and destruction of habitat bridges, and interspecific competition. Figures 5 and 6 outline in diagram form some of the major pathways by which mammals were affected by the onset and the termination of the last Ice Age.

Abiotic Changes

As discussed above, a gradual climatic cooling is what caused the Ice Age. Cooler air automatically leads to increased precipitation, although the exact function of this relationship is complex. A cooler climate also decreases the melting of snow and evaporation of lake water. These effects lead to the formation of continental glaciers in northern North America and valley glaciers on many mountain peaks during the late Pleistocene. A much more consequential effect in the Great Basin was the formation and expansion of numerous lakes, the result of increased precipitation and decreased evaporation in its numerous closed valleys. While these lakes consumed the range of many desert mammals that live on salt flats, they provided an huge increase in range for mammals that require perennial water. At the end of the ice age all of these trends were reversed. As the lakes and glaciers dried up, desert mammals expanded and semiaquatic mammals declined.

When voluminous water was tied up on land in the form of glaciers and lakes, sea level dropped, and the shallow Bering Strait became a wide land bridge. This allowed faunal exchanges between Asia and North America which greatly increased mammalian diversity, especially among large species because of their greater mobility. Many humans also entered North America at this time, and the magnitude of their unique impact is a matter of much controversy. Human hunting and overkill may have been partly responsible for the numerous megafaunal extinctions that occurred at the end of the Ice Age. When the Ice Age ended and sea level rose, the intercontinental land bridge was lost again.

Vegetational Biotic Changes

The Ice Age climatic changes, both in temperature and water availability, had an enormous effect on vegetation. The sparse xeric vegetation of the desert valleys in the Great Basin was replaced by subalpine conifers and other mesic vegetation. This constituted a huge net increase in vegetation and therefore in food available for herbivorous mammals. This lead to a dramatic increase in numbers and diversity of large mammals, and even to an increase in the body size of some species. This increase extended up the food web to carnivores as well. At the end of the Ice Age the vegetational biomass decreased again, and this led to a reduction in diversity of large mammals including many extinctions, extirpations, and range reductions.

Because of the elevational configuration of the Great Basin, the lowering of vegetational thresholds had a unique impact. Boreal conditions are now restricted to isolated mountain islands that small boreal mammals cannot disperse between. But during the Ice Age boreal conditions became continuous across the Great Basin which allowed for dispersal of boreal mammals between ranges. Now that these habitat bridges have disappeared, many species are isolated in mountaintop populations. Some of these small populations have died out on some mountain ranges but still live on others. A few have died out on all ranges in the Great Basin, and fossils are the only evidence that they once lived there. But isolation also provided evolutionary opportunity, and a few populations have diverged so much from their ancestral populations that they are considered new species. Such allopatric speciation resulting from climatic disruption may be a major mode of evolutionary diversification.

The diets of some species of mammals are restricted to particular plant foods, and the distributions of these species closely track the distributions of their favored plants. This occurs most often in desert life-zones that are composed mainly of one or two plant species. These favored plants, such as sagebrush and juniper, now cover vast areas of the Great Basin because they are well adapted to the hot valleys and lower mountain slopes. But during the Ice Age these life-zones were driven southward by the expanding boreal flora or were restricted to small habitat islands, and the mammals depending on these plants were reduced in number. When the climate became warmer and drier after the Ice Age, mesic vegetation could no longer survive in the valleys of the Great Basin, so the ranges of desert plants and and their associated mammals expanded. Range expansion of some such mammals has even been documented in historic times.

Competitional Biotic Changes

Competition between species leads to ecologic specialization, and many species could inhabit a much wider range of environments if those habitats were not occupied by other closely-related species. Many of the range shifts that occurred in the Great Basin during the Quaternary epoch were primarily the result of competition and only secondarily the result of climatic and vegetational changes. This can account for replacement of one species by another among rabbits, wood rats, and foxes for example. As the climate cooled, the mesic species were able to utilize the environment slightly better and thus outcompete the xeric species. As the climate warmed again, the reverse occurred. In a couple of cases the habitat islands of the Great Basin provided ecologic opportunity. On some peaks where a normal boreal competitor was absent, desert species of pocket gopher and chipmunk were able to invade the high elevations and adapt to the colder conditions.

During the Ice Age when vegetation was abundant, the northern hemisphere supported many large-bodied mammals. Continental exchanges across the Bering Land Bridge added greatly to this diversity. It appears that these species coexisted with ease during the Ice Age under favorable conditions and diverse habitats, and interspecific competition was at a minimum. Instead competition was mostly intraspecific, and this lead through sexual selection to the development of large body size and bizarre external features at high metabolic cost. When the Ice Age ended, many habitats deteriorated, and interspecific competition must have increased. With lush vegetation soon restricted to the mountain peaks, it appears that the Great Basin was no longer able to support a diverse population of large mammals. Only the large mammals best adapted to mountainous terrain have survived.

CONCLUSION

The Ice Age has left a portentous mark on the Great Basin. The current distribution of its mammals and many other features cannot be explained in terms of equilibrium conditions, but only by historic considerations. In the past 10,000 years, the broad valleys that comprise almost all of the land area have changed from lush savannas, forests, and lakes to dry, desolate flats. Many species that once dispersed throughout the region are now restricted to mountain peaks or northern latitudes, and a fauna of small desert mammals dominates the valleys. The species most detrimentally affected by the changes are those with large body size, those that prefer boreal conditions, and those that require perennial water. Smaller mammals were replaced with ecological counterparts, but larger ones were not. While the climatic changes at the end of the Pleistocene ultimately have a simple abiotic cause, the pathways by which the mammals were affected are very complex and are both biotic and abiotic. Changes in vegetation and interspecific competition were particularly important. The diverse and fluctuating climatic conditions provided exceptional ecologic and evolutionary opportunity in some cases, and it is possible that most evolutionary change takes place under stressful conditions such as these.

ACKNOWLEDGEMENTS

My work at Crystal Ball Cave was funded by Brigham Young University, the National Speleological Society, and Herbert H. Gersich (an early collector at the cave).

REFERENCES CITED

Adams, D. B. 1979. The cheetah: native American. Science 205, 1155-1158.

Aikens, C. M. 1970. Hogup Cave. Univ. Utah Anthrop. Pap. 93, 296 p.

Barker, M. S., Jr. and T. L. Best. 1976. The wolverine (Gulo luscus) in Nevada. Southwest Nat. 21(1), 133.

Benson, L. V. 1978. Fluctuation in the level of pluvial Lake Lahontan during the last 40,000 years. Quat. Res. 9, 300-318.

Bloom, A. L. 1983. Sea level and costal morphology of the United States through the late Wisconsin glacial maximum. In Late-Quaternary Environments of the United States: Volume I, The Late Pleistocene, Porter, S. C. (ed.). Univ. Minnosota Press, 215-229.

Brooks, R. H. 1956. Appendix III: faunal remains: in The Archaeology of Humbolt Cave, Churchill County, Nevada. Univ. Calif. Publ. Amer. Arch. Ethnol. 47(1), 106-112.

Brooks, R. H. 1963. Deer Creek Cave, unmodified bone material. In Deer Creek Cave, M. E. Shutler and R. Shutler, Jr. (eds.). Nevada State Mus. Anthrop. Pap. 11, 25-30

Brown, J. H. 1971. Mammals on mountaintops: nonequilibrium insular biogeography. Amer. Nat. 105(945), 467-478.

Brown, J. H. 1978. The theory of insular biogeography and the distribution of boreal birds and mammals. In Intermountain Biogeography: a Symposium, Harper, K. T. and J. L. Reveal (eds.). Great Basin Nat. Mem. 2, 209-227.

Carleton, M. D. and R. E. Eshelman. 1979. A synopsis of fossil grasshopper mice, genus Onychomys, and their relationships to Recent species. Pap. Paleont., Univ. Michigan Mus. Paleont. 21, 1-63.

Currey, D. R. 1982. Lake Bonneville: selected features of relevance to neotectonic analysis. U.S. Geol. Surv. Open-File Rept. 82-1070, 29 p.

Currey, D. R. and S. R. James. 1982. Paleoenvironments of the northeastern Great Basin and northeastern Basin Rim region: a review of geological and biological evidence. In Man and Environment in the Great Basin, Madsen, D. B. and J. F. O'Connell (eds.). Soc. Amer. Archaeol. Pap. 2, 27-52.

Davis, J. O. 1978. Quaternary Tephrochronology of the Lake Lahontan Area, Nevada and California. Nevada Archaeol. Surv. Res. Pap. 7.

Downs, T., H. Howard, T. Clements and G. A. Smith. 1959. Quaternary animals from Schuiling Cave in the Mojave Desert, California. Los Angeles Co. Mus. Contrib. Sci. 29, 21 p.

Durrant, S. D. 1952. Mammals of Utah: Taxonomy and Distribution, University of Kansas, Lawrence.

Durrant, S. D. 1970. Faunal remains as indicators of neothermal climates at Hogup Cave. In Hogup Cave, Aikens, C. M. (ed.). Univ. Utah Anthrop. Pap. 93, 241-245.

Durrant, S. D., M. R. Lee and R. M. Hansen. 1955. Additional records and extentions of known ranges of mammals from Utah. Univ. Kansas Mus. Nat. Hist. Pub. 9(2), 69-80.

Eldredge, N. and S. J. Gould. 1972. Punctuated Equilibria: an alternative to phyletic gradualism. In Models in Paleobiology, Schoph, T. J. M. (ed.). Freeman, Cooper and Co., San Francisco, 82-115.

Emslie, S. D. and N. J. Czaplewski. 1985. A new record of giant short-faced bear, Arctodus simus, from western North America with a re-evaluation of its paleobiology. Contrib. Sci. 371, 12 p.

Geist, V. 1971. Mountain Sheep: a Study in Behavior and Evolution. Univ. Chicago Press, 383 p.

Geist, V. 1983. On the evolution of Ice Age mammals and its significnce to an understanding of speciation. Amer. Soc. Biol. Bull. 30(3), 110-133.

Grayson, D. K. 1981. A mid-Holocene record of the heather vole, Phenacomys cf. intermedius, in the central Great Basin and its biogeographic significance. Jour. Mammal. 62, 115-121.

Grayson, D. K. 1982. Toward a history of Great Basin Mammals during the past 15,000 years. In Man and Environment in the Great Basin, Madsen, D. B. and J. F. O'Connell (eds.). Soc. Amer. Archaeol. Pap. 2, 82-101.

Grayson, D. K. 1983. Small mammals. In The Archaeology of Monitor Valley: 2, Gatecliff Shelter, Thomas, D. H. (ed.). Amer. Mus. Nat. Hist. Anthrop. Pap. 59(1), 99-126.

Grayson, D. K. 1985. The paleontology of Hidden Cave: birds and mammals. In The archaeology of Hidden Cave, Nevada, Thomas, D. H. (ed.). Anthrop. Pap. Amer. Mus. Nat. Hist. 61(1), 125-161.

Grayson, D. K. 1987. The biogeographic history of small mammals in the Great Basin: observations on the last 20,000 years. Jour. Mammal. 68(2), 359- 375.

Gruhn, R. 1961. The archaeology of Wilson Butte Cave, south-central Idaho. Idaho State Mus. Occas. Pap. 6.

Hall, E. R. 1941. Revision of the rodent genus Microdipodops. Zool. Ser. Field Mus. Nat. Hist. 27, 233-275.

Hall, E. R. 1946. Mammals of Nevada, University of California Press, Berkeley.

Hall, E. R. 1981. The Mammals of North America, 2nd Ed., Ronald Press, New York.

Harper, K. T., D. C. Freeman, W. K. Ostler and L. G. Klikoff. 1978. The flora of Great Basin Mountain Ranges: diversity, sources, and dispersal ecology. In Intermountain Biogeography: a Symposium, Harper, K. T. and J. L. Reveal (eds.). Great Basin Nat. Mem. 2, 81-103.

Harrington, M. R. 1933. Gypsum Cave, Nevada. Southwest Mus. Pap. 8, 1-197.

Harris, A. H. 1985. Late Pleistocene Vertebrate Paleoecology of the West, University of Tesax Press, Austin.

Harris, A. H. and P. Mundel. 1974. Size reduction in bighorn sheep (Ovis canadensis) at the close of the Pleistocene. Jour. Mammal. 55(3), 678-680.

Hays, , J. D., J. Imbrie, and N. J. Shackleton. 1976. Variations in the Earth's orbit: pacemaker of the ice ages. Science 194, 1121-1132.

Heaton, T. H. 1984. Preliminary report on the Quaternary vertebrate fossils from Crystal Ball Cave, Millard County, Utah. Curr. Res. Pleist. 1, 65-67.

Heaton, T. H. 1985. Quaternary paleontology and paleoecology of Crystal Ball Cave, Millard County, Utah: with emphasis on mammals and description of a new species of fossil skunk. Great Basin Nat. 45(3), 337-390.

Heaton, T. H. 1987. Initial investigation of vertebrate remains from Snake Creek Burial Cave, White Pine County, Nevada. Curr. Res. Pleist. 4, 107- 109.

Imbrie, J. and J. Z. Imbrie. 1980. Modeling the climatic response to orbital variations. Science 207, 943-953.

Jefferson, G. T. 1982. Late Pleistocene vertebrates from a Mormon Mountain cave in southern Nevada. Bull. South. Calif. Acad. Sci. 81(3), 121-127.

Kovach, S. D. 1981. Wolverine, Gulo gulo, records for the White Mountains, California. Calif. Fish Game 67, 132-133.

Kurt‚n, B. and E. Anderson. 1980. Pleistocene Mammals of North America, Columbia Univ. Press, New York.

Kurt‚n, B. and E. Anderson. 1984. Association of Ursus arctos and Arctodus simus (Mammalia: Ursidae) in the late Pleistocene of Wyoming. Breviora 426, 6 p.

Kutzback, J. E. 1983. Modeling of Holocene climates. In Late-Quaternary Environments of the United States: Volume 2, The Holocene, Wright, H. E., Jr. (ed.). Univ. Minnosota Press, 271-277.

Long, W. S. 1940. Notes on the life histories of some Utah mammals. Jour. Mammal. 21(2), 170-180.

Lundelius, E. L., Jr., R. W. Graham, E. Anderson, J. Guilday, J. A. Holman, D. W. Steadman, S. D. Webb. 1983. Terrestrial vertebrate faunas. In Late-Quaternary Environments of the United States: Volume I, The Late Pleistocene, Porter, S. C. (ed.). Univ. Minnosota Press, 311-353.

MacArthur, R. W. and E. O. Wilson. 1963. An equilibrium theory of insular biogeorgaphy. Evolution 17, 373-387.

MacArthur, R. W. and E. O. Wilson. 1967. The Theory of Island Biogeography, Princeton Univ. Press, Princeton.

MacNeish, R. S. 1976. Early Man in the New World. Amer. Sci. 64, 316-327.

Martin, P. S. 1984. Prehistoric overkill: the global model. In Quaternary Extinctions: a Prehistoric Revolution, Martin, P. S. and R. G. Klein (eds.). University of Arizona Press, Tucson, 354-403.

Mawby, J. E. 1967. Fossil vertebrates of the Tule Springs Site, Nevada. Nevada State Mus. Anthrop. Pap. 13, 105-128.

McGuire, K. R. 1980. Cave sites, faunal analysis, and big-game hunters of the Great Basin: a caution. Quat. Res. 14, 263-268.

Mead, J. I. 1987. Quaternary records of pika, Ochotona, in North America. Boreas 16, 165-171.

Mead, J. I. and E. M. Mead. 1985. A natural trap for Pleistocene animals in Snake Valley, eastern Nevada. Curr. Res. Pleist. 2, 105-106.

Mead, J. I., R. S. Thompson and T. R. Van Devender. 1982. Late Wisconsinan and Holocene fauna from Smith Creek Canyon, Snake Range, Nevada. Trans. San Diego Soc. Nat. Hist. 20(1), 1-26.

Mehringer, P. J., Jr. 1967. The environment of extinction of the late- Pleistocene megafauna in the arid southwestern United States. In Pleistocene Extinctions: the Search for a Cause, Martin, P. S. and H. E. Wright, Jr. (eds.). Yale Univ. Press, New Haven, 247-266.

Miller, S. J. 1979. The archaeological fauna of four sites in Smith Creek Canyon. In The Archaeology of Smith Creek Canyon, Eastern Nevada, D. R. Tuohy and D. L. Rendall (eds.). Nevada State Mus. Anthrop. Pap. 17, 271-329.

Mosimann, J. E. and P. S. Martin. 1975. Simulating overkill by paleoindians. Amer. Sci. 63, 304-313.

Nelson, M. E. and J. H. Madsen, Jr. 1978. Late Pleistocene musk oxen from Utah. Trans. Kansas Acad. Sci. 81(4), 277-295.

Nelson, M. E. and J. H. Madsen, Jr. 1979. The Hay-Romer debate: fifty years later. Univ. Wyoming Contrib. Geol. 18, 47-50.

Nelson, M. E. and J. H. Madsen, Jr. 1980. Paleoecology of the late Pleistocene, large mammal community in the northern Bonneville Basin, Utah. Geol. Soc. Amer. Abs. Prog. 12(6), 299.

Nelson, M. E. and J. H. Madsen, Jr. 1983. A giant short-faced bear (Arctodus simus) from the Pleistocene of northern Utah. Trans. Kansas Acad. Sci. 86(1), 1-9.

Orr, P. C. 1952. Preliminary excavations of Pershing County caves. Nevada State Mus. Dept. Archaeol. Bull. 1, 21 p.

Orr, P. C. 1956a. Pleistocene man in Fishbone Cave, Pershing County, Nevada. Nevada State Mus. Dept. Archaeol. Bull. 2, 20 p.

Orr, P. C. 1956b. Dwarf mammoths and man on Santa Rosa Island. Univ. Utah Anthrop. Pap. 26, 74-81.

Orr, P. C. 1969. Felis trumani, a new radiocarbon dated cat skull from Crypt Cave, Nevada. Bull. Santa Barbara Mus. Nat. Hist. Dept. Geol. 2, 1- 8.

Orr, P. C. 1972. The eighth Lake Lahontan (Nevada) expedition, 1957. Natl. Geog. Res. Rept. 1955-1960 Proj., 123-126.

Porter, S. C., K. L. Pierce and T. D. Hamilton. 1983. Late Wisconsin mountain glaciation in the western United States. In Late-Quaternary Environments of the United States: Volume I, The Late Pleistocene, Porter, S. C. (ed.). Univ. Minnosota Press, 71-111.

Reynolds, R. 1980. Kokoweef Cave faunal list. Unpub., 4 p.

Riddle, B. R. and J. R. Choate. 1986. Systematics and biogeography of Onychomys leucogaster in Western North America. Jour. Mammal. 67(2), 233-255.

Romer, A. S. 1928. A "fossil" camel recently living in Utah. Science 68, 19- 20.

Scott, W. E., W. D. McCoy, R. R. Shroba and M. Rubin. 1983. Reinterpretation of the exposed record of the last two cycles of Lake Bonneville, Western United States. Quat. Res. 20, 261-285.

Shutler, R. 1968. Tule Springs: its implications to early man studies in North America. Contrib. Anthrop. East. N.M. Univ. 1(4), 19-26.

Smith, G. W. and D. R. Johnson. 1985. Demography of a Townsend ground squirrel population in southwestern Idaho. Ecology 66(1), 171-178.

Smith, G. I. and F. A. Street-Perrott. 1983. Pluvial lakes in the western United States. In Late-Quaternary Environments of the United States: Volume I, The Late Pleistocene, Porter, S. C. (ed.). Univ. Minnosota Press, 190-212.

Spaulding, W. G., E. B. Leopold and T. R. Van Devender. 1983. Late Wisconsin paleoecology of the American Southwest. In Late-Quaternary Environments of the United States: Volume I, The Late Pleistocene, Porter, S. C. (ed.). Univ. Minnosota Press, 259-293.

Spaulding, W. G. 1985. Ice-age desert in the southern Great Basin. Curr. Res. Pleist. 2, 83-85.

Spiess, A. 1974. Faunal remains from Bronco Charlie Cave (26EK801), Elko County, Nevada. In The Prehistoric Human Ecology of Southern Ruby Valley, Nevada, by Casjens, L., pp. 452-486. Unpub. Ph.D. Dissert., Harvard Univ., Cambridge.

Spurr, J. E. 1903. Descriptive geology of Nevada south of the fortieth parallel and adjacent portions of California. U.S. Geol. Surv. Bull. 208, 229 p.

Stock, A. D. and W. L. Stokes. 1969. A re-evaluation of Pleistocene bighorn sheep from the Great Basin and their relationship to living members os the genus Ovis. Jour. Mammal. 50:805-807

Stock, C. 1931. Problems of antiquity presented in Gypsum Cave, Nevada. Scientific Monthly 32, 22-32.

Stokes, W. L. and K. C. Condie. 1961. Pleistocene bighorn sheep from the Great Basin. Jour. Paleont. 35(3), 598-609.

Stokes, W. L. and G. H. Hansen. 1937. Two Pleistocene Musk-Oxen from Utah. Utah Acad. Sci., Arts, Let. 14, 63-65.

Stokes, W. L., M. Anderson, J. H. Madsen, Jr. 1966. Fossil and sub-fossil bison of Utah and southern Idaho. Utah Acad. Proc. 43(2), 37-39.

Thompson, R. S. and J. I. Mead. 1982. Late Quaternary environments and biogeography of the Great Basin. Quat. Res. 17, 39-55.

Van Devender, T. R. and W. G. Spalding. 1979. Development of vegetation and climate in the southwestern United States. Science 204, 701-710.

Van Devender, T. R. and N. T. Tessman. 1975. Late Pleistocene snapping turtles (Chelydra serpentina) from southern Nevada. Copeia 1975, 249- 253.

Webb, R. H. 1986. Spatial and temporal distribution of radiocarbon ages on rodent middens from the southwestern United States. Radiocarbon 28, 1-8.

Webb, S. D. 1977. A history of savanna vertebrates in the New World. Ann. Rev. Ecol. Syst. 8, 355-380.

Webb, S. D. 1984. Ten million years of mammal extinctions in North America. In Quaternary Extinctions: a Prehistoric Revolution, Martin, P. S. and R. G. Klein (eds.). University of Arizona Press, Tucson, 189-210.

Wells, P. V. 1983. Paleobiogeography of montane islands in the Great Basin since the last glaciopluvial. Ecol. Mon. 53(4), 341-382.

Wells, P. V. and C. D. Jorgensen. 1964. Pleistocene wood rat middens and climatic change in Mohave Desert: a record of juniper woodlands. Science 143, 1171-1173.

Wells, P. V. and R. Berger. 1967. Late Pleistocene history of coniferous woodland in the Mohave Desert. Science 155, 1640-1647.

Ziegler, A. C. 1963. Unmodified mammal and bird remains from Deer Creek Cave, Elko County, Nevada. In Deer Creek Cave, M. E. Shutler and R. Shutler, Jr. (eds.). Nevada State Mus. Anthrop. Pap. 11, 15-24.


Timothy H. Heaton: E-mail, Home page, Phone (605) 677-6122, FAX (605) 677-6121