Chapter I: Part 1

UNITED STATES DEPARTMENT OF AGRICULTURE

BULLETIN No. 1091

Also Technical Bulletin No. 1 of the Agricultural Experiment Station University of Arizona

Washington, D. C. PROFESSIONAL PAPER September 13, 1922

LIFE HISTORY OF THE KANGAROO RAT _Dipodomys spectabilis spectabilis_ Merriam

BY

CHARLES T. VORHIES, Entomologist Agricultural Experiment Station, University of Arizona; and

WALTER P. TAYLOR, Assistant Biologist Bureau of Biological Survey, U. S. Department of Agriculture

CONTENTS

Importance of Rodent Groups 1 Identification 3 Description 5 Occurrence 7 Habits 9 Food and Storage 18 Burrow Systems, or Dens 28 Commensals and Enemies 33 Abundance 36 Economic Considerations 36 Summary 38 Bibliography 40

WASHINGTON GOVERNMENT PRINTING OFFICE 1922

From _Dipodomys merriami_ Mearns and subspecies, which occur over much of its range, this form is easily distinguished by its larger size and the conspicuous white brush on the tail.]

UNITED STATES DEPARTMENT OF AGRICULTURE

BULLETIN No. 1091

Also Technical Bulletin No. 1 of the Agricultural Experiment Station, University of Arizona

Washington, D. C. PROFESSIONAL PAPER September, 1922

LIFE HISTORY OF THE KANGAROO RAT,

_Dipodomys spectabilis spectabilis_ Merriam.

By CHARLES T. VORHIES, _Entomologist, Agricultural Experiment Station, University of Arizona_; and WALTER P. TAYLOR, _Assistant Biologist, Bureau of Biological Survey, U. S. Department of Agriculture_.

CONTENTS.

Page

Importance of rodent groups 1
Investigational methods 2
Identification 3
Description 5
General characters 5
Color 6
Oil gland 6
Measurements and weights 7
Occurrence 7
General distribution 7
Habitat 7
Habits 9
Evidence of presence 9
Mounds 9
Runways and tracks 10
Signals 11
Voice 12
Daily and seasonal activity 12
Pugnacity and sociability 13
Sense developments 14
Movements and attitudes 15
Storing habits 15
Breeding habits 16
Food and storage 18
Burrow systems, or dens 28
Commensals and enemies 33
Commensals 33
Natural checks 34
Parasites 35
Abundance 36
Economic considerations 36
Control 37
Summary 38
Bibliography 40

NOTE.--This bulletin, a joint contribution of the Bureau of Biological Survey and the Arizona Agricultural Experiment Station, contains a summary of the results of investigations of the relation of a subspecies of kangaroo rat to the carrying capacity of the open ranges, being one phase of a general study of the life histories of rodent groups as they affect agriculture, forestry, and grazing.

IMPORTANCE OF RODENT GROUPS.

As the serious character of the depredations by harmful rodents is recognized, State, Federal, and private expenditures for their control increase year by year. These depredations include not only the attacks by introduced rats and mice on food materials stored in granaries, warehouses, commercial establishments, docks, and private houses, but also, particularly in the Western States, the ravages of several groups of native ground squirrels and other noxious rodents in grain and certain other field crops. Nor is this all, for it has been found that such rodents as prairie dogs, pocket gophers, marmots, ground squirrels, and rabbits take appreciable and serious toll of the forage on the open grazing range; in fact, that they reduce the carrying capacity of the range to such an extent that expenditures for control measures are amply justified. Current estimates place the loss of goods due to rats and mice in warehouses and stores throughout the United States at no less than $200,000,000 annually, and damage to the carrying capacity of the open range and to cultivated crops generally by native rodents in the Western States at $300,000,000 additional; added together, we have an impressive total from depredations of rodents.

The distribution and life habits of rodents and the general consideration of their relation to agriculture, forestry, and grazing, with special reference to the carrying capacity of stock ranges, is a subject that has received attention for many years from the Biological Survey of the United States Department of Agriculture. As a result of the investigations conducted much has been learned concerning the economic status of most of the more important groups, and the knowledge already gained forms the basis of the extensive rodent-control work already in progress, and in which many States are cooperating with the bureau. If the work is to be prosecuted intelligently and the fullest measure of success achieved, it is essential that the consideration largely of groups as a whole be supplemented by more exhaustive treatment of the life histories of individual species and of their place in the biological complex. The present report is based upon investigations, chiefly in Arizona, of the life history, habits, and economic status of the banner-tailed kangaroo rat, _Dipodomys spectabilis spectabilis_ Merriam (Pl. I).

INVESTIGATIONAL METHODS.

Some 18 years ago (in 1903) a tract of land 49.2 square miles in area on the Coronado National Forest near the Santa Rita Mountains, Pima County, southern Arizona, was closed to grazing by arrangement between the Forest Service and the Agricultural Experiment Station of the University of Arizona. Since that time another small tract of nearly a section has been inclosed (Griffiths, 1910, 7[1]). This total area of approximately 50 square miles is known as the United States Range Reserve, and is being devoted to a study of grazing conditions in this section and to working out the best methods of administering the range (Pl. II, Fig. 1).

For some years an intensive study of the forage and other vegetative conditions of this area has been made, the permanent vegetation quadrat, as proposed by Dr. F. E. Clements (1905, 161-175), being largely utilized. During the autumn of 1917 representatives of the Carnegie Institution and the Arizona Agricultural Experiment Station visited the Reserve and were impressed with the evidence of rodent damage to the grass cover. The most conspicuous appearance of damage was noted about the habitations of the banner-tailed kangaroo rat (_Dipodomys spectabilis spectabilis_ Merriam), although it was observed also that jack rabbits of two species (_Lepus californicus eremicus_ Allen and _L. alleni alleni_ Mearns), which were very abundant in some portions of the reserve, were apparently affecting adversely the forage conditions in particular localities. Accordingly, the Biological Survey, the Agricultural Experiment Station of the University of Arizona, the Carnegie Institution of Washington, and the U. S. Forest Service have undertaken a study of the relation of the more important rodents to the forage crop of the Range Reserve in Arizona.

The present paper is a first step in this larger investigation.[2] In this work the authors have made no attempt to deal with the taxonomic side of the kangaroo rat problem. It is not unlikely that intensive studies will show that the form now known as _Dipodomys spectabilis spectabilis_ is made up of a number of local variants, some of them perhaps worthy of recognition as additional subspecies. But it is felt that the conclusions here reached will be little, if at all, affected by such developments.

Color descriptions are based on Ridgway's Color Standards and Color Nomenclature published in 1912.

[Footnote 1: References in parentheses are to the Bibliography, p.
40 (the last figure being to the page of the publication).
References to authorities where no citation of literature is
appended relate for the most part to manuscript notes in the files
of the Biological Survey or the University of Arizona Agricultural
Experiment Station.]

[Footnote 2: In addition to assistance rendered by officials of the
Biological Survey and the University of Arizona, which is hereby
acknowledged, the authors are indebted to the following persons for
helpful suggestions and assistance: G. S. Miller and J. W. Gidley,
of the U. S. National Museum; Dr. Frederic E. Clements and Gorm
Loftfield, of the Carnegie Institution; Morgan Hebard, of the
Academy of Natural Sciences of Philadelphia; James T. Jardine and
R. L. Hensel, both formerly connected with the U. S. Forest Service;
and R. R. Hill, of the Forest Service. They are also indebted to
William Nicholson, of Continental, Ariz., for many courtesies
extended in connection with work on the Reserve.]

IDENTIFICATION.

There are only three groups of mammals in the Southwest having external cheek pouches. These are (_a_) the pocket gophers (Geomyidæ), which have strong fore feet, relatively weak hind feet, and short tail, as compared with weak fore feet, relatively strong hind feet, and long tail in the other two; (_b_) the pocket mice (_Perognathus_), which are considerably smaller than the kangaroo rats and lack the conspicuous white hip stripe possessed by all the latter; and (_c_) the kangaroo rats (_Dipodomys_).

_Dipodomys spectabilis spectabilis_ Merriam requires comparison with three other forms of kangaroo rats in the same general region, namely, _D. deserti_ Stephens, of approximately the same size, and _D. merriami_ Mearns and _D. ordii_ Woodhouse, the last two of decidedly smaller size. The range of _deserti_ lies principally to the west of that of _spectabilis_, and the two do not, so far as known, overlap. On the other hand, _merriami_ and _ordii_, and subspecies, occur over a large part of the range of _spectabilis_, living in very close proximity to its burrows; _merriami_ is even suspected of pillaging the stores of _spectabilis_. The range of _merriami_, however, is much more extensive than that of _spectabilis_ (Fig. 1), which argues against a definite ecological dependence or relationship. Separation of the four forms mentioned may be easily accomplished by the following key:

_Key to Species of_ Dipodomys _in Arizona._

_a^1_. Size much larger (hind foot and greatest length of skull more than 42 millimeters); tail tipped with white.

_b^1_. Upper parts dark brownish buffy; tail dark brownish or
blackish with more sharply contrasted white tip; interparietal
broader, distinctly separating mastoids (range in Arizona mainly
southeastern part) =Dipodomys spectabilis.=

_b^2_. Upper parts light ochraceous-buffy; tail pale brownish with
less sharply contrasted white tip; interparietal narrower, reduced
to mere spicule between mastoids (range in Arizona mainly
southwestern part) =Dipodomys deserti.=

_a^2_. Size much smaller (hind foot and greatest length of skull less than 42 millimeters); tail not tipped with white.

_b^1_. Hind foot with four toes =Dipodomys merriami.=

_b^2_. Hind foot with five toes =Dipodomys ordii.=

On account of the small size, _merriami_ and _ordii_ do not require detailed color comparison with the other two. The general color of the upperparts of _spectabilis_ is much darker than that of _deserti_; whereas _spectabilis_ is ochraceous-buff or light ochraceous-buff grizzled with blackish, _deserti_ is near pale ochraceous-buff and lacks the blackish.

The color of the upperparts alone amply suffices to distinguish _spectabilis_ and _deserti_; but the different coloration of the tail is the most obvious diagnostic feature. The near black of the middle portion of the tail, the conspicuous white side stripes, and the pure white tip make the tail of _spectabilis_ stand in rather vivid contrast to the pale-brown and whitish tail of _deserti_.

The dens of the two larger species of _Dipodomys_--_spectabilis_ and _deserti_--can be distinguished at a glance from those of the two smaller--_merriami_ and _ordii_--by the fact that the mounds of the former are usually of considerable size and the burrow mouths are of greater diameter. On the Range Reserve _merriami_ erects no mounds, but excavates its burrows in the open or at the base of _Prosopis_, _Lycium_, or other brush. The mounds of _spectabilis_ are higher than those of _deserti_, the entrances are larger, and they are located in harder soil (Pl. III, Fig. 1). The dens of _deserti_ are usually more extensive in surface area than those of _spectabilis_, and have a greater number of openings (Pl. III, Fig. 2).

A water-hole scene on the U. S. Range Reserve at the base of the Santa Rita Mountains, Ariz., where cooperative investigations are being conducted to ascertain the relation of rodents to forage. This is typical of a large section of country occupied by _Dipodomys spectabilis spectabilis_ and _Dipodomys merriami_. The brush is mesquite (_Prosopis_), cat's-claw (_Acacia_), and paloverde (_Cercidium_).]

An area of the U. S. Range Reserve in the autumn of 1918, showing the result of failure of summer rains. Such a condition is critical not only for the stockmen but also for kangaroo rats and other desert rodents, and results in competition between them as to which shall benefit by what the range has to offer.]

Typical _Dipodomys s. spectabilis_ mound on the Range Reserve, under shelter of desert hackberry (_Celtis pallida_). Most dens on the reserve are located in the shelter of brush plants, the more important being mesquite (_Prosopis velutina_), cat's-claw (_Acacia_ spp.), and the desert hackberry. (See also Pl. VIII Fig. 2.)]

Den of _Dipodomys deserti deserti_, showing typical wide, low mound with numerous entrance holes. This species excavates its den in soft, sandy soil. The tree is a species of _Dalea_.]

DESCRIPTION.

GENERAL CHARACTERS.

Size large; ears moderate, ear from crown (taken in dry skin) 9 or 10 millimeters; eyes prominent; whiskers long and sensitive; fore feet short and weak; hind feet long and powerful, provided with four well-developed toes; tail very long, usually 30 to 40 per cent longer than the body. Cranium triangular, the occiput forming the base and the point of the nose the apex of the triangle, much flattened, auditory and particularly mastoid bullae conspicuously inflated.

COLOR.

General color above, brownish buffy, varying in some specimens to lighter buffy tints, grizzled with black; oblique hip stripes white; tail with dark-brown or blackish stripes above and below, running into blackish about halfway between base and tip, and with two lateral side stripes of white to a point about halfway back; tail tipped with pure white for about 40 millimeters (Pl. I). Underparts white, hairs white to bases, with some plumbeous and buffy hairs about base of tail; fore legs and fore feet white all around; hind legs like back, brown above, hairs with gray bases, becoming blackish (fuscous-black or chætura-black) about ankles, hairs on under side white to bases; hind feet white above, dark-brown or blackish (near fuscous) below.

Color variations in a series of 12 specimens from the type locality and points widely scattered through the range of _spectabilis_ consist in minor modifications of the degree of coloration, length of white tip of tail, and length of white lateral tail stripes. In general the color pattern and characters are remarkably uniform. Young specimens, while exhibiting the color pattern and general color of adults, are conspicuously less brown, and more grayish.

There appears to be little variation in color with season. In the series at hand, most specimens taken during the fall, winter, and spring are very slightly browner than those of summer, suggesting that the fresh pelage following the fall molt is a little brighter than is the pelage after being worn all winter and into the following summer. But at most the difference is slight.

OIL GLAND.

Upon separating the hairs of the middle region of the back about a third of the distance between the ears and the rump, one uncovers a prominent gland, elliptical in outline, with long axis longitudinal and about 9 millimeters in length. The gland presents a roughened and granular appearance, and fewer hairs grow upon it than elsewhere on the back. The hairs in the vicinity are frequently matted, as if with a secretion. In worn stage of pelage the gland may be visible from above without separating the hairs. Bailey has suggested that this functions as an oil gland for dressing the fur, and our observations bear out this view. Kangaroo rats kept in captivity without earth or sand soon come to have a bedraggled appearance, as if the pelage were moist. When supplied with fine, dusty sand, they soon recover their normal sleek appearance. Apparently the former condition is due to an excess of oil, the latter to the absorption of the excess in a dust bath. The oil is doubtless an important adjunct to the preservation of the skin and hair amid the dusty surroundings in which the animal lives.

MEASUREMENTS AND WEIGHTS.

External measurements include: _Total length_, from tip of nose to tip of tail without hairs, measured before skinning; _tail vertebræ_, length of tail from point in angle when tail is bent at right angles to body to tip of tail without hairs; and _hind foot_, from heel to tip of longest claw.

The following are measurements of a series from the U. S. Range Reserve:

[Transcriber note: Next line was corrected per erratum. The
original text was of the following paragraph (_Averages for 17
adult females: Total length, 326.4 millimeters_).]

Average measurements of 30 adult specimens of both sexes: Total length, 326.2 millimeters (349-310); tail vertebræ, 188.4 (208-180); hind foot, 49.5 (51-47); the average weight of 29 adult specimens of both sexes was 114.5 grams (131.9-98.0).

Averages for 17 adult females: Total length, 326.4 millimeters (349-310); tail vertebræ, 188.8 (208-179); weight (16 individuals), 113.7 (131.9-98.0); excluding pregnant females, 13 individuals averaged 112.9 grams (131.9-98.0).

Averages for 13 adult males: Total length, 326 millimeters (345-311); tail vertebræ, 187.8 (202-168); weight, 116.8 grams (129-100).

There appears to be no significant difference in the measurements and weights of males and females, with the possible exception of the comparison of adult males and adult nonpregnant females.

OCCURRENCE.

GENERAL DISTRIBUTION.

_Dipodomys spectabilis spectabilis_ is found in southeastern Arizona, in northwestern, central, and southern New Mexico, in extreme western Texas, in northern Sonora, and in northern and central Chihuahua (Fig. 1). A subspecies, _D. s. cratodon_ Merriam, has been described from Chicalote, Aguas Calientes, Mexico, the geographic range of which lies in central Mexico in portions of the States of Zacatecas, San Luis Potosi, and Aguas Calientes.

HABITAT.

In the Tucson region _spectabilis_ is typically a resident of the Lower Sonoran Zone. This is perhaps the principal zone inhabited over its entire range, but the animal is often found in the Upper Sonoran also, and in the Gallina Mountains of New Mexico Hollister found it invading the yellow pine Transition where the soil was dry and sandy and the pine woods of open character. The same observer found it common in grassy and weed-grown parks among the large junipers, pinyons, and scattering yellow pines of the Bear Spring Mountains, N. Mex. Bailey calls attention to the fact that the animal apparently does not inhabit the lower half of the Lower Sonoran Zone, as it extends neither into the Rio Grande Valley of Texas nor the Gila Valley of Arizona. In extreme western Texas it is common at the upper edge of the arid Lower Sonoran Zone, and in this region does not enter the Upper Sonoran to any extent.

In July, 1914, Goldman found this kangaroo rat common on the plain at 4,600 feet altitude, near Bonita, Graham County, Ariz., and noted a few as high as 5,000 feet altitude on the warm southwestern slopes of the Graham Mountains, near Fort Grant. Apparently _spectabilis_ reaches its upper altitude limit in the Burro Mountains, N. Mex., where Bailey has found it sparingly on warm slopes up to 5,700 feet, and at the western base of the Sandia Mountains, east of Albuquerque, N. Mex., where dens occur at approximately 6,000 feet.

About Tucson it is undoubtedly more common in the somewhat higher portions of the Lower Sonoran Zone, above the _Covillea_ association, than elsewhere (Pl. IV, Figs. 1 and 2). A few scattered dens are to be seen in the _Covillea_ belt, but as one rises to altitudes of 3,500 to 4,000 feet, and the _Covillea_ is replaced by the cat's-claws (_Acacia_ sp. and _Mimosa_ sp.) and scattered mesquite (_Prosopis_), with the _Opuntia_ becoming less abundant, kangaroo rat mounds come more and more in evidence. Here is to be found the principal grass growth supporting the grazing industry, and the presence of a more luxuriant grass flora is probably an important factor in the greater abundance of kangaroo rats, both _spectabilis_ and _merriami_. In this generally preferred environment the desert hackberry (_Celtis pallida_) is one of the most conspicuous shrubs; clumps of this species are commonly accompanied by kangaroo rat mounds.

In order to ascertain whether the banner-tailed kangaroo rat has any marked preference for building its mounds under _Celtis_ or some other particular plant, all the observable mounds were counted in a strip about 20 rods wide and approximately 4 miles long, an area of approximately 160 acres, particular note being taken of the kind of shrub under which each mound was located. Of 300 mounds in this area, 96 were under _Prosopis_, 95 under _Acacia_, 65 under _Celtis_, 11 under _Lycium_, 31 in the open, 1 about a "cholla" cactus (_Opuntia spinosior_), and 1 about a prickly pear (_Opuntia_ sp.). There is apparently no strongly marked preference for any single species of plant. While both desert hackberry and the cat's-claws afford a better protection than mesquite--since cattle more often seek shade under the latter, and in so doing frequently trample the mounds severely--it appears that the general protection of a tree or shrub of some sort is sought by kangaroo rats, rather than the specific protection of the thickest or thorniest species.

The following records indicate particular habitat preferences of _spectabilis_ as noted at different points in its range:

Occurs on open bare knolls exposed to winds, also on gravelly
places at lower edge of foothills (Franklin Mountains, Tex., Gaut);
here and there over the barest and hardest of the gravelly mesas
(Bailey, Tex., 1905, 147); on open creosote-bush and giant-cactus
desert (Tucson, Ariz., Vorhies and Taylor); on firm, gravelly, or
even rocky soil on the grassy bajada land along the northwest base
of the mountains, either in the open or under _Celtis_, _Prosopis_,
_Lycium_, _Acacia greggii_, or other brush (Santa Rita Mountains,
Ariz., Vorhies and Taylor); mounds usually thrown up around a bunch
of cactus or mesquite brush (Magdalena, Sonora, Bailey); in heavy
soil (Ajo, Ariz., A. B. Howell); loamy soil (Gunsight, Ariz., A. B.
Howell); in mesa where not too stony (Magdalena, Sonora, Bailey);
grassy plain (Gallego, Chihuahua, Nelson); in open valley and high
open plains (Santa Rosa, N. Mex., Bailey); in grassy and weed-grown
parks among the larger junipers, pinyons, and scattering yellow
pines (Bear Spring Mountains, N. Mex., Hollister); on sand-dune
strip (east side of Pecos River, 15 miles northeast of Roswell, N.
Mex., Bailey); among _Ephedra_ patches (San Juan Valley, N. Mex.,
Birdseye); in open sandy soil along dry wash (Rio Alamosa, N. Mex.,
Goldman); on sides and crests of bare, stony hills (Mesa Jumanes,
N. Mex., Gaut); in open, arid part of the valley and stony mesas
(Carlsbad and Pecos Valley, N. Mex., Bailey); about the edges of
the plains of San Augustine and the foothills of the Datil and
Gallina Mountains, and in the Transition Zone yellow-pine forest of
the Gallina Mountains (Datil region, N. Mex., Hollister); on hard
limy ridges (Monahans, Tex., Cary).

A. Brazier Howell notes that _spectabilis_ occurs in harder soil than does _deserti_. This observation is confirmed by others, and seems to afford a conspicuous habitat difference between the two, for _deserti_ is typically an animal of the shifting aeolian sands.

Usually, as on the Range Reserve, the rodents are widely distributed over a considerable area. Occasionally, as in the vicinity of Rio Alamosa, N. Mex., as reported by Goldman, they occur only in small colonies.

View on higher portion of Range Reserve, showing type of country where _Dipodomys s. spectabilis_ is most abundant. Good growth of grama and needle grasses in October, following summer growth and before grazing off by cattle and rodents.]

View on lower portion of Range Reserve, where _Dipodomys s. spectabilis_ is less abundant. Vegetation consists principally of _Lycium_, mesquite, rabbit brush, and cactus, there being very little grass.]

HABITS.

EVIDENCE OF PRESENCE.

MOUNDS.

One traveling over territory thickly occupied by the banner-tailed kangaroo rat is certain to note the numerous and conspicuous mounds so characteristic of the species, particularly if the region is of the savannah type, grassy rather than brushy. These low, rounded mounds occupy an area of several feet in diameter, and rise to varying heights above the general surface of the surrounding soil, the height depending rather more upon the character of the soil and the location of the mound as to exposure or protection than upon the area occupied by the burrow system which lies within and is the reason for the mound.

A den in sandy soil in the open may be of maximum size in area occupied and yet scarcely present the appearance of a mound in any sense, due probably both to the fact that the sandy soil will not heap up to such a height over a honeycomb of tunnels as will a firmer or rocky soil, and also to its greater exposure to the leveling action of rains and the trampling of animals. These mounds are in themselves large enough to attract some attention, but their conspicuousness is enhanced by the fact that they are more or less completely denuded of vegetation and are the centers of cleared areas often as much as 30 feet in diameter (Pl. V, Fig. 1); and further that from 3 to 12 large dark openings loom up in every mound. The larger openings are of such size as to suggest the presence of a much larger animal than actually inhabits the mound. Add to the above the fact that the traveler by day never sees the mound builder, and we have the chief reasons why curiosity is so often aroused by these habitations.

On the Range Reserve the mounds are usually rendered conspicuous by the absence of small vegetation, but Nelson writes that in the vicinity of Gallego, Chihuahua, they can be readily distinguished at a distance because of a growth of weeds and small bushes over their summits, which overtop the grass. In the vicinity of Albuquerque, N. Mex., Bailey reports (and this was recently confirmed by Vorhies) that the mounds about the holes of _spectabilis_ are often hardly noticeable. Hollister writes that in the yellow-pine forests of the Gallina Mountains the burrows are usually under the trunk of some fallen pine, both sides of it in some cases being taken up with holes, there being some eight or ten entrances along each side, the burrows extending into the ground beneath the log. In the vicinity of Blanco, N. Mex., Birdseye says that occasionally _spectabilis_ makes typical dens but more often lives in old prairie-dog holes (_Cynomys_), or in holes which look more like those of _D. ordii_.

RUNWAYS AND TRACKS.

Still other features add to the interest in the dwelling places of _spectabilis_. Radiating in various directions from some of the openings of the mounds well-used runways are to be seen, some of them fading out in the surrounding vegetation, but others extending 30, 40, or even 50 or more yards to neighboring burrows or mounds (Pl. V, Fig. 2; Pl. VI, Fig. 1). These runways and the entrances to the mounds are well worn, showing that the inhabitants are at home and are at some time of day very active. The worn paths become most conspicuous in the autumnal harvest season, when they stand out in strong contrast to surrounding grass. One usually finds not far distant from the main habitation one or more smaller burrows, each with from one to three typical openings, connected by the trail or runway system with the central den, and these we have called "subsidiary burrows" (Pl. VI, Fig. 2). These will be again referred to in discussing the detailed plan of the entire shelter system.

Examination of the runways and of the denuded area about a mound discloses an abundance of almost indecipherable tracks. The dust or sand is ordinarily much too dry and shifting to record clear footprints, and there are no opportunities to see footprints of this species recorded in good impressionable soil. Very characteristic traces of kangaroo rats may be readily observed in the dust about the mounds, however, and these are long, narrow, sometimes curving, furrows made by the long tails as the animals whisk about their work or play.

A typical clearing about a mound of _Dipodomys s. spectabilis_, showing the autumnal denudation of the mound and surrounding areas. In this instance about 30 feet in diameter.]

A small mound of _Dipodomys s. spectabilis_ in early autumn, showing runways radiating from the den. Evidences of activity may be noted in and about the surface of the mound.]

Well-traveled path leading from the main den, in the foreground, to a subsidiary burrow (see Fig. 2, below), about 30 feet distant, at apparent end of runway.]

Located at the end of the 30-foot runway shown in Figure 1, above. This has three openings, two in the foreground and the third a little to the rear and indicated by an arrow.]

SIGNALS.

If a scratching or tapping sound be made at the mouth of a burrow, even in the daytime, one is likely to hear a muffled tapping in response, and this may at times be heard while one is engaged in excavating a mound. It has a chirring or fluttering quality, described by Fisher as resembling the noise of a quail flying. Bailey (1905, 148) is of the opinion that it is used as a signal of alarm, call note, or challenge, a view which the present authors believe to be correct. During the winter of 1920-21, however, both Bailey and Vorhies discovered that this sound, or a very similar one, is made by the rapid action of the forefeet in digging. On one occasion in the laboratory the sound was given by one of a pair and was responded to at once by the other, the two being in separate but contiguous cages. This observation, however, could not be repeated. (Vorhies MS.)

One evening, while working in the vicinity of the Burro Mountains, N. Mex., Goldman heard a kangaroo rat near camp making this thumping noise. Taking a lantern, he approached the den, very cautiously, until within 10 feet. The kangaroo rat was just outside the entrance of one of its burrows, and though moving about more or less restlessly at first showed little fear, and kept up the thumping or drumming at intervals. When making the noise the animal was standing with the forefeet on the ground and the tail lying extended. The noise seemed to be made with the hind feet only, and the vibration of the feet could be seen. The tapping was kept up for a second or two at a time, the sounds coming close together and being repeated rhythmically after a very short interval, suggesting the distant galloping of a horse. After continuing in this way for a short time, the animal turned quickly about, with its head in the opposite direction, and began tapping. It appeared to pay little attention to the light, but finally gave a sudden bound and entered one of its holes about 4 feet from the one in front of which it had been standing.

Vorhies has repeatedly noted when watching for the appearance of a kangaroo rat at night that this sound invariably precedes the rodent's first emergence into the open, and often its appearance after an alarm, though when the storage season has begun and the kangaroo rat is carrying loads of grass heads or other material into its den, it regularly comes out without preliminary signaling. Vorhies has also observed it making the sound while on top of the mound, and certainly not digging, but was unable to see how it was made.

VOICE.

No data concerning any call notes or sounds other than those described above are at hand, with the following exception: Price (in Allen, 1895, 213), who studied the habits of the animal in the moonlight, at Willcox, Ariz., says that a low chuckle was uttered at intervals; and Vorhies has had one captive female that would repeatedly utter a similar chuckle in a peevish manner when disturbed by day, and one captive male which, when teased into a state of anger and excitement, would squeal much like a cornered house rat. Vorhies has spent many moonlight hours observing kangaroo rats, but without ever hearing a vocal sound uttered by free individuals.

DAILY AND SEASONAL ACTIVITY.

The kangaroo rat is strictly nocturnal. An observer watching patiently by a den in the evening for the animal's first appearance is not rewarded until darkness has fallen completely, and unless the moon is shining the animal can hardly be seen. Were it not for the white tail-brush of _spectabilis_ and its white belly when upright on the hind legs and tail, one could not as a rule see the animal at all when it makes its first evening appearance. With the first streak of dawn activity usually ceases completely and much more abruptly than it began with the coming of darkness, but on a recent occasion Vorhies observed that a kangaroo rat which did not appear until near morning remained above ground until quite light, but not fully daylight. On removal of the plug from the mouth of a kangaroo rat burrow, one may sometimes see a fresh mass of earth and refuse shoved into the opening from within. As often as not, however, even this unwelcome attention does not elicit any response by day, the great majority of the burrow openings of this species, as observed by the authors, remaining permanently open.

The ordinary activities of the kangaroo rat in southern Arizona can scarcely be said to show any true seasonal variation. The animals are active all the year in this region, there being neither hibernation nor estivation, both perhaps being rendered unnecessary by the storage habit, to be discussed in full later (pp. 15-16), and by the mildness of the winter climate. On any particular night that the weather is rainy, or the ground too wet and cold, activity is confined to the interior of the burrow system, and for this reason one has no opportunity to see a perfect imprint of the foot in freshly wet soil or in snow. On two or three of the comparatively rare occasions on which there was a light fall of snow on the Range Reserve a search was made for tracks in the snow. At these times, however, as on rainy nights, the only signs of activity were the pushing or throwing out of fresh earth and food refuse from within the burrow. This is so common a sight as to be complete evidence that the animals are active within their dens during stormy weather but do not venture outside. Trapping has again and again proved to be useless on rainy nights, unless the rain is scant and a part of the night favorable, in which case occasional individuals are taken. These statements apply to the Range Reserve particularly; the facts may be quite different where the animals experience more winter, as at Albuquerque, N. Mex., although in November, 1921, Vorhies noted no indications of lessened activity in that region.

PUGNACITY AND SOCIABILITY.

So far as their reactions toward man are concerned, kangaroo rats are gentle and make confiding and interesting pets; this is especially the case with _merriami_. This characteristic is the more surprising in view of the fact that they will fight each other so readily and so viciously, and yet probably it is explained in part by their method of fighting. They do not appear to use their teeth toward each other, but fight by leaping in the air and striking with the powerful hind feet, reminding one most forcibly of a pair of game cocks, facing each other and guarding in the same manner. Sometimes they carry on a sparring match with their fore feet. Biting, if done at all, is only a secondary means of combat. When taken in hand, even for the first time, they will use their teeth only in the event that they are wounded. The jaws are not powerful, and though the animals may lay hold of a bare finger, with the apparent intention of biting, usually they do not succeed in drawing blood. As Bailey says (1905, 148), they are gentle and timid, and, like rabbits, depend upon flight and their burrows for protection.

The well-traveled trails elsewhere described (p. 10) indicate a degree of sociability difficult to explain in connection with their pugnacity toward each other. While three or four individuals may sometimes be trapped at a single mound, more than two are seldom so caught, and most often only one in one night. Trapping on successive nights at one mound often yields the larger number, yet in some cases the number is explained by the fact that two or three nearly mature young are taken, and the capture of several individuals at a single mound can not be taken to indicate that all are from the one den. Our investigations tend strongly to the conclusion that only one adult occupies a mound, except during the period when the young are in the parental (or maternal) den. In the gassing and excavating of 25 or more mounds we have never found more than one animal in a den, except in one instance, and then the two present were obviously young animals.

SENSE DEVELOPMENTS.

Without making special investigations through a study of behavior or other special methods, one can speak in only general terms regarding what appear to be the special sense developments of kangaroo rats. The eyes are large, as is very often the case in nocturnal animals, and when brought out into the bright light of day the rats perhaps do not see well. Yet, if an animal leaves a den which is in process of excavation, and follows one runway, even in bright sunlight, it makes excellent speed to the next opening, often a distance of several yards. Whether this is accomplished chiefly by the aid of sight or in large measure by a maze-following ability, such as experiments have shown some rodents to have, can not be stated without precise experimentation. Marked ability to follow a maze would not be at all surprising in view of the labyrinthine character of the underground passages which make up the normal habitation.

When watching beside a mound by moonlight one is impressed with the fact that the rats possess either a very keen sense of hearing or of sight, probably both. The very slightest movement or noise on the part of the observer results, with a timid individual, in an instantaneous leap for safety, a disappearance into the burrow so sudden as to be almost startling. All attempts to obtain flashlight photographs at the mounds were failures, the animal either having gotten completely out of the field before the light flashed following the pull of the trigger, or leaving merely an indistinguishable blur on the plate as it went, and this in spite of carefully hiding the trigger chain behind a screen. A slight noise accompanying the trigger action gave the alarm in one case, and in another the length of time of the flash was sufficient for the get-away. The marvelous quickness of the animal clearly indicates a remarkably short reaction time. Occasionally a bold individual is found, as in the case of one which came out repeatedly, even after being flashed twice in the same night.

Certain peculiar physical characteristics suggest a relationship to sense reactions. On these, however, the authors are not prepared to do more than offer suggestions for future work. The extremely large mastoids found in kangaroo rats suggest a connection in some way with special developments of the sense of hearing or of balance. It may be noted that an intermediate condition between the kangaroo rats and the majority of rodents in respect to this character is to be found in the pocket mice (_Perognathus_), which belong to the same family. Herein lies a field for some interesting experimentation and discovery.

The small, pointed nose might suggest a not overkeen sense of smell, and there appears no reason to believe that this sense is particularly well developed. However, the turbinals are very complex. The vibrissæ are long and sensitive, and may indicate a special development of the sense of touch as an adaptation to nocturnal habits and to life in an underground labyrinth. The long, well-haired tail doubtless serves as an important tactile organ as well as a balance.

MOVEMENTS AND ATTITUDES.

Movements and attitudes are characteristic. As a kangaroo rat emerges from the burrow a reason for the relatively large size of the opening is seen in the fact that, kangaroolike, the animal maintains a partially upright position. Its ordinary mode of progression is hopping along on the large hind legs, or, when in the open and going at speed, leaping. When moving slowly about over the mound, as if searching for food, it uses the fore legs in a kind of creeping movement. It appears to be creeping when pocketing grain strewn about, but close observation shows that the fore feet are then used for sweeping material into the pockets, reminding one somewhat of a vacuum cleaner. When it assumes a partially upright position the fore limbs are usually drawn up so closely that they can be seen only by looking upward from a somewhat lower level than that occupied by the animal. The slower movements of searching or playing about the mound are occasionally interrupted by a sudden leap directly upward to a height of 1-1/2 to 2 feet, often with no apparent reason other than play. This is, however, a fighting or guarding movement, though indulged in for play. The play instinct seems to be well developed, and in evidence on any moonlight night when actual harvesting operations are not going on.

STORING HABITS.

Probably no instinct is of greater importance to the kangaroo rat than that of storing food supplies. When a crop of desirable seeds is maturing the animal's activities appear to be concentrated on this work. During September, 1919, when a good crop of grass seed was ripening following the summer rains, a kangaroo rat under observation made repeated round trips to the harvest field of grass heads. Each outward trip occupied from 1 to 1-1/2 minutes, while the unloading trip into the burrow took only 15 to 20 seconds.

One individual in a laboratory cage, which had not yet been given a nest box, busied itself in broad daylight in carrying its grain supply into the darkest corner of the cage. When a nest box is supplied the individual will retreat into its dark shelter, and will only come forth after darkness has fallen unless forcibly ejected, but will store the food supplied.

In another case an animal escaped while being handled, and sought refuge behind a built-in laboratory table, where it could not be recovered without tearing out the table. For four days and nights it had the run of the laboratory. On the first night of its freedom it found and entered a burlap bag of grass seed that had been taken from a mound. A trail of seed and chaff next morning showed that it had been busily engaged in making its new quarters comfortable with bedding and food. After four nights of freedom it was captured alive in a trap, and later it was found that it had moved from the corner behind the table to the space beneath a near-by drawer, where it had stored about 2 quarts of the grass seed and a handful of the oatmeal used for trap bait.

BREEDING HABITS.

Observations on breeding habits have consisted mainly in taking records from the females trapped at all seasons of the year throughout the course of the investigation, and from examinations made during poisoning operations, and yet from this source the number of pregnant females taken or of young discovered is disappointingly small. The records indicate a breeding period of considerable length, extending from January to August, inclusive. It is possible that the length of the period may be increased by a second litter from the earliest breeding females in summer, but the large percentage of nonpregnant or nonbreeding animals which occurs throughout the season would indicate a wide variation in the time of breeding of different individuals.

Trapping in February and March for the purpose of securing greater numbers of female specimens, begun with the idea that these months were most likely to be the breeding months, has invariably yielded an unsatisfactory number of nonbreeding specimens and males. Unfortunately, the numbers of females secured in some months were not sufficient to be significant if worked out in percentages of breeding and nonbreeding individuals, and this, coupled with the fact that the importance of recording carefully all nonbreeders was not at first recognized, makes it impossible to tabulate such information reliably. The total of females taken in April, for example, is only 3, of which 1 was breeding; while in June, during the course of poisoning operations, 45 females were examined, of which 21 were breeding.

Five breeding females were taken in January, all during the last three days of the month. One of these was a suckling female, the young of which were secured alive and were probably at least a week old when taken. This must have been exceptionally early for young, since of a number of adult kangaroo rats taken during the first week of January none have been found to be breeding. Two records from Vernon Bailey are as follows: May 19-June 8, 1903, young specimen in nest (Santa Rosa, N. Mex.); June 12, 1889, one female, two embryos (Oracle, Ariz.).

The considerable proportion (which we believe to be more than 50 per cent) of nonbreeding females taken during all those months in which breeding has been found to occur may also indicate an extended period of breeding, with a small percentage breeding at any one time. This period also furnishes ample time for the rearing of two litters a year by some females, but we have no evidence as to the occurrence of two litters. Young of the year, practically grown, are taken during and after the month of April.

The mammae are arranged in three pairs, pectoral, 1/1; inguinal, 2/2.

Kangaroo rats are among those rodents in which the vagina becomes plugged with a rather solid material, translucent, and of the consistency of a stiff gelatine, after copulation. This must occur very soon after coitus, since in those individuals taken in this condition no definite evidence of the beginning of development of embryos could be detected by examination.

The length of the gestation period of _spectabilis_ is unknown. The young are born naked, a fact inferred by failure to find any fetus showing noticeable hair development, and from the conditions observed in such young as have been seen. A suckling female was taken by Vorhies, January 31, 1920, and her den immediately excavated in the hope of securing the young. Two juveniles were found in a special nest chamber (see p. 30). These were estimated to be perhaps two weeks old. A serious effort was made to raise the little animals by feeding milk with a pipette and keeping them warm with a hot water bottle, but they survived only 10 days, without the eyes having opened. The uneven temperature as well as the character of the food was probably responsible for their deaths. On February 3 they were measured and weighed, with the following results:

--------------------------------------------------- | | Measurements (in millimetres). | Weight |------------------------------- | (in | Total | Tail | Hind | grams). | length. | vertebrae. | foot. ---------|---------|---------|------------|-------- No. 1 | 13.3 | 90 | 38 | 24 No. 2 | 12.6 | 93 | 38 | 24 ---------------------------------------------------

At this stage the young were partially clothed with a coat of fine velvety fur, more especially on the bodies, the tails being still nearly naked. The body color was dark plumbeous, just the color of the dark underfur of the adult, or a shade darker, while the characteristic white markings of the adult stood out sharply as pinkish-white areas against the dark background (see Pl. IX, Fig. 2, at p. 32). The proportions were much as in the adult, except that the tails were relatively much shorter and the feet relatively longer.

Only one other record of young is at hand, that by Bailey, who secured the young after capture of a suckling female at Santa Rosa, N. Mex. In this case the litter contained only one. This was squeaking when found, but was not large enough to crawl away. Its eyes and ears were closed, and its soft, naked skin was distinctly marked with the pattern of the adult, the colors being as given for the other two. This juvenile lived only a week. Young less than half grown were not trapped or noted in our poisoning operations outside the dens.

Kangaroo rats, if _spectabilis_ be representative, reproduce at a slow rate as compared with many other small rodents. We have records of 67 females with embryos or scars showing the number produced, and of the two litters of young described above. Of the 69 females thus recorded, 15, or 21.7 per cent, had but one offspring each; 52, or 75.3 per cent, but two each; while only 2 individuals, or 2.9 per cent, had three. Three young is the maximum litter recorded. This, taken in connection with the protracted breeding season and lack of sure evidence of the production of two broods a year, gives a surprisingly low rate of reproduction, indicating relative freedom from inimical factors.

Our breeding records for _merriami_ are fewer than for _spectabilis_, but are very similar in every way so far as they go, both as to the time of year and number of young.

FOOD AND STORAGE.

_Dipodomys s. spectabilis_ does not hibernate, so must prepare for unfavorable seasons by extensive storage of food materials. There are two seasons of the year, in southeastern Arizona at least, when storage of food takes place, namely, in spring, during April or May, and in fall, from September to November, the latter being the more important. For the periods between, the animal must rely largely on stored materials. Not infrequently a season of severe drought precludes the possibility of any storage. The summer and fall of 1918 was such a season on the Range Reserve (Pl. II, Fig. 2). If food stores are inadequate at such a time the kangaroo rats must perish in considerable numbers. Fisher found many deserted mounds in the vicinity of Dos Cabezos, Ariz., in June, 1894, which may be accounted for in this way. In 1921 Vorhies found all mounds within 4 or 5 miles of Albuquerque, N. Mex., deserted by _spectabilis_, resulting probably from overgrazing by sheep and goats during a succession of dry years. In the arid Southwest natural selection probably favors the animals with the largest food stores, and it is not surprising that the storing habit has been developed to a remarkable degree.

Some stored material is likely to be found at any time of year in any mound examined, the largest quantity usually in fall and winter, the smallest in July or August (Table 1, dens 1, 2, 14, and 24). Amounts found by different observers vary from a few ounces to several quarts or pecks, and stored materials taken from 22 mounds on the Range Reserve vary in weight from 5 to 4,127 grams (more than 9 pounds). This is exceeded by one lot from New Mexico, which totaled 5,750 grams (12.67 pounds). It is fairly evident that in seasons of scanty forage for stock the appropriation of such quantities of grass seeds and crowns and other grazing materials by numerous kangaroo rats may appreciably reduce the carrying capacity of the range. Studies of cheek-pouch contents and food stores taken from dens show that the natural food of _spectabilis_ consists principally of various seeds and fruits, particularly the seeds of certain grasses. The study of burrow contents has been especially illuminating and valuable.

All of the stored material from 22 dens on the Range Reserve and from 2 near Albuquerque, N. Mex., has been saved and analyzed as to species as carefully as the conditions of storage would permit. Within the mound the food stored is usually more or less segregated by plant species, though the stores of material of any one kind may be found in several places through the mound, and often the material is mixed. In the latter case the quantities of the various species can only be estimated, but in the former the species may be kept separate by the use of several bags for collecting the seeds, and a fairly accurate laboratory weighing can be made later. Very frequently, the explanation of this separation of species lies in the different seasons of ripening, but sometimes where two species are ripe at the same time near the mound, one is worked upon for a time to the exclusion of the other. The one kind is often packed in tightly against the other, but with a very abrupt change in the character of the material.

A number of the more interesting and representative results of the weighing and analyses of burrow contents are presented herewith in tabular form. The data for each den, or lot, shows in grams the quantity of stored material removed and the best estimate it was possible to make of the percentages or weights of the various species. When the weight was less than 5 grams, the mere trace of the species frequently is indicated in the following tables by the abbreviation "Tr."

TABLE 1.--_Analyses of plants stored by _Dipodomys spectabilis spectabilis_ Merriam, obtained from examination of representative dens (all except Den 24 from U. S. Range Reserve, near the Santa Rita Mountains, Ariz.)._

DEN 1.

February 7, 1918. Burrow typical, located on bank of wash in partially denuded grass-land, _Bouteloua rothrockii_ and weed type; soil sandy; burrow photographed in section (Pl. VII, Fig. 1).

Species stored. Grams.

Bouteloua rothrockii 2,205 Bouteloua aristidoides (B. eriopoda and B. rothrockii, Tr.) 1,445 Plantago ignota 442 Eriogonum polycladon 35 ----- Total 4,127

Four species of plants represented in burrow contents (Pl. VII, Fig. 2). Maximum quantity for single burrow in series of 22 from Range Reserve.

DEN 2.