The Theory of Insular Biogeography and the Distribution of Boreal Birds and Mammals James H

Great Basin Naturalist Memoirs

Volume 2 Intermountain Biogeography: A Symposium Article 14

3-1-1978 The theory of insular biogeography and the distribution of boreal birds and mammals James H. Brown Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721

Follow this and additional works at: https://scholarsarchive.byu.edu/gbnm

Recommended Citation Brown, James H. (1978) "The theory of insular biogeography and the distribution of boreal birds and mammals," Great Basin Naturalist Memoirs: Vol. 2 , Article 14. Available at: https://scholarsarchive.byu.edu/gbnm/vol2/iss1/14

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist Memoirs by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. THE THEORY OF INSULAR BIOGEOGRAPHY AND THE DISTRIBUTION OF BOREAL BIRDS AND MAMMALS

James H. Brown 1

Abstract.—The present paper compares the distribution of boreal birds and mammals among the isolated mountain ranges of the Great Basin and relates those patterns to the developing theory of insular biogeography. The results indicate that the distribution of permanent resident bird species represents an approximate equilibrium between contemporary rates of colonization and extinction. A shallow slope of the species-area curve (Z =

0.165), no significant reduction in numbers of species as a function of insular isolation (distance to nearest continent), and a strong dependence of species diversity on habitat diversity all suggest that immigration rates of boreal birds are sufficiently high to maintain populations on almost all islands where there are appropriate habitats. In contrast, the insular faunas of boreal mammals represent relictual populations that receive no significant contemporary immigration. The insular mammal faunas have been derived by extinction from a set of species that colonized the islands when habitat bridges connected them to the continents in the late Pleistocene. A relatively steep species-area curve (Z = 0.326), no effect of isolation on species diversity, and the absence of appropriate species from large areas of apparently suitable habitat all support this conclusion. Measures of habitat diversity that are closely correlated with bird species diversity do not account for much of the variation in number of mammal species among islands. Insular area is the single variable that accounts for most of the variability in both bird and mammal species diversity; this supports the approach of using standard parameters such as area in comparative empirical analyses and general biogeographic theory. The results of this study suggest that extremes of vagility among taxa and a recent history of paleoclimatic and geological changes make it unlikely that equilibrial distributions, of the sort MacArthur and Wilson (1967) propose for the biotas of oceanic islands, are characteristic of the insular distributions of terrestial and freshwater vertebrates of western North America.

Biogeography is an old science that has draw quantitative relationships and derive made recent advances by assimilating new general principles. concepts and data. Within the last two dec- With the publication of their equilibrium ades information on continental drift, pa- model of insular distribution, MacArthur leoclimatology, and ancient sea levels has and Wilson (1963, 1967) contributed not revolutionized our understanding of histori- only a new theory but also a new approach cal events and their effects on plant and an- to biogeography. Prior to their work, most imal distribution. During the same period, biogeographic research had consisted of de- ecologists and evolutionary biologists have scribing the distributions of particular taxa learned much about the processes of popu- and producing ad hoc, historical explana- lation growth, dispursal, extincton, speci- tions. MacArthur and Wilson advocated a ation, and interspecific interactions which quantitative approach designed to build and are the mechanisms that determine distribu- test general models based on ecological pro- tion. Biogeography appears to be entering cesses. The specific model that they pro- an exciting new period in which the volu- posed suggests that the number of species minous data acquired by systematists and inhabiting an island represents an equilib- descriptive biogeographers can be inter- rium between opposing rates of extinction processes preted in terms of recently understood his- and colonization, and that these torical events and ecological processes to are functions of the size of an island and its

'Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Ariz

209 210 GREAT BASIN NATURALIST MEMOIRS No. 2

distance from a source of colonists, respec- 1946, Miller and Bussell 1956, Van Bos- tively. sem 1936; for mammals Durrant 1952, Dur-

MacArthur and Wilsons's choice of is- rant et al. 1955, Grinnell 1933, Hall 1946). lands for their revolutionary approach to The paleoclimatic history of the region also biogeography was not a fortuitous one. Ever is becoming increasingly well understood

since the pioneering work of Darwin and (e.g., Hubbs and Miller 1948, Martin and Wallace, islands have played a preeminent Mehringer 1965, Smith this symposium,

role in the development of the science. It is Wells and Berger 1967, Wells and Jorgen- difficult to do experiments in biogeography, sen 1964). This excellent data base has been but islands and insular habitats represent used in quantitative analyses of the insular natural experiments; they are small, replica- distributions of lacustrine fishes (Barbour ted systems among which species and envi- and Brown 1974) and montane birds (John- ronmental parameters are distributed in dif- son 1975) and mammals (Brown 1971). ferent combinations. Often particular taxa Much more work remains to be done. Some are distributed among insular habitats in of the most interesting contributions can be patterns that imply the operation of general expected when additional kinds of organ- mechanisms of dispersal, extinction, and in- isms and insular habitats are studied, so that

terspecific interaction. The last decade has it is possible to make comparisons among seen several attempts to test the model of taxa which differ in ecological requirements MacArthur and Wilson, using a variety of and dispersal abilities, and among habitats organisms that inhabit both true islands and that differ in environmental parameters and analogous isolated habitats (e.g., Barbour history of isolation. and Brown 1974, Brown 1971, Brown and The present paper discusses the distribu- Kodric-Brown in review, Culver 1970, Cul- tion of boreal birds and mammals among

ver et al. 1973, Diamond 1969, 1970a, isolated mountain ranges in the Great Basin

1971, Johnson 1975, Harper et al. this sym- in relation to the theory of insular bio- posium, Schoener 1974, Seifert 1975, Sim- geography. It attempts to relate distribu- berloff 1974, Simberloff and Wilson 1970, tional patterns to mechanisms of dispersal Simpson 1974, Terborgh 1973, Vuilleumier and extinction and to differentiate relictual

1970, 1973). Not all of these studies have patterns that are the legacy of historical supported the model, but the exceptions events from equilibrial ones that can be at- have contributed importantly to our under- tributed to contemporary ecological pro- standing of the patterns of insular distribu- cesses. The paper tries to develop a general tion and the historical events and ecological conceptual basis for analyzing and predict- processes that produce them. ing insular distributions. It first discusses the Analysis of the distribution of vertebrates current state of insular biogeographic theo- among isolated habitats in the Inter- ry and then utilizes empirical data on the mountain Begion of western North America distribution of boreal birds and mammals to has contributed significantly to the devel- test the theory and search for general mech- opment and testing of biogeographic theory. The dedicated field work of several generations of systematists has documented in de- The Theory of Insular Biogeography tail the species distributions of most verte-

brate groups (e.g., for fishes Miller 1948, Colonization, extinction, and speciation

Hubbs and Miller 1948, Hubbs et al. 1974, are the primary processes that determine Smith, this symposium; for birds Behle the composition of insular biotas. An island 1943, 1955, 1958, this symposium, Grinnell can be defined as a patch of suitable habitat and Miller 1944, Johnson 1965, 1970, 1973, surrounded by unfavorable environment that 1974, 1975, Linsdale 1936, Miller 1935, limits the dispersal of individuals. New spe- 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 211

cies colonize an island either by dispersing tion of island size. The model predicts: first, across the habitat barriers or by immigrat- that relatively constant insular species di- ing sometime when the barriers were tem- versity is maintained by the continual turn- porarily absent. Speciation entirely within over (extinction and colonization) of individ- an island potentially is another source of ual species; second, that the equilibrium new species, but there is no evidence that number of species is positively correlated this process has contributed significantly to with island size and negatively correlated the diversity of boreal birds and mammals with distance to the source of potential col- in the Great Basin region. Extinction, which onists; and third, that the equilibrium turn- may be caused by a variety of factors that over rate is inversely related to both island reduce population size, eliminates species size and distance to a source of species. Al- and reduces insular diversity. The composi- though the MacArthur-Wilson model was tion of an insular biota is determined by the designed specifically to account for the di- interaction of these origination and extinc- versity of organisms on oceanic islands, it tion processes. has proven a useful heuristic device for ana- MacArthur and Wilson (1963, 1967) de- lyzing many kinds of insular distributions veloped a simple, general model of insular because it deals with general processes and biogeography that represents the number of makes robust, testable predictions. species inhabiting an island as an equilib- As the MacArthur-Wilson model has been rium between contemporary rates of coloni- tested using a variety of taxa distributed zation and extinction (Fig. 1). Their model among both oceanic islands and several makes the colonization rate a decreasing kinds of insular habitats (see references function of distance to a source of dis- cited above), it has become increasingly persing species (usually the nearest conti- clear that in its present form it is in- nent) and extinction rate a decreasing func- adequate to account for many empirical ex-

\ COLONIZATION 212 GREAT BASIN NATURALIST MEMOIRS No. 2 amples. Two major problems have arisen in corporate this influence of immigration on attempts to use the model to account for extinction rate (Fig. 1). empirical patterns of animal and plant dis- Neither of these problems detracts from tribution. First, in some insular habitats, his- the utility of MacArthur and Wilson's mod- torical episodes of immigration, speciation, el as a heuristic device or the value of their and extinction have produced numbers of approach. Their work did much to stimu- species that are significantly greater or less late biogeographers to develop precise hy- than the equilibrial number expected on the potheses and test them with appropriately basis of contemporary rates of colonization analyzed quantitative data. Gradually a con- and extinction (Barbour and Brown 1974, ceptual understanding of insular bio- Brown 1971, Culver et al. 1973, Diamond geography is emerging that may not be as 1970). Such historically determined distribu- simple and elegant as the MacArthur- tions should be particularly common in or- Wilson model, but it is hoped it will be ganisms that are poor dispersers (such as the more realistic. In the following sections, I small, nonflying vertebrates) and in geo- hope to use the distribution of boreal birds graphic areas (such as western North Ameri- and mammals among isolated mountain ca) where paleoclimatic and geological ranges in the Great Basin to illustrate this changes have drastically altered the barriers theoretical approach and some of the result- that currently isolate insular habitats. Sec- ing concepts.

ond, it has proven difficult to test the Mac- Arthur-Wilson model's critical predictions Montane Islands about insular turnover. Although it is relatively easy to census accurately the biota of The mountain ranges of the Great Basin

an island, it is much more difficult to assess are islands of coniferous forest and associ- the natural turnover rate. As a result, most ated mesic habitats in a sea of desert. The of the purported measures of insular turn- Great Basin is a vast interior drainage that overs have been criticized (see Diamond lies between two montane continents, the 1969, 1971, Terborgh 1973, Lynch and central mountains of Utah (a part of the Johnson 1974, Simberloff 1974), and the Rocky Mountains) on the east and the critical predictions of how turnover rate Sierra Nevada on the west (Fig. 2). Most of varies with island size and isolation remain this area consists of broad arid valleys, untested. Recently, Brown and Kodric- which lie at an elevation of approximately Brown (1977) obtained empirical evidence 5000 feet and are sparsely covered with a that insular extinction rates may be strongly vegetation dominated by low woody shrubs dependent on the distance of an island from of the genera Artemisia, Chrijsothamntis, a source of colonists, rather than on island and Atriplex. Between the valleys are a size alone as MacArthur and Wilson sug- series of mountain ranges oriented in a gest. Areas that are sources of colonizing north-south direction. Many of these rise to species often also may be sources of immi- over 10,000 feet; on their lower slopes they grant individuals of species already present are covered with juniper-pinyon woodland on the island; the arrival of these immi- and at higher elevations there are forests of grants may refuce the probability that in- mixed conifers, stands of aspens, and some- sular populationd go extinct. We predicted times wet meadows and permanent streams. that, for many islands, species turnover rates The present analysis is based on 19 is- are lower for islands near sources of dis- lands for which the boreal mammal fauna is persing species than for more isolated ones. adequately known; lists of boreal bird spe-

Although this is the opposite of the pattern cies are available for 13 of these. Islands predicted by the MacArthur-Wilson model, were defined by operational criteria applied the model can easily be modified to in- to topographic maps (U.S. Geological Sur- 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 213 vey maps of the states: scale 1:500,000). A species and the presence of riparian wood- montane island was considered to be a land, wet meadow, and aquatic habitats (see mountain range that contains at least one Johnson 1975 for details). Since West (this peak higher than 9800 feet and is separated symposium) has shown that juniper-pinyon from other highland areas by a valley at woodland is absent or poorly developed in least 5 miles across below an elevation of the northern Great Basin, the islands se- 7500 feet. This altitude represents the ap- lected for this analysis lie south of the proximate lower boundary of juniper-pinyon Humboldt River and Great Salt Lake to in- woodland. For each island, area above 7500 sure that they have somewhat comparable feet elevation, distance to nearest continent habitats. (Sierra Nevada or central mountains of Utah), and elevation of highest peak were Boreal Birds and Mammals determined from topographic maps (Table

1). For 13 islands ornithologists (Johnson For the bird and mammal species that 1975, Behle, this symposium) have quan- are restricted to juniper-pinyon woodland tified the diversity of habitats available to and more mesic habitats of higher eleva- boreal birds. This habitat diversity score in- tions, the mountains of the Great Basin are corporates the number of coniferous tree truly islands. Their boreal avian and mam-

Table 1. Data for the boreal habitats used in the present analysis.

Great Basin Montane "Islands" 214 GREAT BASIN NATURALIST MEMOIRS No. 2

BIRDS 20 S= 2.526 A (

r =0.701

2 -

20 50 100 200 500 1,000

MAMMALS 20 0326 = S l 188 A

r =0.846

U. 5

20 50 100 200 500 1,000 AREA ABOVE 7,500 FEET ELEVATION (SO MILES)

Fig. 2. Map of the Great Basin region of western North America showing the location of the isolated mountain ranges used in the present analysis. Numbers refer to individual montane islands and continental sample areas listed in Table 1. Note that islands of varying size and isolation lie in the "sea" of desert habitat between the central mountains of Utah on the east and the Sierra Nevada of California on the west. 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 215

malian faunas are depauperate subsamples is similar to that in Brown (1971), but it

of the Sierra Nevada and Rocky Mountain differs in some details because I have used faunas. All of the bird and mammal species slightly different criteria for designating bo-

that inhabit the isolated peaks are broadly real species and I have included a few more distributed on one or (more frequently) both records of occurrence. The resulting lists of of the continental ranges to the east and boreal bird and mammals contain species of west (see appendix and references cited generally similar body sizes and habitat re- there), but a significant proportion of the quirements, although there are approx- continental species are not present on any imately 50 percent more mammal than bird particular island and some are absent from species on both the large islands and the

all islands. The continental mountain ranges continents. clearly are the source of the insular boreal faunas, and the distribution of species on Species—Area Relationships the islands is the result of their abilities to colonize and avoid subsequent extinction. The number of species (S) inhabiting an

The definition of boreal bird and mam- island usually is positively correlated with

mal species is somewhat arbitrary, and the insular area (A); this relationship takes the Z lists of knowledgable specialists often do not form S = CA , where the values of the

agree precisely (e.g., compare the lists of constant (C) and slope (Z) depend on the boreal bird species of Johnson 1975, and characteristics of the specific taxon and

Behle this symposium). I have followed group of islands under consideration (Mac- Johnson's rather conservative designation of Arthur and Wilson 1967, Preston 1962). The boreal bird species (for a list of these spe- slope of this relationship can indicate the

cies and their distribution see Appendix 1). relative importance of extinction and origi-

I have restricted my analysis to those spe- nation processes in determining the diver- cies that he terms (again conservatively) sity of insular biotas. (Barbour and Brown "permanent resident Boreal species," be- 1974, Brown 1971, Mac-Arthur and Wilson

cause I wanted to include in my com- 1967). Low slopes, Z<0.20, tend to charac- parison with mammals only those species terize samples of different areas within a that maintain sedentary populations continent, and the Z-value varies with envi- throughout the year. This eliminates from ronmental heterogeneity (Harner and Har- the analysis the large number of bird spe- per, 1977). Islands usually have higher Z- cies that breed on the montane islands but values; when the insular biota represents an are likely to migrate or disperse long dis- approximate equilibrium between rates of tances between successive breeding seasons. colonization and extinction, Z-values tend to

The definition of boreal mammal species in- lie in the range 0.20-0.35. When there is no cludes those that inhabit juniper-pinyon contemporary colonization to oppose extinc- woodland or habitats of higher elevation, tion, islands tend to have even higher but not the desert habitats of the Great Ba- Z-values, often > 0.40. The reason for this

sin. I have excluded from the analysis large pattern is straightforward. On continents carnivores and ungulates because their dis- small sample areas contain a significant pro- tributions have been drastically altered by portion of rare species. If these habitats human activity and their original ranges were isolated and no immigration were per- and habitat requirements are poorly known. mitted, rare species would go extinct rapid-

Also I have ignored bats, because their dis- ly on small islands and much more slowly tributions are incompletely documented and on large ones, producing a much steeper they, like the migratory birds, probably are species-area curve. Islands in equilibrium, not permanent residents of the islands. The where a significant rate of colonization op-

list of boreal mammal species (Appendix 2) poses extinction, represent an intermediate 216 GREAT BASIN NATURALIST MEMOIRS No. 2

situation; species that go extinct on small is- clusion can be tested using a somewhat dif- lands tend to be replaced by colonists, but ferent data set for mammals and comparing not at a sufficient rate to produce diversity the mammalian and avian distributions. If comparable to that on continents. There birds are better dispersers and are currently can be exceptions to this pattern. For ex- crossing the desert valleys to colonize the ample, if only a small subsample of the con- isolated peaks, then they should have a sig- tinental biota is able to colonize a group of nificantly lower Z-value than mammals. islands, the Z-value may be lower than ex- This is the case. The number of species of pected because large islands will not ac- both birds and mammals is correlated signif- quire as many species as they can support icantly with area (Fig. 3, Tables 2 and 3). (Barbour and Brown 1974). The Z-value for birds, 0.165, is even lower This conceptual framework can be used than that obtained for most insular biotas to compare the species-area curves for the that are presumed to be in equilibrium and boreal birds and mammals. In my earlier approximates values for continental samples. analysis (Brown 1971), I argued that boreal The Z-value for mammals, 0.326, is less mammals reached all of the islands during than I obtained in my earlier analysis, Z = periods of climatic change in the late 0.428, and lies in the upper range of those Pleistocene; since then there have been ex- observed for the biotas of true islands (Mac- tinctions but no colonizations. This con- Arthur and Wilson 1967). Thus species-area

species (above) and Fig. 3. The relationship of insular area to the number of permanent resident boreal bird the number of small boreal mammal species (below). Note that the slope of the least squares regression line of fitted regres- the species-area curve for birds is much less steep than that for mammals. The equations for the

sions and the correlation coefficients (r) are indicated. 1978 INTERMOUNTAIN HKK.l'ni.HM'HY: A SYMPOSIUM 217

curves for the boreal birds and mammals nearest continent (Table 2, and see next sec- conform qualitatively to theoretical predic- tion) suggest that colonization rates are high tions, but they differ slightly from the quan- and there is little if any effect of insular titative Z-values that might be expected if isolation. This is consistent with Johnson's the avian distribution represents an equilib- (1975) conclusion that the diversity of bo- rium between contemporary colonization real birds is attributable primarily to habi- and extinction, and the insular mammalian tat; the impoverishment of the insular avi- populations are Pleistocene relicts that do faunas is the result of reduced habitat not disperse across the desert valleys. diversity on the isolated peaks and not to These slight deviations from expected Z- any significant extent to low colonization values are not difficult to explain. The low rates. It should be noted that it is not nec- Z-value for birds suggests that the isolation essary to infer from this that the desert val- of the montane islands may not be a signifi- leys do not inhibit dispersal, only that colo- cant barrier to avian colonization. The facts nization rates remain sufficiently high that that a) approximately 80 percent of the var- boreal species which maintain breeding iation in insular bird species diversity can populations on the islands rarely go extinct be accounted for by the combined effects of and, if they do, they recolonize rapidly. area and elevation (Table 2) or by habitat The fact that the Z-value for mammals diversity (Table 3), and b) there appears to also is slightly lower than predicted has a be no impoverishment of species numbers different explanation. As I shall show in the resulting from isolation by distance to the next section, the mammals that have colo-

Table 2. Stepwise multi birds and mammals inhabit 218 GREAT BASIN NATURALIST MEMOIRS No. 2 nized the montane islands are those species species and distance to the nearest conti- of the continental fauna that occur at rela- nent (Tables 2 and 3). For mammals, this tively low elevations within the boreal result is consistent with my 1971 analysis zone; species restricted to high elevations and conclusion that there is virtually no are absent from all of the isolated mountain contemporary immigration to the isolated ranges. The large islands, which also tend to mountains because the desert valleys con- have the highest peaks (Table 3), have ex- stitute almost absolute barriers to dispersal. tensive areas of mixed coniferous forest, wet The islands were colonized in the Pleisto- meadow, and other high altitude habitats. cene when periodic climate changes re- Even with post-Pleistocene extinctions these sulted in shifts in the altitudinal limits of mountains would be expected to support the boreal vegetation. Data from plant some of the mammal species characteristic macrofossils in woodrat middens suggest of these habitats on the continental ranges. that, as recently as 8,000 to 12,000 years Thus, if the islands had an unbiased sample ago, periods of cooler, wetter climate en- of the boreal mammal fauna, the larger is- abled juniper-pinyon woodland to flourish lands should have more species than they at least 2,000 feet (600 m) below its present do and the species-area curve would be lower elevational limit (Wells and Berger steeper. In my 1971 paper, I reported a 1967, Wells and Jorgensen 1964). This was higher Z-value than obtained here. The rea- sufficient to make juniper-pinyon and asso- son is that I included in the present analysis ciated meadow and riparian habitats con- some marginally boreal species (Eutamias tiguous across virtually all of the Great Ba- dorsalis, E. panamintinus, and Sylvilagus sin, and to enable the boreal mammals nuttalli) that are characteristic of juniper- characteristic of these habitats to colonize pinyon woodland and are present on most all of the isolated mountain ranges. With of the montane islands. This had the effect the return of hotter, drier conditions, these of adding an approximately constant num- "habitat bridges" connecting the islands to ber of species to the fauna of all islands, the continental ranges were eliminated and and thus lowering the slope of the exponen- the insular mammalian faunas have been de- tial species-area relationship. This is the rived from the widespread Pleistocene fauna same effect that is produced by the dis- by independent extinctions on each island. proportionate representation of low eleva- Several lines of evidence are consistent tion species in the insular faunas relative to with this interpretation. 1) The relatively continental faunas because of differential steep species-area curve (Fig. 3) suggests colonization in the past. that extinction has played a major role in

determining mammalian diversity. This is Isolation and Paleoclimatic History supported by the discovery of late Pleisto- cene fossils of at least one boreal species

It is a common observation of insular (Marmota flaviventri.s) from an island biogeography that remote islands support (Spring Range) where it no longer occurs fewer species than islands of comparable (Wells and Jorgensen 1964). 2) The lack of size and habitat diversity that are nearer to any correlation between number of mammal a continent (see MacArthur and Wilson species and distance to nearest mainland

1967 and included references). This pattern (Tables 2 and 3) or any other likely measure is attributed to the limited ability of organ- of insular isolation (Brown 1971) suggests isms to disperse so that the rate of immigra- that there is no contemporary immigration tion to an island declines with increasing to the islands. 3) All of the species isolation. In the present analysis, neither known from the islands are found in birds nor mammals demonstrate such a juniper-pinyon or other habitats of com- negative relationship between number of parable elevation. It is possible to account 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 219

for the presence of the entire insular fauna bird species and distance to nearest conti-

in terms of colonization across habitat nent (Table 3), and the distance variable bridges that were known to have existed in does not contribute significantly to account- the late Pleistocene. 4) All of the species of ing for number of bird species using mul-

boreal mammals that are restricted to mixed tiple regression analysis (Table 2). It re- coniferous forest or habitats of higher eleva- mains to explain the difference in species tion on the continental ranges are absent diversity between the continents and islands. from all of the montane islands, even Some of those species present on the conti- though large areas of apparently suitable nents but absent from most of the islands habitat are present on some of the larger is- may be limited in their distributions by lands. These species include Maries ameri- their sedentary natures or their aversion to cana, Aplodontia rufa, Eutamias alpinus, E. crossing inhospitable desert terrain. How- toivnsendi, E. speciosus, Tamiasciurus hud- ever, many of these species such as Lagopus sonicus, T. douglasi, Glaucomys sabrimis, leucurus, Dryocopus pileatus, and Perisoreus Phenacomys intermedins, Clethrionomys canadensis, have specialized habitat require- gapperi and Lepus americanus. Paleobota- ments and low population densities. Their nies evidence indicates that the habitats of habitats are totally lacking from many of these species were not connected across the the islands, and, even where they are pres- Great Basin during the late Pleistocene. ent, they often consist of small patches ob- This is further evidence that small terrestri- viously inadequate to support sustained al mammals usually are unable to cross hab- populations. Two other sources of evidence itat barriers only a few miles in extent and suggest that rates of contemporary immigra-

thus colonize isolated habitats only when tion by most boreal bird species are suffi- bridges of appropriate habitat provide di- ciently high to keep the habitats present on rect access (Brown 1971, 1973). the islands filled with an appropriate com- The explanation for the lack of correla- plement of species. First, such boreal bird tion between number of species and dis- species as Picoides tridactylus and Cyano- tance to nearest continent in birds appears sitta stelleri, which are restricted to high to be the opposite of that in mammals: elevation, well-developed coniferous forests, birds are such good dispersers that they are present on the islands. Since the habi- have colonized virtually all islands with tats of these species have not been con- suitable habitat regardless of their isolation. nected by bridges to the continental ranges

Both Johnson (1975) and Behle (this sym- during the Pleistocene, it must be inferred posium) report a negative relationship be- that at least these species are able to colo- tween their measure of insular isolation (cu- nize across significant barriers of unsuitable mulative width of desert barriers) and habitat. Second, vagrant individuals of some number of permanent resident bird species, boreal species (e.g., Cyanositta stelleri and but both authors included continental Cinclus rnexicanas) infrequently are report- sample areas in their analyses. The apparent ed significant distances from breeding popu- effect of isolation in their analyses can be lations (Johnson 1975, B. Bundick, pers attributed largely to the fact that a signifi- comm.). This evidence indicates that even cantly greater number of species inhabit relatively sedentary permanent resident bo- continental sites than occur on the islands; real birds are much more vagile than small there is little or no effect of isolation by boreal mammals. distance when only islands are considered Determinants of Species Diversity and (see Fig. 2 of Johnson 1975). This is con- Composition sistent with my own analysis of Johnson's

and Behle's data. There is an insignificant Differences in dispersal account in large negative correlation between number of part for differences in the species-area rela- 220 GREAT BASIN NATURALIST MEMOIRS No. 2 tionships between birds and mammals and may be the best parameter for constructing in faunal composition between insular and general theory that can be applied to di- continental mountain ranges (at least for verse taxa and various kinds of islands. In mammals), but the considerable variation in the development of theory some sacrifice of species diversity among the insular bird and precise explanation of specific cases usually mammal faunas must be attributed primari- must be made in order to obtain a desirable ly to ecological characteristics of the is- degree of generality. It is questionable lands. Although a common set of both bird whether it would be practical or profitable and mammal species has had the opportu- to base biogeographic theory on organism- nity to colonize virtually all islands, the iso- specific parameters such as habitat diversity, lated mountain ranges support different food availability, or carrying capacity. Cer- numbers and kinds of species. Species diver- tainly at present we have no accepted, stan- sity of both birds and mammals is closely dardized techniques for measuring these correlated with insular area, but area is variables in the field that are suitable for a only a correlate of factors such as the quan- variety of taxa and habitats. tity and diversity of habitat and food re- Some test of the merit of this approach sources that ultimately determine species di- can be made by comparing the correlates of versity. Power (1972) and Johnson (1975) species diversity for boreal birds and mam- have shown for birds that someone familiar mals. Johnson (1975) developed a quan- with a group of organisms can devise quantitative "habitat diversity score" that ac- titative measures of habitat diversity that counted for most of the variability in the account for significantly more of the vari- number of boreal bird species in his analy- ance in insular species diversity than area. sis. This appears to measure accurately the This approach has great utility for elucidat- requirements of boreal birds in western ing the environmental factors that influence North America, because Behle (this sym- the diversity and distribution of particular posium) tested it and obtained gratifyingly taxa, but it may not be useful for devel- similar results. Since boreal mammals utilize oping general biogeographic or ecological the same forest, meadow, and freshwater theory. habitats as boreal birds, it would be encour-

There are two arguments in favor of bas- aging if Johnson's habitat diversity score ac- ing biogeographic theory on simple, stan- counted for similarly large proportions of dard parameters such as insular area. The the variability in insular species diversity first is practical. Parameters such as area, for both taxa. Unfortunately this is not the elevation (another correlate of habitat diver- case (Table 2). The number of boreal bird sity), and distance to the nearest continent and mammal species inhabiting the same is-

2 are easy to determine from maps. By ob- lands are not very closely correlated (r < taining these measurements and species lists 0.37); whereas 81 percent of the variability from the literature, it is possible to describe in bird species diversity can be attributed to clear patterns of plant and animal distribu- habitat diversity score, the comparable fig- tion without doing all the original fieldwork ure for mammals is only 29 percent. Area is required to quantify accurately more direct by far the best correlate of species diversity environmental variables. Much of the recent for both birds and mammals. When the ap- synthetic work in insular biogeography has propriate log-transformed data are used, been done this way, and even those authors area accounts for 49 and 79 percent of the that have attempted to quantify variables variability in bird and mammal species such as habitat diversity often have used numbers, respectively (Table 3). Inter- data available from the literature or topo- estingly, elevation of highest peak (another graphic maps (e.g., Power 1972, Johnson correlate of habitat diversity readily obtain-

1975). The second argument is that area able from topographic maps) and area taken 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 221

together in multiple regression analysis of particular taxa can contribute to theory (Table 2) account for almost as much varia- by providing information on the mecha-

2 bility in bird species diversity (r = 0.76) nisms and effects of the underlying process-

2 as the habitat diversity score (r = 0.81). es of colonization and extinction. The great Two explanations can be offered for the importance of habitat in determining insular lack of good correlation between Johnson's bird species diversity, coupled with a low habitat diversity score and mammalian spe- slope of the species-area curve and the lack cies diversity. First, the habitat diversity of any significant effect of isolation by dis- score is highly dependent on the species di- tance, suggests that boreal birds are able to versity of coniferous trees. Particular tree disperse at a sufficient rate to maintain species and forest types probably are impor- populations on most islands with suitable tant to the boreal birds, but they are much habitats. The distribution of boreal mam- less likely to influence the diversity of bo- mals presents a dramatic contrast. Mammal real mammals. As I have shown, boreal species are absent from many islands that mammals that depend on the well-developed have suitable habitats, either because they coniferous forests of high elevations have have never colonized or because they have been unable to colonize any of the islands, gone extinct since the last episode of and those mammals that have colonized, if Pleistocene colonization. The interacting ef- they require conifers at all, are primarily fects of habitat requirements and paleocli- species of the juniper-pinyon woodlands matic changes in determining the set of spe- that are well developed on almost all of the cies that colonized the islands already have islands. Second, Johnson's habitat diversity been discussed. Of equal interest are the score is based primarily on the presence or factors that have resulted in extinctions absence of particular tree species or habitat among the original colonists to produce the types. This qualitative approach appears to present insular faunas. work well for birds, because they have suffi- The mammalian faunas of the montane is- ciently high immigration rates to maintain lands have been derived by extinction from populations on almost all islands where suit- a common set of 14 functional species that able habitats are present. In contrast, mam- were widely distributed across the Great mals, which have had to maintain insular Basin during the late Pleistocene when their populations for thousands of years in the habitats were connected by bridges to the absence of any significant immigration, may continental mountain ranges (Table 4). After be much more dependent on the quantity of the islands were isolated by the contraction boreal habitats (particularly those character- of boreal habitats to approximately their istic of low elevations that were connected present position, extinctions reduced the in the Pleistocene) which may be more faunas of each island and thus played a ma- closely correlated with insular area than jor role in determining species composition. with Johnson's more qualitative index. Five large islands have retained at least 10 Whether one or both of these nonexclusive of their original 14 species, but five small explanations turn out to be correct, it is islands have lost all but 3 or 4 of the origi- clear that different factors determine the di- nal set of species in the period of approx- versity of insular species of birds and mam- imately 10,000 years that the boreal habi- mals. These results suggest that parameters tats have been isolated. The distribution of such as area, even though their effects are extinctions among species is highly non- indirect and imprecise, may be the variables random and appears to be related primarily best suited for constructing and testing gen- to population size, as MacArthur and Wil- eral biogeographic theory. son (1967) predicted. Herbivorous species of However, specific factors that influence generalized habitat requirements and small the insular species diversity and composition to intermediate body size have persisted on 222 GREAT BASIN NATURALIST MEMOIRS No. 2

most of the islands (Table 4). In contrast, veloped. Similarly, the chipmunk character- herbivores of large body size and/or spe- istic of higher elevations (Eutamias um- cialized habitat requirements and carnivores brinus) is absent from the Pilot Range, have had higher extinction rates and persist which has only small stands of mixed con- on only a small proportion of the 19 islands. ifers on its single tall peak. On this moun- The frequency of occurrence of boreal spe- tain, in the absence of E. umbrinus, E. dor- cies on the islands (Table 4) corresponds salis has extended its altitudinal range very closely to the relative abundance of upward into the mixed conifers on the these species where they occur together on peak. large islands or continental mountain ranges Conclusions, Predictions, and (personal observations). The dependence of Unanswered Questions extinction on population size probably is

even more precise than is apparent here, The insular distributions of boreal birds because population size is influenced by the and mammals in the Great Basin differ in presence of particular habitats and com- ways that are consistent with both current peting species which vary among islands. biogeographic theory and independent evi- For example, one of the few mountain dence of ecological and historical factors ranges where a juniper-pinyon chipmunk that have affected colonization and extinc-

(Eutamias dorsalis or E. panamintinus) does tion. All data seem consistent with the in- not occur is the Ruby Mountains, where terpretation that insular bird populations juniper-pinyon woodland is very poorly de- represent an equilibrium between contem-

Table 4. Characteristics of the boreal mammal species that inhabit the montane islands of the Great Basin. Species are listed in decreasing order of frequency of occurrence. 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 223 porary rates of colonization and extinction; taxa are unlikely to cross habitat barriers of immigration rates are sufficiently high that even modest extent, and it is well known boreal bird populations inhabit almost all is- that fishes can colonize new areas only lands where there are suitable habitats, re- when suitable habitat bridges are present gardless of their isolation from continents. (Barbour and Brown 1974, Hubbs and Mill-

In contrast, the boreal mammal populations er 1948, Hubbs et al. 1974, Smith, this sym- represent relicts of species that were wide- posium). The distributions of these taxa spread in the Great Basin during the late should be extremely sensitive to paleocli- Pleistocene when habitat bridges connected matic and geological events; where insular the islands and continents; the mammal habitats have been connected and reisolated faunas of the islands have been derived they should show relictual distributions in from a common set of Pleistocene colonists which the identity of the colonists and the by subsequent extinctions in the absence of effects of subsequent extinctions produce immigration. patterns of diversity comparable to those of

Neither the avian nor the mammalian dis- the small boreal mammals. It is interesting tributions fit the equilibrium model pro- to speculate that the extremes of vagility posed by MacArthur and Wilson (1967) to which characterize most of the vertebrates account for diversity on oceanic islands. and the extensive climatic and geological The insular mammal faunas clearly are not changes that have drastically changed the in equilibrium; in the absence of immigra- landscape of western North America within tion they are gradually relaxing toward an the last million years make it unlikely that equilibrium of zero species at rates in- land and fresh water vertebrates in this re- versely related to island size. The distribu- gion will show the sort of equilibrium distri- tion of birds does represent a sort of equi- butions that have been demonstrated for the librium between contemporary colonization biotas of oceanic islands with a long history and extinction, but the immigration rates of isolation and relative environmental sta- are so high that the islands are virtually sat- bility (see MacArthur and Wilson 1967, urated with species for which the appropri- Simberloff 1974). ate habitats are present. The avifaunas of In insular biogeography many important the islands differ from the MacArthur-Wil- questions remain to be answered, and much son model in that there is no significant ef- theoretical and empirical work is yet to be fect of isolation by distance, species compo- done before this promising and vigorous sition is quite precisely determined by young science can afford to rest on its lau- habitat, and rates of faunal turnover prob- rels. The vertebrates that are distributed ably are very low (see Brown and Kodric- among the numerous isolated habitats in Brown 1977). western North America continue to offer The information available on the distribu- great potential as systems for testing theory tion of other vertebrates among insular hab- and developing general concepts. For those itats suggests that the kinds of patterns de- organisms that appear to demonstrate equi- scribed here for birds and mammals may be librial distributions it is particularly impor- widespread. Migratory birds, bats, and very tant to measure turnover rates accurately, large carnivorous and herbivorous mammals and to incorporate the results into an ap- probably are at least as vagile as permanent propriate conceptual framework. For those resident boreal birds. They would be ex- organisms that exhibit relictual distributions pected to colonize most suitable habitats it is important to elucidate patterns of colo- and show little effect of insular extinction nization and extinction and to relate those or isolation by distance. On the other hand, to theory. Three generations of icthyologists amphibians and reptiles probably have dis- (Hubbs and Miller 1948, Hubbs et al. 1974, persal capacities similar to mammals. These Smith, this symposium) have carefully dis- 224 GREAT BASIN NATURALIST MEMOIRS No. 2

sected historical changes in Great Basin Acknowledgments drainages and have related them to the con- This paper could not temporary distribution of fishes. As yet have been written without the field there has been little attempt to relate these work and published species lists of patterns to biogeographic theory, but the numerous collectors, systematists, and biogeographers who have worked possibilities for doing so seem great and it is to document the distributions of hoped it will be attempted soon. The distri- birds and mammals in the Great Basin. butions of many organisms are patchy on I am particularly grateful to N. scales smaller than the gross biogeographic K. Johnson for providing most of the data on birds one considered here. Smith (1974a, b) has and to W. H. Behle for making available obtained some encouraging results by using information presented in this the theory of insular biogeography and the symposium. My wife, A. Kodric-Brown, and students dynamics of opposing colonization and ex- my have done much to tinction rates to account for the distribution encourage my work in insular biogeography. I would like of a small boreal mammal (Ochotona prin- to dedicate this paper to the memory of Stephen D. ceps) among the isolated patches of its spe- Dur- rant, whose knowledge cialized rockslide habitat in the Sierra Ne- of the Great Basin and its mammals was vada. It will be interesting to see to what matched only by his enthusiasm for the outdoor life extent this approach continues to prove use- and his ability as a teacher and storyteller. ful for understanding insular distributions on various scales.

Appendix 1. Records of occurrence ot permanent resident boreal bird species.'

Sample Area

Species 1 2 3 4 8 9 10 11 14 15 16 18 19 20 21 22 23

Dendragapus obscurus Bonasa umbellus Lagopus leucurus Oreortyz picta Glaucidium gnotna Strix occidentals

S. ncltulosa Drijocopus pileatus

Dendrocopus villostis D. pubescens D. albolaruatus Picoides arcticus

P. tridactylus Perisorcus canadensis Cyanositta stellch Varus gambeli

Sitta carolinensis

S. pi/gmaea Cinclus mexicanus Pinicola cnucleator

Total 5 4 6 6 6 5 6 13 15

'Data from Johnson (1975) and Behle (this symposium). 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 225

» 2 = 5 »-~

8. £ = -- ^ 6 | | 2 > = =~' ~ ~ ;£ j: •§ E § | J a. c £ ^11 \%\ llllli | 111 ill 111! lillj ill fill iff 226 GREAT BASIN NATURALIST MEMOIRS No. 2

Literature Cited Hall, E. R. 1946. Mammals of Nevada. Univ. Cali- fornia Press, Berkeley. 1974. Fish spe- Barbour, C. D., and J. H. Brown. Harner, R. F., and K. T. Harper. 1977. The role cies diversity in lakes. Amer. Naturalist 108: of area, heterogeneity, and favorably in plant 473-489. species diversity of pinyon-juniper ecosystems. Behle, W. H. 1943. Birds of Pine Valley Mountain Ecology 57: 1254-1263. Range, southwestern Utah. Univ. Utah Biol. Ser. Harper, K. T., D. C. Freeman, W. K. Ostler, and L. 7(5): 1-85. G. Klikoff. 1978. The flora of Great Basin 1955. The birds of the Deep Creek Moun- mountain ranges: diversity, sources, and dis- tains of central western Utah. Univ. Utah Biol. persal ecology. Great Basin Nat. Mem. 2: 81- Ser. 11(4): 1-34. 103. 1958. The birds of the Raft River Mountains, Hubbs, C. L., and R. R. Miller. 1948. Correlation northwestern Utah. Univ. Utah Biol. Ser. 11(6): between fish distribution and hydrographic his- 1-40. tory in the desert basins of western United 1978. Avian biogeography of the Great Basin States, 17-166. In: The Great Basin with em- and Intermountain Region. Great Basin Nat. p. phasis on glacial and postglacial times. Bull. Mem. 2: 55-80. Univ. Utah 38(2): 1-191. 1971. mountaintops: Brown, J. H. Mammals on Hubbs, C. L., R. R. Miller, and L. C. nonequilibrium insular biogeography. Amer. Hubbs. 1974. Hydrographic history and relics Naturalist 105: 467-478. fishes of the north central Great Basin. Memoirs 1973. Species diversity of seed-eating desert Calif. Acad. Sci. 7: 1-259. rodents in sand dune habitats. Ecology 54: 775- Johnson, N. K. 1965. The breeding avifaunas of the 787. Sheep and Spring ranges in southern Nevada. H., A. Kodric-Brown. 1977. Brown, J. and Condor 67: 93-124. Turnover rates in insular biogeography: effect of 1970. The affinities of the Boreal avifauna of immigration on extinction.' Ecology 58: 445-449. the Warner Mountains, California. Occas. Pap. Culver, D. C. 1970. Analysis of simple cave com- Biol. Soc. Nevada 22: 1-11. munities. I. Caves as islands. Evolution 29: 463- 1973. The distribution of Boreal avifaunas in 474. southern Nevada. Occas. Pap. Biol. Soc. Nevada R. Culver, D. J. R. Holsinger, and Baroody. 36: 1-14. 1973. Toward a predictive cave biogeography: 1974. Montane avifaunas of southeastern Ne- the Greenbrier Valley as a case study. Evolution vada: historical change in species distribution. 2: 689-695. Condor 76: 334-337. Diamond, M. 1969. Avifaunal equilibria and spe- J. 1975. Controls of number of bird species on cies turnover rates on the Channel Islands of montane islands in the Great Basin. Evolution California. Proc. Nat. Acad. Sci. 64: 57-63. 29: 545-657. 1970. Ecological consequences of island colo- birds Pac. Linsdale, J. M. 1936. The of Nevada. nization by Southwest Pacific birds. II. The ef- Coast Avif. 23: 1-145. fect of species diversity on total population den- F., N. K. 1974. Turnover Lynch, J. and Johnson. sity. Proc. Nat. Acad. Sci. 67: 1715-1721. and equilibria in insular avifaunas, with particu- 1971. Comparison of faunal equilibrium lar reference to the California Channel Islands. turnover rates on a tropical and a temperate is- Condor 76: 370-384. land. Proc. Nat. Acad. Sci. 68: 2742-2745. MacArthur, R. H., and E. O. Wilson. 1963. An Durrant, S. D. 1952. Mammals of Utah, taxonomy equilibrium theory of insular biogeography. Ev- and distribution. Univ. Kansas Pnbl. Mus. Natu- olution 17: 373-387. ral Hist. 6: 1-549. 1967. The theory of island biogeography. Durrant, S. D., M. R. Lee, and R. M. Hansen. Princeton Univ. Press, Princeton, N. 1955. Additional records and extensions of J. Martin, P. S., and P. Mehringer. known ranges of mammals from Utah. Univ. J. 1965. Pleistocene pollen analysis and biogeo- Kansas Publ. Mus. Natural Hist. 9: 69-80. graphy of the Southwest, p. 433-451. In: H. E. 1933. of recent Grinnell, J. Review the mammal Wright and D. G. Frey (eds.), The Quaternary fauna of California. Univ. California Publ. Zool. of the United States, Princeton Univ. Press, 40: 71-234. Princeton, N.J. A. 1944. distri- Grinnell, J., and H. Miller. The Miller, A. H. 1935. Some breeding birds of the Pine bution of the birds of California. Pac. Coast. Forest Mountains, Nevada. Auk 52: 467-468. Avif. 27: 1-608. 1946. Vertebrate inhabitants of the pinyon I. life in Grinnell, J., and T. Storer. 1929. Animal association in the Death Valley region. Ecology the Yosemite. Univ. of California Press, Berke- 27: 54-60. ley. 1978 INTERMOUNTAIN BIOGEOGRAPHY: A SYMPOSIUM 227

Miller, A. H. and W. C. Russell. 1956. of pikas: consequences of insular population Distributional data on the birds of the White structure. Ecology 55: 1112-1119. Mountains of California and Nevada. Condor 1974b. The distribution and dispersal of 58: 75-77. pikas: influence of behavior and climate. Ecolo- Miller, R. R. 1948. The cyprinodont fishes of the gy 55: 1368-1376. Death Valley system of eastern California and Smith, G. R. 1978. Riogeography of Intermountain southwestern Nevada. Misc. Publ. Univ. Mich- fishes. Great Rasin Nat. Mem. 2: 17-42. 68: 1-155. 1973. habitat, dispersal igan Mus. Zool. Terborgh, J. Chance, and Power, D. M. 1972. Numbers of bird species on the in the distribution of birds in the West Indies. California islands. Evolution 26: 451-463. Evolution 27: 338-349. distribution of 1936. Rirds of the Preston, F. W. 1962. The canonical Van Rossem, A. J. Charleston

commonness and rarity: Part I. Ecology 43: 185- Mountains, Nevada. Pac. Coast Avif. 24: 1-65. 215. Vuilleumier, F. 1970. Insular biogeography in conti-

Schoener, A. 1974. Experimental zoogeography: nental regions. I. The northern Andes of South colonization of marine mini-islands. Amer. Nat- America. Amer. Naturalist 104: 373-388.

uralist 108: 715-738. 1973. Insular biogeography in continental re-

Seifert, R. P. 1975. Clumps of Heliconia in- gions. II. Cave faunas from Tesin, southern florescences as ecological islands. Ecology 56: Switzerland. Syst. Zool. 22: 64-76.

1416-1422. Wells, P. V., and R. Rerger. 1967. Late Pleisto-

Simberloff, D. S. 1974. Equilibrium theory of is- cene history of coniferous woodland in the Mo- land biogeography and ecology. Ann. Rev. Ecol. jave Desert. Science 155: 1640-1647.

Syst. 5: 161-182. Wells, P. V., and C. D. Jorgensen. 1964. Pleistocene wood rat middens and climatic Simberloff, D. S., and E. O. Wilson. 1969. in the Mojave Desert: a record of juni- Experimental zoogeography of islands. The colo- change per woodlands. Science 143: 1171-1173. nization of empty islands. Ecology 50: 278-296. N., R. Simpson, R. R. 1974. Glacial migrations of plants: West, J. Tausch, K. H. Rea, and P. T. island biogeography evidence. Science 185: 689- Tueller. 1978. Phytogeographical variation 700. within juniper-pinyon woodlands of the Great

Smith, A. T. 1974a. The distribution and dispersal Rasin. Great Rasin Nat. Mem. 2: 119-136.