MAMMALS OF THE ~IULESHOE SANDHILLS IN BAILEY, HALE AND LAMB COUNTIES, TEXAS by RUSSELL JOSEPH PESATURO, B.A. A THESIS IN ZOOLOGY

Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE

Approved

Accepted

December, 1989 ACKNOWLEDGEMENTS

I am indebted to Dr. J. Knox Jones, Jr., for his assistance and advice throughout this study. I am also grateful to Dr. Clyde Jones and

Dr. Kent Rylander for their aid and willingness to serve as members of my committee. For assistance in the field and laboratory, I would especially like to thank Richard Manning. I wish to thank Robert Owen,

Larry Choate, Janie Milner, Beth Henry, and the rest of the NSRL personnel for help during various stages of this project. I am grateful to The Museum of Texas Tech University for logistical support, Dean Clyde

Hendrick of the Graduate School for some support for the cost of field work, the Texas Parks and Wildlife Department for issuing scientific collecting permits, and the ranchers and farmers who gave me permission to trap on their property. Finally, my deepest appreciation goes to my wife Milva, for accompanying me on most field trips and for support and encouragement throughout the study.

ii TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF TABLES ...... iv

LIST OF FIGURES ...... v CHAPTER

I. INTRODUCTION 1

II. METHODS AND MATERIALS 3

III. DESCRIPTION OF AREA 6

Geography 6

Geology 6

Soil 11

Climate 16

History and Usage ...... 19

Vegetation . . . • ...... 21

IV. ACCOUNTS OF SPECIES 35

Species of Possible Occurrence ...... 63

V. DISCUSSION 66

LITERATURE CITED 69

iii LIST OF TABLES

1. Climatological summary of the Muleshoe Sandhills 18

2. Cranial measurements of adults of two species of Perognathus from the Muleshoe Sandhills ...... 46

3. Cranial measurements of adults of two species of Reithrodontomvs from the Muleshoe Sandhills, and comparative measurements of specimens from the Trans-Canadian Panhandle of Texas (after Jones et al., 1988) ...... 53

4. Cranial measurements of adult Peromyscus maniculatus from the Muleshoe Sandhills compared with a sample of adults from the Trans-Canadian Panhandle of Texas (after Jones et al., 1988) . 57

iv LIST OF FIGURES

1. Map of the Muleshoe Sandhills 7

2. Evolution of the physiography of northwest Texas and southeastern New Mexico ...•...... 10

3. General soil associations for Bailey, Lamb, and Hale counties, Texas • ...... • ...... 12

4. Sand dunes on the Texas-New Mexico state border, Bailey County 14

5. Sand dune undergoing fixation, western Bailey County...... 15

6. Plant species found in the Muleshoe Sandhills 23

7. Species composition percentages for selected rodent species from the Muleshoe Sandhills ...... 25

8. Hackberry-shrub series south of Olton, Lamb County...... 27

9. Harvard shin oak-mixed grass series, western Bailey County ... 28

10. Mixed grass-yucca series north of Fieldton, Lamb County 29

11. Sand sage-midgrass series southwest of Muleshoe, Bailey County 30

12. Sand sage-ragweed-yucca disclimax association north of Sudan, Lamb County ...... 31

13. Tallgrass-ragweed-Russian thistle disclimax association west of Hale Center, Hale County ...... •...... 33

14. Sand sage-shin oak community west of Muleshoe, Bailey County . 34

15. Checklist of mammals of the Muleshoe Sandhills 36

16. Prairie dog town west of Muleshoe, Bailey County...... 41

v CHAPTER I

INTRODUCTION

A distinct band of concentrated sandy soils, with varying vegetative cover, stretches from western Roosevelt County, New Mexico, to western Hale County, Texas. We refer to this unique physiographic feature as the Muleshoe Sandhills because Muleshoe, Texas, is the only sizeable town in, or in close proximity to, this landform. Our study was concerned primarily with that part of the Muleshoe Sandhills found in Texas, which comprises approximately 60 percent of the entire belt, located in Bailey, Lamb, and Hale counties. The objectives were to document the distribution and natural history of mammals in this region, and, insofar as possible, to correlate such data with existing environmental parameters, such as plant communities and soil types.

No previous study has dealt specifically with the Muleshoe Sandhills from a mammalogical perspective, although several studies of broader content have encompassed them (see, for example, Bailey, 1905; Davis,

1974; Jones et al., 1987; Jones et al., 1988). Publications detailing

the distribution of particular mammalian taxa on the Llano Estacado include those of Blair (1954), Desha (1967), Judd (1970), Schmidly

(1971,1984), Baumgardner and Schmidly (1981), Hollander et al. (1987), and Manning et al. (1988), among others. In addition, we have consulted all such publications and also general references of the vegetation and

physical environment of the region, including Sellards and Baker (1934),

1 2

Tanner (1939), Tharp (1939,1952), Blair (1950), Webb (1950), Allred and

Mitchell (1955), Gould (1978), Sullivan (1980), Brown (1982), McMahan

~ al. (1984), and the county soil surveys published by the U. S.

Department of Agriculture Soil Conservation Service. CHAPTER II

METHODS AND MATERIALS

This study began in the late autumn of 1987, with actual field work beginning in January of 1988. We made more than 35 field trips to the study site. We trapped and observed mammals in all three counties in every season and both by night and by day. Field studies were terminated in early August 1989.

Mammals were collected by various means. Approximately 2600 trap nights were carried out, including use of Sherman live traps, Victor rat traps, Museum Special snap traps, McAbee gopher traps, and commercial steel traps. Specimens also were caught by hand, found dead along roadways, a few were shot, and some crania were obtained by examination of owl pellets (Pesaturo et al., 1989). Eight pitfall traps were set out 4 mi. N and 5.5 mi. W Cotton Center, Hale County on 19 January, 1989.

These traps were periodically checked, but no mammals were captured in them. Observations by us and recorded sightings by local landowners, farmers, and ranchers were noted as well.

Specimens housed in The Museum of Texas Tech University that were collected by others prior to 1988 were incorporated into our data base, resulting in a total of more than 800 specimens examined. In addition, a study of the New Mexican Llano Estacada (Aday and Gennaro, 1973) provided records of mammals occurring in the Muleshoe Sandhills from

Roosevelt County, New Mexico, which was especially helpful in accounting

3 4 for those species of which we had few or no records. All specimens collected in our survey, approximately 600, also were deposited in The

Museum at Texas Tech University. Most of these were prepared as standard museum study skins accompanied by skulls. A few were preserved in fluid or as skeletal parts only.

Natural history data recorded in the field or laboratory included morphometric measurements, age and seasonal variation in pelage, molt patterns, and reproductive information such as testicular size in males and condition of females (number and size of fetuses, lactation, and the like). When appropriate, standard cranial measurements were recorded.

Most were taken using Fowler Ultra-Cal II digital calipers (to the nearest 0.01 millimeter). Some measurements, however (breadth of interparietal bone and length of upper and lower aveolar toothrows in

Perognathus, for example), were taken using a compound dissecting microscope fitted with an ocular micrometer (calibrated to 0.1 millimeter).

In addition, cranial measurements of three species (Reithrodontomys montanus, R. megalotis, and Peromyscus maniculatus) were compared to specimens from the Trans-Canadian Panhandle of Texas.

Dominant plant species were noted at collecting sites, and vegetative communities were categorized. The occurrence of plants was analyzed for ecological correlation with the distribution and relative abundance of selected mammalian species. Geographic localities were identified by use of a United States Geological Survey topographic map, 1:250,000

(Clovis sheet), as well as county maps distributed by the State of Texas

Highway Department. United States Department of Agriculture aerial maps 5 aided in delineating the boundaries of the Muleshoe Sandhills and differentiating local soil types found therein. CHAPTER III

DESCRIPTION OF AREA

Geography

The Muleshoe Sandhills are located in the southwestern part of the

Panhandle of Texas, on the Staked Plains or Llano Estacado. The area is characterized by small, gently rolling hills and dunes, many of which are covered with wild grasses, shrubs, and yucca. The sandhills are noticeably set apart from other areas of Bailey, Lamb, and Hale counties, the topography of which is nearly flat and under heavy cultivation.

They extend from the west in an east-southeasterly direction. Elevation ranges from about 3900 feet above sea level on the New Mexico border to

3500 feet in Hale County.

In size, the Muleshoe Sandhills constitute one of the largest contiguous concentrations of sandy-soil habitat found on the Llano.

From western Roosevelt County, New Mexico, to Hale County, Texas, they extend approximately 105 miles. In Texas, they measure some 58 miles from west to east, and from north to south they vary from 10 miles wide on the state border, to less than two miles in Hale County (Fig. 1).

Geology

Although there are several sandhill regions of varying size and topographic features on the Great Plains, all probably share some basic origins. It is difficult to describe the early geology of the Staked

Plains due to thick deposits of Cenozoic sediments, hence information

6 Bailey County Lamb County Hale County ,.. - -, • -,------, I Springlake 0 Earth I u ~luleshoej H :>< Olton Plainview I :z~ • :3 I z~ I Sudan Hale Center I :>., Needmore • Fieldton • ~c i-·-·-· I ::I . I 0 I u Cotton Center ! !

~ .--1 ~~ Q) Littlefield Spade > e • I Q) rJl 0 I 0 p:: I L_ 1_ Scale 0 5 10 miles

Fig. 1. Map of the Muleshoe Sandhills.

...... 8 on the development of this area during the Pre-Cambrian and Paleozoic

Eras is incomplete. Only in places where the deeper substrates are exposed can the past be reconstructed.

Inland seas have appeared and disappeared several times in the geologic record in what now is Texas. It is believed that most of the current Texas High Plains was neither deeply submerged nor greatly elevated during the late Ordovician, Silurian, Devonian, and Mississippian periods (Sellards and Baker, 1934). However, marine deposits during this time formed the Permian red beds, the primary underlying strata of the region.

The Amarillo uplift was perhaps an important event for deposition of strata in Bailey, Lamb, and Hale counties. Located to the north,

this uplift resulted from the formation of the Amarillo Mountains during

the Pennsylvanian. Increased flooding by inland seas, continued depression of the Permian Basin, and erosion over time caused these mountains to be leveled, submerged and engulfed by sediments by the close of the Permian

(Sellards and Baker, 1934).

Another factor in the development of the Staked Plains occurred during the Mesozoic Era. Drainage during this era reversed due to the formation of the Appalachian Mountains, which resulted in a rise of parts of the continent, and submergence of the Gulf Coastal Plain. ~1uch of the finer material that had been deposited previously in the Llano region was washed away by streams. This was not, however, of long duration, as sporatic shallow seas again covered the region, leaving sand, clay, and limestone. 9

The most important geologic event relative to development of the

Staked Plains occurred with the formation of the Rocky Mountains during the Cretaceous and early Tertiary. As the mountains rose, streams and rivers flowed east and washed away much surface material on the plains.

It was not, however, until the Rockies began to erode that the Llano

Estacada took on its present form and surface composition, which still basically exists today. As the mountains eroded, the flow of streams and rivers emanating from them slowed. Gravel, sand, and silt was deposited on the flat plains. Finer materials formed alluvial fans that spread over the Llano. Slowly, throughout the Cenozoic, these deposits of sand and gravel, referred to as the Ogallala formation, accumulated

(Girdner, 1963).

With the Pleistocene came great precipitation, the sand and gravel of the Ogallala became saturated and a source for the present underground water supply. In fact, the sandhills are economically important because they provide storage for rainwater while it percolates into the reservoir below, thereby replenishing it (Green, 1951). Near the end of Wisconsin glaciation, growth of the Pecos River system resulted in capture of most of the eastward-flowing streams from the New Mexican mountains. Drastic reduction of streams and rivers on the Llano Estacada followed (Sellards

~ al., 1932). Loss of surface water, wind blown deposits, and an increasingly arid climate is the modern result. See Fig. 2.

Composition of the Muleshoe Sandhills is primarily wind-deposited sands of Holocene origin, and their location is closely tied to relict drainage channels. This drainage system contributed to the building of the sandhills (Ross and Bailey, 1967), and the several draws in the study 10 A

~ \\~~-~~~-;JiL.> ·~ ...... ,:_·· ..... ·. ::- )( ..r~ : : ·· <-.::=.,.;: .. -..: . .· . ·. ~I ...f)-,~;~~,~-~~ : :. _:::_-:_ ,_+-~~;,:~ :· :. _.· .. ·._ ...... : .. ·:­ &r ·. ~(; ..::;....," .. :.·{;;::;,;:, '-:-··-,-...,-

~~ ..... B

o,__ __·AJ.__....;.;I

Fig. 2. Evolution of the physiography of northwest Texas and southeastern New Mexico. A: Area during Pliocene times showing alluvial deposits east of the high mountains in New Mexico. B: Area during present epoch showing capture of eastward-flowing streams of Pliocene times by the head-waters of the Pecos River. (From Sellards et al., 1932.) 11 area are their remnants. Evidence for a local origin of the sandhills was presented by Jones (1959). In any event, several factors have contributed to their formation.

Soil

Soil is an important ecological factor because it directly influences vegetation and thereby indirectly the fauna of any given area. Soil is also important with regard to burrowing mammals. Composition of surface sediments on the Staked Plains is approximately 60 percent sand, 23 percent clay, and 17 percent grit, gravel and other materials (Sellards and Baker, 1934). Soil associations and soil types were mapped for all

three counties of the study site by the U. S. Department of Agriculture

Soil Conservation Service during their soil surveys of the late 1950s and 1960s. A soil complex or association usually consists of two or more major soils and at least one minor soil type, grouped due to shared general characteristics or because the individual soils are too intricately mixed. Soil associations for Bailey, Lamb, and Hale counties are shown in Fig. 3.

The majority of the Muleshoe Sandhills consists of soils of the

Tivoli-Brownfield association. To a minor extent, two other associations are involved. From central Bailey County to Lamb County, north of

Fieldton, a band of the Amarillo loamy fine sand association enters the periphery of the sandhills. The Arch-Drake association is found in

Bailey County west of Muleshoe and surrounding Coyote Lake. Major soil types found in the Muleshoe Sandhills are described as follows:

1. Tivoli fine sand.--Along with the Brownfield fine sand, this soil comprises the majority of the Muleshoe Sandhills. It is a pale GENERAL- SOU.- MAP GENERAL SOU. MAP GENERAL SOU. MAP 8 BAILEY COUNTY, TEXAS LAMB COUNTY, TEXAS HALE COUNTY, TEXAS J

r

D I ~

I

I i I I I I I c I ..... J

··- ---cou;;r;

1. Amarillo fine sandy loam association 1. Amarillo fine sandy loam association 1. Pullman-Olton association 2. Amarillo-Arvana association 2. Olton-Amarillo association 2. Pullman association 3. Tivoli-Brownfield association 3. Tivoli-Brownfield association 3. Olton-Amarillo-Acuff association 4. Amarillo loamy fine sand association 4. Arvana-Kimbrough association 4. Mansker-Bippus-Berda association 5. Ar ch-Drake association 5. Amarillo loamy fine sand association 5. Tivoli-Brownfield association 6. Portales association 6. Portales association 7. Stegall association 7. Berthoud-Mansker association 8. Berthoud-Mansker association B. Portales-Berthoud-Drake association

Fig. 3. General soil associations for Bailey, Lamb, and Hale

counties, Texas. 1--' N 13

brown to yellowish deep, noncalcareous, fine sand occurring primarily

on dunes that support vegetation.

2. Brownfield fine sand.--This sand is found in the large, undulating

plains and on hummocks. It is characterized by a brown to pale brown

deep surface layer followed by a red, sandy-clay loam subsoil.

3. Amarillo loamy fine sand.--Distribution in the sandhills of this

soil type is more extensive than might be supposed. This soil has a

brown surface layer. The subsoil is a reddish-brown, sandy-clay loam

that becomes a pinkish-yellow caliche with depth.

4. Arch loamy fine sand.--The surface composition of this loamy

sand consists partly of wind deposited Brownfield and Tivoli fine sands.

'This soil appears grayish-brown on the surface, becoming lighter and

more calcareous with depth. Occasionally, it is found in association

with Drake, Gomez or Likes soils.

Sand dunes themselves (Fig. 4), found throughout the Muleshoe

Sandhills, are not composed of a single type of soil. Rather, they are

unstabilized areas of varying size composed of sands supporting little

or no vegetation. Due to elevation, semiarid conditions, and flat relief

with virtually few obstacles, wind is the agent of sand dune building,

especially in areas where there are accumulations of sandy soils on the

surface. Eventually, vegetation, fences, or other objects cause deposition

of the wind-blown material. Small knobs or ridges that further act as

barriers cause the eventual formation of dunes. As dunes age, they can

become fixed, that is, they become covered with vegetation and a

fine-grained topsoil develops (Fig. 5). This topsoil can become more

organic and develop into Tivoli or Brownfield soil. 14

Fig. 4. Sand dunes on the Texas-New Mexico state border, Bailey County. 15

Fig. 5. Sand dune undergoing fixation, western Bailey County. 16

Extensive geological and sedimentary studies of the sand dunes of the Muleshoe Sandhills were conducted by Tanner (1939), Sanders (1951),

Green (1951), and Jones (1959). Tanner (1939), stated that the sandhills from Bailey, Lamb, and nearby counties are still in the original pattern layed down as the Staked Plains were formed. A major constructional river probably once flowed in the location of the present sandhills, causing the deposition of parent materials. Soils surrounding the sandhills are predominantly clay or loam types. These soils are not subject to much eolian disturbance and have remained fairly permanent.

Soils within the sandhills, which consist of fine sands, shift and blow but basically remain intact. Although fixation and soil stratification does occur, major changes probably will not take place for some time.

It is possible, then, that alluvial deposits from the formation of western mountains and from local materials were important to the formation of the Muleshoe Sandhills. This was followed by some wind action that deposited sand over the pre-existing alluvial materials during the

Quaternary. The prominant parent material that settled in the Muleshoe

Sandhills region was a quartz sand. In time, this developed into Tivoli soil, a major componant of the sandhills. In addition to alluvian and eolian deposition of parent materials, soil formation in the sandhills was acted upon by other factors. Climate, living organisms, time, and man all have contributed to the present physical condition of the sandhills.

Climate

The Muleshoe Sandhills share the same semiarid, warm climate characteristic of the Llano Estacada. Summer temperatures are moderate 17 to warm, but nights cool. In fact, there is a large range between maximum daytime temperatures and minimum nighttime temperatures throughout the year due to the high elevation, low humidity, and clear skies, which allow maximum solar radiation penetration. Winter temperatures can fluctuate rapidly as Canadian cold fronts sweep down the Great Plains.

However, coldspells usually do not last more than two or three days.

Spring is the most variable season in terms of temperature. Climatic data was obtained from the National Oceanic Atmospheric Administration

(1985). See Table 1.

Winds are predominantly from the south and southwest, and are frequently strong because of the elevation and level terrain. Early spring is the windiest season, causing strong, persistent dust storms.

Strong winds of short duration accompany the frequent thunderstorms of late spring and early summer. Hail and an occasional tornado can accompany thunderstorms during this time.

Monthly and annual amounts of rainfall are also quite variable.

Most rainfall occurs during the thunderstorms, and causes runoff and erosion. Winters are fairly dry as northern cold fronts prevent moisture from the Gulf of Mexico from reaching the Llano. Snowfalls are usually light, and sometimes mixed with rain.

Humidity is usually low, with small seasonal variation. Average annual relative humidity for the Muleshoe Sandhills is between 74 and

78 percent at 6:00 AM and between 35 and 40 percent at 6:00 PM.

Southwesterly winds tend to keep the humidity low. 18

Table 1. Climatological summary of the Muleshoe Sandhills. Data obtained by weather tracking stations in Muleshoe, Bailey County, Texas and Plainview, Hale County, Texas. (From National Oceanic and Atmospheric Administration, 1985.)

Temperature (Fo) Means Precipitation (Inches) Totals Mean number of days E E ::l ::l Q) Q) Q) E E 1-< 1-< 1-< ·r-1 ·r-1 0 0 0 >< c: c: :;: E E E co ·r-1 ·r-1 0 E E >. co c: 1-< 1-< 1-< r-1 1-< U) 0 0 0 >. >. .c r-1 r-1 .j.J c: c: 0 0 0 ·r-1 ·r-1 c: co co ,...... , tn 0 0 Q) Q) . . co co ,...... , 0 0 ::E ::E -,<:.... 0 0

MULESHOE, TX (#416135) 1951-80 34°14'N 102°43'W 3780 FT.

January 52.2 20.0 36.1 0.50 2.7 1 0 0 February 56.1 23.4 39.7 0.53 3.3 2 0 0 March 63.5 29.7 46.6 0.55 1.3 2 0 0 April 73.4 40.3 56.9 0.93 0.3 2 1 0 May 81.3 49.9 65.7 1.93 0.0 4 1 0 June 89.9 59.7 74.8 2.44 0.0 4 2 1 July 91.4 63.7 77.6 2.57 0.0 4 1 1 August 90.4 61.7 75.5 2.42 0.0 4 1 1 September 83.0 54.1 68.6 1. 77 0.0 4 1 0 October 73.9 41.8 57.8 1.43 0.1 3 1 0 November 60.9 29.5 45.2 0.65 1.8 2 0 0 December 54.5 22.5 38.5 0.36 2.0 1 0 0 Year 72.5 41.3 56.9 16.08 ll.5 33 8 3

PLAINVIEW, TX (#417079) 1951-80 34°ll 'N 101°42'W 3370 FT.

January 51.9 22.8 37.4 0.49 2.3 1 0 0 February 56.2 26.4 41.3 0.66 4.2 2 0 0 March 63.6 32.5 48.1 0.79 1.5 2 1 0 April 73.7 42.9 58.3 1.24 0.0 3 1 0 May 81.6 52.5 67.1 3.27 0.0 4 2 1 June 90.1 62.0 76.4 3.07 0.0 4 2 1 July 92.4 65.7 79.0 2.57 0.0 4 2 1 August 90.5 63.8 77.2 2.03 0.0 4 1 0 September 83.9 56.8 70.3 2.03 0.0 3 1 1 October 74.9 45.1 60.1 1.68 0.0 3 1 0 November 61.6 33.1 47.4 0.64 1.1 2 0 0 December 54.8 26.2 40.5 0.50 1.7 1 0 0 Year 72.9 44.2 58.6 18.97 10.8 33 ll 4 19

Historv and Usage

Much of what we would term modern history concerning the Muleshoe

Sandhills and the majority of the Staked Plains has occurred only in the last 115 years. The 1875 Texas Almanac stated: "The Llano Estacada is almost an unexplored region, but little known at present, the tramping

ground of the wild Indian and the buffalo" (Crosbyton Review, 1936:32).

However, the occurrence of peoples in and around the sandhills region

dates back much farther.

Evidence of prehistoric man dating back some 10,000 years, believed

to have been Folsom man, has been found in campsites along Black Water

Draw near Portales, New Mexico, only about 15 miles from Bailey County.

Park (1936) extensively explored the Muleshoe Sandhills in both Bailey

and Lamb counties. He found numerous permanent campsites and trails

belonging to a tribe of prehistoric Indians. Thousands of Indians were

estimated to have lived among the sandhills in enormous camps. Various

artifacts have been found, dating from 1000 BP to a few hundred years

ago.

Pearce (1936) surveyed the sandhill camps as well. He noted that

several campsites were located near a branch of the Yellow House Canyon

in Lamb County. Several artifacts obtained from the sandhill campsites

were examined by the Laboratory of Anthropology at Santa Fe, New Mexico.

Most were of Pueblo origin, the oldest dating back to the Third Pueblo

Period of 900-1300 AD.

The Plains Indians known from recent times, primarily the Comanches and Kiowas, originally migrated into the Staked Plains less than 500

years ago. These Indians were as nomadic as the bison that populated 20

the region, and on which they depended for food, shelter, and hides.

The Muleshoe Sandhills provided excellent camping grounds for the

Comanches due to abundant game, vegetation, protection from the wind,

and availability of water (Dooley and Dooley, 1978).

The area where the three counties of the study site were located,

passed through Spanish, Mexican, and Republic of Texas jurisdiction,

there were no attempts, however, to colonize the area. In 1876, the

State Legislature created 50 counties, including Bailey, Lamb, and Hale,

out of Bexar and Young territories. In an effort to eliminate the Plains

Indians and thereby colonize the Llano, the U. S. Army initiated the

Indian Wars of 1874-75. Throughout the next decade, hunters virtually

eliminated the buffalo from the plains as well. With the buffalo and

Indian gone, permanent colonization began.

The vast plains appealed first to the cattlemen. Parts of Bailey,

Lamb, and Hale counties were incorporated into large ranches such as the

XIT, Muleshoe, Running Water, and Spade. However, by the turn of the century, ranching declined as farming proved to be extremely successful.

Many large ranches were divided and sold to farmers. Railroad land companies and homestead laws provided land for small farms as well.

Today, cotton, sorghum, wheat, hay, and corn are the principal crops grown in Bailey, Lamb, and Hale counties. Both dryland and irrigated farming are practiced. Cattle raising is still important, especially on range lands not suitable for agriculture, the largest portion of which in the three counties is the Muleshoe Sandhills.

Industries of the tricounty region are mainly agriculturally related.

There are cotton gins, compresses, a large denim plant, cattle feeding 21

operations, canneries, and marketing plants. Oil production, caliche

mining, and a large electrical plant are important as well. The towns

located in and around the Muleshoe Sandhills include Muleshoe, Sudan,

Earth, Springlake, Fieldton, Olton, Cotton Center, and Portales, New

Mexico.

Vegetation

To better understand the types and associations of plants found in

the Muleshoe Sandhills, it is best to begin with the largest biogeographical

unit. Several authors have mapped and described ecological-physiographical

areas such as life zones, biotic provinces, and vegetational regions.

Dice (1943) and Blair (1950) studied biotic provinces for North America

and Texas, respectively (see also Webb, 1950). Dice (1943) described a

biotic province as a large geographic area having similarities of vegetation

type, ecological climax, flora, fauna, climate, physiography, and soil.

In his map of North America, the Staked Plains were located in the southern

half of the Kansan biotic province. Short and mid-grasses were the original

climax vegetation in that region, but overgrazing has caused a long

disclimax of short grasses. However, smaller areas with distinct

characteristics cannot be adequately described by such a large unit.

Environmental barriers can create pockets or islands of floral and

faunal uniqueness. Changes in plant communities, physiography, and soil

type are some of the causes. From a soil perspective, the Muleshoe

Sandhills with their sandy substrate would have certain plant and animal affinities, that do not necessarily completely follow the parameters of the Kansan biotic province. Biotic provinces are subdivided into biotic districts, that of the Llano Estacada being the Short-grass Plains district. 22

A biotic province, as previously defined, is a combination of many factors, one of which is vegetation. Several authors, including Bray

(1901), Tharp (1939), and McMahan et al. (1984), divided Texas into vegetational or natural regions. In a study by Allred and Mitchell (1955), the plains region of Texas and Oklahoma was included in the Mixed Prairie zone. The Muleshoe Sandhills were placed in the High Plains Bluestem subdivision, as were other deep sandy areas. Characteristic of this subdivision are taller grasses than found elsewhere on the plains. The occurrence of these tall grass communities within the shortgrass prairie, is due to available moisture and sandy soils. "Consequently, tall-grass communities in the Southwest are restricted to east of longitude 104° on sandhills on the Llano Estacada (e.g., near Portales in Roosevelt County,

New Mexico)," (Brown, 1982:116).

Common plant species found at our trapping sites within the Muleshoe

Sandhills are listed in Fig. 6, nomenclature used here follows Barkley

(1986). This list includes only plants from limited localities and, therefore, is not an inventory of all species from the sandhills. A more extensive analysis of the flora for Bailey County can be found in Rosson

(1971). References of species present on the Llano, or within sandy-soil habitats therein, can be found in Tharp (1939), Blair (1950), and Allred and Mitchell (1955).

A detailed and useful classification system for North American vegetation was developed by Brown et al. (1979). Instead of dividing the continent into physic-vegetational regions based on geography, this system employs a digitized hierarchical structure in which every habitat is classified into increasingly defined natural levels from biogeographic 23

Juniper Juniperus sp. Prickly poppy Argemone sp. Hackberry Celtis sp. Harvard shin oak Quercus havardii Prickly pear Opuntia sp.

Russian thistle (tumbleweed) Salsola iberica Smart weed Polygonum sp. Buffalo-gourd Cucurbita foetidissima Cottonwood Populus sp. Willow Salix sp.

Peppergrass Lepidium sp. Plum Prunus sp. Mesquite Prosopis glandulosa Grape Vitis sp. Soapberry (Chinaberry) Sapindus saponaria (v. drummondii)

Skunk bush Rhus aromatica (v. trilobata) Western ragweed Ambrosia psilostachya Sand sage Artemisia filifolia Broomweed (snakeweed) Gutierrezia sp. Sunflower Helianthus sp.

Western wheatgrass Agropyron smithii Big bluestem Andropogon gerardii Sand bluestem Andropogon hallii Silver bluestem Andropogon saccharoides Little bluestem Andropogon scoparius

Sideoats grama Bouteloua curtipendula Blue grama Bouteloua gracilis Buffalo grass Buchloe dactyloides Sand reed Calamovilfa sp. Sand bur Cenchrus incertus

Windmill grass Chloris sp. Weeping lovegrass Eragrostis curvula Sand lovegrass Eragrostis trichodes Panicum (switchgrass) Panicum sp. Indian grass Sorghastrum nutans

Johnson-grass Sorghum halepense Sand dropseed Sporobolus cryptandrus Giant dropseed Sporobolus giganteus Yucca Yucca sp.

Fig. 6. Plant species found in the Muleshoe Sandhills. Nomenclature and sequence of taxa follows Barkley (1986). 24

realm to plant association. The hierarchy, in descending order, consists

of biogeographic realm, vegetation, formation-type, climactic zone,

regional formation (biome), plant series, and plant association.

Diamond et al. (1987) classified the vegetation of Texas to the

plant series level, a system also utilizing a hierarchical framework.

"Series are named for typical dominant or codominant species or, in a

few cases, genera. [Series] allow for some differences in species

composition and dominance relationships ... " (Diamond et ~-· 1987:205).

Associations are more defined and describe the dominant or codominant

species within a given series.

Plant species were inventoried at all trap site locations (Fig. 6)

in the Muleshoe Sandhills. Forty-five localities were grouped into six

communities based on similar species composition. We assumed four of

these communities to be similar in rank to the series level. The remaining

two communities were based on specific species present and probably should

be referred to as plant associations. However, because they were so

prevalent, we deemed it appropriate to list them. Some sites were

disturbed and represent disclimax communities.

Rodent species caught with Sherman live traps or snap traps at each

of the trapping sites were assigned to one of the six plant communities.

The ratio of each species caught to the total number of individuals from

all species trapped in each habitat is reflected by a species composition

percentage (Fig. 7). These data are analyzed in the discussion chapter.

The six plant communities are defined as follows:

1. Hackberry-shrub series.--This series was located near Blackwater

Draw. It is characterized by hackberry along with shrubs such as skunkbush 25

1. Hackberry-shrub series 2. Harvard shin oak-mixed grass series

R. megalotis 22.7% Dipodomys 76.0% Chaetodipus 19.1% Onychomys 16.0% Dipodomys 15.5% Sigmodon 4.0% Onychomys 13.6% Mus 4.0% Peromyscus 13.6% (25) Perognathus 7.3% Mus 6.4% Sigmodon 1.8% (110)

3. Mixed grass-yucca series 4. Sand sage-midgrass series

Onychomys 28.6% Dipodomys 45.1% Dipodomys 21.4% Onychomys 26.2% Peromyscus 21.4% Perognathus 13.1% Sigmodon 10.7% Peromyscus 9.0% Mus 7.1% Chaetodipus 2.5% Chaetodipus 3.6% Sigmodon 2.5% Perognathus 3.6% R. megalotis 1.6% R. montanus 3.6% (122) (28)

5. Sand sage-ragweed-yucca 6. Tall grass-ragweed-Russian thistle disclimax association disclimax association

Dipodomys 37.0% Onychomys 29.0% Onychomys 22.4% Dipodomys 17.4% R. megalotis 16.1% Mus 13.1% Peromyscus 11.9% Peromyscus 11.6% Sigmodon 5.6% R. megalotis 11.6% Chaetodipus 2.8% Chaetodipus 10.1% Perognathus 2.1% Perognathus 4.3% Neotoma 1.4% R. montanus 2.9% Mus 0.7% (69) (143)

Fig. 7. Species composition percentages for selected rodent species from the Muleshoe Sandhills. Percentages reflect proportions of individuals trapped for each species in relation to total number of individuals captured for all species, per plant community. Sample sizes in parentheses. 26 and sand sage. Also found in this community are soapberry, wild grape, and sunflower, and occasional grasses. See Fig. 8.

2. Harvard shin oak-mixed grass series.--This was the least common plant community, with localities primarily in western Bailey County.

Also present with shin oak was mesquite, yucca, and plum; sand sage was not common. See Fig. 9.

3. ~ixed grasses-yucca series.--Sparsely located throughout Bailey and Lamb counties, this series is characterized by a variety of mid- and short grasses such as silver bluestem, buffalo grass, lovegrass, sand dropseed, sandbur, and various gramas, mixed with yucca (ragweed and broomweed also were present). See Fig. 10.

4. Sand sage-midgrass series.--The sand sage-midgrass community was encountered second in frequency to the sand sage-ragweed-yucca association. It is found throughout Lamb and Hale counties, but in some localities the midgrasses are mostly replaced by a disclimax of weedy plants and yucca, which represents a separate association. Grass species found include windmill grass, sideoats grama, sand bluestem, and little bluestem. Skunkbush, yucca, plum, and sunflower also were present. See

Fig. 11.

5. Sand sage-ragweed-yucca disclimax association.--This association most probably is classified under the sand sage-scrub series; because it was the most commonly encountered community in the tricounty study area, a separate listing is merited. The codominants are sand sage and ragweed, but yucca was always present. Occasional species found in this community include sunflower, panicum, sandbur, Russian thistle, and buffalo-gourd.

See Fig. 12. 27

Fig. 8. Hackberry-shrub series south of Olton, Lamb County. 28

Fig. 9. Harvard shin oak-mixed grass series, western Bailey County. 29

Fig. 10. Mixed grass-yucca series north of Fieldton, Lamb County. 30

Fig. 11. Sand sage-rnidgrass series southwest of Muleshoe, Bailey County. 31

Fig. 12. Sand sage-ragweed-yucca disclimax association north of Sudan, Lamb County. 32

6. Tallgrass-ragweed-Russian thistle disclimax association.--Found primarily in Bailey County, sparingly in Lamb and Hale counties, this community was frequent enough to list separately. Ragweed and Russian thistle obviously are recent invaders to this sandy soil, tallgrass community. Grasses encountered include big bluestem, Johnson-grass, giant dropseed, and lovegrass; sunflower and sand sage also were present, but in relatively small numbers. See Fig. 13.

Some localities in the Muleshoe Sandhills could not be adequately described under one of the previously listed plant communities. These areas either represent a separate series or indicate an intergradation of two or more communities. An example, shown in Fig. 14, illustrates a community of sand sage, shin oak, and mixed grasses. Mesquite and yucca also were present. 33

Fig. 13. Tallgrass-ragweed-Russian thistle disclimax association west of Hale Center, Hale County. 34

Fig. 14. Sand sage-shin oak community west of Muleshoe, Bailey County. CHAPTER IV

ACCOUNTS OF SPECIES

The following 30 accounts are based on all mammals obtained or observed by us in the Muleshoe Sandhills. Also incorporated are species that were recorded by earlier field parties which have deposited specimens in The Museum of Texas Tech University. Arrangement of taxa are in currently accepted phylogenetic sequence except for species in the same genus, which are in alphabetical order. A checklist, including ordinal and familial headings, is found in Fig. 15. Following the accounts is a list of several species of possible occurrence. We follow Jones et al.

(1986) for use of vernacular names.

Didelphis virginiana virginiana Kerr, 1792

Virginia Opossum

Although only a few, narrow and depauperate riparian communities reach the sandhills along Blackwater Draw, this species evidently occurs sparingly in the area. Possibly abandoned farmsteads provide marginal cover in some places. In addition to the specimen listed below, we noted opossums killed along a roadway approximately 9 mi. S Earth, Lamb County, in both 1988 and 1989, and collected or observed individuals at several locations adjacent to the study area. A landowner in eastern Lamb County also reported to us that he had seen opossums in the sandhills on his property.

Specimens examined (1).--BAILEY CO.: 1 mi. N, 6 mi. W Muleshoe, 1.

35 36

ORDER MARSUPIALIA--MARSUPIALS Family Didelphidae--Opossums Didelphis virginiana (Virginia opossum) ORDER INSECTIVORA--INSECTIVORES Family Soricidae--Shrews Cryptotis parva (least shrew) ORDER LAGOMORPHA--LAGOMORPHS Family Leporidae--Hares and Rabbits Sylvilagus audubonii (desert cottontail) Sylvilagus floridanus (eastern cottontail) Lepus californicus (black-tailed jackrabbit) ORDER RODENTIA--RODENTS Family Sciuridae--Squirrels and Allies Spermophilus spilosoma (spotted ground squirrel) Spermophilus tridecemlineatus (thirteen-lined ground squirrel) Cynomys ludovicianus (black-tailed prairie dog) Family Geomyidae--Pocket Gophers Geomys bursarius (plains pocket gopher) Cratogeomys castanops (yellow-faced pocket gopher) Family Heteromyidae--Pocket Mice and Kangaroo Rats Perognathus flavescens (plains pocket mouse) Perognathus flavus (silky pocket mouse) Chaetodipus hispidus (hispid pocket mouse) Dipodomys ordii (Ord's kangaroo rat) Family Cricetidae--New World Rats and Mice Reithrodontomys megalotis (western harvest mouse) Reithrodontomvs montanus (plains harvest mouse) Peromyscus maniculatus (deer mouse) Onychomys leucogaster (northern grasshopper mouse) Sigmodon hispidus (hispid cotton rat) Neotoma micropus (southern plains woodrat) Family Muridae--Old World Rats and Mice Rattus norvegicus (Norway rat) Mus musculus (house mouse) Family Erethizontidae--New World Porcupines Erethizon dorsatum (North American porcupine) ORDER CARNIVORA--CARNIVORES Family Canidae--Canids Canis latrans (coyote) Urocyon cinereoargenteus (gray fox) Family Mustelidae--Mustelids Mustela frenata (long-tailed weasel) Taxidea taxus (badger) Mephitis mephitis (striped skunk) Family Felidae--Cats Felis rufus (bobcat) ORDER ARTIODACTYLA--EVEN-TOED UNGULATES Family Cervidae--Cervids Odocoileus hemionus (mule deer)

Fig. 15. Checklist of mammals of the Muleshoe Sandhills. 37

Cryptotis parva parva (Say, 1823)

Least Shrew

This species recently was reported from the Muleshoe Sandhills in adjacent Roosevelt County, New Mexico, and also from Muleshoe National

Wildlife Refuge to the south of the sandhills in Bailey County (Owen and

Hamilton, 1986). We recovered remains of two individuals from cast pellets of barn owls found in the woodlot of an old homestead 1 mi. N and 5 mi. E Fieldton, Lamb County (Pesaturo et al., 1989). These two shrews probably were captured by owls between early October 1988 and early

January 1989. Evidently, ~- parva has extended its distribution westward in recent years, possibly in response, in part at least, to widespread irrigation on the Llano Estacada.

Sylvilagus audubonii neomexicanus Nelson, 1907

Desert Cottontail

Cottontails are widespread and common in and around the sandhill region. However, despite the fact that S. audubonii is an upland inhabitant, we collected specimens only from Bailey County. Nevertheless, material examined from localities in Hale and Lamb counties near the sandhills indicates that this cottontail occurs throughout the study area.

A female taken on 30 April carried four fetuses that measured 19 in crown-rump length. An immature male obtained on 26 July had testes measuring 27 in length; those of an adult from 28 September measured 34.

Specimens examined (5).--BAILEY CO.: 2 mi. S, 10 mi. W Muleshoe, 1;

8 mi. S, 10 mi. W Muleshoe, 1; 7 mi. S, 2 mi. W Muleshoe, 1; Muleshoe, 2. 38

Sylvilagus floridanus llanensis Blair, 1938

Eastern Cottontail

Rather surprisingly, this species, typical of more mesic habitats, seems to be distributed throughout the study area. However, indications are that it is more likely to be found near irrigated farming operations along the periphery of the sandhills as well as around the few irrigated fields in sandy soil. A female taken on 30 April contained five fetuses measuring 27 in crown-rump length. An adult male captured on 3 May had testes 55 long.

S. floridanus and S. audubonii are similar in appearance and sometimes difficult to distinguish externally at first glance. S. floridanus has shorter ears than does audubonii (usually less than 60), is slightly darker dorsally overall, has somewhat darker ears, and usually has a greater suffusion of buffy-orangish coloration, particularly on the nape.

Cranially, floridanus has much smaller bullae, a noticeably broader mesopterygoid fossa, a smaller external auditory meatus, and a longer maxillary toothrow.

Specimens examined (4).--BAILEY CO.: 0.5 mi. S, 10.5 mi. W Muleshoe,

1. LAMB CO.: 4 mi. S Springlake, 1; 5.5 mi. N Fieldton, 1. HALE CO.:

10.5 mi. WHale Center, 1.

Lepus californicus melanotis Mearns, 1890

Black-tailed Jackrabbit

Jackrabbits were observed in sandhill habitats in all three counties, some killed along roadways, although we did not find this species to be as common as compared to other places in the central part of the Llano 39

Estacada. Our specimens seem best assigned to the subspecies L. £· rnelanotis, although variation in this species on the Llano and details of distribution of rnelanotis with respect to that of L. c. texianus, which is known from the southern part of the region, remain to be studied.

Subadult males taken on 5 August and 4 November had testes that measured 28 and 19 in length, respectively. An adult female shot on 30

April was lactating and carried two fetuses (crown-rump length, 110); another captured on 28 June had a flacid uterus but was not visibly pregnant.

Specimens examined (4).--BAILEY CO.: 1 mi. N, 11 mi. W Muleshoe, 1;

1.5 mi. WNW Muleshoe, 1. LAMB CO.: 8 mi. S Earth, 1; 7 mi. S, 1.5 mi. E

Olton, 1.

Sperrnophilus spilosorna rnarginatus Bailey, 1890

Spotted Ground Squirrel

Although found throughout the Muleshoe Sandhills, the spotted ground squirrel does not appear to be particularly abundant. Four of our specimens from Bailey County carne from the coarse, reddish-brown colored

Amarillo loamy fine sands along the southern edge of the sandhills. Of the two ground squirrels occurring in the region, this species appears to be more common westwardly than is~· tridecernlineatus; we never observed the two species at the same place.

Adult males collected on 30 April (two) and 1 July had testes 22,

24, and 13, respectively, in length. Nonpregnant females were taken on

1 April and 30 April, whereas a lactating individual was recorded on 3

June and one captured on 4 August had completed lactation. 40

Specimens examined (7).--BAILEY CO.: 6.5 mi. S, 6 mi. W Muleshoe,

1; 3.2 mi. S, 5.5 mi. W Muleshoe, 1; 6.5 mi. S, 2.5 mi. W Muleshoe, 2;

6 mi. S, 1 mi. W Muleshoe, 1. LAMB CO.: 7 mi. S Olton, 1; 4.1 mi. NE

Hart Camp, 1.

Spermophilus tridecemlineatus arenicola (A. H. Howell, 1928)

Thirteen-lined Ground Squirrel

This sciurid was collected along roadside rights-of-way in Bailey and Lamb counties. While nowhere especially common, ~· tridecemlineatus seemed to be more widely distributed in the eastern part of the region than in the west, and often was found in peripheral areas of sandy habitat.

Adult males collected on 1 April (two) and 22 April had testes measuring 20, 19, and 20 in length. Females taken on 1 and 22 April were not visibly pregnant. Young animals were captured on 13 June and 4 July.

Specimens examined (7).--BAILEY CO.: 3.5 mi. S, 10 mi. W Muleshoe,

1. LAMB CO.: 10 mi. N, 0.2 mi. W Sudan, 1; 12 mi. N Sudan, 1; 9.5 mi. N

Sudan, 3; 2.6 mi. N Fieldton, 1.

Cynomys ludovicianus ludovicianus (Ord, 1815)

Black-tailed Prairie Dog

Colonies of black-tailed prairie dogs were found at scattered locations on shallow sandy soils in Bailey and Lamb counties, in flat, short-grass habitats with occasional small mesquite trees. Most "towns" were small, no more than a hundred yards or so in diameter, but at least one in Bailey County (Fig. 16) was a quarter of a mile or more in extent along both sides of a roadway and at least as wide on both sides. Colonies were observed as follows: 2 mi. S, 10 mi. W Muleshoe, 6 mi. S, 10 mi. W 41

Fig. 16. Prairie dog town west of Muleshoe, Bailey County. Notice how the occupants of the colony have kept the surrounding grasses low to the ground. 42

Muleshoe, and 2 mi. S Muleshoe, Bailey County; 9 mi. N Sudan, 7 mi. S,

2.5 mi. E Earth, and 7.5 mi. S Springlake, Lamb County. A male with abdominal testes (12 x 5) and a nonpregnant female, obtained on 1 April and 30 April, respectively, were in process of spring molt.

Specimens examined (4).--BAILEY CO.: 2 mi. S, 10 mi. W Muleshoe,

1; 6 mi. S, 10 mi. W Muleshoe, 1; Muleshoe, 1; 2 mi. S Muleshoe, 1.

Geomys bursarius major Davis, 1940

Plains Pocket Gopher

This pocket gopher is common, in some places abundant, in the Muleshoe

Sandhills. Most of our specimens came from rights-of-way along highways and roadways, and from sandy bar ditches, although the species is distributed generally throughout the region. Presently, it is known to come into contact with another gopher, Cratogeomys castanops, only at one place as described in the account of that species, but the two also may occur parapatrically elsewhere, particularly along the periphery of the sandhills. As has been found in other studies, there are more females among our specimens than males, 21 as opposed to 12.

A female trapped on 8 April carried two fetuses (39 in crown-rump length); another obtained on 25 March had three placental scars but no longer was lactating. Nonpregnant females in adult pelage were taken in

January, February, June, July, and October. Animals in juvenile pelage or molting from that pelage were collected in March, May, June, July, and

September. Adult males had testicular lengths of 25 in January, 24 and

23 in February, 23 and 18 in April, 12 in July, and 13 and 9 in October.

From these rather meager data, it appears that G. bursarius breeds from mid-winter to late spring or perhaps early summer in the Muleshoe Sandhills. 43

Specimens examined (33).--BAILEY CO.: 3 mi. S, 18 mi. W Muleshoe,

1; 6.5 mi. S, 2.5 mi. W Muleshoe, 1; 6-6.2 mi. S Muleshoe, 3; 6-7 mi. S,

1 mi. E Muleshoe, 2; 2-3.6 mi. SE Muleshoe, 5. LAMB CO.: 6 mi. N Sudan,

1; 4.5-5 mi. S Earth, 3; 4 mi. S, 2.5 mi. E Earth, 4; 2-3.5 mi. N Fieldton,

5; 1-1.5 mi. N, 5 mi. E Fieldton, 4; 6.5 mi. S Olton, 2. HALE CO.: 4 mi.

N, 5.5 mi. W Cotton Center, 2.

Cratogeomys castanops perplanus Nelson and Goldman, 1934

Yellow-faced Pocket Gopher

This species enjoys a widespread distribution on the Llano Estacada,

including areas immediately surrounding the Muleshoe Sandhills. It

ordinarily is not found in sandy soils on the Llano, however, these being

occupied by Geomys bursarius. Distribution of the two gophers is

parapatric where their ranges abut.

In August and October of 1975, R. C. Dowler trapped pocket gophers

at a Southwest Public Service Company station south of Earth. Here, sand

had been cleared away down to caliche from a considerable area in order

to construct buildings on the site. Cratogeomys burrows were in this

altered habitat. Both Dowler and our parties took Geomys in sandy banks along a roadside just opposite the power plant. The only other C.

castanops in our material came from the edge of the sandhills where we would expect to find interspersed populations of the two gophers.

Two females captured on 5 October were lactating. A nonpregnant animal was trapped on 20 August. Molting adults were obtained in both

August and October.

Specimens examined (6).--LAMB CO.: 4.8 mi. S, 0.3 mi. W Earth, 5;

5.5 mi. S Olton, 1. 44

Perognathus flavescens copei Rhoads, 1894

Plains Pocket Mouse

The plains pocket mouse is common in the sandhills, most frequently in relatively open areas where Dipodomys ordii, Peromyscus maniculatus, and Onychomys leucogaster are common associates. For example, on 2 July

1989, in 140 Sherman traps set in Hale County in an area dominated by sand sage, skunkbush, yucca, and sunflower, and with some grasses and weeds, we took seven P. flavescens, seven D. ordii, two 0. leucogaster, one P. maniculatus, and one Chaetodipus hispidus. In Lamb County on 5

September 1988, at a place where sand sage, yucca, skunkbush, and mid-grasses prevailed, we trapped five P. flavescens, 10 ~· ordii, three

0. leucogaster, and one f. maniculatus in 108 Shermans.

Many of our mice were trapped along fencerows. Except for August, we have specimens from the months of May through December. Small pocket mice, presumably f. flavescens, were one of the most common small mammals taken by barn owls at the edge of the sandhills in Lamb County (Pesaturo g_ al., 1989).

Pregnant females taken on 5 June and 2 July (two), carried four, five, and five fetuses, respectively, with crown-rump lengths of 7, 4, and 12. Nonpregnant adults were collected in May, July, September, and

November. Testicular lengths of adult males measured 3-6 in May, 4-6 in

June, 5-7 in July, 3-6 in September, and 3 (abdominal) for two in October.

Animals in juvenile pelage or molting from that pelage were trapped in

May, June, July, and October.

Reed and Choate (1986) recently reviewed the systematics of this species on the Great Plains, referring all Texas specimens they examined 45 to the subspecies P. f. copei. In northwestern Texas, this subspecies seemingly occurs in isolated or semi-isolated populations wherever sandy or, sometimes, sandy loam soil prevails. Average (and extreme) external measurements of 20 adults from our series (14 males and six females) from the Muleshoe Sandhills are as follows: total length, 122.6 (117-134); length of tail, 56.3 (51-61); length of hind foot, 16.1 (15-17); length of ear, 6.8 (6-8). Weights of 14 males from this series averaged 9.4

(8.1-10.7). Three nonpregnant females weighed 8.2, 9.0, and 9.9. For cranial measurements, see Table 2.

Specimens examined (43).--BAILEY co.: 0.5 mi. s, 18.2 mi. w Muleshoe,

2; 2 mi. s, 7.5 mi. \.J Muleshoe, 1. LAMB CO.: 4. 6 mi. s, 2.6 mi. E Earth,

2· 5 mi. s, 2.5 mi. E Earth, 3; 6.5 mi. S Springlake, 7· 8 mi. S Springlake, ' ' 1. 3 mi. N, 1 mi. w Fieldton, 1. 3-4 mi. N Fieldton, 16; 5.5-7 mi. S Olton, ' ' 2. HALE CO.: 4-4.5 mi. N, 5.5 mi. W Cotton Center, 8.

Perognathus flavus gilvus Osgood, 1900

Silky Pocket Mouse

This species typically does not occur on fine, sandy soils. Thus, although it is found all around the Muleshoe Sandhills, we did not expect to find that it occurred within the area, where P. flavescens is the common small pocket mouse. However, an adult female P. flavus was trapped

3.2 mi. N Fieldton, Lamb County (about three-fourths of a mile within the sand), in association with flavescens, on 2 December 1967. It clearly differs from specimens of the latter in being smaller and darker, less huffy dorsally and especially on flanks, having larger postauricular patches, larger tympanic bullae, and a narrower interparietal bone.

External measurements: total length, 105; length of tail, 51; length of 46

Table 2. Cranial measurements of adults of two species of Perognathus from the Muleshoe Sandhills. Sample sizes in parentheses, superscript numbers indicate fewer specimens than sample size.

,_., ,_., ~ ,_., ,_., ;3:: ,_., 0 ::l co ;3:: 0 ::l ·.-! ~ +-l 0 1-< ~ +-l UJ Q) 1-< ..c UJ u ·.-! ..c +-l ·.-! 4-1 1-< +-l 0 4-1 1-< 0 co 0 0 0 E +-l 0. 0 +-l ;:::l UJ ..c 1-< +-l ..c 1-< ~ +-l Q) . +-l +-l 0 "0 +-l "0 00 UJ u co ~ ~ ~ ~ 0 Q) ·.-! co co Q) 1-< ,_., 1-< E E UJ ,_., co ..0 4-1 u 4-1 +-l 0 4-1 "--' ·.-! +-l 0 •.-! +-l 0 0 ~ UJ ..0 UJ ..c UJ Q) ..c 1-< Q) +-l ..c ..c •r-i +-l +-l 0 +-l "0 +-l +-l +-l co 00 +-l co co 00 00 co Q) ~ UJ Q) Q) ~ ~ +-l 1-< Q) 0 1-< 1-< Q) Q) C/) c., .....::1 0... c., t:Q .....::1 .....::1

Perognathus flavus gilvus, Lamb County, Texas

TTU 6032, 20.49 6.98 4.42 12.30 3.4 3.3 3.2

Perognathus flavescens copei, Muleshoe Sandhills, Texas (35)

31 31 Ave. (23 , 12 ) 21.48 7.53 5.13 12.31 4.6 3.3 3.2

Min. 20.44 6.80 4.64 11.60 3.9 2.9 2.9

Max. 22.53 8.23 5.51 13.17 5.5 3.5 3.4 47 hind foot, 14; length of ear, 6; weight (nonpregnant), 7.3. See Table 2 for cranial measurements. It is noteworthy that Aday and Gennaro (1973) also recorded P. flavus from several localities in the sandhills near the

Texas border in eastern New Mexico.

Populations now assigned (Hall, 1981) to P. flavus from western Texas are in need of taxonomic review. We tentatively refer our specimen to flavus and to the subspecies gilvus following Wilson (1973), who presented evidence that flavus and merriami (with which gilvus previously was associated) represent a single species.

Chaetodipus hispidus paradoxus (Merriam, 1899)

Hispid Pocket Mouse

Chaetodipus hispidus is generally distributed throughout the sandhills region and in adjacent areas, but we found it only locally common. The largest number taken at a single locality was 15 (in 118 Sherman traps) on 8 October 1988 northwest of Cotton Center, Hale County. Our impression is that this species occupies relatively thick, successional, brushy-weedy habitats, in company with Reithrodontomys megalotis, to a much greater degree than do other heteromyids. Hispid pocket mice commonly were taken by barn owls at a place near the edge of the sandhills in Lamb County

(Pesaturo et al., 1989).

A pregnant female carrying seven fetuses (crown-rump length 10) was taken on 8 June. Animals in juvenile pelage (or molting from that pelage) were collected in March, May, July, October, and November. Testes of adult males measured 9 (March), 10-16 (May), 10-13 (June), 9 (July), and

3-9 (October). 48

Twenty adults (12 males and eight females) had the following average

(and extreme) external measurements: total length, 219.0 (204-240); length of tail, 103.7 (93-117); length of hind foot, 25.4 (23-27); length of ear, 12.7 (11-15). Weights of 12 males from this series averaged 51.0

(36-69); seven nonpregnant females had an average weight of 41.9 (36-51).

Specimens examined (44).--BAILEY CO.: 8.5 mi. S, 16.5 mi. W Muleshoe,

1; 2 mi. S, 7.5 mi. W Muleshoe, 1; 3.2 mi. S, 7.5 mi. W Muleshoe, 5; 4 mi.

S, 1 mi. W Muleshoe, 1; 6 mi. S, 1-2 mi. E Muleshoe, 2; 7 mi. S, 3.5 mi.

E :1uleshoe, 1. LAMB CO.: 7.2 mi. S Earth, 1; 6.5-8.5 mi. S Springlake,

3; 14.8 mi. N Littlefield, 1; 5.5-7 mi. S Olton, 7. HALE CO.: 4.5 mi.

N, 5.5 mi. W Cotton Center, 2; 4 mi. N, 5.5 mi. W Cotton Center, 19.

Dipodomys ordii medius Setzer, 1949

Ord's Kangaroo Rat

This heteromyid is probably the most abundant, certainly the most conspicuous, small mammal of the Muleshoe Sandhills. We trapped or observed sign of kangaroo rats in all habitats sampled, but they were most frequently taken along sandy roadways, in blowout and dune areas, and other places where bare patches of sandy soil prevailed. Their numerous tracks frequently were visible by day as were their burrows in sandy banks and hillsides, at the base of dunes, and beneath shrubs.

Kangaroo rats also were common prey of owls in the region as judged by remains found in pellets from Lamb (Pesaturo et al., 1989) and Bailey counties. We examined more specimens of D. ordii in this study than of anv other species.

We tentatively refer our material to the subspecies D. o. medius, with type locality at Santa Rosa, Guadalupe Co., New Mexico, to the 49 northwest, following Setzer (1949:520). Its relationship to the older-named D. £· richardsoni, which occurs to the north, has not been resolved in detail. Although Schmidly (1971) studied variation in ordii in western Texas (see also Desha, 1967) and concluded that rats from

Hemphill County (north of the Canadian River in the Panhandle) differed from those in populations to the south (including one from Needmore,

Bailey County) to a greater degree than the latter differed from each other, he drew no taxonomic conclusions regarding the variation. Jones et al. (1988) opined that "the Canadian River and its breaks could provide an area where gene flow is, in part at least, discontinuous between populations distributed to the north [richardsoni] and south [medius]."

Also, populations on the Llano Estacada may differ from those to the east

(because both Setzer, 1949:520, and Baumgardner and Schmidly, 1981:25, listed as richardsoni a few specimens from the Rolling Plains to the east of the Llano). In any event, variation in D. ordii on the Llano Estacada and in adjacent areas of western Texas and eastern New Mexico, and the taxonomic inferences that can be drawn therefrom, will be a fruitful area for future study.

Gravid females were captured in the months of January, February,

March, April, May, July, September, October, and November. Two lactating animals were taken in March. Number of fetuses in 21 pregnant animals plus one instance of newborn young in a trap ranged from two to four, with a mode of two and an average of 2.6. Two neonates born in a Sherman live trap, a male and female, had the following respective external measurements: total length, 65, 64; length of tail, 22, 22; length of hind foot, 12, 12; length of ear, 2, 2; weight, 7.0, 6.4. Testicular so lengths of 63 adult males from our series measured 12-15 in January,

10-14 in February, 4-13 in March, 8-13 in April, 12-15 in May, 9-14 in

June, 7-13 in July, 9-12 in September, 11-13 in October, and 12-14 in

November.

Specimens examined (198).--BAILEY CO.: 0.5-2.5 mi. S, 18.2 mi. W

Muleshoe, 4; 5 mi. S, 18 mi. W Muleshoe, 4; 8.5 mi. S, 16.5 mi. W Muleshoe,

2; 9.5 mi. S, 13 mi. W Muleshoe, 15; 2 mi. S, 10 mi. W Muleshoe, 2; 5.5 mi. S, 10 mi. W Muleshoe, 4; 9.5 mi. S, 8.5-9 mi. W Muleshoe, 2; 3.2 mi.

S, 2.7-3.1 mi. W Muleshoe, 5; 4 mi. SW Muleshoe, 1; 4 mi. S, 1-2 mi. W

Muleshoe, 15; 4 mi. S Muleshoe, 1; 4 mi. S, 1 mi. E Muleshoe, 1; 6 mi. S,

1-2 mi. E Muleshoe, 24; 6 mi. S, 3 mi. E Muleshoe, 2. LAMB CO.: 6-7.5 mi. N Sudan, 7; 7.2-7.5 mi. S Earth, 5; 4 mi. S, 2.5 mi. E Earth, 5; 5-5.5 mi. S, 2.5 mi. E Earth, 25; 6-6.5 mi. S Springlake, 9; 14.1-14.7 mi. N

Littlefield, 5; 3-4 mi. N Fieldton, 35; 6.5 mi. N, 0.5 mi. E Fieldton, 1;

7 mi. S Olton, 6; 10.6 mi. N Spade, 2; 8.5 mi. S, 1 mi. E Olton, 3. HALE

CO.: 4-4.5 mi. N, 5.5 mi. W Cotton Center, 12; 0.5 mi. N, 12.5 mi. W

Hale Center, 1.

Additional records (Setzer, 1949:520).--BAILEY CO.: 9 mi. SW

Muleshoe; 6 mi. SW Muleshoe.

Reithrodontomys megalotis aztecus J. A. Allen, 1893

Western Harvest Mouse

We found this harvest mouse to be common in the Muleshoe Sandhills, restricted primarily to weedy, brushy, or grassy fencerows. Frequently it was taken in association with Chaetodipus hispidus, but was trapped along with Dipodomys, Onychomys, and Peromyscus as well. On one occasion, we took a western harvest mouse and a kangaroo rat in the same live trap. 51

Our largest series, 23 specimens, was collected in January, May, and

October, 4 mi. N and 5.5 mi. W Cotton Center, in the small extension of

the sandhills into Hale County. Here, the mice were taken in dense weedy

cover atop a burm and along fencerows overgrown with weeds, grasses, and

some yucca and mesquite.

~· megalotis apparently breeds throughout the warm months in

northwestern Texas. We took pregnant females in March, April, May,

October, and November. In seven pregnancies, number of fetuses ranged

from three to six (average, 4.1). Three neonates (born in a live trap)

had the following ranges of external measurements: total length, 38-40;

length of tail, 9-10; length of hind foot, 5-6; weight, 1.0-1.1.

Nonpregnant adult females were captured in the months of January, February,

March, April, May (including one postlactating), and June. A lactating

individual was trapped in March.

Testes of adult males measured 4-7 in January, 4-6 in February, 8-9

in March, 4-8 in April, 6-9 in May, 7-9 in June, 9 in July, and 8 (two)

in October. Juvenile animals were trapped in May, June, and October.

We follow Jones and Mursaloglu (1961) in referring our material to

R. m. aztecus as did Jones et al. (1988) for specimens from the northern

Texas Panhandle. In size, particularly in the ratio of length of tail

to length of head and body, mice from the Muleshoe Sandhills resemble aztecus. Average (and extreme) measurements of 30 adults (14 males, 16

females) from the sandhills are: total length, 141.6 (134-155); length of tail, 67.8 (61-74); length of hind foot, 16.9 (16-18); length of ear,

14.3 (13-15); ratio of tail to head and body, 92.1 (78.2-110.8). Weights 52 of the males averaged 10.4 (8.0-12.0), whereas those of 11 nonpregnant females averaged 10.6 (9.5-12.0). Cranial measurements of two species of Reithrodontomys from the Muleshoe Sandhills are compared with samples from the Trans-Canadian Panhandle of Texas in Table 3.

Specimens examined (67).--BAILEY CO.: 8.5 mi. S, 16.5 mi. W Muleshoe,

5; 6 mi. s, 1-2 mi. E Muleshoe, 8. LAMB CO.: 4 mi. S, 2.5 mi. E Earth,

9; 5 mi. S, 2.5 mi. E Earth, 2; 6 mi. s Springlake, 2· 3.2 mi. N Fieldton, ' 4· 6.5 mi. N, 0.5 mi. E Fieldton, 1· 5. 5-7 mi. S Olton, 7· 10.6 mi. N ' ' ' Spade, 3. HALE CO.: 4 mi. N, 5.5 mi. W Cotton Center, 23; 0.5 mi. N,

11.5-12.5 mi. WHale Center, 3.

Reithrodontomys montanus griseus Bailey, 1905

Plains Harvest Mouse

This species, which ordinarily avoids deep sandy substrates, has been taken only in the southwestern part of the study area, southwest of

Muleshoe in Bailey County, where it is uncommon on sandy loam soils.

Possibly it occurs only in this part of the sandhills, although this species is widespread on the Llano Estacada.

A pregnant female (five fetuses, 11 in crown-rump length) was trapped on 4 October, whereas a nonpregnant individual was obtained on 5 November.

A March-taken male had testes that measured 5 in length.

Hooper (1952:39) noted that specimens of this species from northwestern Texas, including some from Bailey County, showed evidence of intergradation between R. m. griseus and R. m. montanus to the west.

He assigned them, however, to griseus. Our three specimens had the following external measurements (male followed by two females): total 53

Table 3. Cranial measurements of adults of two species of Reithrodontomys from the Muleshoe Sandhills, and comparative measurements of specimens from the Trans-Canadian Panhandle of Texas (after Jones et al., 1988). Sample sizes in parentheses; superscript numbers indicate fewer specimens than sample size.

,...... , ,...... , c: 0 ~ ;:l ·r-1 0 ~ +J 1-< rn a> u ..c:: CJJ ·r-1 +J ~ co 1-< E 0 0 ..c:: u +J E ;:l 0 +J c: rn E ;:::l 1-< +J ..c:: "0 ·r-1 c: ;:l 1-< +J +J co co 0 ·r-1 +J rn . 00 a> 1-< u c: rn 0 >< c: 1-< ,.0 ...... , co 0 1-< co ,...... ,a> ,.0 1-< 1-< E rn ~ co u ~ u u 0 +J '-+-< 0 ~ ...... , +J ·r-1 ·r-1 ~ 0 0 .;...; rn +J .c ,.0 0 .c CJJ a> co +J 1-< .c +J ..c:: ·r-1 +J E "0 0 .c +J "0 +J +J co 0 co +J +J 00 co 00 co a> 00 a> rn 0.. c: a> c: +J 1-< :>-. 1-< 0 a> a> 1-< a> C/) c..? N c:o 0.. A .....::1 c:o .....::1

Reithrodontomys montanus griseus, Bailey County, Texas

TTU 53055, 20.25 10.46 9.13 2.82 7.70 6.95 3.14 3.28

TTU 53357, 19.27 10.41 9.59 3.03 7.49 6.48 3.63 3.37

TTU 55988, 19.76 10.32 9.75 3.03 7.61 6.59 3.23 3.13

Trans-Canadian Panhandle, Texas (15)

14 14 14 Ave. (7 , 8) 19.74 10.41 9.49 3.01 7.48 6.77 3.58 3.30

Min. 19.14 9.84 9.19 2.88 7.23 6.41 3.35 3.16

Max. 20.25 10.85 9.77 3.15 7.87 6.99 3.78 3.49 54

Table 3. (continued)

rl c:: rl 0 :;: ;=j ·r-1 0 ~ .j..J I-< rn (!) u ..c:: rn ·r-l .j..J ~ co I-< E 0 0 ..c:: u .j..J E ;=j 0 .j..J c:: rn E ;=j I-< .j..J ..c:: '"d ·r-l c:: ;=j I-< .j..J .j..J co co 0 •r-l .j..J rn . 00 (!) I-< u c:: rn 0 >< c:: I-< ..c co 0 I-< co (!) ..c rl I-< I-< E rn rl ~ co u ~ u u 0 .j..J ~ 0 4-1 ·r-1 .j..J ·r-1 ·r-l ~ 0 0 .j..J rn .j..J ..c:: ..c 0 ..c:: rn (!) co .j..J I-< ...c: .j..J ...c: ·r-1 .j..J E '"d 0 ..c:: .j..J '"d .j..J .j..J co 0 co .j..J .j..J 00 co 00 co (!) 00 (!) rn 0.. c:: (!) c:: .j..J I-< >.. I-< 0 (!) (!) I-< (!) C/) c..? N r:o p... 0 .....:l r:o .....:l

Reithrodontomvs megalotis aztecus, Muleshoe Sandhills, Texas (25)

Ave. (12 , 13 ) 21.33 10.79 10.24 3.22 7.98 7.59 3.42 3.35

Min. 20.52 10.41 9.90 3.02 7.59 7.05 3.06 3.18

Max. 22.15 11.37 10.71 3.49 8.50 8.07 3.76 3.59

Sherman County, Texas (19)

Ave. (7 , 12 ) 20.62 10.65 10.08 3.17 7.82 7.15 3.50 3.30

Min. 19.83 10.30 9.62 3.03 7.60 6.80 3.22 3.03

Max. 21.23 11.30 10.44 3.44 8.09 7.69 3.88 3.53 55 length, 121, 120, 141; length of tail, 57, 51, 64; length of hind foot,

16, 16, 15; length of ear, 14, 13, 13; weight, 8.5, 6.9, 11.0 (pregnant).

For cranial measurements, see Table 3.

Specimens examined (3).--BAILEY CO.: 8 mi. S, 11 mi. W Muleshoe, 1;

5.5 mi. S, 10 mi. W Muleshoe, 1; 3.2 mi. S, 7.5 mi. W Muleshoe, 1.

Peromyscus maniculatus luteus Osgood, 1905

Deer Mouse

The deer mouse is a ubiquitous and common small mammal of the sandhill region, being exceeded in numbers on the basis of our trapping results only by Dipodomys ordii and Onychomys leucogaster. We took specimens in all habitats except in Harvard shin oak. These mice appeared to be most common in brushy and grassy areas of sand sage. We were somewhat surprised that the related Peromyscus leucopus did not occur in shin oak and other brushy habitats, but we took none of these mice.

The deer mouse probably breeds throughout the year in western Texas.

In this study, we examined gravid females taken in January, February,

July, October, and November. Number of fetuses ranged from three to six

(average, 4.4) in eight pregnancies.

A female and 10 young were found in a nest of shredded paper and feathers under a piece of plywood in a sandy dump area in Bailey County on 22 April. Total weight of the young was 47.7, whereas the female weighed 17.0. Six of the young had their eyes open, were conspicuously larger (total length, 90-94; weight, 4.8-5.6) than the other four (total length, 73-76; weight, 3.8-4.3), and were attached to the female's mammae.

We suppose this was a communal nest, that the two groups of young represent separate litters, and that a second female escaped us without being seen. 56

Judd (1970) referred P. maniculatus from the Llano Estacado to the subspecies P.m. luteus, with type locality at Kennedy, Cherry Co.,

Nebraska. We tentatively follow Judd, but doubt that our specimens ultimately will prove assignable to luteus. They average larger and darker in color than typical specimens of that subspecies, and most closely resemble P.m. pallescens (type locality, San Antonio, Bexar Co.,

Texas) among races known to occur in Texas and adjacent New Mexico.

Specimens from the Muleshoe Sandhills and elsewhere on the Llano Estacado obviously are in need of critical systematic review. Cranial measurements of a series of adults from the sandhills and a series from the

Trans-Canadian Panhandle of Texas are presented in Table 4.

Specimens examined (76).--BAILEY CO.: 8.5 mi. S, 16.5 mi. W Muleshoe,

2; 2 mi. S, 10 mi. W Muleshoe, 3; 3.2 mi. S, 1.5 mi. W Muleshoe, 11; 6 mi.

S, 2 mi. E Muleshoe, 2. LAMB CO.: 7.5-10.5 mi. N Sudan, 3; 3.5 mi. S

Earth, 1; 4 mi. S, 2.5 mi. E Earth, 3; 4.6 mi. S, 2.6 mi. E Earth, 2;

5-5.5 mi. S, 2.5 mi. E Earth, 13; 6-8 mi. S Springlake, 12; 14.7 mi. N

Littlefield, 1; 3.5-4 mi. N Fieldton, 4; 6.5 mi. N, 0.5 mi. E Fieldton, 1;

5.5-7.2 mi. S Olton, 7. HALE CO.: 4-4.5 mi. N, 5.5 mi. W Cotton Center,

5; 0.5 mi. N, 12.5 mi. WHale Center, 6.

Onychomys leucogaster arcticeps Rhoads, 1898

Northern Grasshopper Mouse

This obligate dust-bather is widely distributed and abundant in the

Muleshoe Sandhills. In our trapping efforts, we found it second in numbers only to Dipodomys ordii, a species with which it almost invariably was associated. We trapped grasshopper mice in all vegetative habitats 57

Table 4. Cranial measurements of adult Peromyscus maniculatus from the Muleshoe Sandhills compared with a sample of adults from the Trans-Canadian Panhandle of Texas (after Jones et al., 1988). Sample sizes in parentheses; superscript numbers indicate~ewer specimens than sample size.

...-i c:: ...-i 0 ::;: ::l ·r-1 0 ...:.:: .j...l ,.... rJ'J Q) u ..c:: rJ'J ·r-1 .j...l 4-1 co ,.... E 0 0 ..c:: u .j...l E ::l 0 .j...l ,.... c:: rJ'J E ,....::l .j...l ..c:: "0 ·r-1 c:: ::l .j...l .j...l co co 0 ·r-1 .j...l rJ'J . 0!) ,.... Q),.... u c:: rJ'J 0 >< c:: ..c co 0 ,.... co Q) ..c ...-i ,.... ,.... E rJ'J ...-i 4-1 co u 4-1 u u 0 .j...l 4-1 0 4-1 ·r-1 .j...l ·r-1 •r-1 4-1 0 0 .:...; rJ'J .j...l .c ..c,.... 0 .c rJ'J Q) co .j...l .c .j...l ..c:: ·r-1 .j...l E "0 0 ..c:: .j...l "0 .j...l .j...l co 0 co .j...l .j...l 0!) co 0!) co Q) 0!) Q) rJ'J 0.. c:: Q) c:: .j...l ,.... >. ,.... 0 Q) Q) ,.... Q) Cl) 0 N >:0 p... ~ .....:I >:0 .....:I

Muleshoe Sandhills, Texas (20)

Ave. (11 , 9 ) 24.26 12.71 11.16 3.79 8.52 9.01 4.47 3.71

Min. 23.42 11.84 10.68 3.47 8.00 8.51 4.05 3.44

Max. 25.22 13.26 11.57 4.07 9.01 9.60 4.93 4.13

Sherman County, Texas (20)

19 19 Ave. (10 ' 10 ) 24.78 12.86 11.03 3.90 8.71 9.42 4.54 3.83 Min. 23.90 12.19 10.52 3.71 8.25 9.05 4.22 3.42

Max. 25.97 13.91 11.41 4.29 9.31 10.10 4.92 4.28 58 sampled, and found them to be the most frequently collected small mammal in grassy-weedy habitats with adjacent sandy areas.

Judging from our data, 0. leucogaster breeds from late winter to early autumn in the sandhill region. Pregnant females were collected in the months of March through July and in September. A postlactating individual with placental scars was captured on 4 October. Females obtained in October, November, January, and February were nonpregnant.

Fourteen gravid animals carried an average of 4.4 (range, three to six) fetuses.

Males taken in October and November had abdominal testes. Length of those of adult males taken in other months measured as follows: 9-16 in January, 6-22 in February, 9-22 in March, 20 in April, 15-23 in May,

13-20 in June, 6-20 in July, 20 in August, and 25 in September. Juvenile mice were collected in May, June, July, October, and November.

Specimens examined (122).--BAILEY CO.: 0.5 mi. S, 18.2 mi. W

Muleshoe, 3; 4.7 mi. S, 18.2 mi. W Muleshoe, 1; 5 mi. S, 18 mi. W Muleshoe,

1; 9.5 mi. S, 13 mi. W Muleshoe, 2; 8 mi. S, 11 mi. W Muleshoe, 1; 2 mi.

S, 10 mi. W Muleshoe, 2; 9.5 mi. S, 8.5 mi. W Muleshoe, 1; 2-3.2 mi. S,

7.5 mi. W Muleshoe, 5; 7 mi. S, 3.5 mi. W Muleshoe, 1; 3.2 mi. S, 1.3-3.1

mi. W Muleshoe, 10; 4 mi. S, 1 mi. W Muleshoe, 2; Muleshoe, 2; 6 mi. S,

1-3 mi. E Muleshoe, 7. LAMB CO.: 6-7.5 mi. N Sudan, 7; 4-5.5 mi. S,

2.5 mi. E Earth, 17; 6.5-8.5 mi. S Springlake, 15; 14.7 mi. N Littlefield,

1; 3.2-4 mi. N Fieldton, 16; 6 mi. SW Olton, 2; 5.5-8.5 mi. S Olton, 15;

10.6 mi. N Spade, 1. HALE CO.: 4-4.5 mi. N, 5.5 mi. W Cotton Center, 7;

0.5 mi. ~. 12.5 mi. WHale Center, 3. 59

Sigmodon hispidus texianus (Audubon and Bachman, 1853)

Hispid Cotton Rat

Cotton rats are generally distributed throughout the Muleshoe

Sandhills, but are not one of the common small mammals of the region.

Many of our specimens were trapped along fencerows, usually where weeds, mid-grasses, shrubs, and yucca prevailed.

Six young were born to a female in a live trap on 4 July. A pregnant

female carrying seven fetuses (5 in crown-rump length) was taken on 5

June. Three more collected on 8 June had five, six, and 11 fetuses that measured 38, 34, and 5 respectively. Adult males captured in February,

May, June, and July (two), had testes that measured, respectively, 15,

18, 22, 20, and 22 in length.

Specimens examined (24).--BAILEY CO.: 9.5 mi. S, 8.5 mi. W Muleshoe,

1; Muleshoe, 1. LAMB CO.: 10 mi. N, 0.2 mi. W Sudan, 8; 7.2 mi. S Earth,

2; 4 mi. S, 2.5 mi. E Earth, 4; 5.5 mi. S Olton, 1; 7 mi. S Olton, 4;

10.6 mi. N Spade, 1; 8.5 mi. S, 1 mi. E Olton, 1. HALE CO.: 4 mi. N,

5.5 mi. W Cotton Center, 1.

Neotoma micropus canescens J. A. Allen, 1891

Southern Plains Woodrat

This woodrat apparently is uncommon in the sandhill area. Specimens are on record only from Lamb County. Few of the characteristic nests were observed, and no remains were recovered from owl pellets analyzed from Bailey and Lamb counties. Of the three rats we collected, two juveniles were taken in Sherman traps on 5 June in dense shrub with yucca, whereas an adult female was shot on 22 April after being chased from a nest in a drainage channel into which some refuse had been thrown. 60

Several other nests were observed in the same area amid yucca, sparse grasses, and some mesquite; one was built atop a large tractor tire, another at the base of a juniper. The female carried three large

(crown-rump length, 44) fetuses.

Specimens examined (6).--LAMB CO.: 5.5 mi. S Earth, 1; 4.6 mi. S,

2.6 mi. E Earth, 2; 14.8 mi. N Littlefield, 1; 3 mi. N Fieldton, 2.

Rattus norvegicus (Berkenhout, 1769)

~orway Rat

Our only record of this introduced rodent from the Muleshoe Sandhills comes from remains of an individual found in owl pellets from 1 mi. N and 5 mi. E Fieldton, Lamb County (Pesaturo et al., 1989). Probably this rat will be found around feedlots and at places of other human activities elsewhere in the region.

Mus musculus Linnaeus, 1758

House Mouse

We took specimens of this Old World native, primarily a commensal of man, in both Hale and Lamb counties, and found remains in owl pellets from Bailey County. Feral populations were found in thick, grassy-weedy situations, and also along the grassy borders of the rangeland. These mice seem to avoid open, sandy areas. A female taken on 14 July carried nine fetuses that were 15 in crown-rump length.

Specimens examined (16).--LAMB CO.: 10 mi. N, 0.2 mi. W Sudan, 2;

6 mi. S Springlake, 1; 14.7 mi. N Littlefield, 1; 4 mi. N Fieldton, 1;

5.5 mi. S Olton, 2. HALE CO.: 4 mi. N, 5.5 mi. W Cotton Center, 5;

0.5 mi. N, 11.5-12.5 mi. WHale Center, 4. 61

Erethizon dorsatum couesi Mearns, 1897

Porcupine

Porcupines evidently range sparingly throughout the sandhill region, although we obtained no specimens. We recorded individuals killed along roadways as follows: 1.5 mi. S Muleshoe, Bailey County, on 4 November

1988; 2.8 mi. SE Muleshoe, Bailey County, on 28 February 1988; 7.5 mi.

N Sudan, Lamb County, on 14 October 1988. We follow Hall (1981) in use of the subspecific name couesi; geographic variation remains to be studied in this species.

Canis latrans latrans Say, 1823

Coyote

This canid is common throughout the study area. We frequently heard coyotes calling in the evening and observed them crossing roadways at night, in one instance an adult followed by three or four young. Ranchers in the sandhills routinely shoot coyotes and hang the carcasses on fencerows.

Specimens examined (5).--LAMB CO.: 4-6 mi. N Fieldton, 5.

Urocyon cinereoargenteus scottii Mearns, 1891

Gray Fox

One of us (Pesaturo) found remains of a gray fox on Farm-to-Market

Road 746 at a place 6.5 mi. Sand 7.5 mi. W Muleshoe, Bailey County, on

27 February 1988. We know of no others from the Muleshoe Sandhills. 62

Mustela frenata neomexicana (Barber and Cockerell, 1898)

Long-tailed Weasel

This weasel seems to be relatively common elsewhere on the central

Llano Estacada, but we neither took nor observed it in the sandhill region. We have examined, however, a single specimen from the sandhills in Lamb County that was obtained in 1968.

Specimens examined (1).--LAMB CO.: 3.3 mi. N Fieldton, 1.

Taxidea taxus berlandieri Baird, 1858

Badger

Although we took no specimens, we observed badgers crossing roadways or killed on them in all three sandhill counties as follows: 9.5 mi. S,

13 mi. W Muleshoe, and 6.5 mi. S, 6.5 mi. W Muleshoe, Bailey County;

7.5 mi. N Sudan, Lamb County; and 4.5 mi. N, 5.5 mi. W Cotton Center,

Hale County. This species evidently is widespread, if uncommon, in the tricounty sandhill area.

Mephitis mephitis varians Gray, 1837

Striped Skunk

This skunk rivals the coyote as the most visible carnivore in the

Muleshoe Sandhills. Individuals commonly were seen along roadways at night. We have observations from all three counties in the study area.

A road-killed female found on 30 April carried six fetuses that measured 31 in crown-rump length. A male trapped at the entrance to its burrow on 11 February had testes that measured 20 x 14.

Specimens examined (4).--BAILEY CO.: 1 mi. N, 7.5 mi. W Muleshoe, 1;

6 mi. S Muleshoe, 1. LAMB CO.: 2 mi. S Earth, 1; 5.5 mi. S, 2.5 mi. E

Earth, 1. 63

Felis rufus texensis (J. A. Allen, 1895)

Bobcat

The bobcat is generally distributed on the Llano Estacada, especially in broken country. We observed an individual 1 mi. N and 5 mi. W Muleshoe in June 1988. Schmidly and Reed (1986) referred bobcats from throughout

Texas to the subspecies texensis.

Odocoileus hemionus crooki (Mearns, 1897)

Mule Deer

Mule deer are uncommon in the sandhill area, being mostly limited to the vicinity of the few draws and their adjacent vegetative cover.

One buck was seen in Lamb County in an area of tall grass near dunes covered with low shrubs. Several landowners reported the occurrence of this deer to us as well, one near the New Mexican border noting that a herd of approximately 20 ranged on his ranch.

Species of Possible Occurrence

The following list contains species that are known to occur in Texas adjacent to the Muleshoe Sandhills or, in several cases, have been reported from the New Mexican part of the region, and well may be found in the tricounty study area. Not included are species like the gray wolf,

Canis lupus, or the bison, Bison bison, which once occupied this country but were long ago extirpated by man.

Notiosorex crawfordi (desert shrew).--This shrew occurs throughout the Llano Estacada and thus no doubt occupies the Muleshoe Sandhills.

Use of pit-fall traps and examination of owl pellets may be the best means to establish its presence. 64

Bats.--There are few places in the tricounty sandhills where bats could find suitable refuge other than an occasional abandoned building or isolated stand of trees. However, some species no doubt migrate over the region in season and it always is possible that strays could be collected there.

Neotoma albigula (white-throated woodrat).--Although we have examined a specimen (TTU 2863) of this rat from Muleshoe National

Wildlife Refuge, south of the sandhills in Bailey County, it doubtfully occurs in the sandhill area.

Vulpes velox (swift fox).--This small canid is widespread and relatively common on parts of the Llano Estacada. Aday and Gennaro

(1973) reported a specimen from the sandhills 4 mi. N and 1 mi. E

Portales, Roosevelt Co., New Mexico, some 18 miles west of the Texas

(Bailey County) border.

Vulpes vulpes (red fox).--A rancher in eastern Lamb County reported to us that he once had seen a red fox. We would expect this species to be sparingly distributed at least along the periphery of the sandhills.

Bassariscus astutus (ringtail).--We have no records of this species from the Texas part of the Muleshoe Sandhills, and we doubt it occurs there with the possible exception of an occasional stray animal. Aday and Gennaro (1973) reported a specimen taken 6.4 mi. N and 6 mi. W

Portales, Roosevelt Co., New Mexico, 11 miles west of the Bailey County border.

Procyon lotor (raccoon).--Several ranchers reported to us that they had seen "coons" in the tricounty study area, but that they were "rare." 65

We observed no evidence of the presence of raccoons, but they may occur infrequently in the Muleshoe Sandhills.

Spilogale interrupta (spotted skunk).--This small mustelid has been reported from nearby Hockley and Lubbock counties, Texas, and possibly occurs in the Muleshoe Sandhills. CHAPTER V

DISCUSSION

While the Muleshoe Sandhills are not a particularly large physiographic feature, the diversity of mammals found there was typical for the Great Plains. Of the 30 species of mammals recorded, 18 were rodents, as was expected for a sandy grassland habitat. Bats, which were not recorded by us, no doubt occur here, and would have increased our species diversity if there were suitable places to trap them.

Based on 19 months of work in the Muleshoe Sandhills, Dipodomys ordii was by far the most common mammal found with Onychomys leucogaster appearing second in frequency. Also common were Peromyscus maniculatus and Reithrodontomys megalotis. We observed or obtained six species of carnivore. Additionally, five other species possibly occur in the sandhill region. Only two of the recorded species, Mephitis mephitis and Canis latrans, could be considered common.

Natural history data such as reproductive, morphometric, and environmental correlations were noted for most species. In several cases,

Geomys, Chaetodipus, Reithrodontomys, and Onychomys, for example, reproductive data (size of testes in males, presence and size of fetuses in females) fluctuated with seasonal changes.

Placement of taxa at the subspecific level was based on geographic variation as established in the literature and physical analysis of specimens from different populations which are housed in The Museum of

66 67

Texas Tech University. Several species are currently in need of taxonomic

review concerning geographic variation and distribution on the southern

Great Plains.

Results obtained from trapping using Sherman live traps and snap

traps were analyzed for the proportion of selected rodent species found

in each of the six plant communities within the Muleshoe Sandhills

(Fig. 7). Two species, D. ordii and 0. leucogaster, were the most

frequently encountered species trapped in five out of the six communities.

The hackberry-shrub series, characterized by sparse sand and dense

vegetation, was the exception. In this community, however, R. megalotis

was the most common (22.7%), probably due to this mammal's propensity

for thick brushy habitats. This species was absent or uncommon in the

mixed grass-yucca series and the sand sage-midgrass series, habitats

abundant in grasses and hence more open. Conversely, Reithrodontomys

montanus was captured only in the mixed grass-yucca series and the

tallgrass-ragweed-Russian thistle association, indicating a preference

for sandy-grassy areas.

Although taken in all but one habitat, Perognathus flavescens was

not especially common except in the sand sage-midgrass series (13.1%),

affirming its attraction to sand sage-dominated sandy areas. Chaetodipus

hispidus was not a common inhabitant of the Muleshoe Sandhills. This

species was trapped second in frequency (19.1%) toR· megalotis in the

hackberry-shrub community, and in fact was found in proportions similar

to R· megalotis in most communities. Peromyscus maniculatus was collected

in moderate numbers in most of the communities. In summary, from our 68 sample size of approximately 500, the species composition proportions in the plant communities generally coincide with habitat preferences found in the literature.

The Muleshoe Sandhills have remained mainly unaltered by human intervention as opposed to the majority of the Llano Estacada, which is a rich agricultural region. Ranching, still commonly practiced in the area, does not alter the original habitat as drastically as farming.

This ensures that the unique physical, floral, and faunal characteristics of the Muleshoe Sandhills will be preserved. LITERATURE CITED

Aday, B. J., Jr., and A. L. Gennaro. 1973. Mammals (excluding bats) of the New Mexican Llano Estacada and its adjacent river valleys. Stud. Nat. Sci., Eastern New Mexico Univ., 1(5):1-33, 44 maps.

Allred, B. W. and H. P. Mitchell. 1955. Major plant types of Arkansas, Louisiana, Oklahoma and Texas and their relation to climate and soils. Texas J. Sci., 7:7-19.

Anonymous. 1936. Spanish explorers, Indian wars, buffalo slaughter in Crosby County history. Pp. 28-37, in The South Plains of Texas historical sketches (Texas Hwy. Dept., Div. 5), 43 pp. [Article originally published in 1936 Anniversary Edition, Crosbyton Review.]

Bailey, V. 1905. Biological survey of Texas. N. Amer. Fauna, 25:1-222.

Barkley, T. M. (ed.). 1986. Flora of the Great Plains. Univ. Press Kansas, Lawrence, vii+1392 pp.

Baumgardner, G. D., and D. J. Schmidly. 1981. Systematics of the southern races of two species of kangaroo rats (Dipodomys compactus and D. ordii). Occas. Papers Mus., Texas Tech Univ., 73:1-27.

Blair, W. F. 1950. The biotic provinces of Texas. Texas J. Sci., 2:93-117.

1954. Mammals of the Mesquite Plains Biotic District of Texas. Texas J. Sci., 4:230-250.

Blakely, E. R. and W. M. Koos. 1974. Soil survey of Hale County, Texas. U.S. Dept. Agric., Soil Conserv. Serv., 1-54 pp. +maps.

Bray, W. L. 1901. The ecological relations of the vegetation of western Texas. Bot. Gazette, 32:195-217, 262-291.

Brown, D. E. 1982. Biotic communities of the American Southwest-United States and Mexico. Desert Plants, 4:1-342.

Brown, D. E., C. H. Lowe and C. P. Pace. 1979. A digitzed classification system for biotic communities of North America with community (series) examples for the Southwest. Appendix 1. Arizona-Nevada Acad. Sci., 14:1-16.

69 70

Davis, W. B. 1974. The mammals of Texas. Bull. Texas Parks Wildlife Dept., 41:1-294.

Desha, P. G. 1967. Variation in a population of kangaroo rats, Dipodomys ordii medius (Rodentia:Heteromvidae) from the High Plains of Texas. Southwestern Nat., 12:275-289.

Diamond, D. D., D. H. Riskind, and S. L. Orzell. 1987. A framework for plant community classification and conservation in Texas. Texas J. Sci., 39:203-221.

Dice, L. R. 1943. The biotic provinces of North America. Univ. Michigan Press, Ann Arbor, viii+78 pp.

Dooley, C. and B. Dooley. 1978. Why stop? (A guide to Texas historical roadside markers.) Gulf Publishing Co., Houston, x+557 pp.

Girdner, C. L., Jr. 1963. Soil survey of Bailey County, Texas. U. S. Dept. Agric., Soil Conserv. Serv., 21(ser. 1959):1-67 +maps.

Gould, F. W. 1978. Common Texas grasses. Texas A&M Univ. Press, College Station, x+267 pp.

Green, F. E. 1951. Geology of sand dunes, Lamb and Hale counties. M. S. thesis, Texas Tech. Call., Lubbock, vi+74 pp.

Hall, E. R. 1981. The mammals of North America. John Wiley & Sons, New York, 2nd ed., 1:xv + 1-600 + 90 and 2:vi + 601-1181 + 90.

Hollander, R. R., J. K. Jones, Jr., R. W. Manning, and C. Jones. 1987. Noteworthy records of mammals from the Texas Panhandle. Texas J. Sci., 39:97-102.

Hooper, E. T. 1952. A systematic review of the harvest mice (genus Reithrodontomys) of Latin America. Misc. Publ. Mus. Zool., Univ. Michigan, 77:1-255, 9 pls.

Jones, B. R. 1959. A sedimentary study of dune sands, Lamb and Bailey counties, Texas, and White Sands National Monument, New Mexico. M. S. thesis, Texas Tech. Call., Lubbock, vi+95 pp.

Jones, J. K., Jr., and B. Mursaloglu. 1961. Geographic variation in the harvest mouse, Reithrodontomys megalotis, on the central Great Plains and in adjacent regions. Univ. Kansas Publ., Mus. Nat. Hist., 14:9-27.

Jones, J. K., Jr., C. Jones, and D. J. Schmidly. 1988. Annotated checklist of Recent land mammals of Texas. Occas. Papers Mus., Texas Tech Univ., 119:1-26. 71

Jones, J. K.. Jr., R. W. Manning, R. R. Hollander, and C. Jones. 1987. Annotated checklist of Recent mammals of northwestern Texas. Occas. Papers Mus., Texas Tech Univ., 111:1-14.

Jones, J. K., Jr., R. W. Manning, C. Jones, and R. R. Hollander. 1988. Mammals of the northern Texas Panhandle. Occas. Papers Mus., Texas Tech Univ., 126:1-54.

Jones, J. K., Jr., D. C. Carter, H. H. Genoways, R. S. Hoffmann, D. W. Rice, and C. Jones. 1986. Revised checklist of North American mammals north of ~1exico, 1986. Occas. Papers Mus., Texas Tech Univ., 107:1-22.

Judd, F. D. 1970. Geographic variation in the deer mouse, Peromyscus maniculatus, on the Llano Estacada. Southwestern Nat., 14:261-282.

Manning, R. W., C. Jones, J. K. Jones, Jr., and R. R. Hollander. 1988. Subspecific status of the pallid bat, Antrozous pallidus, in the Texas Panhandle and adjacent areas. Occas. Papers Mus., Texas Tech Univ., 118:1-5.

McMahan, C. A., R. G. Frye, and K. L. Brown. 1984. The vegetation types of Texas including cropland. Texas Parks and Wildlife Dept., Austin, ii+40 pp. + map.

National Oceanic and Atmospheric Administration. 1985. Climate of the states. Gale Research Co., Detroit, 3rd ed., 2:v+759-1572.

Newman, A. L. 1962. Soil survey of Lamb County, Texas. U. S. Dept. Agric., Soil Conserv. Serv., 7(ser. 1959):1-69 +maps.

Owen, R. D. and M. J. Hamilton. 1986. Second record of Cryptotis parva (Soricidae:Insectivora) in New Mexico, with review of its status on the Llano Estacada. Southwestern Nat., 31:403-405.

Park, C. H. 1936. Indians on the South Plains. Pp. 23-24, in The South Plains of Texas historical sketches (Texas Hwy. Dept., Div. 5), 43 pp. [Article originally published in 1936 Anniversary Edition, Crosbyton Review.]

Pearce, W. M. 1936. A survey of the sand-hill camp sites of Lamb and Bailey counties. Bull. Texas Arch. and Paleo. Soc., 8:184-186.

Pesaturo, R. J., R. W. Manning, and J. K. Jones, Jr. 1989. Small mammals captured by barn owls in Lamb County, Texas. Texas J. Sci., 41:433-434.

Reed, K. M. and J. R. Choate. 1986. Geographic variation in the plains pocket mouse (Perognathus flavescens) on the Great Plains. Texas J. Sci., 38:227-240.

Ross, W. J. and 0. F. Bailey. 1967. Soil survey of Roosevelt County, ~ew Mexico. U.S. Dept. Agric., Soil Conserv. Serv .. 1-72 pp. +maps. 72

Rosson, T. C. 1971. Vascular flora of Bailey County, Texas. M. S. thesis, Texas Tech Univ., Lubbock, iv+145 pp.

Sanders, M. K. 1951. Sand dunes of Bailey County, Texas. M. S. thesis, Texas Tech. Coll., Lubbock, iv+43 pp.

Schmidly, D. J. 1971. Population variation in Dipodomys ordii from western Texas. J. Mamm., 52:108-120.

1984. The furbearers of Texas. Bull. Texas Parks Wildlife Dept., 111:viii+1-55.

Schmidly, D. J., and J. A. Read. 1986. Cranial variation in the bobcat (Felis rufus) from Texas and surrounding states. Occas. Papers Mus., Texas Tech Univ., 101:1-39.

Sellards, E. H., and C. L. Baker. 1934. The geology of Texas. Vol. 2, Structural and economic geology. Bull. Univ. Texas, 3401:1-884.

Sellards, E. H., W. S. Adkins, and F. B. Plummer. 1932. The geology of Texas. Vol. 1, Stratigraphy. Bull. Univ. Texas, 3232:1-1007.

Setzer, H. W. 1949. Subspeciation in the kangaroo rat, Dipodomys ordii. Univ. Kansas Publ., Mus. Nat. Hist., 1:473-573.

Sullivan, J. C. 1980. Differentiation of sand shinnery oak communities in west Texas. M.S. thesis, Texas Tech Univ., Lubbock, vii+100 pp.

Tanner, W. F. 1939. A study of the characteristics and sedimentation of certain sand dunes in Lynn, Lamb and Bailey counties. M. A. thesis, Texas Tech. Coll., Lubbock, viii+71 pp.

Tharp, B. C. 1939. The vegetation of Texas. Anson Jones Press, Houston, xvi+74 pp.

1952. Texas range grasses. Univ. Texas Press, Austin, xiv+125 pp.

Webb, W. L. 1950. Biogeographic regions of Texas and Oklahoma. Ecology, 31:426-433.

Wilson, D. E. 1973. The systematics of Perognathus merriami Allen. Proc. Biol. Soc. Washington, 86:175-191.