Evolutionary Relationships of Heteromyid Rodents John C

Home , Chaetodipus, Dark kangaroo mouse, Dipodomyinae, Geomyoidea, Gopher, Heteromyidae

Great Basin Naturalist Memoirs

Volume 7 Biology of Desert Rodents Article 2

8-1-1983 Evolutionary relationships of heteromyid rodents John C. Hafner Moore Laboratory of Zoology and Department of Biology, Occidental College, Los Angeles, California 90041

Mark S. Hafner Museum of Zoology and Department of Zoology and Physiology, Louisiana State University, Baton Rouge, Louisiana 70893-3216

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

Recommended Citation Hafner, John C. and Hafner, Mark S. (1983) "Evolutionary relationships of heteromyid rodents," Great Basin Naturalist Memoirs: Vol. 7 , Article 2. Available at: https://scholarsarchive.byu.edu/gbnm/vol7/iss1/2

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]. EVOLUTIONARY RELATIONSHIPS OF HETEROMYID RODENTS'

John C. Hafner and Mark S. Hafner'

Abstract.— The rodent superfamily Geomyoidea is an old, undoubtedly inonophyletic lineage having only ob- scaire affinities with other rodent groups. Geomyoid rodents, autochthonous in North America, experienced major evolutionary diversification in the Mio-Pliocene coincident with the development of the Madro-Tertiary Geoflora and the climatic trend toward increasing aridity and coolness. Extant geomyoids are divisible into two groups: (1) the Geomyidae, all members of which are fossorial, and (2) the Heteromyidae, whose members display an adaptive con- tinuum from bipedal, xeric-adapted forms to scansorial, mesic-adapted forms. These moieties, although recognizable on biochemical criteria, become particularly difficult to distinguish when paleontological data are considered. Nev- ertheless, most lines of evidence indicate that the families Heteromyidae and Geomyidae are distinct, monophyletic lineages.

The extant heteromyids comprise three main lineages (including six genera) that diverged during the Eocene: (1) subfamily Perognathinae (Chaetodipus and Perognathus); (2) subfamily Dipodomyinae (Dipodoinys and Micro- dipodiips); and (.3) subfamily Heteromyinae {Lioimjs and Heteromijs). Protein differentiation has occurred at hetero- geneous rates among these major lineages. Based on available karyotypic data, the main direction of chromosomal evolution in the Heteromyidae appears to be toward increasing chromosome number. Cladistic analysis of morphological characters used in previous studies supports biochemical evidence allying Microdipodops with Dipodoinys. A model is introduced to describe how heterochronic changes in ontogeny may explain the great breadth of morphological diversification within the superfamily. Taxonomic recommendations at the subfamilial, generic, and sub- generic levels are provided.

The most important point to be emphasized is that relationships within the Heteromyidae was "Parallelism, parallelism, more parallelism and still Wood's exhaustive treatment, now aged one- more parallelism" is the evolutionary motto of the ro- half century. However, during the past 50 dents in general and of the heteromyids in particular. volume of literature per- This extends to all parts of the body. It makes the task of years a tremendous determining interrelationships particularly difficult, and taining to heteromyid evolution has accumu- renders exceptionally dangerous any postulates as to lated, justly reflecting the immense interest what the relationships of a given form may really be, if in these mammals. Some of the questions hill evidence does not exist to clear the maze of parallel posed by Wood and the others have been an- adaptations for us. This shows the insuperable diffi- culties awaiting anyone who attempts a classification swered to satisfaction, whereas the answers to based on a single character or on a group of characters other queries still elude us and await extrica- with a common cause. tion by future research. Albert Elmer Wood (19.35:250) It is the intent of this contribution to pre- A trio of monographs on the evolutionary sent a compendium of the evolution of heter- biology of heteromyid rodents appeared in omyid rodents, wherein we attempt to in- the early 1930s and, subsequently, has hall- tegrate the classic morphological studies of marked this specialized area of scientific in- the 1930s with the more recent systematic quiry. Hatt (1932) and Howell (1932) pro- treatments. As a definitive statement on het- vided definitive accounts of the morphology eromyid relationships, this effort may appear of the ricochetal forms, and Wood (1935) inchoate in a few years. However, the assimi- synthesized the then available data, gleaned lation of earlier ideas with those of the pres- is neces- from fossil and recent forms, into a coherent ent, coupled with due introspection, of summarization. Interestingly, the last com- sary in any field of science. The study prehensive statement of the evolutionary heteromyid evolution is no exception and

of Mammalogists, hosted by Brigham Young 'From the symposium "Biology of Desert Rodents," presented at the annual meeting of the American Society University, 20-24 June 1982, at Snowbird, Utah. 90041. 'Moore Laboratory of Zoology and Department of Biology, Occidental College, Los Angeles, California Louisiana 70893-3216. •Museum of Zoology and Department of Zoology and Physiology, Louisiana State University, Baton Rouge, Great Basin Naturalist Memoirs No. 7 through this reflection of the present to the myoids, and it was not mitil 1848 that Water- past we hope to gain a profitable avenue for house recognized the Saccomyina ( = modern further investigations. Geomyoidea) as a distinct group of New The evolution of the Heteromyidae (kan- World rodents (Waterhouse 1848). In 1872, garoo rats, pocket mice, and their allies) is Gill recognized two closely related families closely associated with that of the Geo- within Waterhouse's Saccomyina, the Geo- myidae (pocket gophers) and, consequently, myidae ( = modern Geomyidae) and the Sac- tlieir taxonomic histories are necessarily in- comyidae (= modem Heteromyidae). Gill tertwined. Together these two families form (1872) imited these families under the super- an internally cohesive superfamily (the Geo- familial nomen, Saccomyoidea. Weber myoidea) whose members are miited by the (1904), recognizing that Saccomys Cuvier presence of externally opening, fur-lined 1823 was a junior synonym of Heteromys cheek pouches (among other features). Geo- Desmarest 1817 (see Gray, 1868 for details), myoid rodents underwent major phyletic first used the superfamily name Geomyoidea. radiation from the Oligocene to Pliocene of The phyletic position of the Geomyoidea North America, in step with the global trend within the order Rodentia has long been a toward a cooler and drier climate (Flint matter of debate. Coues (1877) considered 1971) and the diversification and migration the geomyoids to be "myomorphs" {sensii of geofloras (Axelrod 1950, 1958, 1976). His- Brandt, 1855). However, Miller and Gidley toric biogeographic considerations have been (1918) and many recent workers place the presented elsewhere (e.g.. Wood 1935, geomyoids near, or within, the "sci- uromorphs" (e.g., Simpson, Most Reeder 1956, Genoways 1973, Hafner, J. C., 1931, 1945). 1981a) and will not be repeated here; for recently. Wood (1965:128) suggested that the general reviews see Stebbins (1981; coe volu- muroid and geomyoid rodents may have tion of grasses and herbivores), and Cole and shared a common ancestor within the primi- Armentrout (1979; Neogene paleogeog- tive, protrogomorphous rodent family Sciura- raphy). It is interesting to note, however, that vidae, and he thus placed the Geomyoidea heteromyids and geomyids represent two of within the suborder Myomorpha (see also the three families of living mammals autoch- Wood 1955, Wahlert 1978). thonous in continental North America (the Rodents of the family Eomyidae represent third being Antilocapridae). As a con- a third, wholly extinct group of geomyoid ro- sequence of the similarities in heteromyid dents present from late Eocene to Plio- and geomyid biogeographic histories and Pleistocene in Europe and North America. their intimate phyletic relationships, we have According to Wilson (1949a, 1949b), who found it illuminating to include relevant geo- first placed the eomyids within the Geo- myid information in this review of the Heter- myoidea, the eomyid skull shows many sim- omyidae. Indeed, in order to appreciate fully ilarities to both primitive heteromyids and the history of evolutionary diversification cricetids. Wood (1955) concurs with Wilson within the Heteromyidae, it is necessary first in recognizing the Eomyidae as a primitive to view this family within the broader frame- geomyoid group, perhaps ancestral to the work of the superfamily Geomyoidea. Heteromyidae. However, none of the known eomyid forms appears to be directly ancestral Review of Geomyoid Systematics to living geomyids or heteromyids (Wahlert 1978). The taxonomic history of the Geomyoidea The phyletic propinquity of geomyids and began with the description of the "tucan" or heteromyids, and the fact that both groups "Indian mole" (probably a pocket gopher, evolved imder similar environmental condi- Tliomomys) by Fernandez in 1651. According tions in western North America, may account to Merriam (1895:201), both Fernandez and, for the remarkable level of evolutionary par- later, Kerr (1792, not seen) believed the tu- allelism in the two groups as evidenced in the can to be a large, aberrant species of mole fossil record. Due, in large part, to the con- {Sorex mexicantis Kerr 1792). Systema Na- founding effects of parallelism, the taxonomy turae (Linnaeus 1758) did not mention geo- of extant geomyids has vacillated between a 1983 Biology of Desert Rodents

single family classification (either the Sacco- and Pappogeomys extend from the early myidae, Geomyidae, or Heteromyidae; Baird Blancan (late Pliocene) to the Recent of both 1858, Cams 1868, Gray 1868, Alston 1876, the Great Plains and southwestern United Shotwell 1967, Lindsay 1972) or a classifica- States. Orthogeomys (subgenus Heterogeomys) tion composed of two families equivalent to is known from the Rancholabrean (late the modem Geomyidae and Heteromyidae Pleistocene) of Nuevo Leon, Mexico. Ortho- (Gill 1872, Cones 1877, Merriam 1895, Wood geomys (subgenera Orthogeomys and Macro- 1931, 1935, Rensberger 1971, 1973). Recent- geomys) are known only from Recent mate- ly, M. S. Hafner (1982) used biochemical evi- rial (Russell 1968). dence to demonstrate that the extant genera commonly placed in the separate families do Kangaroo Rats, Pocket Mice and indeed represent inclusive, monophyletic Their Allies: Family Heteromyidae lineages, thus supporting the traditional two- family classification. The heteromyid rodents constitute a morphologically and ecologically diverse assem- Pocket Gophers: Family Geomyidae blage of geomyids whose fossil history dates back to middle Oligocene of western North For almost a century, Merriam's (1895) America (see review by Wood 1935). Wood monograph on the Geomyidae has stood as arranged the heteromyids into three sub- the definitive statement on systematic rela- families (Perognathinae, Dipodomyinae, and tionships among extant members of the fam- Heteromyinae) and later recognized a fourth ily. Merriam's work has been modified to subfamily, the Florentiamyinae, based on the varying degrees by Hooper (1946), Russell early Miocene genus Florentiamys (Wood (1968), M. S. Hafner (1982), and Honeycutt 1936). More recent, suprageneric systematic and Williams (1982). Living geomyids are treatments of fossil heteromyids include the represented by five genera (six according to works of Reeder (1956) and Lindsay (1972). Honeycutt and Williams 1982), and 33 nomi- Living heteromyids are traditionally sub- nal species, all of which are fossorial divided into five genera and approximately herbivores. 66 species. In this paper, we will recommend

Within early geomyids ( = Geomyinae of that a sixth genus, Chaetodipus, be recog- Shotwell 1967, and Lindsay 1972), there was nized. Analyses of suprageneric relationships a Miocene radiation in western North Ameri- among extant heteromyids include Kelly's ca of forms that ranged from semi-ricochetal (1969) study of bacular and penile variation, to fossorial in habitus as inferred from both Homan and Genoway's (1978) study of hair

cranial and postcranial structure (Rensberger stmcture, and M. S. Hafner 's (1982) biochem- 1971, Munthe 1975). These forms are as- ical analysis of intrafamilial relationships. Re- signed to the geomyid subfamilies Entopty- sults of each of these studies, and others, will chinae and Pleurolicinae, which appear to be incorporated into the following accounts. have been early, independent geomyid side- Kangaroo Rats: Genus Dipodoinys.-The branches not directly ancestral to later 24 species of kangaroo rats currently recog- geomyids. nized are spread throughout much of western Following the radiation and eventual ex- North America from south central Canada to tinction of the Miocene entoptychines and central Mexico. Fossil material referable to

pleurolicines, there was a Pliocene radiation Dipodoinys is known from the Barstovian of geomyine pocket gophers in western (late Miocene) to the Recent of western North America. Most workers agree that all North America. The most recent diagnosis of subsequent geomyids had their roots in this fossil kangaroo rats is provided by Zak- Pliocene radiation. Details are very scarce; rzewski (1981; see included references). however Shotwell (1967) and Lindsay (1972) The phyletic position of Dipodomys within tentatively derive all living pocket gophers the Heteromyidae is an area of controversy directly from the Mio-Pliocene form Para- that will be addressed in the present study. pliosaccomys. The geological time ranges of Morphologically, kangaroo rats are truly bi- the genera Tfiomomys, Geomys, Zygogeomys, zarre, and they show no obvious affinity to Great Basin Naturalist Memoirs No. 7 any other heteromyid genus. Although kan- ters and have recognized only two of five results of garoo rats show a gross overall resemblance species recognized by Hall. Specific to kangaroo mice (Microdipodops), Wood their analysis will be discussed in a later sec- C. Hafner (1978) attributed tion of this paper. A comprehensive study of (1935) and J. shared features such as inflated auditory bul- chromosomal and biochemical variation in lae and elongated hind feet to evolutionary the genus is now in progress (D. S. Rogers, convergence. In contrast, Reeder (1956) in- pers. comm.). Heteromys shows close phyletic terpreted the fossil evidence to suggest a true ties with spiny pocket mice of the genus phyletic link between Dipodomys and Micro- Liomys (Goldman 1911, Wood 1935, M. S. dipodops, the same conclusion reached by M. Hafner 1982), and the two genera are placed

S. Hafner (1982) using biochemical evidence. together in the subfamily Heteromyinae. The

It is hoped that our reanalysis of Wood's relationship of the Heteromyinae to other (1935) data, as well as new information pres- heteromyid subfamilies remains obscure. ented herein, will clarify this issue. Spiny Pocket Mice: Genus Liomys.— Relative to other heteromyids, kangaroo Five extant species of Liomys are currently rats have received considerable attention recognized, ranging from northern Mexico to from systematic biologists. Studies of inter- Panama. The fossil record of the genus ex- specific relationships based on morphology tends back to late Pliocene of Kansas (Hib- include the works of Grinnell (1921, 1922), bard 1941). Goldman's (1911) revision of Wood (1935), Burt (1936), Setzer (1949), Liomys has been substantially updated by Lidicker (1960), Kelly (1969), Best and Genoways (1973), who subducted 6 of 11 spe- Schnell (1974), and Schnell et al. (1978). cies recognized prior to his analysis. As dis- Phylogenies resulting from these analyses are cussed previously, Liomys shows its closest far from concordant (see Schnell et al. 1978). phyletic affinity to Heteromys, and the place- Studies of interspecific relationships in Di- ment of these genera into the subfamily Het- podornys based on karyology (Stock 1974) eromyinae has never been seriously and protein electrophoresis (Johnson and Se- contested. lander 1971) present still different pictures of Kangaroo Mice: Genus Microdipodops.— kangaroo rat relationships. It is not within There are but two living species of kangaroo the scope of this study to reevaluate Dipod- mice (M. megacephahts and M. pallidus), ornys species relationships; we only wish to both forms restricted to the arid Great Basin call attention to the need for a thorough, region of western North America. Fragmen- comprehensive analysis utilizing a broad tary fossil material (referred to M. mega- spectrum of approaches. In view of the com- cepliahis) is known from late Pleistocene of plexity of the situation, we suggest that Nevada (Miller 1979). The genus was de- chromosomal banding studies, DNA hybridi- scribed by Merriam (1891) and subsequently zation, arid protein sequencing analyses may revised by Hall (1941) and J. C. Hafner provide new insights to this old problem. (1981a). D. et al. J. Hafner (1979) affirmed Spiny Pocket Mice: Genus Heteromys.— the specific status of the two living forms us- Spiny pocket mice of the genus Heteromys ing morphological, chromosomal, and bio- are, by far, the least studied of all hetero- chemical evidence.

myids. To date, the genus has no fossil rec- The genus Microdipodops is certainly the ord. According to Hall (1981), Heteromys is most problematic heteromyid in terms of our

represented by 10 Recent species ranging understanding of its phylogenetic affinities from southern Mexico to northwestern South within the family. Wood' (1935: 107-1 17) dis- America. Until very recently, Goldman's eased at length the morphology of Micro- (1911) revision of Heteromys stood as the dipodops relative to Dipodomys and Pe- most recent taxonomic work focusing on in- rogruithus (subgenus Perognathus). Although terspecific relationships in the genus. Rogers he pointed out a close morphological resem- and Schmidly (1982) have reevaluated inter- blance between Microdipodops and Pe- -specific relationships in Hall's (1981) H. des- rogmithus (see also J. C. Hafner 1976. 1978), murestiamis group (exclusive of H. gaumeri) Wood remained equivocal as to its phyletic using external, cranial, and bacular charac- placement (see Wood 1935:78). Both Reeder 1983 Biology of Desert Rodents

(1956) and Lindsay (1972) derived Micro- Comparative Analysis of the dipodops and Dipodomys from the Mio- Male Reproductive System Pliocene form Cupidinimtis and, thus, suggest large dipodomyine affinities for kangaroo mice. In A body of data concerning mor- contrast, phology of the glans penis, baculum, sperma- J. C. Hafner (1976, 1978) derived Microdipodops from perognathine stock while tozoa, and accessory glands of the male re- recognizing the possibility that kangaroo productive system in heteromyids is now mice may represent an independent hetero- available and may be brought to bear on the issue of myid lineage with no close relatives in the phyletic relationships within the family. Much of the evidence extant faima. M. S. Hafner (1982) assessed the presented here is previously phylogenetic affinities of Microdipodops us- unpublished; in particular, we wish ing electrophoretic and immunological evi- to call attention to the work of Kelly dence and concluded that kangaroo mice are (1969), which is a particularly thorough analysis of penile genetically somewhat closer to kangaroo rats and bacular variation in heteromyids. Figure illustrates than to other living heteromyids. In this 1 representative phalli, bacula, study we will reexamine the phyletic position and spermatozoa from eight geomyoid taxa. In the of Microdipodops relative to other hetero- interest of brevity, we myids and introduce new evidence relevant have omitted detailed descriptions of the var- ious structures in individual to the issue. species; for this, Pocket Mice: Genus Perognathus.— The we refer the reader to the original literature. genus Perognathus, as defined prior to this study, includes 25 nominal species spanning Glans Penis much of western North America from south- em Canada to central Mexico. The genus ex- The glans penis has been shown to be a hibits an vmusiially long and complete fossil valuable tool in systematic studies of rodents record extending from Lower Miocene of (Hooper 1961, 1962, and included references, western North America (Reeder 1956, Lind- Hershkovitz 1966, Lidicker 1968, and others) say 1972). but the glans penis of heteromyids has re- In traditional treatments dating from Mer- ceived httle attention. Noteworthy excep- riam (1889), Perognathus is subdivided into tions include Kelly's (1969) study of penile two subgenera, Perognathus and Chaetodipus. variation in Dipodomys, Genoways' (1973) For reasons detailed beyond, we recommend study of the glans penis in Heteromys and Hafner's analysis of that CJiaetodipus be elevated to full generic Liomys, and J. C. (1976) status and, thus, reduce the number of extant penile variation in Dipodomys, Perognathus, Perognathus species to a total of nine. and Microdipodops. Following is a synthesis Species groups within the genus Pe- of the results of these studies, emphasizing rognathus (sensu lato) are not clearly de- salient differences in penile morphology lineated using the evidence available at pres- among heteromyids. ent. Studies of interspecific relationships Kelly (1969) and Genoways (1973) re- based on morphology include works by Mer- ported that the glans penes of Heteromys and overall riam (1889), Osgood (1900), Wood (1935), Liomys (Fig. 1 C, D) are similar in and Caire (1976). Systematic treatments structure and share several characteristics based on karyology include studies by Patton imique among heteromyids. These features (1967a, 1967b) and Williams (1978).' Protein (spineless phalli; glans long relative to bacu- variation in the genus (with emphasis on lar length; unique urethral lappet mor- Chaetodipus) has been analyzed by Patton et phology) may be viewed as shared-derived of these gen- al. (1981). It appears that a complete elucida- characters supporting the union tion of interspecific relationships within Pe- era into the subfamily Heteromyinae. Out- (Kelly rognathus and Chaetodipus must await future group comparison with Thomomys studies involving finer considerations of 1969) supports the derived nature of penile chromosomal evolution (banding studies) and characters used to define the Heteromyinae. with geo- molecular change (DNA hybridization; pro- Based on outgroup comparison tein sequencing). myids, the male phalli of Dipodomys, Micro- Great Basin Naturalist Memoirs No. 7

A DIPODOMYS B MICRODIPODOPS

C HETEROMYS D LIOMYS

E PEROGNATHUS F CHAETODIPUS

G THOMOMYS

Fii;. 1. I{c-presentative plialli, l);Kiila, and siHMiiiato/oa troiu srvoii lu-k-iomv id ami one ueoiinid t;L\a. Phalli and bacula an- shown in lateral viow. To latilitatc foniparison. all illnstiations art- drawn to different scales. Sperm from Uomijs salvini (right) and lAomtjs pictus (left) are illustrated. Sperm from Hctcwimis and Thoiuoiiit/s are as vet nn- described. (Illustrations modified from Kelly, 1969, Genoways, 197.3, and J. C. Hafner, 1976). 1983 Biology of Desert Rodents dipodops, and pocket mice of the subgenus A combination of three bacular features Perognathus (Fig. 1 A, B, E) have retained a (bulbous base, stout midregion, sharply up- broad spectnim of primitive geomyoid char- turned distal end) clearly distinguishes Di- acters. The phaUus of Thomomijs (Fig. 1 G) podomys species from other heteromyids most closely resembles that of perognathine (Fig. 1 A). The bacula of Microdipodops, Het- pocket mice (Kelly 1969), and in this and eroniys, Liomijs, Tliomomys, Geomys, and other respects (see beyond) members of the Pappogeomys also have bulbous bases, and subgenus Perognathus appear to be extremely this feature appears to be primitive for the conservative morphologically. Geomyoidea. The bacula of certain pocket

The glans penis of Microdipodops is similar mice of the subgenus Chaetodipus (Fig. 1 F) in most respects to that of Dipodomys (Kelly have moderately to sharply upturned distal 1969). Importantly, several phallic characters ends (Anderson 1964), a feature which ap- shared between Microdipodops and Di- pears to have been derived independently in podotnys (including morphology of urethral Dipodomys and CJiaetodipus. lappets and external spines) are unique with- Pocket mice of the subgenus Chaetodipus in the Geomyoidea and, hence, are of phy- are clearly distinguished from members of logenetic signficance. Those features used by the subgenus Perognathus using bacular mor- C. Hafner to suggest a close rela- J. (1976) phology (Burt 1936, J. C. Hafner 1976). The tionship between Microdipodops and mem- baculum in chaetodipine pocket mice is long bers of the subgenus Perognathus (cylindrical, relative to body length such that the soft tis- nonelongated phalkis; dorsal groove) are now sue of the penis terminates approximately known to be present in the geomyid genus midway along the length of the Ijaculum. In Thomomys (Kelly 1969) and are thus re- perognathine pocket mice, the baculum is garded as shared-primitive characters. The much shorter and soft tissue extends approx- sharply upturned distal portion of the phallus imately two-thirds of the length of the bacu- in Dipodomys (Fig. 1 A) clearly distinguishes lum. The baculum of Perognathus (Chaeto- kangaroo rats from all other heteromyid dipus) hispidus (Fig. 1 H) possesses an ornate, genera. trifid tip seen nowhere else in the The male phallus in pocket mice of the Geomyoidea. subgenus Chaetodipus (Fig. 1 F) is unique among all geomyids examined thus far (Kelly Spermatozoan Morphology

1969, J. C. Hafner 1976). The chaetodipine presented a rather de- penis is long and slender, lacks urethral lap- Genoways (1973) of the spermatozoa of Liomys pets, and the rim of the terminal crater forms tailed study statement as to the rela- a ventlike urethral opening. The structure of and provided a brief tive shape of the sperm head in Perognathus the chaetodipine phallus is, doubtlessly, de- was not examined). C. rived within the Heteromyidae. The terminal pemix {Heteromys J. analyzed gross sperm mor- portion of the phallus in Perognathus Hafner (1976) phology in Dipodomys, Microdipodops, and (CJiaetodipus) hispidus (Fig. 1 H) is markedly Perognathus. The head different from that of other chaetodipine spe- additional species of (minus the acrosomal tip) of cies in possessing a distinctly ornate tip. and neck region representative heteromyid spermatozoa are illustrated in Figure 1. Os Baculum The spermatozoa of Microdipodops (Fig. 1 Bacular variation in heteromyids has re- B) are characteristically large, with especially ceived considerable attention, including long heads. The head is roughly triangular in studies by Burt (1936, 1960), Schitoskey shape, with rounded vertices. The sperm tail to other hetero- (1968), Kelly (1969), Genoways (1973), Best is of medium length relative Hafner myid species. Species of the subgenus Pe- and Schnell (1974), and J. C. (1976). sim- Because of considerable intrageneric varia- rognathus (Fig. 1 E) possess spermatozoa those of tion in bacular morphology, the use of this ilar in general morphology to small- structure in delineating higher-level hetero- Microdipodops, except that the head is These similarities do myid relationships is very limited. er and the tail shorter. 10 Great Basin Naturalist Memoirs No. 7

not indicate a close phyletic relationship be- taken from M. S. Hafner (1979). The male ac- tween Microdipoclops and Perognathus (sub- cessory reproductive gland complements of genus Perognathus), inasmuch as the sperm of 11 geomyoid genera are listed in Table 1. Mi- certain species of Liomys also share these fea- crotus is included in Table 1 to represent the tures (Genoways 1973). most common muroid (outgroup) condition,

The spennatozoa of chaetodipine pocket i.e., all gland complements present (Arata mice are easily distinguished from those of all 1964). other heteromyids (Fig. 1 F, H). Here, the Loss of glandular complements appears to sperm head resembles a somewhat elongated be much more widespread among geomyid isosceles triangle with acute (unrounded) ver- than among heteromyid taxa. Patterns in loss tices. The sperm of Perognathus {Chaeto- or retention of gland complements agree in dipus) hispidus is peculiar in some respects general with patterns seen in muroid rodents because the tail is very long and the neck re- (Arata 1964); for example, it is not uncom- gion is not discernible. mon to find the preputial glands absent in In Dipodomys, the spermatozoan head ap- taxa representing either group of rodents proximates an equilateral triangle (Fig. 1 A) (geomyoids or muroids), whereas the bulbo- and this feature alone distinguishes kangaroo urethral glands are present thus far in all mu- rat sperm from that of other heteromyids; the roid and geomyoid taxa examined. apex of the sperm head is rather acute and By far the most striking changes in the the tail long. male reproductive tract of geomyoid rodents involve changes in the morphology of the ve- Accessory Glands of the sicular glands (seminal vesicles). In the large Male Reproductive Tract majority of geomyoids (and in rodents in general), the vesicular glands are elongate, hook- With the exception of Gunther's (1960) de- or cane-shaped, translucent structures. In scription of the male reproductive tract in contrast, most specimens examined thus far Thomomys, no studies of geomyoid reproduc- of the genus Thomomys (subgenus Mega- tive tract morphology have appeared prior to scapheus) have tubular, translucent vesiculars this analysis. The descriptions below are L. Fatton, pers. all (J. comm.), and specimens

Table 1. Male accessory reproductive gland complements in 14 geomyoid taxa (+ indicates gland complement present). Micwtus is representative of the typical nuiroid condition (Arata, 1964). Terminology as per Arata (1964).

Gland complements' Taxa examined vmp vlp ap dp bu Tlxomomijs bottae Thomojmjs umbriniis Geomijs hursarius Ztjgogecmiijx trichopus O. (Orthogeotnys) grandis O. {Macrogeomys) heterodus rappogeoiny.s gymnurus lU'tcnmiys dcsinarc.stianus Liomys pictits

Pcrogiiatit us s))iii(itus Perognathus parvus

Mi( n nlipodops megm cph a I us Dipixloiuys ordii Dipodomys merriami Microtus sp.

'"""°"'"'"'' -«"»^°''""''> P™"'"'^^ ''P = prostate: ap = anterior prostate; dp = dorsal prostate; = urlral;7'="p;;prial.° a =^;;;pullary; bu bulbo' 'Numbers refer to shape of vesicular gland complement: (1) elongate; (2) round; and (3) short, tubular '''""" '"'"""' ""'"'' "' '" ''*"'"'''* "" ''"""" '" ''""^'"" ''" '" '^'^ ''"'^^"'^'' -"-'" preplltia''^"n." °' '"^g" '^ "^ -'"«''- '--- -^ '«' - 'he 1983 Biology of Desert Rodents 11 examined belonging to the subgenus Chaeto- phology of the glans penis, baculum, dipus of the genus Perognathiis have round, and spermatozoa. smooth, yellow- to gray-colored vesicular 5. Species of the pocket mouse subgenus glands (pinkish and granular in appearance in Perognathus show extreme morpho- fresh specimens). The morphologies of the re- logical conservatism with respect to the spective vesicular glands in Thonioinys and male reproductive system. Chaetodipus are unique among those described thus far for rodents. This unusual vesicular morphology in the subgenus Chaeto- Biochemical Variation dipus is particularly striking in that all Analyses of intrapopulation protein specimens examined belonging to the sub- varia- tion in Microdipodops (D. Hafner et al. genus Perognathus have more typical rodent J. 1979) and Perognathus, subgenus vesicular glands. Typical Perognathus and Chaetodipus (Patton et al. have revealed levels Cliaetodipus male reproductive tracts are 1981), of genetic polymorphism and heterozygosity contrasted in Figure 2. approximately equivalent Among the heteromyid genera examined to averages summarized for 46 species of mammals by Nevo (Table 1), Heteromys shows an unusually high In contrast, the average values of level of evolutionary loss of glandular com- (1978). polymorphism and heterozygosity measured plements (4 of 8 glands are absent). In con- in populations of Dipodomys and trast, all glandular complements are present (Johnson in Perognathus parvus (subgenus Pe- rognathus). In Heteromys and Microdipodops, PEROGNATHUS the absence of ventromedial prostates is certainly a derived condition; however, it would be dangerous to link these two genera phylogenetically based solely on the absence of a single gland complement, especially when loss of ventromedial prostates is seen in the outgroup (Geomyidae). Similarly, the presence of preputial glands in Perognathus parvus and Microdipodops should not be used to infer a special phylogenetic relationship; the genera have merely retained a gland complement normally present in nongeomyoid rodents. CHAETODIPUS

A conservative assessment of the data reviewed above leads to five conclusions:

1. Lioinys and Heteromys show close phylogenetic relatedness based primarily on morphology of the glans penis. 2. Dipodomys and Microdipodops are also suggested to be phylogenetically allied, again based largely on morphology of the glans penis. 3. Pocket mice of the subgenus Chaeto- dipus show only remote morphological

similarities to species of the subgenus Fig. 2. Ventral (left) and dorsal (right) views of the Perognathus. All aspects of the male re- male reproductive tracts of Perognathus (Perognathus) spinatus. The urinary productive system support this longimembris and P. (Chaetodipus) bladder (b) has been excised to reveal underlying acces- conclusion. sory gland complements. Abbreviations as per Table 1 4. Within Cliaetodipus, P. (Chaetodipus) with following additions: caput epididymis (cap); cauda hispidus is unique with respect to mor- epididymis (cau); testis (t); vas deferens (vd). 12 Great Basin Naturalist Memoirs No. 7

Selander 1971, Patton et al. 1976) are only tion has proceeded at the same rate within half as large as those measured in most mam- the perognathine and heteromyine lineages. malian populations. Johnson and Selander Thus, due to extreme exomorphological con- (1971) suggest that a combination of phyloge- servatism, Perognathiis and Chaetodipus are, netic and ecological factors may explain de- in a sense, "cryptic genera." pressed levels of intrapopulational genetic Studies of albumin (M. S. Hafner 1982) and variation in kangaroo rats relative to other transferrin (Sarich 1975) immimology in het- mammals. Indeed, future genetic analyses of eromyids show Microdipodops to be extreme- kangaroo rat populations may find a causal ly conservative with respect to rate of pro- Unk between ecological amplitude (niche tein change through time. In immunological width) and genetic variation, but a clear con- comparisons with nonheteromyid outgroups, nection was not evident in Johnson and Se- Microdipodops shows consistently lower lev- lander 's (1971) study. els of protein change relative to other hetero- Interpopulation genetic differentiation in myids. According to Sarich (pers. comm.), the Dipodoniys and Perognathiis {Chaetodipus) high degree of protein conservatism seen in has been shown to be approximately com- Microdipodops is unique among taxa thus far mensurate with values measured between examined immimologically belonging to sev- populations of other rodents, and mammals eral different mammalian orders. At present, in general (Johnson and Selander 1971, Csuti we have no explanation, or even hypotheses, 1979, Patton et al. 1981). Genetic studies at to accomit for extreme protein conservatism the intra- and interpopulation level are cur- in Microdipodops. rently in progress for the genera Lioniys and Heteroniys (D. S. Rogers, pers. comm.). Karyotypic Evolution Estimates of interspecific protein differen- genera tiation within the three heteromyid Chromosomal Variation thus far examined {Micwdipodops, Di- podomys, and Perognathiis, subgenus Chaeto- Heteromyid and geomyid rodents have at- dipus) reveal two divergent patterns. First, tracted a great deal of attention from evolu- Micwdipodops megacephahis and M. pallidus, tionary biologists and systematists. Interest in which are sibling species on morphological these rodents stems, among other reasons, criteria (D. Hafner et al. are ge- from the great of varia- J. 1979), degree chromosomal netically differentiated at the appropriate tion across the group. Whereas karyotypic in- "sibling species level" as defined by Zimmer- fonnation has been applied commonly to man and Nejtek (1977) based on data from 10 questions at the species level, such informa- species of mammals. Similarly, genetic dis- tion has not been used adequately in assessing tances .measured between morphologically evolutionary relationships at higher levels. well-differentiated species of Dipodomys Further, the fusion paradigm of karyotypic (Johnson and Selander 1971) are approx- reorganization has been invoked to the near imately equivalent to those measured be- exclusion of other possible mechanisms that tween well-differentiated species in other alter diploid number (e.g., Thaeler 1968, mammalian taxa (Zimmerman and Nejtek Davis et al. 1971, Genoways 1973, Selander 1977). In Perognathiis, on die other hand, et al. 1974, Stock 1974, Williams and Gen- Sarich (1975), M. S. Hafner (1979, 1982), and oways 1975, Williams 1978). The Rob- Patton et al. (1981) have shown that the ge- ertsonian fusion model (Robertson 1916),

netic distance measured between perog- whereby the karyotype is reduced in diploid nathine and chaetodipine pocket mice is ex- number through the joining of uniarmed tremely large, in fact larger than that mea- chromosomes to form biarmed elements, has sured between Lioniys and Heteroniys. reigned as the predominant view in inter- Clearly, the rate of exomorphologic change pretations of geomyoid chromosomal evolu- between Perognathiis and Chaetodipus has tion despite there being no convincing argu- lagged far behind that of biochemical ment to support its occurrence. In geomyoid change. Moreover, all indications from com- studies, empiricism has lagged well behind parative rate tests suggest that protein evolu- theoretical considerations; and other hypoth- 1983 Biology of Desert Rodents 13

eses, alternative to the fusion paradigm, have Directionality in Chromosomal Evolution not been seriovisly considered (see also Imai Geomyoid karyotypic data do and Crozier 1980). warrant a close evaluation of directionality Diploid numbers are now known for repre- of chromosome number change within the superfamily, sentatives of all genera and subgenera within irrespective of the mechanism(s) that the Geomyoidea (Heteromyidae: Dipodomys, may be involved in the change C. Hafner 1981b). Cross 1931, Matthey 1952, 1956, Csuti 1971, (J. The Geomyoidea, with its high degree of Dingman et al. 1971, Fashing 1973, Stock diploid number variation and doubtless et al. 1974; Microdipodops D. J. Hafner 1979, monophyletic origin, is 1981a; Chaetodipus, Patton an appropriate group J. C. Hafner in which to examine polarity in diploid num- 1967a, 1969, 1970; Perognathus, Patton ber change. Are high numbers primitive with 1967b, Williams 1978; Liomijs, Genoways a general reductional trend, or are low num- 1973; Heteromys, Genoways 1973, A. L. bers primitive with an increasing trend? If no Gardner, pers. comm., D. S. Rogers, pers. general trend is discernible, we must consider comm., M. S. Hafner and J. C. Hafner, un- tlie possibility that both trends operate, but publ. data; Geomyidae: Geomys, Davis et al. in separate geomyoid lineages or at different 1971, Baker et al.' 1973, Selander et al. 1974, times in the history of a single lineage. Williams and Genoways 1975, Hart 1978, Honeycutt and Schmidly 1979; Pappogeomys, In this analysis we have considered 12 ma-

Laguarda-Figueras et al. 1971, Berry and jor taxonomic groups (genera and subgenera Baker 1972, Hart and Patterson 1974, Smolen listed above). Note that we include both sub- (Tho- et al. 1980, Honeycutt and Williams 1982; genera within the genus Thomomys Thomomys, Thaeler 1968, 1972, 1973, 1974a, momys and Megascapheus, sensu Thaeler 1974b, 1976, 1977, 1980, Thaeler and 1980) as well as the subgenera Chaetodipus Hinesley 1979; Thornomys, subgenus Mega- and Perognathus of the genus Perognathus. A scapheus, Patton and Dingman 1968, 1970, histogram plot of species numbers versus (including intraspecific Patton et al. 1972, Patton 1973, 1980, Patton diploid numbers C. chromosomal variation) reveals that the dis- and Yang 1977, Patton and Feder 1978, J. karyotypes is rather Haftier et al. 1983; Orthogeomys, M. S. Haf- tribution of geomyoid exhibit a damped os- S. trimodal and seems to ner 1979, M. Hafner and J. C. Hafner, un- (Fig. Therefore, the array publ. data; Zygogeomys, M. S. Hafner, 1979, cillatory pattern 4). trisected into S. Haftier data). of diploid numbers can be M. and J. C. Hafner, unpubl. states: Figure 3 summarizes the chromosomal var- chromosomal subgroups or character (2n = 34-50), me- iability for geomyoid species. If Robertsonian low chromosome numbers and high diploid chromosomal events (fusion and/ or fissions) dium numbers (about 60), were chiefly responsible for effecting diploid numbers (68-88). number change in the six major groups of the Invoking the criterion of character state Geomyidae, the dots representing karyotypes distribution, we first attempt to assign po- within a species and species within the major larity by using in-group analysis. According groupings would be expected to be aligned in to the criterion of in-group analysis, the char- vertical patterns. It can be seen (Fig. 3A) that acter state that occurs most frequently within this does not appear to be the general case. the group under study (the common state) is In the Heteromyidae (Fig. 3B), a similar situ- considered to be the primitive state (Stevens ation exists. Within the major heteromyid 1980:335-37, Criterion lA). The low groups, karyotypes are not, for the most part, chromosome number character state contains aligned in vertical arrays. Lacking G- and G- 47 karyotypes, the medium state 44 karyo- banding information for most groups in the types, and the high number state has 27 kar- evidence may Geomyoidea, it is premature to speculate on yotypes (Fig. 4). Although the the specific mechanisms, or actual mechan- not be strong, the criterion of in-group analy- that lower ics, of chromosomal rearrangement. More- sis would lead to the conclusion are over, even banding data will not allow us to chromosome numbers (2n = 34-64) are assess directionality of change, vis-a-vis the primitive and higher numbers (2n > 64) earlier fusion /fission controversy. derived, which is contradictory to No. 7 14 Great Basin Naturalist Memoirs

GEOMYIDAE

MEGASCAPHEUS

ORTHOGEOMYS

• n = 1

• n = 2

ZYGOGEOMYS • n = 3 PAPPOGEOMYS n = 4

60 80 100 120 140 ARM NUMBER HETEROMYIDAE 1983 Biology of Desert Rodents 15 published opinions. in the low diploid number character state de-

To view tliis in-group analysis at a higher fined earlier; the mode for the Muridae is 2n taxonomic level, we next examine the num- = 48. Selecting another out-group for com- ber of major geomyoid groups versus diploid parison, the Cricetidae, again we see that the numbers (Fig. 5). Again, the criterion would vast majority of cricetid karyotypes fall into maintain that the most common character the low chromosome number character state state across the major groupings is primitive. (mode: 2n = 48; Fig. 6B). Further, out-group The low chromosome nvmiber character state comparison with the Sciuridae (Fig. 6C; contains 10 of tlie 12 major geomyoid taxa. mode: 2n = 38) and the Aplodontidae (2n = In comparison, the medium character state 46; Matthey 1973), two families commonly contains 8 of the major taxa and the high believed to represent primitive rodents, again chromosome number state contains only corroborates the earlier assigned polarity. three major taxa {Dipodomys, Geomys, and Thus, all out-group comparisons indicate that

Megascapheus; Fig. 4). The conclusion to be the low character state is primitive. drawn from this refined in-group treatment, One other criterion that can be used to de- once again, is that the lower diploid numbers termine the polarity of character states is are primitive and the higher character state that of correlation among character states derived. (Stevens 1980:345-348, Criterion 5). This ar-

Such evolutionary polarity, assigned by in- gument is based on the assumption that group analysis must be confirmed using other primitive character states frequently occur criteria. We therefore must consider out- together. As Stevens (1980) aptly noted, the group analysis (see Stevens 1980:337-340, criterion of correlation should be used in Criterion IB). Following the out-group meth- combination with other criteria (e.g., in- od, the character state that is shared with the group and out-group analyses) as has been out-group is taken to be primitive. It has done here. been suggested by some workers that the mur- One morphological character whose char- ids stand as a distant sister group to the Geo- acter state polarity is reasonably well sup- myoidea. The range of diploid numbers in ported is that of sulcation in the upper in- the Muridae (Fig. 6A) is almost entirely with- cisors. It seems that grooved upper incisors is

LOW MEDIUM HIGH -^ 12-

WlO- O UJ 8- co

uj 4- OQ g2H K K F F F I r 30 40 50 60 70 80 90 DIPLOID NUMBER

triniodal and Fig. 4. Frequency distribution of the chromosome numbers in the Geomyoidea. The distribution is Ortho- skewed to the right. Odd diploid numbers have been omitted for clarity. Geomyidae: A, Zygogeomys; B, geomys; C, Geomys; D, Pappogeomys\ E, Thoniomys; F, Megascapheus. Heteromyidae: G, Chaetodipus; H, Pe-

rognathiis; I, K, L, Dipodomys. Microdipodops; J, Heteromys; Liomys; No. 16 Great Basin Naturalist Memoirs 7 plesiomorphic for groups within the Geo- numbers of Dipodomys. Kangaroo rats are myoidea, with loss of a sulcus being the de- distributed chromosomally in the medium rived state (see Merriam 1895, Wood 1935, and high diploid number character states and Russell 1968). Within the Geomyidae, only exhibit the highest diploid number for the one extant group totally lacks sulci in the up- Heteromyidae. This correlation criterion per incisors. This is the smooth-toothed pock- again supports our initially assigned polarity. et gophers, which include Thomomys and We shall argue beyond that Microdipodops, Megascapheus. Interestingly, pocket gophers which exhibits relatively low diploid numbers of the subgenus Megascapheus have the high- (2n = 40,42), is much less derived morpho- est diploid numbers in the superfamily (Fig. logically than is Dipodomys.

4), and Megascapheus is a major taxon that is Further evidence that can be used to sup- entirely restricted to the high diploid number port this mode of directionality comes from character state. Although it cannot be dem- recent information on Thomomys bottae et onstrated that this instance of character state pocket gophers in Colorado (J. C. Hafner

it is al. C. Hafner et al. 1983). Through- correlation is not due to chance alone, 1980, J. nevertheless in accord with the aforemen- out its extensive range, T. bottae is known to tioned polarity and the general trend for in- have, almost without exception, 76 chromo- crease in chromosome number. somes. However, a chromosomal race of T. Within the Heteromyidae, kangaroo rats bottae recently discovered in Colorado had (Dipodomys) display a surfeit of derived mor- 88 chromosomes (the second highest known phological characters. Kangaroo rats are gen- for the Mammalia); yet allozymically it is erally large in body size, have large eyes, nearly identical with the parental 2n = 76 long hind feet, expanded auditory bullae, a form. Clearly, this anomalous race, character- middorsal sebaceous gland, a long tail, and ized by a diploid number of 88, is derived elaborate pelage markings; doubtless, kan- from the common 2n = 76 form. Had the 88 garoo rats are the most morphologically de- chromosomal race been primitive, one would rived members of the Heteromyidae (see have expected its level of genie divergence to beyond). Correlated highly with the derived be equal to or greater than levels of major nature of these characters is the high diploid protein differentiation seen within the 2n =

O §6- CD oCC < F. 4- o oLi- 0:2-

30 V 50 V DIPLOID NUMBER

Fig. 5. Number of genera or subgenera versus diploid number. XuMibers above lirackets indi ijor taxa represented in the three chromosomal character states. 1983 Biology of Desert Rodents 17

2n = 48

CO LU ^20 CO

a: 10

DIPLOID NUMBER

Sciu- Fig. 6. Frequency distributions of diploid numbers for out-group comparisons: A, Muridae; B, Cricetidae; C, ridae. Data are taken from Hsu and Benirschke (1967-1977) and Matthey (1973). Modal numbers are indicated in the figures. 18 Great Basin Naturalist Memoirs No. 7

description of Micwdipodops 76 form; this is not the case (see Patton and riam's (1891) C. Hafner et al. 1983). Impor- there has been much controversy as to Yang 1977, J. closely re- tantly, this study represents the only thor- whether kangaroo mice are most oughly documented (chromosomal, allo- lated to kangaroo rats or pocket mice (for re- of Hafner 1978, M. S. Hafner zymic, morphologic) demonstration view see J. C. chromosomal directionality in the 1982). Kangaroo rice are phenetically most C. Hafner 1978), Geomyoidea. similar to pocket mice (J. In view of the foregoing, it seems that yet biochemical data indicate that they are

there is certainly ample justification for an somewhat more closely related to kangaroo alternative viewpoint with respect to the di- rats (M. S. Hafner 1982). One might therefore rection of chromosomal number change in hypothesize that Micwdipodops, although the Geomyoidea. Several statements in the being a cladistically old group, is patristically literature claiming high diploid numbers to primitive and the similarities it shares with be primitive for particular heteromyid pocket mice are actually symplesiomorphous. groups and for the entire Geomyoidea (e.g., Thus, it is necessary to perform a cladistic Williams 1978:605) are without support. We study in which shared-derived characters (hy- fully realize that reductional trends may in- pothesized on the basis of out-group com- deed have occurred in specific geomyoid parisons) are used to unite taxa. Accordingly, lineages; however, arguments used to support herein we evaluate this central issue in heter- the diploid number reductional trend are omyid phvlogeny by analyzing the phenetic utilize general bio- weak and commonly data of Wood (1935) and J. C. Hafner (1978), geographical and ecological explanations using cladistic procedures. (e.g., Williams and Genoways 1975). As Wood (1935) revisited.— Fifty-three pointed out by numerous authors, such rea- characters were used by Wood (1935) in his

soning is fraught with problems of circularity treatment of the evolutionary relationships of

and is inadiuissible in the assignment of po- kangaroo mice. Wood (1935:108) tabulated larity to character states. In contrast, in- the character-state distribution for each of group and out-group analyses and correlation these characters across the heteromyid gen- among character states support an alternative era and concluded that kangaroo mice were

view that low diploid numbers seem to be most closely related to pocket mice. It is primitive for the Geomyoidea, and the over- these tabulated data (Wood 1935:108) that all direction of karyotypic change has been are available for the present cladistic analy-

that of increase in chromosome number. We sis; we have made no effort to reexamine shall not speculate on the mechanism of in- Wood's characters and, for present purposes, creased numbers at this time, but will men- accept his selection of characters and his in- tion that fission may be involved, as well as terpretations concerning character ho- other mechanisms such as chromosome dupli- mologies. In our analysis of Wood's data, the cation and addition of euchromatin or heter- Heterominae (Liomijs and Heteromys) was ochromatin to centromeric fragments. A full chosen for out-group comparison because

discussion of karyotypic interconversions most authors agree that it is an evolutionarily within the Geomyoidea must await the time independent lineage quite removed from Pe- when more karyotypes have been banded and rognathus, Micwdipodops, and Dipodomys. we have a more firm understanding of the Gonsequently, characters shared with Liomys complexities of chromosomal evolution. and Hetewmys (see Wood 1935:108) are hy- pothesized to be primitive. Ten of Wood's MoRl'HOLOCICAL EvOLUTION original characters were omitted from the analysis (all three genera shared seven char- (^ladistic Analyses of acters with the out-group and three other Heteromyid Relatioaships characters were ambiguous), reducing the to-

tal number to 43 characters (Table 2). A prominent issue in the study of hetero- Tlie three possible phylogenetic hypoth- myid relationships is the question of the sub- eses, showing the apportionment of the 43 familial affinity of kangaroo mice. Since Mer- morphological characters, are presented in 1983 Biology of Desert Rodents 19

Table 2. Characters used in our reanalysis of Wood (1935). Ten of the 53 characters listed by Wood (1935:108) have been omitted in the reanalysis (see text). 20 Great Basin Naturalist Memoirs No. 7 eromyine data from Genoways 1973). A total Table 3. Seventeen characters used in the reanalysis ot Hafner (1978). Quantitative characters C. Hafner of 17 characters was available for cladistic (J. 1978:363-364) and characters whose character-state analysis (Table 3) after paring down Hafner's polaritv was indeterminable were omitted in our (1978:363-364) data matrix of 40 characters analysis. (quantitative characters were omitted from our analysis as well as those characters with Character indeterminable character-state polarity). Throughout our analysis, we have accepted as a working hypothesis that Perognathus and Chaetodipus form a clade relative to the other lineages; this assumption is supported by biochemical evidence (e.g., Sarich 1975, M. S. Hafner 1982).

Figure 8 shows the three phylogenetic hypotheses with the 17 characters (Table 3) ap- portioned among them. The results of this analysis, in accord with our above treatment of Wood's (1935) data, suggest that the most parsimonious phylogenetic hypothesis is that which links Microdipodops and Dipodomys as a clade independent of the Perognathus- Clmetodipus clade (Fig. 8C). However, the analysis is less than robust; the Perognathus- CJiuetodipus clade lacks an autapomorphic character, and tlie characters that define the Microdipodops-Dipodomys association (Fig. 8C) are dubious (fully haired soles of hind 1983 Biology of Desert Rodents 21 feet and ricochetal mode of locomotion). Figs. 7 and 8), the pocket mice are the least Conclusions about the Microdipodops derived, and Microdipodops exhibits an inter- CONTROVERSY.— The above cladistic analyses mediate number of derived features. tlie C. of data of Wood (1935) and J. Hafner (1978) are necessary exercises, yet the results Evolution of Geomyoid Morphotypes: are not imequivocal and provide us with only An Hypothesis a preliminary insight to the problems of as- certaining phylogenetic relationships within Tlie diversity of morphological forms, or the Heteromyidae using morphological char- morphotypes, within the superfamily Geo- acters. Both reanalyses demonstrate modest myoidea eclipses that seen within any other support for the recognition of the Micro- mammalian group of comparable size. The dipodops-Dipodomys clade; thus, they are in evolution of forms as stnicturally divergent agreement with biochemical evidence. How- as kangaroo rats, pocket mice, and pocket ever, the treatments are inchoate and there is gophers is commonly believed to be the re- a need for a detailed phylogenetic analysis sult of orthoselection; i.e., directional selec- wherein character homologies and character- tion acting on ancestral species favoring cer- state transformations are determined. As tain adaptations which are accentuated in noted by many workers (e.g., Hennig 1966), descendant species. The origin of each struc- the crux of any analysis and its attending ar- tural novelty is thus explained in terms of its gmnents relies on the characters under study. present day fimctional advantage to the ani-

Several others points emerge from the mal. Although it cannot be denied that a con- above analyses. Importantly, these cladistic spicuous morphological feature present in a treatments tabulate quantitatively the tre- living animal may, and probably does, have mendous amount of parallelism in the hetero- an adaptive fimction today, we agree with myids. That the morphology of these rodents Gould and Vrba (1982:13) that "current util- is rife with parallelisms is documented con- ity carries no automatic implication about vincingly in Figure 7A-C. The analyses also historical origin." As Lewin (1982:1212) provide a comparative measure of the degree points out, "it is folly to infer without cau- of morphological divergence among the het- tion the historical genesis of a feature from eromyid genera. In accord with intuition, its current utility." Accordingly, we herein these results document that Dipodomys is the introduce a causal hypothesis, devoid of func- most derived genus in terms of morphology tional explanations, to account for the evolu- (highest number of autapomorphic features; tionary origin of morphological diver-

Fig. 8. Cladistic analysis of Hafner (1978) for Pewgnathus (P), Chaetodipus morphological data presented by J. C. (C), Microdipodops (M),' and Dipodomys (D). Cladograms A, B, and C reflect the three possible ways that Micro- dipodops and Dipodomys may be linked with the Perognathiis-Chaetodipus assemblage. Liomys and Heteromys are treated as an outgroup (data from Genoways, 1973). See Table 3 for list of characters. 22 Great Basin Naturalist Memoirs No. sification v/ithin the Geomyoidea. This and maturation is accelerated. Gould (1977) hypothesis is expanded and supportive evi- argues that the key to vmderstanding the im- Hafner and significance of shifts in devel- dence is detailed elsewhere (J. C. mediate M. S. Hafner, in manuscript). opmental timing (heterochrony) lies in the Several years ago we observed that the theory of r and K selection (life history strat- conspicuous morphological features of adult egies). Gould (1977:293) predicts that pro- kangaroo rats and kangaroo mice, most no- genesis will be associated with r-strategists, tably the large head and eyes, enlarged brain, and, indeed, early kangaroo mouse evolution long hind feet, and delicate, weakly fused Ls postulated to have occurred in an obvious skeleton, were traits commonly seen in the r-selected environment (ephemeral sand-dune Desert; C. juvenile state of other animals. Further exam- habitats in the Great Basin J. Haf- ination revealed that kangaroo rats and kan- ner 1978); kangaroo mice possess many of the garoo mice possess many of the classical fea- classical attributes of an r-strategist. It ap- tures characteristic of paedomorphic forms. pears that in kangaroo mice there has been This initial observation, coupled with the progenetic truncation by precocious matura- subsequent discovery that neonatal pocket tion and tliat this early maturation is the gophers look remarkably like mature pocket principal object of selection. That is, there mice, prompted further investigation culmi- has been a "redirection of selection" (Gould nating in our hypothesis that geomyoid mor- 1977) away from morphology, per se, and to- phological transformations through phy- ward precocious maturation as a life history logeny may be the result of evolutionary strategy; juvenilization may have been en- epigenetics (for recent reviews see Gould tirely incidental. 1977, Alberch et al. 1979, Alberch 1980, Both morphology and general life history Lc^vtnip 1981a, 1981b, Rachootin and Thom- of kangaroo rats lead us to suggest that son 1981). In other words, regulatory changes paedomorphosis in this group is the result of in ontogeny may have affected the timing of neoteny. Neoteny is fundamentally different gene action and rates of morphogenesis and from progenesis and involves retardation in growth and thus have led to morphological growth rate resulting in juvenilization of the phyletic evolution in this group. adult animal. According to Gould's (1977) hy-

Most authors would agree that pocket pothesis, neoteny is a common occurrence mice, including Perognothus, Liomijs, and and may result from direct selection for juve- Heteromijs, exliibit a generalized body plan nile features and/or larger body size in envi- and probably represent a good approx- ronmental regimes that are more iC -selected. imation of the ancestral geomyoid morpho- Indeed, kangaroo rats are all medium to large type (e.g., Eisenberg 1981:90). Our hypoth- heteromyids possessing many of the classical esis suggests that perturbations in features associated with tlie neotenic syn- developmental "control parameters" in an- drome (e.g., long life span, slow devel- cestral (pocket-mouse-like) geomyoids, in- opment, small litters, brain; enlarged J. C. cluding changes in the time of onset of Hafner and M. S. Hafner, in manuscript). The growth, cessation of development, and rate of early evolution of kangaroo rats appears to growth, could deform the ancestral ontoge- have been confined to grasslands (Reeder netic pathway and effect phylogenetic trans- 1956), and several living forms inhabit fairly mutations in morphology leading to such di- stable, nondesert environments; those tliat verse forms as kangaroo mice, kangaroo rats, are strictly desert-dwelling forms tend to buf- and pocket gophers. fer an otlierwise luistable environment by According to our hypothesis, paedomor- subsisting on a fairly stable resource (leaves) phosis in kangaroo mice and kangaroo rats and/or utilizing large stores of hoarded food may have originated via different kinds of de- during periods of food scarcitv. velopmental perturbations. Hoth morphology Lastly, we suggest that pocket gophers and general life history of kangaroo mice sug- may be hypermorphic descendants of a gest that they are progenetic descendants of pocket-mouse-like ancestor. Hypermorphosis pocket-mouse-like ancestors. Progenesis is is a process wherein heterochronic per- the process whereby ontogeny is truncated turbation in ontogeny has led to a length- 1983 Biology of Desert Rodents 23

ened growth period producing a "per- suspect and in need of thorough reevaluation. amorphic" (as opposed to paedomorphic; Finally, we provide a synonymy for Chaeto- Alberch et al. 1979) organism. The marked dipus, which we herein evaluate to full ge- similarities between neonatal pocket gophers neric rank. Our views on supraspecific rela- and adult pocket mice suggest that reproduc- tionships within the family are summarized tive maturation in pocket gophers has been in Figure 9. retarded relative to exomorphological maturation, allowing for extreme development of Subfamily Dipodomyinae Coues 1875 somatic features prior to reproduction. Adult pocket gophers possess many of the classical Genus Dipodomys Gray 1841.— Hall features characteristic of late mammalian on- (1981) recognizes 22 species of kangaroo rats, togeny, most notably a rugose, heavily os- partitioned into six species groups following sified skeleton that results from prolongation Setzer (1949; see Schnell et al. 1978, for dis- of somatic growth late in life. cussion of interspecific taxonomy in the In summary, we suggest that many, per- genus). Best (1978) has used morphologic cri- haps most, of the flamboyant morphological teria to suggest that three species recognized modifications seen in heteromyids and geo- by Hall, D. antiquarius Huey 1962, D. para- myids may have originated as incidental by- liits Huey 1951, and D. penisularis (Merriam products of heterochronic shifts in ontogeny. 1907), are conspecific with D. agilis Gambel Instead of acting separately on each mor- 1848. Biochemical (Johnson and Selander phological feature (e.g., large head, long legs, 1971) and morphological (Schmidly and delicate skeleton, etc.), natural selection may Hendricks 1976) evidence has been used to have acted at the developmental level favor- suggest that D. compactus True 1899 (not ing the new life history strategy associated recognized by Hall 1981) warrants full spe- with a progenetic, neotenic, or hypermorphic cific status. Similarly, Patton et al. (1976) animal. The paedomorphic traits invariably have used biochemical evidence to document associated with these developmental changes specific status for D. californicus Merriam (peramorphic traits in the case of hypermor- 1890. phosis) may have been "tolerated" by natural Genus Microdipodops Merriam 1891.— selection initially and subsequently modified Herein, kangaroo mice are placed provision- (adapted) for fimctional purposes. On the ally in the subfamily Dipodomyinae, based other hand, certain novel traits may provide largely on biochemical evidence (M. S. Haf- little or no functional improvement over the ner 1982) and comparative analysis of the ancestral condition, but are maintained (as male reproductive system (this study). We re- long as they are not deleterious) because they gard the many morphological similarities be- are developmentally mandated. tween Microdipodops and Perognathus (Wood to shared primi- 1935, J. C. Hafner 1978) be tive features. Nevertheless, more detailed Taxonomic Comments on the analyses the molecular level are needed to Heteromyidae Allen and Chapman 1893 at confirm or refute the placement of Micro-

The following is an up-to-date statement dipodops within the Dipodomyinae. Two spe- on the taxonomy of the Recent Hetero- cies of Microdipodops are recognized by Hall myidae, necessarily reflecting the views and (1981). Hall's suggestion (1981:560) that M. biases of the authors. We use as our point of rnegacepJiahis leticotis may warrant specific departure Hall (1981) and included refer- status is not supported by chromosomal or ences. In interest Hafner 1981a). the of brevity, those hetero- protein evidence (J. C. myid species recognized by Hall (1981) are not listed, except where we feel taxonomic Subfamily Heteromyinae Coues 1875 comments are warranted. For species groupings within each genus we again refer the Genus Heteromys Desmarest 1817.— Hall reader to Hall (1981), but we hasten to add (1981) recognizes 10 species of Heteromys that species groupings within Dipodomys, partitioned into two subgenera, Heteromys Hetemmys, Perognathus, and Chaetodipus are and Xylomys. Heteromys nigricaudatus 24 Great Basin Naturalist Memoirs No. 7

Goodwin 1956 (recognized by Hall 1981) was ing the problematic species P. formosus Mer- synonymized under H. lepturus Merriam riam 1889 (see Osgood 1900). Patton et al. 1902 by Goodwin (1969). Rogers and (1981) have shown that P. formosus is clearly Schmidly (1982) have analyzed phenetic rela- referable to Chaetodipus, thus reducing the tionships in Hall's (1981) desmarestianus number of Recent taxa of Perognathus to group (exclusive of H. gaumeri) and have syn- nine. onymized H. longicaudatus Gray 1868, H. Genus Chaetodipus Merriam 1889.— temporalis Goldman 1911, and H. lepturus CJmetodipus is here elevated to full generic under H. desmarestianus Gray 1868. status. Following is a synonymy and a rediag- Genus Liomijs Merriam 1902.— We have nosis of the taxon. reviewed a large body of evidence supporting the phyletic affinity of Liomijs and Heter- Genus Cluietodipus Merriam, new status omijs. Hall (1981) follows Genoways (1973) in Chaetudipus Merriam, 1889, N. Amer. Fauna 1:5 (25 Oc- recognizing five species in the genus Liomys. tober). Type-species: Perognathus spinattis Mer- Future biochemical and chromosomal band- riam, original designation. Subgenus elevated to ing studies may serve to identify species generic level. groups within the genus. Diagnosis.— Size medium to large (total length 152-230; Hall 1981). Pelage generally harsh, often with spiny bristles on rump. In Subfamily Perognathinae Coues 1875 those species lacking harsh pelage and rump Genus Perognathus Wied-Neuwied spines, tail crested distally and longer than 1839.— Our elevation of Chaetodipus to full head and body. Hair flattened in cross section generic rank (see beyond) results in the in- with distinct trough having well-developed clusion of only "silky pocket mice" (formerly ridges on dorsal surface (Homan and Gen- referred to the subgenus Perognathus) within oways 1978). Soles of hind feet naked. In- the genus Perognathus. Hall (1981) recog- terparietal width equal to or greater than in- nizes 10 species of silky pocket mice, includ- terorbital width. Auditory bullae not

EOCENE .OLIGOCENE , 1983 Biology of Desert Rodents 25

markedly inflated as in rcwgnatiuts. Male ing the course of this study and for critical phallus long and slender, lacking urethral evaluations of the manuscript. We thank M. lappets; rim of terminal crater forms ventlike D. Carleton, A. L. Gardner, and R. W. Thor- inethral opening. Baculum long relative to ington, Jr., for helpful discussions. It is our body length (Burt, 1936), having a narrow pleasure to thank P. M. Hafner, who gener- proximal end and sharply upturned distal end ously provided encouragement and clerical

(the baculum of C hispidus is unique in pos- assistance. The Smithsonian Institution is sessing an ornate, trifid tip). Head of sperma- gratehilly acknowledged for support to JCH tozoa resembling elongated isosceles triangle during the early phases of this study. Finally, with acute, imrounded vertices. Vesicular we dedicate this contribution to our valued glands roimd or bulb shaped, yellow to gray friend and mentor. Professor James L. Patton. in color (pinkish in fresh specimens). Included taxa.— Chaetodipus anthonyi, C. arenarius, C. artus, C. baileyi, C. califor- Literature Cited niciis, C. dalqucsti, C. fcdlax, C. forrnosus, C. Alberch, p. 1980. Ontogenesis and morphological diver- goldmani, C. hispidus, C. intermedins, C. sification. Amer. Zool. 20:653-667.

P., S. lincatus, C. nelsoni, C. penicillatus, C. pernix, Alberch, J. Gould, G. F. Oster, and D. B. Wake. 1979. Size and shape in ontogenv C. spinatus. Chaetodipus anthonyi Osgood and phvlogenv. Paleobiology 5;296-317. 1900 is most likely an island race of C. fallax .Alston, E. R. 187o. On the classification of the Order L. Patton, pers. comm.). Merriam 1889 (J. Glires. Proc. Zool. Soc. London 1876:61-98. The specific status of C. lineatus Dalquest .\>;derson, S. 1964. The systematic status of Perognathus artus and Perognathus (Rodential. 1951 is dubious (Patton et al. 1981), and fur- goldmani .\mer. Mus. Novitates 2184:1-27. ther study is necessary to confirm the validity Arata, \. A. 1964. The anatomy and taxonomic signifi- of C. dalquesti a kar- Roth 1976, which has cance of the male accessory reproductive glands yotype identical to C. arenarius Merriam of nniroid rodents. Bull. Florida State Mus. 1894 S. 9:1-42. (J. C. Hafner and M. Hafner, unpubl. .\xelrod, D. I. 1950. Evolution of desert vegetation in ms.). According to Patton et al. (1981), no western North America. Carnegie Inst. Wasliing- strongly delineated species groupings are ap- ton, Publ. 590:217-306. parent within Chaetodipus. However, C. his- 1958. Evolution of the Madro-Tertiarv Cieotlora. pidus is biochemically divergent from other Bot. Rev. 24:433-509. members of the group. 1976. History of the coniferous forest, California and Nevada. Univ. California Publ. Bot. 70:1-62. Comments.— Extreme evolutionary con- Baird, S. F. 1858. Mammals. Reports of explorations and servatism in in exomorphology the genera smvey to ascertain the most practicable and eco- Perognathus and Chaetodipus belies their nomical route for a railroad from the Mi.ssi.ssippi long history of evolutionary independence as River to the Pacific Ocean 8:1-757. Barer, R. S. L. Williams, and C. P.atton. 1973. documented by fossil (Wood 1935) and bio- J., J. Chromosomal variation in the plains pocket chemical evidence (Sarich 1975, Patton et al. gopher, Geomijs bursarius major. J. Mammal. 1981, M. S. Hafner 1982). M. S. Hafner 54:765-769.

L., R. 1972. Chromosomes of (1982) has shown that the genetic distance Berry, D. and J. Baker. between Perognathus and Chaetodipus ex- pocket gophers of the genus Pappogeomijs, subgenus Cratogeomijs. Mammal. 53:303-309. ceeds that measured between Liomys and J. Best, T. L. 1978. Variation in kangaroo rats (genus Di- Heteromys and approaches that measured be- podomtjs) of the heermanni group in Baja Califor- tween Dipodomys and Microdipodops. Tax- nia, 59:160-175. Mexico. J. Mammal. onomic consistency within the Heteromyidae Best, T. L., and G. D. Schnell. 1974. Bacular variation in kangaroo rats (genus Dipodomys). .\mer. .Midi. demands that the high degree of chromosom- Nat. 91:257-270. al, anatomical, and biochemical differen- F. 1855. Beitrage zur nahern kenntni.ss der Br\ndt, J. tiation between Perognathus and Chaetodipus saugethiere Russlands. Mem. Acad. Imp. Sci. St. be recognized at the generic level. Petersbourg, ser. 6, 9:1-375. Burt, W. H. 1936. A study of the baculum in the genera

Perognathu.-i and Dipodomys. J. Mammal. Acknowledgments 17:145-156. 1960. Bacula of North American mammals. are most grateful Zool.. Univ. Michigan 113:1-76. We to D. J. Hafner and J. Misc. Publ. Mus. L. Patton for their advice and criticism dur- Caire, W. 1976. Phenetic relation,ships of pocket mice No. 7 26 Great Basin Naturalist Memoirs

origin and homology of the in the subgenus Chaetodipus (Rodentia: Hetero- GuNTHER, W. C. 1960. The

57:375-378. male reproductive system in Tliomomys bottae. J. myidae). J. Mammal. V. 1868. Handbuch der Zoologie. Saugethiere Mammal. 41:243-250. Carus, J. 1:39-191. C. Hafner, and M. S. Hafner. 1979. Rafner, D. J., J. .\rme.ntrout. 1979. Neogene pa- of kangaroo mice, genus Micro- Cole, M. R., and J. M. Systematic status

leogeography of the western United States. In J. dipodcrps: morphometric, chromosomal, and pro- R. Cole, and H. Terbest, Jr., 60:1-10. M. Armentrout, M. tein analyses. J. Mammal. the western eds. Cenozoic paleogeogrpahy of C. 1976. Specific and evolutionary relation- Rafner, J. of Economic Paleontolo- United States. Society ships of kangaroo mice, genus Microdipodops. (Pacific Section), Los gists and Mineralogists Unpublished thesis. Texas Tech. Univ., Lubbock. Angeles. 139 pp. Sacconiyidae. Monographs of North CouES, E. 1877. 1978. Evolutionar) relationships of kangaroo .\merican Rodentia. Rept. U.S. Geol. Surv. Terr. mice, genus Microdipodops. J. Mammal. 9:481-542. 59:354-366. C. 1931. A comparative study of the chromo- Cross, J. 1981a. Evolution, systematics, and historical somes of rodents. Morphol. 52:373-401. J. biogeography of kangaroo mice, genus Micro- 1971. Karyotypes of kangaroo rats from CsuTi, B. A. dipodops. Unpublished dissertation, Univ. of Cal- southern California, Mammal. 52:202-206. j. ifornia, Berkeley. 269 pp. 1979. Patterns of adaptation and variation in the 1981b. Directionality in chromosomal evolution Great Basin kangaroo rat (Dipodomifs microps). of geomyoid rodents. No. 124, in Abstracts of Univ. California Publ. Zool. 3:1-69. technical papers presented at the 61st annual Davis, B. L., S. L. Williams, and G. Lopez. 1971. meeting of the American Society of Mammalo- Chromosomal studies of Geoimjs. J. Mammal. gists, Miami Univ., Oxford, Ohio. .52:617-620. R^fner, D. Hafner, and L. Patton. 1980. E., Davis, a.\d D. IL 1971. J. C, J. J. Di.\G.M.\.N, R. R. J. Hunsaker, Anomalous karyotypic variation and hybridiza- Karyotypes of the panamint kangaroo rat (Di- tion in gopher, Thoinomys bottae. No. podoinys panamintinus [MeTna.m]). Experientia the pocket in Abstracts of papers posters, 60th an- 27:1491. 177, and meeting, American Society of Mammalo- F. 1981. Tlie mammalian radiations: an nual EisENBERG, J. gists, Univ. of Island, Kingston. analysis of trends in evolution, adaptation, and Rhode behavior. Univ. of Press, Chicago, 610 L. F. Chicago Hafner, J. C, D. J. Hafner, J. Patton, and M. pp. Smith. 1983. Contact zones and the genetics of N. 1973. Implications of karyotypic variation differentiation in the pocket gopher. Fashinc;, J. Tliomomys rat, in the kangaroo Dipocloinys heermanni. J. bottae (Rodentia: Geomyidae). Svst. Zool. Mammal. 54:1018-1020. 32:1-20.

R. 1971. Glacial geology. S. 1979. relationships Fli.nt, J. and Quaternary John Hafner, M. Evolutionary and sys- Wiley and Sons, Inc., New York. 892 pp. tematics of the Geomyoidea (Mammalia: Ro- Genoways, H. H. 1973. Systematics and evolutionary dentia). Unpublished dissertation. Univ. of Cali- relationships of spiny pocket mice, genus Liomys. fornia, Berkeley. 97 pp. Special Publ., Mus. Texas Tech. Univ. 5:1-368. 1982. A biochemical investigation of geomyoid Gill, T. H. 1872. Arrangement of the families of mam- systematics (Mammalia: Rodentia). Z. Zool. Svst. mals, with analytical tables. Prepared for the Evolut.-forsch. 20:118-130. Smithsonian Institution. Smiths. Misc. Cloll., no. Hall, E. R. 1941. Revision of the rodent genus Micro- 2.30:1-198. dipodops. Field Mus. Nat. Hist., Zool. Ser.

Goldman, E. A. 191 1. Revision of the spiny pocket mice 27:233-277. (genus Heteromys and Lioniys). N. Amer. Fauna 1981. The mammals of North .\merica. John

.34:1-70. Wiley and Sons, New York. 2 vols. 1 181 pp. Goodwin, G. G. 1969. Mammals from the state of Oa- Hart. E. B. 1978. Karyology and evolution of the plains xaca, Mexico, in the -Xmerican Museum of Natu- pocket gopher Ceomys bursarius. Occas. Papers ral Historv. Bull. Amer. Mus. Nat. Hist. Mus. Nat. Hist., Univ. of Kansas, 71:1-20. ' 141:1-270. R\RT, E. B., and R. R. Patterson. 1974. Chromosomes (i(JiLD, S. 1977. Ontogeny and phylogeny. Harvard J. of the yellow-faced pocket gopher, Pappogeomys Univ. Press, Cambridge, Massachusetts. 501 pp. castanops, in Kan.sas. Trans. Kansas Acad. Sci. Ck)i;LD, S. S. J., AND E. Vrba. 1982. Exaptation— a miss- 77:239-243. ing term in the science of form. Paleobiologv llArr, R. T, 1932. The vertebral columns of ricochetal 8:4-15. rodents. Bull. Amer. Mus. Nat. Hi.st. 63:599-738. E. Gray, J. 1868. Synopsis of the species of Sacconiyidae

Hi-.NMc, W . 1966. Phylogenetic systematics. Unix, of Il- or pouched mice in the collection of the British linois Press, Urbana. 263 Museum. Proc. Zool. Soc. London 1868:199-206. pp. (iRiNNELL. Hkhsiikovitz, p. 1966. South fos- J. 1921. Revised list of the species in the American swamp and

genus Dipodomys. ]. Mammal. 2:94-97. sorial rats of the scapteromyine group (Cri- 1922. A geographical study of the kangaroo rats cetinae, Miuidae) with conunents on the glans of California. Univ. C;aiiforiiia Publ. Zool. penis in nuirid taxonomx. /. Saugetier. 24:1-124. 31:81-149. 1983 Biology of Desert Rodents 27

HiBBARD, C. \V. 1941. New niamnuils from the Rexroad 1981b. The epigenetic utilization of the ge- nomic message. In G. G. E. Fauna, Upper Pliocene of Kansas. Anier. Midi. Scudder and J. L. Re- Nat. 26:337-368. veal, eds. Evolution today. Hunt Inst, for Bot. A., H. H. Genoways. 1978. An analysis Documentation, Carnegie-Mellon Univ., HoMAN, J. AND of hair structure and its phylogenetic implica- Pittsburgh. Matthey, R. 1952. Chromosome de Muridae (II). Expe- tions among heteromyid rodents. J. Mammal. 59:740-760.^ rientia 8:389-390.

L., 1979. Chrouio- 1956. Nouveau apports a la cytologic comparee HoNEYCUTT, R. AND D. J. ScHMiDLY. sonial and morphological variation in the plains des rongeurs. Chromosoma 7:670-692. pocket gopher, Geomys bursarius, in Texas and 1973. The chromosome fornnilae of eutherian adjacent states. Occas. Papers Mus., Texas Tech. mammals. In A. B. Chiarelli and E. (^apanna, Univ. 58:1-54. eds. Cytotaxonomy and vertebrate evolution. Ac- HoNEYCuTT, R. L., AND S. L. WiLLiAMS. 1982. Gcnic dif- ademic Press, New York. ferentiation in pocket gophers of the genus Pap- Merrl\m, C. H. 1889. Revision of North American pock- pogeomys, with comments on intergeneric rela- et mice. N. Amer. Fauna 1:1-29. 1891. Description of a new genus and species of tionships in the subfamily Geomyinae. J. Mammal.63:208-217. dwarf kangaroo rat from Nevada (Microdipodops Hooper, E. T. 1946. Two genera of pocket gophers megacephalus). N. Amer. Fatma 5:115-117. congeneric. 27:397-399. 1895. Monographic revision of the pocket go- should be J. Mammal. 1961. The glans penis in Proechimys and other phers, family Geomyidae (exclusive of the species caviomorph rodents. Occas. Papers Mus. Zool., of Tliomomys). N. Amer. Fauna 8:1-258. 1918. Synopsis of the Univ. of Michigan 623:1-18. Miller, G. S., and J. W. Gidley.

1962. The glans penis in Sigmodon, Sigmomys, supergeneric groups of rodents. J. Wa.shington and Reithiodon (Rodentia, Cricetinae). Occas. Acad. Sci. 8:341-448.

Papers Mus. Zool., Univ. of Michigan 625:1-11. S. 1979. The archaeological fauna of four sites Miller, J. Howell, A. B. 1932. The saltatorial rodent Dipodomys: in Smith Creek Canyon. Nevada St. Mus., An- the fimctional and comparative anatomy of its thro. Papers, No. 17. muscular and osseous systems. Proc. Amer. Acad. MuNTHE, K. 1975. The osteology of the Miocene rodent Arts and Sci. 67:377-536. Schizodontomys. Unpublished thesis. Univ. of Hsu, T. C., AND K. Benirschke. 1967-1977. An atlas of California, Berkeley. mammalian chromosomes. Vols. 1-10. Springer- Nevo, E. 1978. Genetic variation in natural populations: York. Verlag. New patterns and theory. Theor. Pop. Biol. 13:21-177. Imal H. T., and R. H. Crozier. 1980. Quantitative anal- Osgood, W. F. 1900. Revision of the pocket mice of the ysis of directionality in mammalian karyotype genus Perognathus. N. Amer. Fauna 18:1-63. evolution. Amer. Nat. 116:537-569. L. Chromosomal studies of certain Patton, J. 1967a. Johnson, VV. E., and R. K. Selander. 1971. Protein vari- pocket mice, genus Perognathus (Rodentia: Het- systematics in kangaroo rats (genus ation and 48:27-37. eromyidae). J. Mammal. Dipodomy Syst. Zool. 20:377-405. si 1967b. Chromosomes and evolutionary trends in Kelly, T. S. 1969. The comparative morphology of the the pocket mouse subgenus Perognathus (Ro- male phallus in the genus Dipodomys. Unpub- dentia: Heteromyidae). Southwestern Nat. lished thesis. San Fernando Valley State College, 12:429-438. Los Angeles, California. 133 pp. 1969. Chromosome evolution in the pocket Laguarda-Figueras, a., Romero-Jarero, Paulete- J. J. mouse, Perognathus goldmani Osgood. Evolution VaNRELL, and S. SCAGLIA DE PaULETE-VaNRELL. 23:645-662. 1971. Cytogenetic analysis of Pappogeomys 1970. Karyotypes of five species of pocket mice, (Cratogeomys) merriami merriaini. Mammal. genus Perognathus (Rodentia: Heteromyidae), Chromo. Newsletter 12:129-130. and a summary of chromosome data for the Lewin, R. 1982. Adaptation can be a problem for evolu- genus. Mammal. Chromo. Newsletter 11:3-8. tionists. Science 216:1212-1213. 1972. A complex system of chromosomal varia- Lidicker, W. Z., 1960. An analysis of intraspecific Jr. tion in the pocket mouse, Perognathus baileyi variation in the kangaroo rat Dipodomys mer- Merriam. Chromosoma 36:241-255. riami. Univ. California Publ. Zool. 67:125-218. 1973. An analysis of natural hybridization be- 1968. A phylogeny of New Guinea rodent gen- tween the pocket gophers, Thomomys bottae and era phallic based on morphologv. J. Mammal. Thomomys unibrinus, in Arizona. J. Mammal. 49:609-&43. 54:561-.584. Lindsay, E. H. 1972. Small mammal fossils from the 1980. Chromosomal and genie divergence, pop- Barstow Formation, California. Univ. California ulation structure, and speciation potential in Publ. Geol. Sci. 93:1-104. Tliomomys bottae pocket gophers. In O. A. Reig, Linnaeus, C. 1758. Systema Naturae. Regnum Animale animal. ed. Ecologia y genetica de la especiacion (10th ed. tomus I; L. Salvii, Holminae). Univ. of Simon Bolivar, Caracas, Venezuela. 295 LcivTRUP, S. 1981a. Introduction to evolutionary epige- pp. netics. In G. G. E. Scudder and L. Reveal, eds. L., R. E. Dincman. 1968. Chromosome J. Patton, J. and Evolution today. Hunt Inst, for Bot. Documenta- studies of pocket gophers, genus Thomomys. I. tion, Carnegie-Mellon Univ., Pittsburgh. Tlie .specific status of Thomomys umbrinus (Rich- 28 Great Basin Naturalist Memoirs No. 7

in 49:1-13. Selander, R. K., D. W. Kaufman, R. Baker, and S. L. ardson) Arizona. J. Mammal. J. 1970. Chromosome studies of pocket gophers, Williams. 1974. Genie and chromosomal differ-

genus Tftomomys. II. Variation in T. bottae in the entiation in pocket gophers of the Geomijs bur- American southwest. Cytogenetics 9:139-151. sariiis group. Evolution 28:.557-564. Patto.v, L., H. Feder. 1978. Genetic diver- ScHNELL, G. D., T. L. Best, and M. L. Kennedy. 1978. J. .\ND J. gence between populations of the pocket gopher, Interspecific morphological variation in kangaroo Tliomomtjs unihrinu.s (Richardson). Z. Saugetierk. rats (Dipochmys): degree of concordance with 43:17-30. genie variation. Syst. Zool. 27:34-48.

L., a.nd S. Y. 1976. Setzer, H. W. 1949. Subspeciation in the kangaroo rat, Patto.n, J. H. MacArthur, Yang. Systematic relationships of the four-toed popu- Dipodomys ordii. Univ. Kansas Publ., Mus. Nat. Hist. 1:473-573. lations of Dipodonnjs heermanni. J. Mammal. A. 1967. Tertiary geomyoid rodents 57:159-163. Shotwell, J. Late L., R. K. Sel.\nder, 1972. of Oregon. Univ. Oregon Mus. Nat. Hist. Bull. Patton, J. and M. H. Smith. Cenic variation in hybridizing populations of 9:1-51. gophers (genus Thomonnjs). Svst. Zool. Simpson, G. G. 1931. A new classification of mammals. 21:26,3-270. Bull. Amer. Mus. Nat. Hist. 59:259-293.

L., S. S. 1945. The principles of classification and a clas- Patton, J. W. Sherwood, and Y. Yang. 1981. Biochemical systematics of chaetodipine pocket sification of mammals. Bull. Amer. Mus. Nat. mice, Hist. 85:1-350. genus Perognathus. J. Mammal. R. 62:477-492. Smolen, M. J., H. H. Genoways, and J. Baker. 1980. L., S. Y. Demographic and reproductive parameters of the Patton, J. a.nd Yang. 1977. Genetic variation in nwrnomys bottae pocket gophers: macrogeo- yellow-cheeked pocket gophers (Pappogeomys castanops). graphic patterns. Evolution 31:697-720. J. Mammal. 61:224-236. Rachootin, S. p., a.nd K. S. Thomson. 1981. Epigenet- Stebbins, G. L. 1981. Coevolution of grasses and herbi- ics, paleontology, and evolution. In G. G. E. vores. .Ann. Missouri Bot. Card. 68:7.5-86. Scudder L. eds. Stevens, P. F. 1980. Evolutionary polarity of character and J. Reveal, Evolution today. Hunt Inst, for Bot. Documentation, Carnegie- states. Ann. Rev. Ecol. Syst. 11:333-358. Mellon Univ., Pittsburgh. Stock, A. D. 1974. Chromosome evolution in the genus

Reeder, W. G. 1956. A review of Tertiarv rodents of the Dipodomys and its taxonomic and phylogenetic family implications. Heteromyidae. Unpublished dissertation. J. Mammal. 55:505-526. Univ. of Michigan, Ann .\rbor. 618 pp. Thaeler, C. S., Jr. 1968. Karyotypes of sixteen popu-

Rensberger, J. M. 1971. Entoptychine pocket gophers lations of the TJiomoniys talpoides complex of (Mammalia, Geomyoidea) of the Early Miocene pocket gophers (Rodentia-Geomyidae). Chromo- John Dav Formation, Oregon. Univ. California soma 25:172-183. Publ.Geol.Sci. 90:1-209. 1972. Taxonomic status of the pocket gophers, 1973. Pleurolicine rodents (Geomyoidea) of the Viomomys idahoensis and Thoniomys pygmueus John Day Formation, Oregon. Univ. (Rodentia, Geomyidae). 53:417-428. California J. Mammal. Publ. Geol. Sci. 102:1-95. 1973. The karyotype of Thomomys townsendii

Robertson, W. R. B. 1916. Chromosome studies. I. Tax- similis and comments on the karyotypes of Tho- onomic relationships .shown in the chromosomes momys. Southwest Nat. 17:327-331. of Tettigidae and Acrididae: V-shaped chromo- 1974a. Four contacts between ranges of differ- somes and their significances in Acrididae, Lo- ent chromosome forms of the Thomomys tal- custidae, and Gryllidae: chromosomes and evolu- poides complex (Rodentia: Geomvidae). Svst. tion. J, .Morph. 27:179-331. Zool. 23:.343-.3,54. Rogers, D. S., and D. J. Schmidly. 1982. Systematics of 1974b. Karyotypes of the Thomomys talpoides spiny pocket mice (genus Heteromys) of the des- complex (Rodentia: C^eomyidae) from New Mexi- marestianus species group from co. Mexico and J. Mammal. 55:855-859. northern Central .America. J. Mammal. 1976. Chromosome polymorphism in Thomomys R3:375-.386. talpoides agrestis Merriam (Rodentia-Geo- Russell, R. J. 1968. Evolution and cla.ssification of the myidae). Southwest. Nat. 21:105-116. pocket gophers of the subfamily Geomvinae. 1977. Taxonomic status of Thomointis talpoides Univ. Kansas Publ., Mus. Nat. Hist! 16:47.3-579. confintis. Murrelet 58:49-50. Sarioh, V. M. 1975. Geomyoid systematics: albumin and 1980. Chromo.sonie numbers and systematic re- transferrin immunology. In Abstracts of technical lations in the genus Thomomys (Rodentia: Geo- papers presented at the 55th annual meeting of myidae). J. Mammal. 61:414n422. the American Society of Mammalogists, Univ. of Thaeler, C. S., Jr., and L. L. Hinesi.ey. 1979. Tho- Montana. 117 pp. momys clusius, a rediscovered species of pocket F., gophers. Schitoskey, Jr. 1968. Notes on morphological varia- J. Mammal. 60:480-488.

tion in the dark kangaroo mouse. Southwestern Wahlert, J. H. 1978. Cranial foramina and relation- Nat. 13:24,3-248. ships of the Eomyoidea (Rodentia, Geomorpha). SrHMn)LY. D. Skull and upper teeth of Kansasimys. .Amer. J.. AND F. S. Hendric:ks. 1976. Systematics Mu.s. Novitates 2645:1-16. of the southern races of Ords kangaroo rat, Di- W'.vterhouse, (;. R. pocUmnjs onlii. Bull. Soutluru California .Acad. 1848. A natural historv of the mam- Sci. 75:225-2.37. 1983 Biology of Desert Rodents 29

malia. Hippolyte Bailliere, London, 2 vols. 500 1935. Evolution and relationships of the hetero- pp. myid rodents with new forms from the Tertiary Weber, M. 1904. Die Saugetiere. Einfiihiiag in die Ana- of western North America. Ann. Carnegie Mus toniie und Systematik der Recenten and Fossilen 24:73-262. Mammalia. Jena, Giistav Fischer. 866 pp. 1936. A new subfamily of heteromyid rodents Williams, D. F. 1978. Karyological affinities of the spe- from the Miocene of western United States. cies groups of silky pocket mice (Rodentia: Het- Amer. Sci. 31:41-49. 59:599-612. J. eromyidae). J. Mammal. Williams, S. L., and H. H. Genoways. 1975. Karyotype 1955. A revised classification of rodents. J. of Geomijs pinetis (Mammalia: Geomyidae), with Mammal. 36:165-187. a discussion of the chromosomal relationships 1965. Grades and clades among rodents. Evolu- within the genus. Experientia 31:1141-1143. tion 19:115-130. Wilson, R. W. 1949a. On some White River fossil ro- Zakrzewski, R. 1981. dents. Carnegie Inst. Washington, Publ. J. Kangaroo rats from the Borchers 584:27-50. Local Fauna, Blancan, Meade County, Kan.sas. 1949b. Early Tertiary rodents of North America. Trans. Kansas Acad. Sci. 84:78-88. Carnegie Inst. Washington, Publ. 584:67-164. Zimmerman, E. G., and M. E. Nejtek. 1977. Genetics Wood, A. E. 1931. Phylogeny of the heteromyid ro- and speciation of three semispecies of Neotoma. dents. Amer. Mus. Novitates 501:1-19. J. Mammal. 58:391-402.