In the southern part of the range of _taylori_, intergradation occurs between _B. t. taylori_ in western Tamaulipas and _B. t. fuliginatus_ in the mountains of San Luis Potosi.

Dalquest (1953:156) found no indication of intergradation between the two species, _B. taylori_ and _B. musculus_, in San Luis Potosi. After examination of specimens from San Luis Potosi, I am in agreement that they are all referable to the species _taylori_.

_Specimens examined._--Total 435. TEXAS: _Montague County_: 3 mi. N Bowie, 1. _Dallas County_: 2 mi. NE Cedar Hill, 2. _Travis County_: 8 mi. NW Austin, 2[60]; _Austin_, 2[60]; _4 mi. E Austin_, 4[60]; _5 mi. E Austin_, 3[60]; _6 mi. E Austin_, 16[60], 1; _7 mi. E Austin_, 1[60]; _15 mi. E Austin_, 1[60]; _4 mi. S Austin_, 1[60]. _Bastrop County_: 25 mi. E Austin, 2. _Kendall County_: Boerne, 1[61].

_Bexar County_: _1 mi. N Randolph Field_, 3[64]; _5 mi. ENE_ (_on U. S. Highway 81_) _San Antonio_, 1; _3 mi. NE San Antonio_, 1; San Antonio, 26[61], 11[62], 1[63]; _5 mi. E San Antonio_, 11; _4-1/2 mi. E Sayers_, 3. _Gonzales County_: 7 mi. S Luling, 2[60].

_Wilson County: 4 mi. W LaVernia_, 3; 12 mi. W Floresville, 1.

_Atascosa County_: 9 mi. SW Somerset, 1. _Goliad County_: 8 mi.

NE Goliad, 1[60]. _Bee County_: Beeville, 1[61]. _Aransas County_: Aransas (Wildlife) Refuge, 1[65]; _5 mi. E Copana Bay_, 1[65]; _4.6 mi. NE Rockport_, 5[60]; _4.5 mi. NW Rockport_, 2[60]; 3 mi. N, 2 mi. E Rockport, 4; _Rockport_, 1[60], 1[61], 1[63]; _1-1/2 mi.

SW Rockport_, 1[60]; _2 mi. SW Rockport_, 2[60]; _13.4 mi. SW Rockport_, 1[60]; _14 mi. SW Rockport_, 1. _San Patricio County_: Welder Wildlife Refuge, 7. _Duval County_: type locality, 2[61], 1[66]. _Nueces County_: Corpus Christi (south Nueces Bay), 1[64] (Cleveland Mus. Coll.). _Kleberg County_: 2 mi. S Riviera, 3[65]. _Brooks County_: 3 mi. S Falfurrias, 2[65]. _Hidalgo County_: 6 mi. S McAllen, 17[60]. _Willacy County_: 28 mi.

E Raymondville, 10[65]. _Cameron County_: Brownsville, 31[61], 23[62], 5[64]. COAHUILA: 6 mi. SW San Geronimo, 32. NUEVO LEoN: Santa Catarina, 1[61]; 14 mi. N Monterrey, 1950 ft., 2[67]; Monterrey, 1[61]; 20 km. N General Teran, 3[64]. TAMAULIPAS: _Near Headwaters Rio Sabinas, 8 km. W, 10 km. N El Encino_, 400 ft., 1; Camargo, 5[61]; Charco Escondido, 20 mi.

S Reynosa, 3[67]; Matomoras, 5[61]; _Ejido Santa Isabel, 2 km. W Inter-American Highway_, 2000 ft., 7; Hidaglo, 7[61]; _Hda. Station Engracia_, 4[63]; 4 mi. N La Pesca, 1; 29 mi. N Ciudad Victoria, 1[67]; Ciudad Victoria, 6[61], 3; Jaumave, 2400 ft., 6[64], 10; Sierra de Tamaulipas, 3[64]; _25 mi. N El Mante, 3 km. W Inter-American Highway_ (_on Rancho Pano Ayuctle_), 300 ft., 4; _6 mi. N Gomez Farias_ (_on Rancho Pano Ayuctle_), 1; _5 mi. NE Gomez Farias_, 12[64], 1[62]; 70 km.

(by highway) S Ciudad Victoria, 2 km. W El Carrizo, 5[62], 2; Antigua Morelos, 5[64]; _6 mi. N, 6 mi. W Altamira_, 31; _5 mi. N, 5 mi. W Altamira_, 4; _Alta Mira_ (_Altamira_), 2[61]; 1 mi. S Altamira, 6; _10 mi. NW Tampico_, 1. SAN LUIS POTOSi: Ebano, 5[68]; _4 km. NE Ciudad Valles_, 1; Ciudad Valles, 1; _3 km. W Tamuin_, 1[68]; _Tamuin_, 6[68]; _Pujal_, 300 m., 1[64]. VERACRUZ: Tampico Alto, 50 ft., 1; Potrero Llano, 350 ft., 1; Ozulama, 2; Cerro Azul, 350 ft., 1.

_Marginal Records._--TEXAS: 3 mi. N Bowie; 2 mi. NE Cedar Hill; 25 mi. E Austin; 7 mi. S Luling; 8 mi. NE Goliad; Aransas (Wildlife) Refuge; 3 mi. N, 2 mi. E Rockport; Corpus Christi (South Nueces Bay); 2 mi. S Riviera; 28 mi. E Raymondville; Brownsville. TAMAULIPAS: Matomores; 4 mi. N La Pesca; 1 mi. S Altamira. VERACRUZ: Tampico Alto; Ozulama; Cerro Azul; Potrero Llano. SAN LUIS POTOSi: Ciudad Valles. TAMAULIPAS: Antigua Morelos; 70 km. S Ciudad Victoria, 2 km. W El Carrizo; Jaumave; Hidalgo.

NUEVO LEoN: 20 km. N General Teran; Santa Catarina. COAHUILA: 6 mi. SW San Geronimo. TEXAS: 9 mi. SW Somerset; Boerne; 8 mi. NW Austin.

[60] Coll. University of Texas.

[61] U. S. Nat. Museum (Biol. Surv. Coll.).

[62] American Museum of Natural History.

[63] Chicago Natural History Museum.

[64] Univ. Michigan, Museum of Zoology.

[65] Texas A & M Coop. Wildlife Res. Coll.

[66] Carnegie Museum.

[67] Univ. California, Mus. Vert. Zool.

[68] Museum of Natural History, Louisiana State University.

EVOLUTION AND SPECIATION

The history of the genus dates back to the early late Pliocene, but morphological change since then has been slight insofar as can be judged from lower jaws. _Baiomys_ seems to have been relatively conservative also in types of habitat occupied.

According to Wilson (1937:59), the late Pliocene was a time of decided expansion of myomorph rodents, more particularly cricetines.

Furthermore, at this time, the climate in the interior basin of southwestern North America presumably was becoming arid, if we can judge from the spread of elements of the Madro-Tertiary flora. Axelrod (1950:266) points out that the drier, continental climate initiated in the early Tertiary probably had its culmination in middle Pliocene time.

Some floras of early late Pliocene of the southwestern United States reflect a climate slightly cooler and more moist than the climates of the middle Pliocene. However, late Pliocene times reflect an arid climate. The flora of the southwestern interior basin of North America in early late to late Pliocene was intermediate between the previous gra.s.sland floras of the middle Pliocene and the savannah flora of upper Pliocene. Axelrod (_loc. cit._) suggests that this intermediate flora of the interior basin of southwestern North America resulted from the folding of the Cascades and uplifting of the Sierra Nevada and Peninsular ranges to the south. The development of these mountains produced greater aridity to the lee of the mountains, thus accounting for the gra.s.sland-savannah flora. Pygmy mice probably originated in that time, I judge in Mexico, and moved northward and southward in a gra.s.sland-savannah habitat that seemingly existed as far north as what is now Meade County, Kansas (where the Sawrock fauna lived). Further evidence for occupancy of a gra.s.sland-savannah habitat by ancestral pygmy mice stems from the distribution of the living species, _B.

taylori_, that at present occupies territory adjacent to parts of the Sonoran and Chihuahuan deserts. _B. taylori_ seems to be morphologically more specialized for life in an arid gra.s.sland than was _B.

sawrockensis_.

The geographic range of ancestral pygmy mice possibly extended farther south in late Pliocene time than the range of _B. musculus_ does now.

Anyhow, _B. sawrockensis_ of the early late Pliocene dwelt in a more mesic type of habitat than _B. musculus_ does, and such habitat may have existed from the Pacific lowlands of Central America to the Caribbean lowlands of northern South America (see Duellman, 1958:136, and Dunn, 1940:156) during late Pliocene times. An ancestral stock of hesperomine mice, not greatly different from _Baiomys_, may have emigrated from the North American continent into South America across the continuous land connection, which Simpson (1950:395) suggests was formed in the Chapadmalalan age (= Blancan age of North American terminology). The length of time of interchange of genes between northern and southern populations of mice across the Central American land connection probably was brief. Duellman (_op. cit._:129) pointed out that once the Panamanian portal was closed, the warm counter equatorial current, El Nino, combined with the uplifting of the Andes, began to produce heavy rain forests in Central America and northern South America in late Pliocene or early Pleistocene times. These forests presumably isolated the stock in North America from that in South America where the latter probably evolved rapidly into kinds that differed from one another and from _Baiomys_ in shape of body, type of pelage, and shape of skull.

Internal structures such as hyoid apparatus, auditory ossicles, and baculum remained almost unchanged, as for example in _Calomys_ now living in South America. The present resemblance in internal morphological features between it and _Baiomys_, I judge, reflects taxonomic relationships more accurately than do shape and conformation of body and skull that seem to respond more rapidly to external environmental changes. The cranial characters distinguishing _Baiomys musculus_ from _Calomys laucha_ are as follows: posterior lacerate foramina between second, rather than first, upper molars; parapterygoid fossa shallower; mesopterygoid fossa as wide or wider, instead of narrower, than parapterygoid processes; burr for attachment of superficial ma.s.seter muscle hypertrophied instead of well-developed. In other cranial characters studied, the two genera closely resemble each other. Such similarities of crania between _Calomys_ and _Baiomys_ may reflect convergence, but the total of internal and external morphological characters shared, I think reflects true relationships.

_Peromyscus_ has a large number of living and extinct species and exhibits a wide range of morphological variation, whereas _Baiomys_ has a small number (7) of species and exhibits a narrow range of morphological variation. The small number of known species of pygmy mice suggests their conservatism in elaboration of morphological characters.

Possibly this is because the habitat, or even the ecological niche, occupied in geological time by these mice was restricted, geographically and in kind. If the habitat of the pygmy mice oscillated between savannah and arid gra.s.sland, then an hypothesis can be made possibly accounting for the origin of species of these mice. My idea is that the geographical distribution of _Baiomys_ today reflects a predilection on the part of these mice for a relatively uniform warm climate. Therefore, in the past, in times of warmer continental climate, these mice moved toward favorable habitat northward from an area in central and northern Mexico. In cooler periods, the mice moved southward as habitats to the north became unfavorable.

Dr. W. B. Davis (_in. litt._) informs me that _B. taylori_ was uncommon in Brazos County, Texas, approximately 15 years ago, and suggests that the abundance there now of this mouse and my taking it in 1958 northward nearly to the southern border of Oklahoma reflects a definite movement northward. Movement in the same direction in late years has been suggested for the nine-banded armadillo and the hispid cotton rat (Hall, 1959:373) that are a.s.sociated with warm climates to the south. These movements possibly reflect only minor fluctuations of climate, but in a long period of warmth movements northward would be expected to be p.r.o.nounced and extensive.

Extinct species of _Baiomys_ may have originated as a result of extension northward of the geographical range and subsequent retreat southward of the northern populations, as follows: (1) the range of the genus moved northward in a warm period; (2) in cooler times, most of the mice in the north disappeared and only isolated colonies remained in small patches of remaining habitat still favorable to the mice; (3) the small populations of isolated pygmy mice after a time changed through mutations, recombinations and subsequent selection to a degree that prevented crossbreeding once populations from the south again moved northward and came in contact with previously isolated stocks; (4) then compet.i.tion caused further divergence in morphological characters. Such an hypothesis would account for the morphological differences between the extinct _B. kolbi_ and _B. rexroadi_. The extinct _B.

brachygnathus_, presumably a dweller of a xerophytic gra.s.sland, may have had its origin from a _B. minimus_-like stock in the manner outlined.

FORMATION OF THE RECENT SPECIES

The morphological difference between the extinct _B. minimus_ and the living _B. musculus_ is not great, and musculus seems to be the product of the _B. sawrockensis-B. minimus_ line of development. Morphological characters of the parental stock of the two living species, _musculus_ and _taylori_, may have been intermediate between those of _B. minimus_ and those of _B. musculus_. The princ.i.p.al part of the range of _Baiomys_ today is in Mexico, and probably was there through much of Pleistocene time. Extension northward of the species and retreat southward of those northern populations of pygmy mice would not only have left isolated populations in the north, but would have allowed the mice that retreated south to share a common gene pool. Therefore, populations of pygmy mice occurring to the south in central Mexico might be expected to maintain a relatively high degree of heterozygosity in morphological and behavioral characters. The occurrence of any physical or biotic barrier that would have separated this h.o.m.ogeneous group would be conducive to speciation.

There is evidence that a barrier occurred in the Pleistocene in central Mexico sufficient to separate the supposed interbreeding, relatively h.o.m.ogeneous populations of pygmy mice. According to Sears (1955:529) and De Terra _et al_. (1949:51), parts of the higher regions in the Valley of Mexico, and the transverse volcanic zone in central Mexico were glaciated. On the mountain Ixtaccihuatl, De Terra (_op. cit._:52) found evidence of four marked advances of ice, from oldest to youngest, as follows: Salto, ice advanced to 3100 meters; Xopano, ice at 3200-3300 meters; Trancas, ice to 3400 meters; Ayolotepito, ice to 4350 meters.

The Salto advance is correlated by De Terra (_loc. cit._) with the Iowan glacial period. The advance of ice down the mountain sides in the transverse volcanic zone was accompanied by cool moist climates or pluvial periods. Such climates probably altered habitat formerly suitable for _Baiomys_. There is no record of _Baiomys_ known to me exceeding 8000 feet in elevation, although the lower edge of the ice on Ixtaccihuatl is at approximately 15,300 feet (4600 meters, Sears, _loc.

cit._). Presumably, the advance of ice down the mountains forced the pygmy mice to move to lower alt.i.tudes. Pluvial conditions possibly rendered the habitat even at lower alt.i.tudes uninhabitable for the mice, with the result that none continued to live in the transverse volcanic zone, but only north and south thereof. Long-continued separation of these northern and southern segments allowed species formation to occur.

As climatic and habitat conditions became more favorable in central Mexico, the two species moved back toward each other, and eventually their geographic ranges overlapped.

An a.n.a.lysis of external and cranial characters of pygmy mice (see Figure 12) reveals that both species are essentially largest to the north and smallest to the south. There are exceptions to this cline in both species. For example, _B. taylori a.n.a.logous_ is a large subspecies; it lives allopatrically in the southern part of the range of the species.

_B. musculus pallidus_ is not the largest subspecies; it lives allopatrically in the northern part of the range of the species. In west-central Mexico, where the two species are sympatric, _B. taylori_ is smaller than elsewhere and _B. musculus_ is larger than elsewhere.

_B. t. a.n.a.logous_ lives in the mountains of the transverse volcanic zone in central Mexico. Its large size may be a result of the cooler climate in the mountains. _B. t. allex_, the smallest subspecies, lives sympatrically with _B. musculus musculus_ at lower elevations in west-central Mexico. The small size of _allex_ could be a result of the warmer climate of the lower elevations. _B. m. pallidus_, at lower elevations in southern Oaxaca, is smaller than other subspecies of _musculus_ to the south at higher elevations. _B. m. musculus_ lives at low elevations along the coast of west-central Mexico. Unlike _B. m.

pallidus_, _B. m. musculus_ is large at lower elevations. It occurs sympatrically with _B. t. allex_. It is my idea that during the period of separation, when the two species were evolving, larger subspecies evolved to the north or at higher alt.i.tudes where climates were cooler; smaller subspecies evolved to the south or at lower elevations; the two cognate species, _musculus_ and _taylori_, made contact at lower elevations where individuals of _taylori_ may have been smallest, but individuals of _musculus_ were not the largest of the species. The differences, therefore, between the two species in their initial contact probably were slight. Hybrids, if they occurred, were probably inviable, sterile, or ill-suited for occupancy of the habitat of either of the parental stocks. The occurrence of hybrids, therefore, would result in what geneticists call "gamete wastage," and any further divergence in the parental stock, either in external characters (size and shape of body and head), or behavior, useful in recognition of species, would be favored by natural selection (see Dobzhansky, 1951:225; and Koopman, 1950:147). The two species seem to have diverged more in external characters where they occur together than in areas where they live separately (see Figure 12). The two species could be confused if a sample of adults of _taylori_ from 7 mi. S La Belle, Jefferson County, Texas, were compared to a sample of adults of _musculus_ from Tehuantepec, Oaxaca (see Figure 12). No confusion in species ident.i.ty would arise, however, if a sample of adults was taken from the area where the two species live together (see Figure 12). Brown and Wilson (1956:49) pointed out that where two closely related species occur together, characters (morphological, ecological, physiological, or behavioral) of each species are easily distinguished. However, where the two species are allopatric, the two closely related species so resemble one another that the species are not easily distinguished. This phenomenon has been called "character displacement" by Brown and Wilson (_loc. cit._).

In the area where the two species of pygmy mice occur together, there seems to be a disparity in numbers between them. Hooper (1952a:91) has recorded the collection of both _B. musculus_ and _B. taylori_ in a single trap line. A series of pygmy mice collected from San Gabriel, Jalisco, contained one _taylori_ and 33 _musculus_; another sample from La Resolana, Jalisco, had a ratio of 25 _taylori_ to 6 _musculus_. The disparity in numbers where the two species occur together has been further substantiated by collections of the University of Kansas.

Possibly this disparity in numbers is a result of interspecific compet.i.tion. Hooper (_op. cit._:90) pointed out that where the range of _B. musculus_ (typical of arid tropical lowlands) meets that of _B.

taylori_ (typical of arid temperate highlands), the two geographic ranges interdigitate with parts of the range of _musculus_ extending into the highlands and parts of the range of _taylori_ extending into the lowlands. In the lowlands, _musculus_ may be better adapted to environmental conditions and, therefore, more successful in compet.i.tion with _taylori_ for available habitat. The reverse situation may exist in the highlands. Also, the fact that _musculus_ is more of a diurnal animal than is _taylori_ may account for the difference in numbers of individuals of the two species taken in trap lines. Many collectors set their traps in late afternoon or evening and retrieve them in early morning. Such a schedule might not yield many _musculus_. If interspecific compet.i.tion does occur in the area where the two species occur, any change in habits or microhabitat by either species that would reduce this compet.i.tion would be favored by natural selection (see Mayr, 1949:518; Lack, 1944:262-263; and Brown, 1958:154-155). Brown (_op.

cit._:154), as I understand him, pointed out (taking account of Gause's principle) that when two species having similar ecological valences move into the same niche in the same locality, one of three things must eventually happen: (_a_) the two species occupy different geographic ranges; (_b_) they compete and one is eventually eliminated; (_c_) the two species, because of differentiation or specialization, exploit different aspects of the niche. In _Baiomys_, (_c_) seems to apply.

Natural selection probably would favor a continuation of diurnal activity in _musculus_ and nocturnal activity in _taylori_, thereby preventing frequent meeting of the two species.

AREAS OF PRESENT DIFFERENTIATION

In both species of _Baiomys_, the most distinct subspecies, _B. t.

allex_ and _B. m. musculus_, occur in the area where the two species are sympatric. Seven subspecies, or 44 per cent, occur either in or adjacent to the transverse volcanic zone. This area is the major area of active differentiation. Incipient subspecies are also evident in these areas. A secondary area of differentiation is indicated within the range of _B.

musculus_ in Guatemala, El Salvador and Honduras. Three subspecies occur in this area (_grisescens, handleyi_ and _nigrescens_) and incipient subspeciation is in evidence there.

ZOOGEOGRAPHIC POSITION

Hooper (1949:25) regards _Baiomys_ as a member of the rodent fauna of the arid, western Sonoran region, whereas Hershkovitz (1958:609) suggests that _Baiomys_ is a nearctic-neotropical varicant (a kind that occurs in contiguous zoogeographic regions without our knowing in which region the taxon originated). The findings from my study do not contradict either of the above suggestions. Because of the close resemblance of _Baiomys_ to certain hesperomine mice of South America, it is postulated that _Baiomys_, in more primitive form than now, occurred farther south in past times than it does now. Fossils show that primitive stocks of the genus in late Pliocene or early Pleistocene times occurred also north of the present range of the genus. The belt in west-central Mexico between nearctic and neotropical regions is the current center of distribution of the genus and probably has been for a considerable time.

[Ill.u.s.tration:

FIG. 12. Averages of the occipitonasal lengths of skulls of adults at 19 localities of occurrence (solid symbols) of _Baiomys taylori_, and at 17 localities of occurrence (open symbols) of _Baiomys musculus_. Note that the occipitonasal length decreases from north to south in each of the two species, and that in the region where the two species occur together, west-central Mexico, _B. taylori_ is smallest and _B. musculus_ is largest. Average, extremes, number of specimens averaged (in italic type), and name of locality, from north to south for each species, are as follows:

_Baiomys taylori_

18.0 (17.5-18.6) _15_, 9-1/2 mi. W New Mexico state line, Ariz.