University of Kansas Publications

Museum of Natural History

Vol. 4, pp. 1-466, plates [1]-[41], 31 figures in text

December 27, 1951

[AMERICAN WEASELS]

BY

E. RAYMOND HALL

University of Kansas

Lawrence

1951

University of Kansas Publications, Museum of Natural History

Editors: E. Raymond Hall, Chairman, A. Byron Leonard, Edward H. Taylor, Robert W. Wilson

Vol. 4, pp. 1-466, plates [1]-[41], 31 figures in text

December 27, 1951

University of Kansas

Lawrence, Kansas

PRINTED BY

FERD VOILAND, JR., STATE PRINTER

TOPEKA, KANSAS

1951

23-3758

Plate 1.

Coloration of head and foreparts in ten subspecies of long-tailed weasel, Mustela frenata. All figures are of males, approximately × 1/2.

In regions of heavy rainfall (see figs. [2] and [3]) there is an increase in pigmentation and extent of blackish color backward over the neck and a decrease in extent of the white facial markings. In regions progressively more arid (see figs. [3] to [7]) there is a decrease in pigmentation and extent of blackish color and an increase in extent of the white facial markings.

As shown by rearing mammals from humid regions in arid regions, and vice versa, the color is not visibly altered in one or a few generations; the color is an hereditary character. Beginning with the southernmost subspecies (fig. [1]) and continuing northward to the northern subspecies (fig. [10]) there is a darkening, next a lightening, and finally a darkening closely conforming to amounts of precipitation in the geographic regions concerned. A fuller discussion of this correlation is given on page [51].

Fig. 1. Map showing localities of capture of specimens depicted in plate [1].

American Weasels

BY

E. RAYMOND HALL

[CONTENTS]

American Weasels

By E. Raymond Hall

[INTRODUCTION]

The weasel's agility and speed take it in and out of retreats, over obstacles and across open places in amazingly rapid fashion and are responsible for the animal's actions being described as "quick as a flash." The common long-tailed weasel of the United States measures approximately a foot and a half in length, of which the tail comprises a third; but the round, slender body is scarcely more than an inch and a half in diameter. Brown above and whitish below in summer dress, the animal is sleek as well as lithe and graceful. It is easy to understand, therefore, why the Bavarian name Schönthierlein (pretty little creature) and the Italian name donnola (little lady) were bestowed upon it. The Spanish name is comadreja (godmother).

In the winter, in temperate and northern regions, the coat becomes pure white except for the black tail-tip. In this dress the correct name for the animal is ermine, a mammal whose fur is known to all and justly esteemed, especially for its luster in artificial light, where it is scarcely excelled in enhancing the beauty of gems and their feminine wearers.

In relation to its weight, the weasel is thought to be unsurpassed, and perhaps it is unequalled among mammals, in the effectiveness with which it exercises its carnivorous heritage; it kills with speed and strength a wide variety of animals including many much larger than itself; and it has been known to attack even man himself when he stood between the weasel and its intended prey. In structure and temperament it is so highly specialized for offense that, when opportunity affords, it sometimes kills, for storage in its larder, far more than enough to meet its immediate needs. After speaking of this tendency, Elliott Coues (1877:129) has said:

"A glance at the physiognomy of the weasels would suffice to betray their character. The teeth are almost of the highest known raptorial character; the jaws are worked by enormous masses of muscles covering all the side of the skull. The forehead is low and the nose is sharp; the eyes are small, penetrating, cunning, and glitter with an angry green light. There is something peculiar, moreover, in the way that this fierce face surmounts a body extraordinarily wiry, lithe, and muscular. It ends in a remarkable long and slender neck in such a way that it may be held at right angle with the axis of the latter. When the creature is glancing around, with the neck stretched up, and flat triangular head bent forward, swaying from one side to the other, we catch the likeness in a moment—it is the image of a serpent." Although Coues' colorful description more closely links the weasel with the symbol of evil than pleases me, his description does emphasize the raptorial character of the weasel.

Even though most weasels are intractable as pets, they have a value to man, as, for instance, when he is plagued by mice. In a field where mice and other small rodents are so abundant as to damage cultivated crops, the weasel is the farmer's best friend. A weasel may inhabit one den until the rodents thereabouts are almost exterminated in an area two or three hundred yards across; in this way the weasel acts as a control, locally, as well as a check more widely, on the increase in size of populations of kinds of rodents upon which it preys. The smaller species are mousers of remarkable efficiency and can, if necessary, follow a mouse to the end of the mouse's burrow. The slender body allows the weasel to pass through any burrow or hole into which it can thrust its head. This ability in an organism as highly specialized for killing other animals as is the weasel, has earned for it a bad name in connection with poultry yards. Authentic instances are recorded in which a weasel, gaining entrance through a knot-hole to a coop of young chickens, killed several dozen of the fowls. In other instances, however, weasels have lived under buildings close by a poultry yard without even molesting the birds in the slightest; in the latter instances the weasels probably were present because there was an abundant supply of rats and mice. At least three poultry raisers (see page [214]) have encouraged weasels to live in their poultry yards feeling that the good they do by destroying rats outweighs the damage caused by the occasional weasel which turns to the fowls; the idea is that the individual weasel can be eliminated if he becomes destructive.

Although tending to be nocturnal, weasels are almost as active by day as by night. Their young, numbering 4 to 9, are born in a nest in a burrow and as with other members of the Order Carnivora, are blind, and incapable of looking after themselves at the time of birth. In Mustela frenata of Montana, breeding occurs in July and August, and the young are born in the following April and May. Wright (1948A:342) showed that the gestation period could not have been less than 337 days in one individual and that it averaged 279 (205-337) days in 18 instances. Findings of the same author (1942B:109) showed that the embryos are implanted only 21 to 28 days before the young are born. In the preceding part of the "long gestation period, the embryos lie dormant in the uterus as un-implanted blastocysts. The young female weasel [of M. frenata] mates when 3 or 4 months old." Consequently, in the spring, all females of this species may produce young (Wright, 1942A:348). The circumboreal species Mustela erminea likewise has been shown to have a delayed implantation of the ova. Each of these two species, M. frenata and M. erminea, has only one litter per year; but the weasel, Mustela nivalis, of the Old World seems to lack the delayed implantation, in this respect resembling the ferret (subgenus Putorius) as it does also in its ability to have more than one litter per year (see Deanesly, 1944). The manner of reproduction in the South American species M. africana and the circumboreal species M. rixosa at this writing is unknown.

The genus Mustela includes the true weasels, the ferrets and minks. The ferrets commonly are treated as a subgenus, Putorius, along with the Old World polecat. The minks usually are accorded subgeneric distinction under the name Lutreola, and the true weasels comprise the subgenus Mustela, the three subgenera together, along with some other subgenera which are mostly monotypic, comprising the genus Mustela. Considered in this way, the group of true weasels, subgenus Mustela, has a geographic range roughly coextensive with that of the genus Mustela. This range includes Asia and Europe, Northern Africa, North America and northern South America. Java has its weasel. Australia and nearly all the oceanic islands lack weasels, and the animals are absent from roughly the southern half of Africa and the southern half of South America. Other small mustelids, weasellike in shape and with corresponding habits and dentition, take the place of true Mustela in the southern half of Africa and in the corresponding part of South America.

In America the subgenus Mustela occurs from the northernmost land in Arctic America southward to Lake Titicaca in the Andes of South America, a distance of approximately 6900 miles. Felis, I think, is the only other genus of land mammals in the western hemisphere that has a geographic range as extensive from north to south. Felis does not range so far north but does range farther south. The one species, Mustela frenata, ranges from Lake Titicaca northward to about 57° N in British Columbia or for approximately 5000 miles in a north to south direction and from within the Alpine Arctic Life-zone through the Tropical Life-zone. In North America, weasels occur in almost every type of habitat, being absent only in the extremely desert terrain of western Arizona and western Sonora and in adjoining parts of California and Baja California. Even this area, along the Colorado River, may support some weasels; evidence suggesting that it does so is given in the account of Mustela frenata neomexicana.

[PALEONTOLOGICAL HISTORY]

The paleontological record fails to show the precise ancestry of Mustela. The genus has been found in deposits of Pleistocene age, but, so far as I can ascertain, not in deposits of earlier times. The Pleistocene remains are not specifically distinct from Recent (living) species, and in only a few instances (see M. f. latirostra and M. e. angustidens) are they even subspecifically distinct from the Recent weasel living in the same area today. It is true that fossil remains from deposits of several stages of the Tertiary beds have in the past been identified in the literature as Mustela, but most of these identifications were made many years ago when the generic name Mustela was used in a far broader and more inclusive sense than it is today and much of the fossil material was so fragmentary that the generic identity could not be ascertained, at least at that time. Because the generic identity could not be ascertained, the fossil material was tentatively assigned to the genus Mustela, the "typical" genus of the family Mustelidae instead of to some other more specialized or less well-known genus of the family. To satisfy my curiosity about these species of "Mustela" of a geological age earlier than the Pleistocene I have personally studied nearly all of the original specimens from North America and have found each to be of some genus other than Mustela. Also, such study as I have been able to make of the Old World fossils themselves that have been referred to the genus Mustela up to 1938, and my study of the illustrations and descriptions of the others from there lead to the same conclusion; that is to say, none that is true Mustela is known up to now from deposits older than the Pleistocene.

When, in 1930 (pp. 146-147), I wrote about the taxonomic position of three American genera of fossils (known only from lower jaws), each of which had been previously referred to the genus Mustela, I said that they pertained "to that section of the weasel family (Mustelidae) which comprises the polecats, true weasels, ferrets, minks and martens. The fossil specimens . . . are smaller than any other later Tertiary members of the group yet described, and are more primitive than any of the above mentioned Recent relatives. Of the three extinct genera . . . Miomustela [Lower Pliocene or Upper Miocene of the Lower Madison Valley, Montana] is the most primitive and Martinogale [Pliocene, 18 mi. SE Goodland, Sherman County, Kansas] is the most advanced. This view rests largely on the character of M₌₁ which in Miomustela has a deeply basined, short, narrow talonid with a thick, high metaconid situated partly posterior to the protoconid. In Martinogale the talonid is incipiently trenchant, long, broad, and it has a lesser developed metaconid which is situated more anterior [ly]. Pliogale [Lower Pliocene, Humboldt County, Nevada] is intermediate in this respect.

"These three forms are of special interest as possible ancestors of the subgenus Mustela, true weasels. No members of this subgenus, nor related forms which can with any degree of certainty be regarded as directly ancestral to them, have yet been described from Miocene or Pliocene deposits. Palaeogale of the Old World and Bunaelurus of North America, each of Oligocene age, have been placed by Schlosser (1888, p. 116) and Matthew (1902, p. 137) as members of the primitive group of mustelids ancestral to Mustela. This course seems logical; and with no truly intermediate links between these forms of the Oligocene on the one hand, and Mustela which first appears in the Pleistocene, on the other, more definite statements about ancestral positions of the small Oligocene forms can hardly be made. The deciding considerations for authors who placed Palaeogale and Bunaelurus as ancestral to Mustela were the absence of a metaconid on M₁ and the trenchant talonid of that tooth. These characters are found also in Mustela. On the other hand certain structures in the basicranial region of Palaeogale and more especially of Bunaelurus indicate that these genera possibly are not close to the ancestral form of Mustela . . . Martinogale may stand near the ancestral form of Mustela and . . . Pliogale may be ancestral to Martinogale. Pliogale, in turn, may have had an ancestor similar to Miomustela. If this should prove to be the case, Palaeogale and Bunaelurus might be regarded as an independent branch which displays merely a parallelism to Mustela in the loss of the metaconid on M₁ and the development of a trenchant talonid on that tooth. The writer would make it clear that he does not hold such to be the case. The ancestral relation of Martinogale to Mustela is presented merely to show the possibility, and not the special probability, of such an origin for Mustela. Knowledge of the tympanic bullae and other structures of the basicranial region would go far toward answering the question and until these structures are known [in mustelids of the Later Tertiary,] some uncertainty will remain."

At the present writing I can add to the above statement only a few facts. The discovery of better material of Bunaelurus than was available to previous workers led Simpson (1946), correctly I think, to synonymize Bunaelurus with Palaeogale. Simpson figures the cranial foramina in Palaeogale. The differences, between Palaeogale and Mustela, in cranial foramina, possibly are only the result of the elongation of the tympanic bullae. The bullae of the subgenus Mustela are seen to be much elongated posteriorly if comparison is made with the bullae of earlier mustelids. Consequently, it might be concluded that there is nothing in the arrangement of the cranial foramina which would preclude the derivation of Mustela from Palaeogale. However, the anterior situation of the carotid foramen—well forward along the medial margin of the tympanic bulla—is a character typical of other mustelids and the posterior location of this foramen in Palaeogale might indicate that it was not ancestral to Mustela.

[SKELETON AND DENTITION]

The outstanding features of a weasel's skeleton are its length and slenderness. Whereas the length of the vertebral column measured from the atlas (the first cervical vertebra) to the last sacral vertebra is 175 per cent of the length of the hind leg (as measured from the head of the femur to the tip of the longest claw), the corresponding percentage is only 116 in the raccoon. Stated in another way, the vertebral column and the hind leg are of approximately equal length in a raccoon, but in a weasel the vertebral column is one and three-fourths times as long as the hind leg.

VERTEBRAE

The vertebral column consists of 7 cervicals, and ordinarily 14 thoracics, 6 lumbars, 3 sacrals and, depending on the species, 11 to 23 caudals. For the three species of which skeletons were examined, variations from the normal number of vertebrae are noted in the following table:

Table 1
Data on vertebrae in three species of the subgenus Mustela
(Numerals in parentheses indicate number of specimens)

Variation according to the species is evident in the number of caudal vertebrae, but in the other categories of vertebrae no consistent difference in number according to species was found in the material examined. Apparently there is also some geographic variation in the number of caudal vertebrae within a species. For example, the one skeleton seen of Mustela rixosa eskimo (no. 219036, U. S. Nat. Mus., from St. Michaels, Alaska) has only 11 caudal vertebrae, whereas in the 11 Mustela rixosa rixosa from Roseau County, Minnesota, the usual number is 15 with extremes of 14 and 16. Similarly specimens of Mustela frenata from Idaho and California almost always have 1 or 2 more caudal vertebrae than do individuals of the shorter-tailed subspecies of the same species from eastern Kansas.

Of the vertebrae, only the cervicals, of which there are 7, were found to be constant in number. In M. erminea, two of the seven individuals in which the anticlinal vertebra was the 12th (instead of the 11th) had 15 instead of the customary 14 thoracic vertebrae. In M. frenata, seven of the twenty-seven individuals in which the anticlinal vertebra was the 12th (instead of the 11th) had 15 instead of 14 thoracic vertebrae. The one M. erminea with a pseudosacral vertebra had only two instead of the customary 3 sacral vertebrae but the same individual had 15 thoracic vertebrae. Of the six M. frenata with a pseudosacral vertebra, two animals had only two instead of three sacral vertebrae. Conceivably, therefore, the pseudosacral vertebra in each of the three instances mentioned may represent merely an unfused sacral vertebra, instead of a true pseudosacral as occurs in four individuals of M. frenata.

TEETH

In American weasels, for example in Mustela frenata, the permanent dentition normally is

I 3 C 1 P 3 M 1
-, -, -, -, -, -, -, -
i 3 c 1 p 3 m 2

or 34 teeth in all. In most respects the dentition is typical for post-Tertiary mustelids but in several parts is highly specialized for a diet of flesh, the degree of this specialization being second only to that of the cats, family Felidae. The outstanding specialization is in the first lower molar, in which, as in the cats, the internal cusp (metaconid) is completely suppressed and the heel (talonid) forms an elevated blade for cutting food rather than a basin for crushing it. In one sense the tooth is simplified since it owes its distinctive form to a reduction in number of parts; nevertheless, the distinctive form of the lower molar clearly is correlated with a diet of flesh, and the tooth is correctly to be thought of as the lower blade of a pair of shears; the upper blade is the fourth upper premolar. The reduction in size of the second (last) lower molar and small size of the inner lobe of the one remaining upper molar probably are additional modifications for a diet of flesh.

The absence of the last two upper molars and last molar in the lower jaw would be expected in any mammal as highly specialized for a diet of flesh as is the weasel, but these teeth are absent also in other Quaternary members of the family Mustelidae, many of which are substantially less specialized for a diet of flesh than is the weasel. Therefore, in the weasel, it is reasonable to regard the absence of these teeth more as a heritage than as an indication of a special adaptation. The absence of a first premolar above and below, as in the weasel, is to be expected in any carnivore that has the first lower molar and fourth upper premolar highly specialized for shearing, but the loss of these premolars and the small size of the second premolars may be as much the result of a slight shortening of the face as it is a result of a lengthening of the third and especially the fourth premolars. The lengthening of these more posteriorly-situated teeth would appear to be an adaptation to a diet of flesh. The cause of the lengthening of the mentioned teeth and the reason for the absence of the first premolars probably will be unknown until the fossil record is more complete.

The teeth of American species vary little except in size. The absence of P2 in Mustela africana is the only difference of a qualitative (presence or absence) nature that was detected. Also, the Central American subspecies of Mustela frenata exhibit a tendency to early loss of P2 and thus foreshadow the condition typical of M. africana.

As a whole the dentition of the weasel exhibits a high degree of specialization for a diet of flesh and this specialization is fully as evident in the deciduous dentition as in the permanent dentition.

The deciduous, or milk, dentition, of Mustela frenata, as known from immature specimens of Mustela frenata noveboracensis and Mustela frenata frenata available for this study, is comprised of canines, one on each side above and below, and 3 cheek teeth on each side above and below. See figures [2-9]. The upper cheek teeth from anterior to posterior are: a minute peglike tooth in general similar to the first premolar of the permanent dentition; a shearing tooth in general similar to P4 of the permanent dentition; and an anteroposteriorly compressed tooth in general similar to M1 of the permanent dentition. In the lower jaw, behind the canine, there is first a minute peglike tooth, second a two-rooted tooth similar in general outline to a permanent third premolar, and finally a shearing tooth corresponding in function to m1 of the permanent dentition.

No postnatal specimens which show deciduous incisors have been examined.

Selected, outstanding differences between the permanent teeth and the deciduous teeth are as follows: In the deciduous teeth the canine above has on the posterior face a well-defined ridge extending from the tip to the cingulum. This ridge is absent or at most faintly indicated in the permanent tooth. The lower deciduous canine, in cross section is seen to have a marked indentation on the anteromedial border in the region of the cingulum; this indentation is lacking in the permanent tooth. The anterior one of the deciduous cheek teeth, both above and below, is single rooted and its crown-surface is only about one-fifteenth as much as that of the anterior premolar of the permanent dentition. The second deciduous cheek tooth below has two roots, usually fused, and differs from p4 of the permanent dentition in having the tip of the principal cusp more recurved, in having the anterior basal cusp better developed and the posterior heel less well developed.

The second deciduous cheek tooth above corresponds in function and general plan of construction to P4 of the permanent dentition but differs from that tooth in the more pronounced protostyle, longer tritocone, more posteriorly located deuterocone and as noted by Leche (1915:322) separation of the protocone and tritocone by a notch. The third upper deciduous tooth has a single cusp internally and two cusps laterally. Thus it reverses the relation of parts seen in M1 where the internal moiety is larger than the lateral or buccal moiety. The third deciduous tooth below differs from m1 in very much shorter talonid and separation of the paraconid from the protoconid by a deeper notch.

All the features in which the last two deciduous teeth, both above and below, are described as differing from their functional counterparts in the permanent dentition, are features found in the permanent teeth of primitive fossil mustelids and certain fossil and Recent viverrids. Even so, taking into account Leche's (1915) work, which shows that the milk teeth of some carnivores have structures lacking in the corresponding permanent teeth of the same individual animal and also in the teeth of genera that seem to be ancestral, a person suspects that some of the structural features mentioned above are not inheritances of ancestral conditions but rather specializations of the milk dentition.

Figs. 2-9. Views of permanent and deciduous teeth of Mustela frenata nigriauris. Incisors not shown. In each instance teeth are of the left side.

Permanent dentition × 3. No. 32421, Mus. Vert. Zoöl., ♂, adult; Berkeley, Alameda County, California; obtained October 4, 1921, by D. D. McLean.

Deciduous dentition × 5. No. 132158, U. S. Nat. Mus., ♂, juvenile; Stanford University, Santa Clara County, California; obtained May 7, 1898, by W. K. Fisher.

Figs. 2-3. Lateral views of upper teeth, of adult and juvenile respectively.

Figs. 4-5. Occlusolingual views of upper teeth of adult and juvenile respectively.

Figs. 6-7. Lateral views of lower teeth of adult and juvenile respectively.

Figs. 8-9. Occlusolingual views of lower teeth of adult and juvenile respectively.

In other deciduous teeth there is clearer evidence of more specialization for a diet of flesh in the deciduous teeth than in the permanent teeth. For example, the upper carnassial of the milk dentition is even more highly sectorial than is the permanent tooth and strikingly like that of some of the cats. The lower tooth that is effective in the shearing action bears no more trace of the metaconid than does the permanent first lower molar. These features of the deciduous dentition suggest that it is more specialized for a diet of flesh than is the permanent dentition. If this be the fact, it may seem especially remarkable because the commonly employed term "milk teeth" suggests that the animal makes but little or no use of these teeth in the short time that they are in place. Accordingly, the student may credit the form of these teeth more to some indirect effects of inheritance than to natural selection acting directly upon the teeth. But, after all, natural selection probably is responsible for the form of these teeth as is indicated by the observations of Hamilton (1933:318-325). He found that these milk teeth are used for eating solid food as soon as the principal shearing teeth are in place. This is three weeks after birth and before all of the deciduous teeth have broken through the gums. These shearing teeth are used for almost two months before being replaced by the permanent teeth and it is, therefore, evident that natural selection could operate to fully as great a degree in determining the form of the deciduous teeth as it may with the permanent teeth.

Hamilton (1933:325-326) found that the permanent dentition was complete at 75 days after birth in captive specimens of Mustela frenata noveboracensis. In the same subspecies, he noted 28 days after birth that the canines and carnassial teeth [second deciduous cheek tooth above and third below] had erupted through the gums. Animals 45 days old, Hamilton found, were losing the milk dentition, and had the gums broken through by several of the permanent cheek teeth.

Study of the cleaned skulls available of juveniles indicates that the deciduous teeth which persist longest are, on each side of the mouth, the second cheek tooth above and the third cheek tooth below. These teeth persist until after the permanent P4 and m1 have come into use. These permanent teeth are situated immediately behind their functional counterparts of the milk dentition. P3 and p4 are the teeth of the permanent dentition which ultimately push out the last milk teeth to be lost. Accordingly, in the permanent dentition, P4 and M1 appear before P3 does, and m1 and m2 make their appearance before p4.

[DISPARITY IN NUMBERS OF MALES AND FEMALES (IN ZOOLOGICAL COLLECTIONS)]

The question has frequently been asked why twice as many male as female weasels are captured. This is the proportion in research collections, as may be seen from table no. 2, and I am convinced that the specimens in these collections are saved in approximately the same proportion as that in which they are caught. Although it might be assumed, upon first consideration, that there are twice as many males as females in nature, selective factors enter into the catch. For example, because a male weasel is approximately twice as heavy as a female, it may be necessary for him, in a given length of time, to travel twice as far as the female to obtain the required amount of food with the result that a given number of traps or snares will catch twice as many males as females. Indeed, Glover (1943B:8) shows that, on the average, in Mustela frenata noveboracensis in Pennsylvania, the male actually does travel slightly more than twice as far as the female (704 feet versus 346 feet). From table no. 2, it may be seen that in most winter months the ratio is 3 males to one female. This ratio is reasonable enough, in view of what has been said, if it is considered also that the lighter weight of the female permits her safely to step on the pans of traps that would be sprung by heavier males.

If in the breeding season, which is April through August in M. frenata, the female is passive and if the male is restlessly searching for her, he may thus increase still more his chances of being caught in traps set for weasels.

My own studies of live weasels in nature indicate that in the season when females are attending young which are half grown, or larger, the adult male weasels live singly in dens of their own, separate and apart from the females and their young (Hamilton, 1933:328, records adult males living with the female and her young, but possibly this was when the young were less than half grown). Perhaps these males at that time travel no farther than is necessary to obtain food for themselves. Females, at this time, forage not only to meet their own needs, but for food to supply their young as well. At this time, in May and June, as may be seen from table no. 2, almost as many adult females as adult males are caught. The reason why only relatively more females than in other months, instead of actually more females than males, are caught at this time probably is that the adult males also are extraordinarily active at this time because they are in breeding condition. Perhaps the explanation in part is to be found in the lesser weight of the female (approximately half of the male's weight) which, as indicated above, permits her to step on the pan of a steel trap without springing it whereas the heavier male does spring the trap and as a consequence is caught. Hamilton (1933:299-300), who mentions this selective factor, found an equal number of males and females in the three newly born litters that came under his observation.

Table 2
Specimens of Mustela frenata (north of the range of M. f. frenata) arranged by sex and under each sex by age

I suppose that in nature there are approximately equal numbers of male and female weasels and further suppose that the selective factors which cause more males than females to be caught are the greater distances traveled by the males and their greater weight.

[MATERIALS, ACKNOWLEDGMENTS AND METHODS]

At a late stage in the preparation of this manuscript a total of 5,457 specimens had been examined. For the most part these were conventional study-specimens; that is to say, they were stuffed skins with the skulls separate and each was accompanied by the customary data as to locality of capture, date of capture, name of collector, external measurements and sex recorded on the labels by the collectors. Skulls unaccompanied by skins, nevertheless, comprised a large share of the total and a small proportion was made up of skins unaccompanied by skulls, mounted specimens, skeletons, and entire animals preserved in liquid.

It was the recognition of this need for specimens from extensive areas from which no specimens previously had been collected that influenced me, approximately a year after the study was begun, to allot for it a long span of time. The procedure adopted, in general, was to study the weasels of one species from a given geographic area in so far as the material warranted, then lay this aside until additional critical material could be obtained, and finally, some months or a year later, complete the account. In this fashion the manuscript of the American weasels received my attention in each of the past twenty-five years (September, 1926 to date of publication). This is a confession of fact rather than a recommendation of procedure. This type of procedure unduly delays the diffusion of knowledge and for a variety of reasons justifiably annoys other students of the subject. Nevertheless, many gaps have been filled that otherwise would have remained open. Although specimens to solve several problems still remain to be collected and studied, it seems that a point of diminishing returns has now been reached, which, in fairness to all concerned, calls for publication of the results so far obtained.

For assistance in the entire undertaking, I am more indebted to Miss Annie M. Alexander than to any other one person; she provided the means by which specimens from critical areas were obtained, made it possible to examine the European collections, and assisted in other ways. The late Professor Joseph Grinnell and Mr. Charles D. Bunker, among others, gave truly valuable encouragement and assistance.

Collections containing weasels which were examined in the study here reported upon were as follows:

Acad. Nat. Sciences of Philadelphia American Mus. Nat. History Baylor University Berlin Zoological Museum Boston Society of Natural History Brigham Young University British Museum of Natural History California Academy of Sciences Carnegie Museum Charleston Museum Coe College Collection of J. Arnold Collection of Stanley C. Arthur Collection of Rollin H. Baker Collection of William Bebb Collection of R. H. Coleman Collection of Ian McTaggart-Cowan Collection of Stuart Criddle Collection of John Cushing Collection of Walter W. Dalquest Collection of William B. Davis Collection of J. M. Edson Collection of Ralph Ellis Collection of John Fitzgerald, Jr. Collection of Mr. Green Collection of Ross Hardy Collection of Donald V. Hemphill Collection of L. M. Huey Collection of R. W. Jackson Collection of Stanley G. Jewett Collection of E. J. Koestner Collection of J. E. Law Collection of A. H. Miller Collection of Lloye H. Miller Collection of R. D. Moore Collection of J. A. Munro Collection of O. J. Murie Collection of Robert T. Orr Collection of Arthur Peake Collection of Kenneth Racey Collection of William B. Richardson Collection Rocky Mt. Spotted Fever Lab. Collection of Victor B. Scheffer Collection of William T. Shaw Collection of O. P. Silliman Collection of W. E. Snyder Collection of Frank Stephens Collection of T. C. Stephens Collection of D. D. Stone Collection of Myron H. Swenk Collection of Joe and Dean Thiriot Collection of John Tyler

Collection of Jack C vonBloeker Collection of Alex Walker Collection of Edward R. Warren Colorado Museum of Natural History Charles R. Conner Museum Cornell University Donald R. Dickey Collection Field Museum of Natural History Florida State Museum Fresno State Junior College Humboldt State Teachers College Illinois Natural History Survey Iowa State College Iowa Wesleyan College Kansas State Agric. College Leland Stanford Junior University Leningrad Academy of Science Los Angeles Mus. Hist. Art and Sci. Louisiana State University Mt. Rainier Nat'l Park Collection Museum of Comparative Zoölogy Mus. Polonais d'Hist. Nat., Warsaw Mus. Vert. Zoöl., Univ. California Museum of Zoölogy, Univ. Michigan National Museum of Canada Naturhistoriska Ricksmuseum, Sweden Neuchatel University Museum New York State Museum Ohio State Museum Oklahoma Agric. and Mech. College Ottawa University, Kansas Paris Museum Provincial Museum of British Columbia Royal Ontario Museum of Zoölogy San Diego Society of Natural History State Hist. and Nat. Hist. Soc. Colo. State Normal School, Cheney, Wash. Texas Cooperative Research Collection United States National Museum University of Arkansas Univ. California Mus. Palaeo. University of Idaho Univ. Kansas Mus. Nat. History University of Minnesota University of Notre Dame University of Oklahoma University of Oregon University of South Dakota University of Utah Univ. Washington Museum of Zoölogy University of Wisconsin Univ. Zool. Mus., Copenhagen

The largest single collection is in the United States National Museum, where the specimens of the National Museum proper and the United States Biological Surveys Collection, together, provide essential materials including a large share of the holotypes. Specimens in all of the North American collections including Canada and México have been made available, by loan, and in 1937 materials were examined in the principal collections of northern and central Europe. After the materials in North American collections were assembled, special effort, with considerable success, was made in each of several winters, to obtain specimens from areas not previously represented in collections.

To the many persons who were in charge of the collections consulted, to those who at my request sought critical specimens, and to those who assisted in various stages of assembling data and in preparation of the manuscript, I am grateful indeed. Likewise, I am deeply appreciative of the grants-in-aid received from the Carnegie Institution of Washington, the University of California Chapter of Sigma Xi, the John Simon Guggenheim Memorial Foundation and the Kansas University Endowment Association. I am mindful also of an obligation to those who appropriated funds, by legislative action, for research use by The University of California and The University of Kansas.

For assistance with the illustrations I am indebted to the late Major Allan Brooks for [Plate 1] , to Mrs. Mary Blos for figures 25-31, to Miss Ann Murray for figures [11-13], to Mr. W. C. Matthews for all the photographs, to Mrs. Freda L. Abernathy for figures 2-9, 18-22, 24, and for retouching all the photographs except the following which were retouched by Mrs. Virginia Unruh: figs. d of plates [2], [3], [4], [9], [10], [11], [16], [17]; figs. i of plates [5], [6], [7]; figs. h, j, k of plate [7]; figs. f and g of plates [12] and [13]; and figs. c and d of plate [14]. To Mrs. Unruh I am further indebted for figures [1], [16], [17] and [23] and for much terminal assistance with preparing most of the illustrations for the engraver.

• Acad. Nat. Sciences of Philadelphia
• American Mus. Nat. History
• Baylor University
• Berlin Zoological Museum
• Boston Society of Natural History
• Brigham Young University
• British Museum of Natural History
• California Academy of Sciences
• Carnegie Museum
• Charleston Museum
• Coe College
• Collection of J. Arnold
• Collection of Stanley C. Arthur
• Collection of Rollin H. Baker
• Collection of William Bebb
• Collection of R. H. Coleman
• Collection of Ian McTaggart-Cowan
• Collection of Stuart Criddle
• Collection of John Cushing
• Collection of Walter W. Dalquest
• Collection of William B. Davis
• Collection of J. M. Edson
• Collection of Ralph Ellis
• Collection of John Fitzgerald, Jr.
• Collection of Mr. Green
• Collection of Ross Hardy
• Collection of Donald V. Hemphill
• Collection of L. M. Huey
• Collection of R. W. Jackson
• Collection of Stanley G. Jewett
• Collection of E. J. Koestner
• Collection of J. E. Law
• Collection of A. H. Miller
• Collection of Lloye H. Miller
• Collection of R. D. Moore
• Collection of J. A. Munro
• Collection of O. J. Murie
• Collection of Robert T. Orr
• Collection of Arthur Peake
• Collection of Kenneth Racey
• Collection of William B. Richardson
• Collection Rocky Mt. Spotted Fever Lab.
• Collection of Victor B. Scheffer
• Collection of William T. Shaw
• Collection of O. P. Silliman
• Collection of W. E. Snyder
• Collection of Frank Stephens
• Collection of T. C. Stephens
• Collection of D. D. Stone
• Collection of Myron H. Swenk
• Collection of Joe and Dean Thiriot
• Collection of John Tyler

• Collection of Jack C vonBloeker
• Collection of Alex Walker
• Collection of Edward R. Warren
• Colorado Museum of Natural History
• Charles R. Conner Museum
• Cornell University
• Donald R. Dickey Collection
• Field Museum of Natural History
• Florida State Museum
• Fresno State Junior College
• Humboldt State Teachers College
• Illinois Natural History Survey
• Iowa State College
• Iowa Wesleyan College
• Kansas State Agric. College
• Leland Stanford Junior University
• Leningrad Academy of Science
• Los Angeles Mus. Hist. Art and Sci.
• Louisiana State University
• Mt. Rainier Nat'l Park Collection
• Museum of Comparative Zoölogy
• Mus. Polonais d'Hist. Nat., Warsaw
• Mus. Vert. Zoöl., Univ. California
• Museum of Zoölogy, Univ. Michigan
• National Museum of Canada
• Naturhistoriska Ricksmuseum, Sweden
• Neuchatel University Museum
• New York State Museum
• Ohio State Museum
• Oklahoma Agric. and Mech. College
• Ottawa University, Kansas
• Paris Museum
• Provincial Museum of British Columbia
• Royal Ontario Museum of Zoölogy
• San Diego Society of Natural History
• State Hist. and Nat. Hist. Soc. Colo.
• State Normal School, Cheney, Wash.
• Texas Cooperative Research Collection
• United States National Museum
• University of Arkansas
• Univ. California Mus. Palaeo.
• University of Idaho
• Univ. Kansas Mus. Nat. History
• University of Minnesota
• University of Notre Dame
• University of Oklahoma
• University of Oregon
• University of South Dakota
• University of Utah
• Univ. Washington Museum of Zoölogy
• University of Wisconsin
• Univ. Zool. Mus., Copenhagen

The methods of study, after specimens were assembled, included first comparisons of specimens of like age and sex from each of several localities to ascertain the constant features by which full species were distinguishable, one from the other. For example, it was found that in every individual from Trout Lake, Washington, of the species here designated Mustela erminea, the postglenoidal length of the skull amounted to more than 47 per cent of the condylobasal length whereas it was less than 47 per cent in all individuals here designated as Mustela frenata, from the same locality. Testing of specimens from other localities by means of this and other selected characters permitted the outlining of the geographic ranges of the full "species-groups." By comparing specimens of other nominal species and by examining specimens from localities geographically intermediate between the nominal species, I found intergradation and therefore arranged the nominal species as subspecies of a single species. Intergradation here is understood to be the result of crossbreeding in nature between two kinds of animals in the area where the geographic ranges of the two kinds meet. Presence of intergradation between two kinds of weasels was basis for according them subspecific rank. Absence of intergradation in nature at every place where the geographic ranges of two kinds met or overlapped, and absence of intergradation by way of some other kind, or chain of kinds, was basis for according each of the two kinds full specific rank. By thus applying the test of intergradation, or lack of it, I found that there were four full species of weasels, of the subgenus Mustela, in all of the Americas.

Next, the specimens of one species were arranged in trays in a geographic sequence. The specimens from any one locality were segregated by sex and under one sex from one place were arranged from oldest to youngest, that is to say by age. The four series with the largest numbers of individuals of a given age were selected. Seventeen cranial measurements and three external measurements were recorded for each individual of each of these four series. For each measurement, the coefficient of variation, standard deviation and probable error were computed. The four samples subjected to such analysis were a series of adult males, one of adult females, one of subadult males and one of subadult females. Also, studies of each sex were made to ascertain seasonal changes in pelage. After data were obtained on ontogenetic (age) variation, secondary sexual variation, seasonal variation, and degree of individual variation by studying specimens in the manner described above, tests were made for subspecific (geographic) variation by comparing series of specimens of like sex, age and season, from different localities. For each one of several geographically variable features noted, a map was prepared for animals of each sex. When all the data thus obtained were codified, subspecific ranges were, in a sense automatically, obtained. On the resulting map showing geographic ranges of subspecies for a species, a type locality was accurately plotted for each name that had been applied to the species, and names then were applied in accordance with the international rules of zoölogical nomenclature.

[VARIATION]

Variation with Age

The kind of variation which results from increasing age has been dealt with extensively for the skull (of the Old World Mustela erminea) by Hensel (1881) and for the external features and to some extent for the skull by Hamilton (1933) in the North American forms M. erminea cicognanii and M. frenata noveboracensis.

The young of both erminea and frenata are hairless and blind at birth. In M. frenata noveboracensis, the eyes open on approximately the 37th day. When 2 to 4 months old, the tail is pointed at the tip. This is because the terminal hair of the tail, including the black tip, is short and lies flat on the tail. In subadults and adults the hair on the terminal part of the tail is as long as that on the basal part, and the tail appears to be of uniform diameter all the way out to the end.

In the western subspecies of M. frenata, and in its tropical subspecies, animals so young as to have pointed tails commonly have the underparts of the body more intensely colored than do adults. The young may have salmon-colored instead of yellowish fur on the underparts.

Otherwise, in animals that have attained approximately adult proportions—which appears to be at approximately 6 months of age in males—there are no variations which are ascribable to increasing age in the color-pattern or pelage that cause the systematist to confuse species or subspecies.

Of the several parts of the skull in juvenal animals, the braincase and width of the posterior part of the palate are most nearly of the size attained in the adult, the facial part of the skull at birth is the least developed, and the interorbital region is, in relation to its ultimate adult size, intermediate in stage of development. The permanent teeth are acquired when the animal is approximately eleven weeks old.

Four age groups, based on characters of the dentition and skull, have been recognized. They are:

Juvenile.—One or more deciduous (milk) teeth present. Birth to three months of age.

Young.—Sutures widely open between the maxillae and nasals and between the premaxillae and nasals. Three to seven and a half months of age.

Subadult.—Sutures between maxillae and nasals visible but indistinct. Seven and a half to ten months of age.

Adult.—Bones of rostrum coalesced with no traces of sutures visible to the naked eye. More than ten months old.

The skull as a whole increases in size until the animal is two-thirds of the way through the stage designated as young. After this time the width of the rostrum, as measured across the hamular processes of the lacrimals, increases until approximately a third of the way through adulthood. The interorbital breadth decreases from late subadulthood to adulthood and even in adults there appears to be a slight decrease in this part of the skull with increasing age.

The average zoölogist will readily distinguish skulls of juveniles and young from adults but usually fails to distinguish subadults from adults. Nevertheless, subadults must be distinguished from adults if geographic variation is to be measured accurately. The reason for this is that such differences in the form (not size) of the skull as result from increasing age equal and often exceed the differences of a geographic sort which serve for distinguishing subspecies that have adjoining geographic ranges. All sutures in the skull, except those between the tympanic bulla and the braincase, and those on the dorsal face of the rostrum, are obliterated while the animal is a subadult. Most kinds of mammals retain sutures throughout life or until the animals are well into adulthood. Therefore, skulls of weasels offer fewer features for estimating age than do those of most mammals and the skulls of weasels that are subadults or older are more difficult to classify accurately as to age than are the skulls of most other mammals. More reliance on shape of entire skull and less reliance on extent and shape of any individual bone is necessary in estimating the age of a weasel. Wright (1947:344) shows that the weight of the baculum (os penis) is a certain means of differentiating adults from males of lesser age. When approximately eleven months old, Mustela frenata oribasus of western Montana molts from the white winter coat into the brown summer coat. At that time spermatogenesis starts for the first time and the weight of the baculum increases from less than 30 milligrams to more than 52 milligrams.

In the autumn and early winter, most of the specimens are subadults. Ordinarily the few adults obtained in these seasons can easily be segregated from the subadults because ontogenetic development in the twelve additional months of life of each of the older animals has obliterated the sutures on the rostrum, heightened (vertically) and lengthened (anteriorly) the sagittal crest, widened the rostrum, and produced still other changes in form that are revealed by direct comparison of specimens of the two ages.

Secondary Sexual Variation

The secondary sexual variation, which has been detected, is in size of the animal, relative length of the tail and shape of the skull. The female is the smaller. In the small Mustela rixosa and apparently in Mustela africana the secondary sexual difference in size is relatively slight. In Mustela frenata and Mustela erminea, males are approximately twice as heavy as females, the degree of difference very definitely depending upon the subspecies. For example, in M. e. richardsonii the recorded weights are 175 and 69 grams as opposed to 81 and 54 grams in M. e. cicognanii. In general, within one species the greatest difference in size of males and females is in those subspecies in which the animals are of large size. The secondary sexual variation in size is much more than the individual variation in either sex. The same is not true of secondary sexual difference in length of the tail (relative to the length of the head and body), which in eighteen subspecies of M. erminea is from 1 to 7 per cent longer in males than in females. In two subspecies, M. e. haidarum and M. e. olympica, the tail is a fraction of a per cent the longer in females if we may rely upon the few specimens for which collectors' measurements are available.

In both M. erminea and M. frenata the skull of the female is approximately 45 per cent lighter than that of the male, or put in the opposite way, the skull of the male is 83 per cent heavier than the skull of the female. The difference in this respect varies greatly depending on the subspecies. For example, the skull of the male is 127 per cent heavier than that of the female in M. e. richardsonii but only 33 per cent heavier in M. e. anguinae. In Mustela frenata, the subspecies noveboracensis shows most sexual dimorphism in weight of skull (3.6 and 1.7 grams) and olivacea the least (5.3 and 3.8 grams). In general, the difference in this respect is less in subspecies the individuals of which are of small size.

Therefore, as might be expected, the secondary sexual variation in weight of the skull is less in M. rixosa, individuals of which are of small size, than in M. erminea or than in M. frenata, in general of larger size. Nevertheless, in M. africana, in which the individuals are of large size, there appears to be less sexual dimorphism in weight of the skull than in M. frenata or than in M. erminea, although it should be remarked that there are too few data for M. africana to allow of forming a trustworthy conclusion concerning the amount of secondary sexual variation in that species.

The secondary sexual variation in shape of the skull consists of a slenderness in the female. In relation to the basilar length the spread of the zygomatic arches is more in males and, except in the one subspecies M. f. altifrontalis, the rostrum is broader. Also the interorbital region is relatively broader in males of most subspecies. In most subspecies of both M. frenata and M. erminea the tympanic bullae are relatively (to the basilar length) longer in females. The maximum sexual dimorphism occurs in M. erminea arctica and the minimum dimorphism in M. e. haidarum, M. e. anguinae and M. e. muricus. Taking into account all of the subspecies of each of the North American species, the shape of the skull differs most in M. erminea and least in M. frenata. In the latter species the greatest difference in shape of the skull, as was true also of its weight, is in the subspecies M. f. noveboracensis. In these two subspecies, M. f. noveboracensis and M. e. arctica, in addition to the secondary sexual variation already mentioned in the skull, females have the braincase smoother and more rounded, the postorbital-, mastoid-, and lacrimal-processes relatively smaller, and the ventral face of the tympanic bulla at its anterior margin more nearly flush with the floor of the braincase.

In the weasels, subgenus Mustela, the disparity in size of the two sexes is almost or quite as much as in any other fissiped carnivore. It is because of this large degree of difference that the skulls of the two sexes are described separately in the following systematic accounts. The need for such treatment was recognized by Reinhold Hensel (1881:127) more than sixty years ago when he wrote in the introduction to his "Craniologische Studien," of Mustela, as follows: ". . . die Geschlechtsdifferenzen am Schädel vieler Säugethiere . . . so gross sind, dass man diese wie Schädel verschiedener species behandeln muss, während in anderen Ordnungen (Rosores, Edentaten) die Schädel solche Unterschiede nichtzeigen." In the past, failure to appreciate the large amount of secondary sexual variation has resulted in erroneous deductions as regards characters of certain geographic races and has been the cause of some nomenclatural confusion, as for example, in Mustela frenata macrura, where the female was named as a separate species (Mustela jelskii).

Individual Variation

Individual variation is here considered to be the variation in one species which can occur between offspring of a single pair of parents, after variation ascribable to differences in age, sex, and season is excluded. Individual variation, therefore, is a term here used in a composite sense; it includes variations which probably represent different genetic strains within certain populations and variations induced within one generation by environmental factors.

In skulls of weasels, the individual variation in size is more than it is in relative proportions. Hensel (op. cit.) has stressed that weasels, like other carnivores, produced "dwarfed" individuals more than do herbivorous mammals. I cannot vouch for the accuracy of this view, but can say that individual variation is not greater than in some other fissiped carnivores. Impressions to the contrary probably result largely from failure to recognize age-variation. When skulls of a large series from any one locality are arranged first by sex, and under each sex according to probable age on the basis of extension anteriorly of the sagittal crest and of degree of postorbital constriction, individual variation is seen to be less than a cursory examination, even of only one sex, would suggest.

Study of a large series of one age of one sex of one species from one locality shows that some parts, of the skull for example, vary more than other parts. In illustration, among 22 male topotypes of Mustela frenata washingtoni the least interorbital breadth varied 25 per cent (9.0 mm. to 12 mm.) whereas the length of the tooth-rows varied only 13.3 per cent (15.6 mm. to 18.0 mm.). In color the individual variation definitely is more in areas of intergradation between subspecies than in other areas. Details of one such instance of intergradation are given in the account of Mustela frenata spadix.

Statements to the effect that there is much individual variation in the color of weasels, were made mostly fifty years or so ago by writers who had but few specimens from widely separated localities. Where marked climatic differences exist between localities only a few miles apart, marked differences occur in coloration of the weasels from the different localities. Much of what formerly was mistaken for individual variation now proves to be geographic variation. Individual variation actually is of slight amount in comparison with that in mammals generally. Differences in size and relative proportions of parts usually are correlated with geographic differences in color. The color does fade slightly in the period between molts. Also as a result of the seasonal color change, in autumn along the upper margin of the Austral Life-zone, some individuals become white whereas others become white on only the underparts, the upper parts changing only to lighter brown. Probably it would be correct to say that this variation was a combination of seasonal and individual variation rather than either one alone.

As might be supposed, individual variation is not the same in all species or subspecies. For example, p2 is always absent in Mustela africana and always present in certain subspecies of M. frenata. In some other subspecies of M. frenata, p2 is absent approximately as often as present. In the writer's experience, when only a few specimens are available for comparison, individual variation is more difficult to distinguish from specific and subspecific (geographic) variation than is age-variation or secondary sexual variation.

Among the larger series of specimens examined, only one instance of what might be called a mutation in the old sense of a large, sudden change, was detected. That was the loss of the second lower molar in many (less than a third) of the specimens from Newfoundland. The six instances of abnormal coloration described on pages 41 to 43, might be regarded as mutations of large magnitude but no evidence was found of repetition of an abnormality in any one population. Otherwise, in every instance where plotted, the manifestations of a variation arranged themselves about the mean in such a way as to form a smooth, unimodal curve.

Seasonal Variation

When subspecific and specific variations are the objectives of study, seasonal variation must be understood, in order to be excluded from consideration, in the same way that variations ascribable to age, sex and individualism must be understood in order to be excluded from consideration. In weasels, change in color of the pelage is the seasonal variation most important for the systematist to understand. Other seasonal variations in the pelage are hairiness versus nakedness of the pads of the feet, length of the pelage on the body, and possibly the density of the pelage on the body. In the northern half of North America, roughly speaking, seasonal change in color is so pronounced (white in winter and brown in summer) as to be easily recognized. South of this area, in the Austral and Sonoran life-zones, the color of the winter pelage differs only slightly from that of the summer pelage. In these more southern latitudes the winter pelage in almost all subspecies is of lighter color than the summer pelage and has a smoky suffusion. With material of the two seasons in hand for comparison, close attention to the variation will permit the systematist to recognize the difference in shade of brown as seasonal variation and not geographic or specific variation. Farther south still, in the Tropical Life-zone, seasonal difference in color was not detected in the material studied. Seasonal change in color is discussed in the section immediately following.

Variation in Coloration and Molt

In all American weasels (subgenus Mustela) the color, at least in summer, is brown with more or less white or whitish on the underparts. In one species, Mustela africana, there is a longitudinal stripe of brown on the middle of the light-colored underparts; this stripe is absent in each of the other three American species. Two species, M. erminea and M. frenata, always have a black tip on the tail. Of the other two species, M. africana lacks the black tip and M. rixosa may or may not have a few black hairs in the tip of its tail. White or light yellowish facial markings occur in subspecies of M. frenata from the southwestern United Stated to Central America. Subspecies having the most extensive light-colored facial markings have the remainder of the upper part of the head black. In weasels without light facial markings the upper parts of the head all are brown. In the two species, M. erminea and M. frenata, the extent to which the light color of the underparts extends down the insides of the legs and out on the underside of the tail, or the absence of light color on these parts, is a matter of geographic variation. The same can be said for M. rixosa except that first its tail is unicolored and second individual variation as well as geographic variation accounts for the color pattern on the underparts and legs in animals from the southeastern part of the range of the species.

The most remarkable feature of the coloration of weasels is the winter whitening. This occurs in the northern part of North America in each of the three species of weasels found on that continent. The black tip of the tail in M. erminea and M. frenata remains black in winter. If an individual of M. rixosa has black hairs on the tip of its tail in summer, there are thought to be black hairs there also in winter. Otherwise the winter pelage is all white in northern areas in each of the three species. In this white winter coat the animal is known as ermine.

The underlying cause seems to be protective coloration. At any rate, weasels are always white in winter if they are from areas where snow lies on the ground all winter, every winter, or almost every winter; and they are always brown if from areas where there is never, or rarely, snow in winter. The changes in color are effected by molt, one in autumn and one in spring. Animals that are brown in winter undergo the same two molts as do those that are white in winter. The capacity to acquire a white coat or a brown coat in winter is an hereditary matter just as one man grows red hair and another grows black hair. In the weasels, however, all individuals in the north turn white in winter and if one that was born there is kept through successive winters in the warmer south where there is no snow, he will still turn white each winter. A weasel born in a southern area, where all are brown in winter, molts into a brown (not white) winter coat even when kept in a cold, snowy, northern area where native weasels of the same species all turn white. Obviously, therefore, neither snow nor temperature is an immediate cause and, as we have said, the color in winter is a matter of heredity. The time of the molt, we now know, is determined by the amount of light. When nights grow longer and days shorter, a point is reached at which the lesser light received through the eyes causes the pituitary gland to cease producing a gonadotropic hormone. Directly or indirectly, the lack of this hormone stimulates molt and, probably enzyme action, or the lack of it, causes the melanoblasts of the cells in the hair follicle to be without pigment. Hence the hair grown from a follicle under such conditions lacks pigment (melanin) and is white. In spring, as the days grow longer and the nights shorter, the increasing amount of light received day by day through the eyes stimulates the pituitary gland to produce the gonadotropic hormone which directly or indirectly, stimulates molt and, probably by enzyme action, the melanoblasts are caused to be present in cells of the hair follicle and the melanoblasts provide granules of melanin pigment which are incorporated in cells of the growing hair. These granules of pigment give the hair its color.

Evidence in support of this hypothesis is given below.

Along the Pacific Coast from British Columbia southward, M. erminea (see fig. [25] on page [95]) is brown in winter. This is an area where snow rarely falls and the temperature in winter ordinarily is above freezing. In the remaining part of the American range of this species the temperature in winter is below freezing much of the time and snow remains throughout the winter or for long periods. In this colder part of the animal's range, only white coats occur in winter. M. frenata likewise has a white coat in winter in the part of its geographic range where snow and freezing temperatures prevail throughout most of the winter and a brown coat in warmer, snowless areas to the southward and along the Pacific Coast. The third species, M. rixosa, exhibits a corresponding correlation between coat color and climate. On the Asiatic continent, several species, including M. erminea, provide parallel correlations and nowhere are there any exceptions for the subgenus Mustela. These data are an important part of the material on which we have based the induction that the underlying cause of seasonal change in color is a need for protective coloration.

As regards molt, most naturalists who have written upon the subject regard it as responsible for the change from the white winter coat to the brown summer coat. However, the change from brown summer coat to white winter coat has been thought by several writers to be effected by change in coloration of the individual hairs. Among those holding this opinion there may be cited Bell (1874:197) in reference to Mustela erminea, and Coues (1877:123) in reference to American specimens to which he applied the same name. More lately Hadwen (1929) has taken this same view, and Gunn (1932) also discusses the possibility of the hairs changing color. Bachman (1839:228-232), Macgillivary (1843?:158), Audubon and Bachman (1851 (vol. 2):62), Schwalbe (1893:538), Pearson et al. (1913:447), Miller (1930, 1931A), Hamilton (1933:300) and Rothschild (1942), among others, have been inclined to the opinion, or positively affirm, that the color change in autumn is the result of a molt. The papers cited above contain, in turn, references to many other printed accounts dealing with this question.

To my mind, it has not so far been demonstrated that the change in color of weasels in autumn is accomplished without a molt. Also so far as I am aware, no explanation has been given of how the pigment may disappear from the hair of weasels. Metchnikoff's (1901:156) idea that the senile whitening of the hair in man is accomplished by phagocytes which remove the pigment granules would hardly seem to explain the relatively sudden and complete autumnal change occurring in weasels. Anyhow, Danforth (1925:108), and some other students have thought that the action of these phagocytes was at most a factor of slight importance in the whitening of hair. Whatever be the complete answer to the question of how the weasel changes color in autumn, at least one specimen of long-tailed weasel, which is in process of color change in autumn, presents clear evidence of molt of the overhairs. This specimen of M. f. longicauda is no. 188408, U. S. Nat. Mus., taken on November 12, 1897, at Rapid City, South Dakota. Other specimens of M. erminea which were taken in autumn similarly show molt to be in progress. For these and other reasons, I am inclined to the opinion that the autumnal change in color, like the one in spring, is effected by molt. During the period of the autumnal color change, Noback (1935:27) had a captive M. f. noveboracensis and, each morning, found clumps of brown hair on the floor of its cage; this was strong indication that molt was responsible for the color change in this instance.

However, I freely admit that the evidence does not prove that the change from brown to white can be accomplished only by molt; in the present state of knowledge it would be unscientific to deny that the change were possible of accomplishment by other means. Also, it is true that the fifteen specimens before me of Mustela frenata, subspecies included, in process of change from brown to white, with the exception of the one from Rapid City, South Dakota, if taken individually, do not, in macroscopic examination, show definite molt lines or other absolutely convincing evidence of molt. However, these same specimens, insofar as examined microscopically, do show overhairs all white, or overhairs pigmented throughout. The lighter color of the proximal parts of the overhairs in itself should not be accepted as evidence of color change, for in the fresh summer pelage, the same condition exists. Also, careful macroscopic examination suffices to show that in the transitional pelage of autumn, the brown overhairs generally are longer than the intermixed white overhairs.

Whether the underfur behaves in exactly the same way as the overhair, I have not myself definitely ascertained, but I assume that the underfur is molted twice each year, at least in the northern populations of Mustela frenata and in the other species of more northern distribution. Schwalbe's (1893) work, including sectioning of the skin and study of the hair follicles, led him to conclude that the underfur was molted twice each year in Mustela erminea.

In Mustela frenata noveboracensis, M. f. nevadensis, and M. f. nigriauris, measurements taken on adult males show the overhairs to be longer in the winter pelage than in the summer pelage of specimens from the same locality. For example, in M. f. nigriauris from Berkeley, California, the overhairs of the summer coat (July and August) average 8 millimeters in length on the hinder back and 7 mm. on the belly, but average 9.5 mm. and 8 mm. respectively in January-taken specimens possessing the full winter coat. At Ann Arbor, Michigan, in the summer coat, the longest hairs on the hinder back average approximately 12 mm., and those on the belly, 9.5 mm., against 13 mm. and 9.5 mm. respectively in winter. Although general observations initially led me to believe that the black, terminal hairs of the tip of the tail are longer in the winter pelage than in the summer pelage, actual measurements fail to show a difference in length.

The change from one coat to the other in the long-tailed weasel has been described among others by Miller (1930, 1931A), Hamilton (1933) and Glover (1942) on the basis of captive specimens. In a general way, the progress of the molt in their specimens agrees with that which I have been able to make out from examination of skins taken in the wild. There is, however, this difference: Their specimens show a more spotted pattern when in process of hair-change than do specimens taken in the wild. Probably the more or less unnatural conditions under which these captive animals lived modified the normal progress of molt.

In wild-taken specimens of the species Mustela frenata, subspecies included, the spring molt begins on the mid-dorsal line and proceeds laterally, producing, at almost any given time, a relatively sharp molt line separating the white winter hair from the incoming brown summer coat. However, in autumn the change takes place first on the belly, then on the sides, and finally makes its appearance over all the upper parts at about the same time, with the result that the upper parts have a salt-and-pepper appearance without at this time any sharply defined molt lines. In general, the molt pattern can be said to be reversed in the two seasons; in spring, it begins on the back and in autumn, on the belly. The difference in spring and autumn color pattern is better illustrated on plate [39] than by additional description. Swanson and Fryklund (1935:123) have observed that the "spring molt proceeds differently" than the fall one in Mustela rixosa, and Barrett-Hamilton (1903:309) in commenting on the European hare (and the stoat?) remarks, "In spring the moult, and with it the brown colour, progresses in exactly the opposite order . . ." as compared with the white color of autumn, which that particular writer thought resulted from removal of pigment from the hairs rather than from molt.

The tail, excepting the black tip, lags in the molt in many instances, with the result that, especially in spring, it may retain a few white hairs as late as does the belly. In autumn it is less tardy and so far as I have observed, becomes white at about the same time that the general area of the back changes color. On the tail, the black tip itself, as clearly shown in more than a score of specimens, is molted at approximately the same time in autumn as is the pelage of the body. However, the long black hairs, which appear in, say, November, appear to increase in length until January. In spring, the long black hairs of the tip of the tail seem not to be shed at the same time as the rest of the winter pelage, but remain approximately six weeks longer and then are replaced by long black hairs of the summer coat. At any rate, this is the picture presented by a half dozen specimens of M. f. nevadensis and M. f. longicauda which do show a spring molt to be in progress on the black tip of the tail. Schwalbe similarly (1893:536-537) has suggested that the black tip of the tail in Mustela erminea in spring is not molted until about two months after the pelage on the rest of the body is changed. Schwalbe (loc. cit.) thinks also that in M. erminea studied by him, the black tip of the tail in autumn is replaced approximately one month in advance of the pelage on the rest of the body. As indicated above, my specimens of Mustela frenata, subspecies longicauda and nevadensis, do not show this discrepancy in autumn. I have considered the possibility that the black tip of the tail, in some species of Mustela, is molted only once while the remainder of the coat was undergoing two molts. My inconclusive data lend but little support to this possibility.

The difference in pattern of color between specimens taken in autumn and spring is known to some fur-trappers of my acquaintance who have suggested that molt occurs in spring, whereas the individual hairs change color in autumn. Reference to plate [39] will show how gross comparisons might lead one to this erroneous explanation of the color change.

As to time of molt: In eight subspecies of Mustela frenata, namely, noveboracensis, occisor, primulina, spadix, longicauda, arizonensis, nevadensis and effera, material is available to indicate that the autumnal molt begins in October and is completed in November, and that the spring molt occurs in March or April. A condensed list of specimens providing basis for this statement is as follows:

M. f. noveboracensis: 26 specimens in transitional pelage taken in autumn and 14 taken in spring; M. f. occisor: One topotype has acquired one-fifth of the winter pelage on October 22, 1896; M. f. primulina: 2 in November, one in March, and 2 in April are in process of change; M. f. spadix: 6 autumnal specimens and one in April show pelage change; M. f. longicauda: 7 autumnal specimens and one in April show pelage change; M. f. arizonensis: 12 specimens in autumn and 3 in spring are in process of molt; M. f. effera: One November-taken male has acquired four-fifths of the winter coat and another taken on April 21 at Fort Rock, Oregon, is half finished with the spring molt.

It may be added that no marked difference in time of either autumnal or spring molt is apparent as between the more northern and more southern localities from which the mentioned specimens come. With more complete material I would expect to find a difference in this regard.

The material of the other, more southern, subspecies of Mustela frenata has not been adequate to show the time of molting or the number of molts which occur in one year.

Animals in the northern part of the range of Mustela frenata acquire a white winter coat, whereas those in the southern part acquire a brown winter coat, and in an intervening area the winter coat may be either brown or white. By plotting on a map the localities of capture of all specimens examined in the winter coat, it was possible to outline this intervening area as shown in figure [10] on page [37]. However, Dearborn (1932:36) shows that in Michigan some animals have a brown coat in winter at places farther north than figure [10] shows to be the case. Hamilton's (1933-306) map for New York shows the same to be true in that state. Accordingly, the boundaries of the area shown in figure [10], in which both brown and white long-tailed weasels occur in winter, are known to be only approximate; with full information available the belt would be represented as wider.

Fig. 10. Map showing the region (in black) where both the brown and white winter pelage is found in the long-tailed weasel, Mustela frenata.

Hamilton (1933:302) has pointed out that "Where half of the weasels remain brown, these brown winter specimens are always males." The results of my own examination of specimens not studied by Hamilton, in a general way provide confirmatory data. More exactly, my examination reveals that at the most northern localities where brown specimens occur, only males are in this coat. In explanation, it may be said that in plotting on a map localities of capture of specimens in the winter coat, thirteen places were found where both sexes were represented and where both brown and white winter coats were found. With the two sexes, it is theoretically possible to have nine different combinations of coat color. With males all brown, there might occur females (1) all brown, (2) all white, or (3) some brown and some white. In addition to these three combinations, we might have three more by finding the mentioned types of female coat color repeated where all males are white, and three more, or nine in all, by substituting a population of males some of which were brown and some of which were white. Seven of these possible combinations actually were found. The two combinations not found were all white males with all brown females, and all white males with females both brown and white. In the three instances where the males all were brown and the females all were white, the localities of capture were in the northern part of the variable area. This indicates that where the brown winter coat occurs at northern localities, the brown individuals are all males. Farther south, of course, the females, too, acquire the brown winter coat.

Stated in another way, there is a broad belt across North America from the Atlantic to the Pacific in which males of Mustela frenata at any one locality may be either brown or white in winter. Inside this broad belt there is a narrower one, approximately half as wide, in which females at any one locality may be either brown or white.

In support of the idea that color of the winter coat is an hereditary matter and that it is not dependent on temperature, the following evidence derived from my transplanting specimens of Mustela frenata supports the idea that color of the winter pelage is dependent on heredity and not on temperature or snowfall.

A male captured on June 24, 1937, in the brown summer coat in Salt Lake City, Utah, was received by me at Berkeley, California, five days later and kept in captivity almost six months. On November 17, 1937, half the pelage was white and on December 27, 1937, when next examined, the animal was in the full, white, winter coat as it was on January 25, 1938, when it died. Native weasels all turn white in winter in Salt Lake City, but in Berkeley native weasels always are brown in winter.

A juvenile or young animal, a male, captured in May, 1936, at Lafayette, Contra Costa County, California, was kept there until August 13, 1936, when transferred to Calneva at the north end of Lake Tahoe, California. The weasel was kept at Calneva until its death on December 23, 1937. In both the winter of 1936-'37 and in that of 1937-'38, the winter coat was brown as in animals from its place of origin (Contra Costa County) and unlike weasels of the Tahoe region nearly all of which turn white in winter.

Two females, each approximately two months old, captured on May 1, 1936, at James Landing, 4 miles northwest of San Pablo, Contra Costa County, California, were kept in Berkeley, California, until August 13, 1936, when they were transferred to the mouth of Blackwood Creek, on the west side of Lake Tahoe, California. On October 25, 1936, both weasels escaped. On December 25, 1936, the headless body of one of these was found approximately 300 yards south of the mouth of Blackwood Creek. The animal had been dead at most a few days when found and was in the brown winter coat. At the place of its origin all weasels are brown in winter but at the mouth of Blackwood Creek only 2 of 60 weasels caught there in the winter coat were brown; the other 58 were white. The headless weasel was identified, as one of the two formerly in captivity, by means of certain short toes, the ends of which had been clipped off when the animal was a captive. No trace of the second female was found.

A female of unknown age, in white winter pelage, captured 4 miles southeast of Tahoe City, California, and kept there until April 3, 1937, on which date it was brought to Berkeley, California, molted to brown in the spring. The first signs of the brown coat were noted on April 14. On May 24 or 25 she gave birth to 4 young which lived less than ten days. In the following winter this animal acquired a white coat. As previously noted, weasels native to the Berkeley area, where this female was kept, have brown coats in winter.

The weasels were in every instance kept in cages out-of-doors. The sides of the cages were open to the elements. A nest box in each cage provided shelter. All were of the species Mustela frenata.

The significant results, it seemed to me, were that the winter coat was the kind found in the area where the weasel originated instead of the kind found in weasels native to the areas in which the specimens were held in captivity.

That the time of molt is determined by the amount of light has clearly been shown by Bissonnette (1944:223) for American weasels of the two species Mustela erminea and M. frenata. In his words (op. cit.:246) "Reducing the daily periods of light induced molting and regrowth of new fur. . . . In the Bonaparte weasels [Mustela erminea], white replaced brown. . . . Increasing daily light-periods caused molting and change to dark brown. . . . Incomplete molts in both directions (toward white or toward brown) were produced as a result of early reversal of increase or decrease of daily light-time. . . . That this stimulus is received through the eyes and acts through the anterior pituitary gland is indicated by Bissonnette's [1935:159] studies on ferrets, a nearly related animal. That the thyroids and sex-glands are not essential is at least suggested . . . by Lyman's (1942) study on the varying hare [Lepus americanus]." It can be added that Lyman (1943:451) demonstrated in Lepus americanus that the effect of light is received through the eyes. He demonstrated this by masking the animals. To Wright (1942B:109) who studied the two American weasels, M. erminea and M. frenata, it seemed likely that the pituitary produced or released gonadotropic hormone at about the time of the spring molt and that this molt and the spring changes in the reproductive tracts of the weasels might be caused by a stimulus from a common source. Later, Wright (1950:130) injected a gonadotropic hormone into long-tailed weasels which had recently acquired their white winter pelage and thereby caused them to lose the white pelage and acquire the brown pelage. It is Lyman (1943:450) who says, in relation to Lepus americanus, "When in the physiologically white condition, the melanoblasts of the regenerating guard- and pile-hair follicles contain no melanin-forming enzyme (dopa-oxidase), which may be the reason for the lack of pigment." Schwalbe (1893) by sectioning the skin and microscopically examining the hair-follicles of M. erminea learned that the basal cells producing hairs lacked pigment granules in autumn when the European ermine (M. erminea) was acquiring its white winter coat and that the cells contained granules of pigment in spring when, as we know, the granules are incorporated in the growing hair and give it its color.

The above material, then, is basis for the account on pages 31 and 32 of what causes the weasel of northern areas to have a white coat in winter. The discerning student will instantly perceive that although some parts of the account on pages 31 and 32 are precisely accurate, other parts are the result of inferences which need to be proved. More careful work of the kind that Schwalbe (1893) and Wright (1942B) did is needed. The account on pages 31 and 32 is merely the best that can be given with the information now available.

Many writers have commented on the yellowish color, sometimes with a greenish tinge, found on the fur of weasels in the white winter coat. The stain is more often found on the tail and hinder-parts of the body than elsewhere. Possibly, partly on this account, some have ascribed this color to the smearing of the fur with urine. Still others have thought it resulted from the smearing of the fur with secretions from the anal scent glands. Schumacher (1928) takes this point of view, and while it may be that he has not proved his point, still his conclusions fit the known facts and seem sound to me. Schumacher points out that the same soiling of the fur is present in summer as well as in winter, but that on the summer pelage the stain can be detected only on the light-colored underparts. It is from this point of view that he criticizes the systematic worth of white versus yellowish-white underparts in the summer pelage of geographic races of Mustela erminea and Mustela nivalis. Although in the long-tailed weasels (Mustela frenata) the underparts of all the races are pigmented with some form of red, orange or yellow, it seems probable to me that the additional color resulting from the soiling effect of this glandular secretion explains the greater variation, found at a single locality, in the color of underparts than of upper parts in the summer pelage.

I have neither seen nor heard of a black weasel in any part of the New World or of the Old World. I have found only one albino among American specimens. It is an adult female, no. 121424, American Museum of Natural History, of Mustela erminea richardsonii, taken on August 30, 1935, at Hot Springs, Northwest Territory. This place, I am told by G. G. Goodwin who obtained the animal, is on the "Nahanni River where the rugged mountain ridges rise abruptly from the low mud flat lands, latitude 61, longitude 125." The shortness and coarseness of the hair corresponds to that of the summer pelage and not winter pelage. The pelage is everywhere white, even the tip of the tail. True, all except the nape and top and sides of the head has a faint yellowish-green tinge which has been supposed to result from staining by secretion of the anal scent glands but there is no pigment in the hair as in erythristic specimens. From the Old World, Farurick (1873:17) has recorded what he regards as an albino of Mustela vulgaris since it had no black hairs on the tip of its tail. Flintoff (1935:228, 229) records what may have been an albino Mustela vulgaris from Yorkshire and an albino M. erminea from an unstated locality. Jäckel (1873:459) mentions specimens of Mustela erminea and Mustela vulgaris, which were partly "albinistic" or "erythristic." Among the American specimens of M. erminea I have not recorded any which appeared to be either partly or wholly erythristic or only partly albinistic. Among the 1550 skins of M. frenata which were in summer pelage or brown winter pelage, five, described below, show marked abnormalities in color.

Two of these five are partly albinistic. One is an adult male, no. 223880, U. S. Nat. Mus., from Billy's Island, Okefinokee Swamp, Georgia, which has the nose as well as the area between the eyes white. Also there is a tuft of white hairs at the anterodorsal margin of each ear, scattering white hairs suggesting a postorbital bar on each side of the head, and a patch of white hairs on the mid-dorsal line behind the ears. Markings of this kind are not abnormal in M. f. peninsulae, the subspecies adjoining on the south, except for the white nose which clearly is an instance of partial albinism. The second specimen is a subadult male, of M. f. noveboracensis, no. 177679, U. S. Nat. Mus., in process of acquiring the brown winter coat, taken on November 27, 1911, at Gaylordsville, Connecticut. It has white markings on the nose, on the right side of the neck, on the right hind foot and right forefoot, and on the tip of the tail. The white area of the nose on the left side extends back to the eye, but on the right side barely encircles the nose-pad. On the right side of the neck, all that area between the foreleg and ear is white from the mid-dorsal line (including 7 or 8 millimeters to the left of the mid-dorsal line) down to the throat, which is white as it is also in normal individuals. The toes of the right hind foot are more extensively white than in normal specimens of noveboracensis, and all of the right forefoot as well as the wrist is white. The tail is of striking appearance because of its tricolor pattern. The proximal part is of the normal brown color. The black terminal part commences proximally at the usual place, but the distal 11 millimeters of the fleshy part of the tail bear only pure white hairs producing a terminal white pencil 35 millimeters long.

The three other specimens abnormally colored are erythristic individuals. An adult male of M. f. latirostra, no. 7574, coll. D. R. Dickey, taken on April 14, 1918, at Covina, Los Angeles County, California, has the color of the upper parts greatly restricted, and, in addition, has spots and blotches of the color of the underparts distributed over the back and rump. A spot of this same color occurs above each ear. Incidentally, this and other subspecies of Mustela frenata from the Pacific Coast of North America obviously have the factor for erythrism operating over a larger part of the body than it does in M. erminea or than in M. f. noveboracensis, where the underparts sometimes are white. In M. f. latirostra and in other subspecies from the Pacific Coast the light color of the underparts always is tinged with this reddish color.

Another erythristic specimen is a young male of M. f. nevadensis, no. 23493, U. S. Nat. Mus., taken on August 6, 1890, at Birch Creek, Idaho. It has all of each foreleg, the axillary regions, and a saddle-shaped area over the shoulders of the same buff-yellow color as the underparts.

The third erythristic specimen is a subadult female, of M. f. oregonensis, no. 47149, Mus. Vert. Zoöl., taken on December 20, 1930, at Carlotta, Humboldt County, California. This specimen appears to be white and initially was thought to be merely an individual in the white winter coat. Closer examination, however, shows that it has a light wash of ochraceous or faint reddish color. Also, other specimens taken in winter at Carlotta show that weasels there do not acquire a white winter coat. The only normally brown area is approximately three millimeters in diameter at the anterodorsal margin of the pinna of the right ear. The tip of the tail is black as in a normal specimen. The specimen in question is actually pure white only on top of the head from a short distance behind the ears on over the forehead nearly to the eyes, and on the inside of the ears. In a normally colored animal this area is the dark area of the head. In this freak, the other parts of the head, which, in individuals of normal coloration are the white or light orange facial markings, have the reddish cast of the remainder of the body, although the color is less intense than on the back. The collector noted that the specimen had eyes of normal color. A possible explanation for the coloration of this specimen is that this species has three factors for color, one for the black tail tip, one for the reddish color, and a third, missing in the specimen in question, for the blackish brown.

For some more exact knowledge concerning this erythristic type of coloration, we are indebted to Pitt (1921:99), who describes a population of polecats, Mustela putorius, in Cardiganshire, England, in which this erythristic variation is maintained in a state of nature. In ferrets, Mustela furo, Pitt (op. cit.:114) notes that ". . . erythrism is certainly dependent on a Mendelian factor, being dominant to albinism and recessive to the black-brown coloration. Both in the ferret and polecat, erythrism seems to be correlated with increased size, and certainly in the ferret is usually accompanied by a quick temper and general increase in vitality."

Variations of Taxonomic Worth

Variations of taxonomic worth usually are referred to as characters. For example, shortness of the tympanic bulla is a character, and the opposite condition, long tympanic bulla, is another character. Specific variations, that is to say specific characters, are provided by the color-pattern, length of tail, number of premolar teeth, shape of the tympanic bullae, and length of the braincase in relation to the length of the tooth-bearing parts of the skull. Subspecific characters are provided by color-pattern, color itself, size as measured by weight of the animal, and its linear measurements, size of the skull, and size and shape of parts of the skull. The characters distinguishing subspecies from one another are not of a different nature from those distinguishing species from one another.

Given any one of the above structural features, say, dorsal outline of the skull, several characters may be provided by it. For example, weasels of the species Mustela frenata have the dorsal outline of the skull convex in southern Louisiana, straight in Missouri and concave in North Dakota, thus providing three characters. This is geographic variation. These variations, characters in zoölogical parlance, when plotted on maps, reveal the geographic occurrence of, say, the convex shape of the skull. In combination with other characters, for example, dark color and short tail, basis is provided for recognizing a subspecies, in this instance Mustela frenata arthuri of Louisiana. Because the change from convex to flat skull takes place geographically at about the same place (in eastern Texas) as does the change from short tail to long tail, and the change from dark color to light color, it is easy to draw a line there marking the western geographic limit of occurrence of the M. f. arthuri. This same line marks also the eastern margin of the geographic range of the subspecies Mustela frenata frenata, the subspecies next adjacent to the westward. On this line and for several miles to either side of it weasels show varying combinations of these three characters or an intermediate condition as regards one or more of the characters, or both. For example, from a locality in eastern Texas a weasel may have (1) a facial pattern exactly intermediate between that of the unicolored face of arthuri and that of the bicolored face of frenata, (2) the long tail of frenata and (3) the convex skull of arthuri. In the sum of its characters this specimen is exactly intermediate between typical arthuri and typical frenata. Another specimen from the same place may differ from the first specimen only in having the tail slightly shorter. The total "score" for the two specimens is, therefore, by a very slight margin in favor of arthuri. Let us suppose that we obtain a third specimen from the same place and that it has the face marked like that of arthuri but the tail fully as long, and the skull as lacking in dorsal convexity, as in frenata. Now the score is definitely for frenata. For convenience of handling, the population is referred to frenata, providing that the average of specimens from a nearby locality to the westward is not in favor of arthuri. In event the average of specimens from a locality next adjacent to the westward is in favor of M. f. arthuri, the total evidence from the two localities may be weighed together and appropriate decision as to subspecific status of weasels from the area is made according to what the average is for the area as a whole.

The three individual animals of an intermediate sort are ordinarily termed intergrades. This implies that their characters are the result of mixed parentage—perhaps a female of M. f. arthuri and a male of M. f. frenata but probably each parent itself was an intergrade and the offspring, of which we examined three, owe their characters to reproductive processes operating in obedience to Mendelian laws of inheritance.

The two kinds of animals, Mustela frenata arthuri and Mustela frenata frenata, are identified as subspecies because of the intergradation between them. If at this and all other places where the geographic ranges of arthuri and frenata met there was no crossbreeding (no intergrades), the two kinds would be treated as distinct species. Intergradation, and the lack of it, are accepted as the criteria of subspecies and species, respectively.

These criteria suffice for animals, in this instance weasels, which have a continuous geographic distribution. Some kinds of weasels are confined to islands, as for example the islands off the coast of Alaska and British Columbia. Because weasels are land animals, crossbreeding in nature between the weasels of two islands is, of course, impossible. A modified test (used in the study here reported upon) in deciding on specific versus subspecific status in these instances can be made as follows: On the adjacent mainland, ascertain the degree of difference between two subspecies whose geographic ranges meet (for example, M. e. richardsonii and M. e. alascensis). Next ascertain the degree of difference between the insular kind of animal and the kind on the mainland. If the degree of difference is greater when the insular kind is compared than when only the kinds of the mainland are compared, the insular kind is to be regarded as a species. If the degree of difference is no greater between the insular kind and the mainland kind than it is between the two adjacent mainland kinds, the insular kind is to be regarded as a subspecies. In short, for insular kinds, the criterion is degree of difference, with the limitation of geographic adjacency, rather than intergradation.

The geographic variation (subspecific characters) found could be spoken of as two kinds: First, there is the variation which is expressed in a general trend for a long distance, producing, in general, a cline of even slope; and second, that of inconstant trend in any one direction. In his "The Rabbits of North America" Nelson (1909:34-35) has commented on the latter type of variation as follows: "While studying series of specimens from all parts of the vast range occupied by the geographic races of such species as Sylvilagus floridanus and S. auduboni, I have been impressed with evidences of fluctuation of both external and skull characters. These fluctuations are somewhat wavelike in character and rise to central points of extreme development and then sink away to intermediate borders beyond which new waves rise. Where the waves of differentiation are pronounced they mark recognizable geographic races. Within the area covered by the larger or geographically broader waves of differentiation (recognized as of subspecific value), smaller waves of differentiation are included, which may represent local variations in intensity of characters of the subspecies, or these characters may diminish and the variation tend in other directions, sometimes even closely reproducing the characters of another subspecies occupying a distinct area." In Mustela frenata, much of the geographic variation at first inspection appears to be of this nature. Closer scrutiny, however, reveals that the repetition, at geographic intervals, of several features of color and structure are closely correlated with environmental features which are repeated only at these same places.

In Mustela erminea, much of the variation is of the first kind, namely, that which can be expressed as long clines of relatively even slope. As several authors have said, zoölogical classification based on this kind of variation is like dividing the spectrum and depends largely upon the standards set, for, theoretically, the possibilities of subdivision are unlimited. Actually, however, none of the clines has an even slope and the possibilities for subdivision therefore are limited. Also, when several features are used, instead of only one feature, the classification is more satisfactory even if the basis is more complex.

Some features of structure which provide subspecific characters are mentioned below.

Total length, of males, ranges from 598 to 360 mm. in M. frenata and from 336 to 228 mm. in M. erminea. There is no cline of sustained slope in M. frenata but in M. erminea there is a progressive decrease in total length from north to south.

Length of tail varies from as little as a half to as much as seven-tenths of the length of the head and body in M. frenata, the subspecies neomexicana having the long tail and the two subspecies arthuri and primulina having short tails. The geographic ranges of primulina and neomexicana are contiguous. In M. erminea there is likewise no variation of a clinal nature in length of tail and furthermore the variation is much less than in M. frenata.

In length of hind foot, which in males varies from 49 mm. in northern populations of M. erminea to 28 mm. in southern populations, the same cline is seen as in the total length of animals of this species. In M. frenata, however, there are several decreases and increases along any straight line which can be drawn through the geographic range of the species. The range of variation in males is 41 mm. (M. f. arizonensis) to 59 mm. (M. f. macrophonius).

Weight of the entire animal is an excellent measure of size but weights are unavailable for many subspecies. In M. frenata, the two subspecies texensis and macrophonius probably are the heaviest and effera, arizonensis and helleri probably are the lightest. Geographically the variation in weight behaves in approximately the same way as does the measurement of total length. In M. erminea the variation in weight of males is from 206 grams in northern animals to 58 grams in southernmost populations, there being a relatively constant gradient geographically.

Degree of hairiness of the foot-soles in M. frenata clearly is linked with the temperature; in regions of high average temperature the hairiness is least and in regions of low average temperature it is most. The decrease in hairiness is accomplished in two ways, namely, smaller breadth and decreased length of individual hairs and decrease in number of hairs on a given area of dermal surface. This correlation holds throughout the entire north to south range of the species. Corresponding differences are found on the same latitude where topographic diversity in an east to west direction produces northern conditions at high altitudes and southern conditions at low altitudes. The conclusion seems unavoidable that climate, directly or indirectly, determines the degree of hairiness. Less careful observations were made on the hairiness of the soles of the feet in other species but it is clear that the northern species M. erminea has the most hair on the foot-soles and that M. africana, the tropical weasel, has the least. In this regard, M. frenata is intermediate as it is also in geographic position.

Figs. 11-15. Dorsal views of adult skulls of each sex of five subspecies of the ermine, Mustela erminea, to show secondary sexual variation and geographic variation in size of the skull. Males on the left and females on the right. All × 1.

Note especially the geographic variation in decreasing size of the skull from north to south in each sex, and that the secondary sexual variation in size of skull is less in ermines with small skulls than in those with large skulls.

Fig. 16. Map showing the localities where the skulls, represented in figures [11-15], were obtained.

The maximum length of facial and carpal vibrissae is attained in M. erminea in the far north. In weasels from north of the Arctic Circle the longest facial vibrissae extend posteriorly beyond the posterior border of the ear. In the tropical weasel, M. africana, the facial vibrissae do not extend posteriorly beyond the ear and the carpal vibrissae are not so long as the distance between their bases and the apical pad of the first digit. The correlation of long vibrissae with low temperature, is mentioned here merely because length and density of pelage were under consideration.

The most obvious and most exact correlation between change in climate and change in the animal is furnished by color. This is well shown in the one species, Mustela frenata, to which the following remarks apply unless indication is given to the contrary. The color of the upper parts varies from bay (blackish brown) in M. f. panamensis to buckthorn brown (light brown) in M. f. neomexicana. The color of the head varies from solid brown (white chin excepted) to contrasting black and white markings.

Dark color of the upper parts is associated with a large area of this color; the enlargement of this area is at the expense of the area of light color on the underparts. In the weasels of darkest color the upper parts occupy four-fifths of the circumference of the body (as measured in the anterior lumbar region) but in the lightest-colored weasels the upper parts comprise only two-thirds of the total circumference. In these light-colored animals the color of the underparts extends onto the underside of the tail and down the insides of the legs and over the feet whereas in the animals with the darkest upper parts the entire tail, feet, and legs below the knees ordinarily are of the same dark color as the upper parts. The length of the black tip on the tail varies inversely with the length of the tail, probably because the lightest-colored weasel has the longest tail. In some subspecies the black brush is almost half as long as the tail-vertebrae but in others is less than a fourth as long as the tail-vertebrae.

The extent of the color of the head, as well as the intensity of the color there, varies markedly and is correlated with climatic conditions. The extent and intensity of this dark color is greater in weasels inhabiting regions of heavy rainfall than in those inhabiting regions of sparse rainfall. Considering the geographic range of each subspecies of Mustela frenata, that of M. f. panamensis has the maximum of rainfall. Reference to the colored plate (1) will show that in M. f. panamensis (2) the black of the head is extended over all of the upper parts. M. f. macrura (1) of Perú, to the southward, is from an area of lesser rainfall and is correspondingly lighter colored. Returning to panamensis (2) as a starting point and proceeding northward to the range of nicaraguae (3), which also has lesser rainfall, thence another step northward to Guatemala, which has still less rainfall, the weasel there, M. f. goldmani (4) has the black extending posteriorly only to the shoulders. M. f. leucoparia (5) from Michoacán, and M. f. frenata (6) from Tamaulipas are from progressively more northern and also progressively drier regions. In M. f. frenata (6) the dark color extends posteriorly only to the ears and is blackish rather than black. In M. f. neomexicana (7) of the extremely arid parts of Durango, Arizona, and New Mexico the dark marking of the head is confined to a brown spot on the nose. Its geographic range is the most arid of those of all of the subspecies. The contrast between neomexicana (7) and panamensis (2) illustrates the great range of geographic variation in color which occurs in the one species. Continuing from the geographic range of neomexicana (specimen from Safford, Arizona) northwesterly 480 miles to Riverside, California (see 8, latirostra), 430 miles north to Point Reyes, California (see 9, munda), and finally 570 miles north to Tillamook, Oregon (see 10, altifrontalis), each place with more rainfall than the one farther south, another correlation of increasingly dark coloration with increasing amount of rainfall is illustrated.

This geographic variation, it should be remembered, is all within one species. It is the more significant still when we remember that the same correlation, with never an exception, occurs at hundreds of places within the geographic range of the species. A particular feature of climate, namely rainfall, and possibly therefore humidity, is concerned in this correlation. The same correlation, heavy rainfall and dark color, is shown also in the other species of North American weasels. The conclusion is unavoidable that climate, directly or indirectly, determines or influences the color of weasels.

The light facial markings appear in American weasels in two separate geographic areas. One is the southwestern United States, México and northern Central America. The second area is in the same latitude, in Florida and adjoining parts of Georgia and Alabama. In the western weasels the markings are white south of latitude 32° N. North of this latitude, the facial markings, if at all extensive, usually are of the same yellowish color as the underparts of the body. Weasels of southern California and its interior valley usually have these yellowish instead of white facial markings. The light facial markings, in this instance, white markings, attain their maximum extent in M. f. leucoparia of the southwestern margin of the tableland of México, at latitude 19° N. A gradual decrease in area of the light facial markings occurs both to the north and south; they disappear at 10° N in M. f. costaricensis and at 35° N at approximately the southern limits of range of M. f. arizonensis and M. f. nevadensis. In the mild climate of California the light (yellowish) facial markings are found at still higher latitudes. These light facial markings crop up as vestiginal remnants, consisting of a few white hairs, in some individuals of nearly all races of weasels.

In certain parts of the skull there are trends, in size and shape, which continue for long distances geographically. In other words, clines can be recognized. Changes in size and shape in some other parts of the skull are wavelike; change toward narrower rostrum, for example, is not progressive in a given geographic direction for any great distance. Length of the upper tooth-rows and zygomatic breadth, when expressed as percentages of the basilar length, and also the actual length of individual teeth vary geographically in the same wavelike fashion as does the width of the rostrum.

Size of the skull, on the other hand, shows a sustained trend for a long distance; it becomes progressively smaller from the southern United States southward to Columbia, South America. This clinal variation can be demonstrated by plotting on a graph, the basilar length, the zygomatic breadth, or the weight of the skull. Beginning at Mérida, Venezuela, and proceeding southward to increasing elevations in the mountains of South America, there is a reversal of the direction of the variation in this cline; weight of skull, for example, increases to the southward from Mérida for a considerable distance. A cline of decreasing width of the postorbital constriction of the skull is evident from Panamá north into Texas.

Variations in the tympanic bullae provide many characters useful in distinguishing weasels from different localities. Most of these characters have to do with degree of inflation of the bullae. Indirectly correlated with degree of inflation is first the extent of removal of the anterior margin of the bulla from the glenoid fossa and foramen ovale, and second the form (convex, flat, or concave) of the part of the squamosal bone between the foramen ovale and the anterior margin of the tympanic bulla. As one proceeds southward from, say, southwestern Kansas through the geographic range of the species Mustela frenata, there is a progressive deflation of the bulla, an increase in length of the space between its anterior margin and the foramen ovale, and the floor of the braincase in front of the bulla changes from ventrally concave to ventrally convex. (See figs. e and h of pl. 24 and figs. e and f of pl. 27.)

One extreme of this variation in bulla is shown in Mustela frenata neomexicana (fig. e of pl. 24), in which the anterior margin of the bulla (viewed from the ventral side) rises vertically from the floor of the braincase to form a 90-degree angle. The other extreme, the uninflated bulla, is in Mustela frenata panamensis (fig. e of pl. 27), in which the anterior margin of the bulla is not raised above the floor of the braincase. This variation is remarkable because it occurs within a single species. Otherwise, in the family Mustelidae, differences in the tympanic bullae as great as that between the two subspecies M. f. neomexicana and M. f. panamensis, occur only between genera. The need for caution in inferring the limits of variation for a particular structure in one species or genus, on the basis of variation in another group, is therefore obvious.

Speaking now of full species, the most inflated tympanic bullae in American weasels are in Mustela frenata, and more restrictedly in those subspecies of it which occur in the temperate region. Subspecies of M. frenata in Central and South America, as already noted, have less inflated bullae. The tropical weasel, Mustela africana, of the Amazon drainage of South America has the bullae still less inflated (see fig. i of pl. 39 and fig. f of pl. 40). The bullae are less inflated even than in the mink, subgenus Lutreola. In M. africana the cleidomastoideus, omotrachelian, levator scapulae, and rhomboideus profundus muscles take origin from a fossa on the mastoid bone, whereas in the forms with greatly inflated bullae these muscles take origin from a raised ridge or tubercle. Using Mustela frenata of the temperate region as a starting point and proceeding northward, a reduction in inflation of the tympanic bulla is seen also in that direction in that Mustela erminea has less inflated bullae. The bullae are less inflated in southern than in far northern (arctic) populations of Mustela erminea. In erminea the lesser inflation is real enough but at the same time there appears to be less inflation than actually exists, for the squamosal floor of the braincase is "pushed down." This places the anterior end of the tympanic bulla farther in the braincase than it otherwise would be. Although the anterior end of the bulla is flattened to the extent that it resembles the sharp edge of a splitting-wedge, inspection of the lateral and medial edges shows that in its central part the bulla is more inflated than it is in the weasels of Central and South America.

For reasons set forth later, M. erminea is judged to resemble the ancestral stem form more closely than does any one of the other three American species of weasels. If this judgment is correct, the shape of the tympanic bullae of the American weasels may be explained as follows: In the subspecies of Mustela frenata of the temperate regions of North America the bullae have most nearly been pushed out of the braincase and at the same time have undergone some enlargement. The subspecies of this same species in Central and South America represent an earlier stage in the evolution of American weasels and retain less inflated bullae—less inflated even than those of the southern subspecies of erminea. M. africana probably separated from the stem form at a still earlier time if we may judge by the lesser inflation of its tympanic bullae. There are other reasons for thinking that africana separated from the stem form earlier than M. frenata did. During the time that elapsed since the separation of M. frenata from the stem form, the tympanic bullae of M. erminea probably increased slightly in size, as probably also did the brain but without shoving the auditory complex forward from its former position.

[DISTRIBUTION AND SPECIATION]

Weasels of the subgenus Mustela are known from the Pleistocene but not from deposits laid down at an earlier time (see page [10]). The Pleistocene weasels from Rancho La Brea of southern California and from Potter Creek Cave and Samwel Cave, both of northern California, are subspecifically indistinguishable from the weasels living in those same localities today. The other notable occurrence of weasels in the Pleistocene is in the Conard Fissure of Arkansas. Brown (1908:181, 182, pl. 17) names two kinds from the Fissure. One is an extinct subspecies (Mustela frenata gracilis) possibly of the species which occurs in the same region today and the other, Mustela erminea? angustidens, is an extinct subspecies of a species which occurs only farther north today. M. erminea came south, probably in front of one of the ice sheets, as did several other species of American mammals, now of more northern distribution, that left their remains in Conard Fissure. Mustela rixosa is not recorded as a fossil in America although it is known from the "Diluvial" deposits of the Old World; see Woldrich (1884:1000), who employs the name "Foetorius minutus n. sp.," and see also Zimmerman (1943:295-296).

The ermine, Mustela erminea, is the most generalized of the full species. For example, the number of teeth is as large as in any other species and greater than in certain species. The teeth are sharp-pointed, uncrowded, and individually less specialized than in any other American weasel. M1 has the inner half, or lobe, of approximately the same size as the outer lobe instead of much larger than the outer lobe (the outer lobe is the larger in several other species). The tympanic bullae are less inflated and less protruded from the braincase. The skull is rounded, and has no marked crests and ridges whereas the skulls of the other species are more pronouncedly modeled and sculptured. Therefore, it is possible to think of these other species as derived from M. erminea. A derivation in the reverse direction would be more difficult. From the foot soles of an ermine, or a weasel closely resembling an ermine, the more complex soles of Mustela africana could have been derived by a decrease in hairiness, although it would be necessary to suppose that the thenar pad has been retained in africana and has been lost in the living erminea. The alternate possibility, namely, that the thenar pad was a relatively recent acquisition in the africana line seems less probable. The tail of erminea is of "average" length and in size of entire animal erminea is intermediate between the other American weasels. Structurally, Mustela erminea appears to be nearest the stem form from which all of the living weasels ascended. Its present holarctic distribution is in harmony with the view that it is a direct descendant from the stem form because the stem forms of most of the known kinds of mustelids appear to have lived in the holarctic region. To be sure, Mustela erminea is regarded as having undergone some progressive change in structure, but less than the other weasels, in the period of time when the weasels were evolving from the stem form.

The least weasel, Mustela rixosa, seems to be an ancient type and to judge from the size and proportions of its parts, was differentiated from the erminea stem at a time earlier than were the other American Recent species of weasels. In size, in reduction of the tail, and in proportions of the skull, M. rixosa is, in each instance, the most aberrant of all the weasels, Mustela nivalis of Europe and western Asia included. This aberrancy results from the retention of certain primitive features, in the teeth and basicranial region, and from specialization in proportions of the skull. The skull is long, deep, and narrow. These proportions probably are adaptations permitting the animal to follow the smaller kinds of mice into their burrows. In most of that part of North America where erminea and rixosa occur together, erminea is a much larger animal and takes as prey almost all kinds of land vertebrates that it is powerful enough to kill. These include varying hares and ptarmigans. The least weasel, rixosa, can hardly manage such large prey and lives on the smaller rodents. Mustela rixosa may eat numbers of insects (see page [176] beyond),—a kind of food which Mustela erminea is not known to eat. Apparently the two species are able to live in the same areas because each eats a somewhat different kind of food than does the other and hence they do not compete to the point where one is crowded out by the other. This is the case in the latitudes where the two species of weasels are of different bodily size, but in the southernmost latitudes where these two species occur, erminea becomes almost as small as rixosa and only one of the species, to the exclusion of the other, occurs in a given area. All through the Rocky Mountains, south of Montana and in the territory west of these mountains all the way to the Pacific Coast, only the small subspecies of erminea is to be found. In the Alleghenies of the eastern United States only rixosa occurs. In New England where erminea approaches the size of rixosa, the latter is unknown. Probably this exclusiveness results from competition for food, although competition for dens, safe breeding places and other requirements of life may be involved.

The species erminea invaded the western United States and in the process of invasion probably developed there the small size appropriate to permit erminea to live in that latitude before it could do the same thing in the Appalachian region. Later than erminea, the least weasel, Mustela rixosa, which was small to begin with, also spread southward from the holarctic region, stopped short in the western United States at the northern boundary of the area in which erminea was of small size, but in the Appalachian region of the eastern United States continued on southward to the limits of temperature tolerant for it because erminea had not yet penetrated into that region and no other small carnivore was there to offer competition.

The long-tailed weasel, Mustela frenata, occurs mostly south of the regions inhabited by the ermine, and mostly south of the region inhabited by the least weasel which appears to live as well with frenata as with erminea. It is true that erminea and frenata occur in the same region, but this is a relatively narrow belt across the United States; and from within it a person cannot go far either north or south without reaching a region in which only one of the two species occurs. Exception has to be made for the Rocky Mountains and the Sierra Nevada, where erminea is of exceptionally small size. In these mountains and in the boreal mountainous parts of the intervening region of the United States, erminea and the large-sized frenata occur together over a wide area. Presumably the two occupy different ecologic niches, much as rixosa and frenata probably do where they occur together.

Most of the geographic range of the long-tailed weasel, M. frenata, is in the temperate region. Structurally, this species is the most advanced of the American weasels. Its dentition is the most highly specialized for cutting. M1 is relatively small and the inner lobe is slightly larger than the outer lobe. The skull, throughout, is more modeled than in the other species; the rostrum, the lower jaws and the teeth—all parts of the offensive equipment—are well developed relative to the corresponding structures in other weasels; the basicranial region exhibits an advanced stage of development in that the tympanic bullae show the maximum degree of inflation. Also, they are thrust far out of the braincase, thereby providing more room for the relatively larger brain which is protected by a more solidly built braincase than in erminea.

Several subspecies of Mustela frenata occur in the tropics, that is to say, south of the Mexican tableland and on the coastal plain to the east of it. Each is structurally more primitive than subspecies of the temperate region. As compared with Mustela frenata frenata of the temperate Mexican tableland the size in these tropical subspecies is smaller; the tail is shorter; the braincase and entire skull are less modeled; the postorbital breadth is more; the teeth are smaller; the deuterocone of P4 is not so far anterior to the protocone; the tympanic bullae are less inflated, are farther removed from the foramen ovale, and a larger proportion of each bulla is contained within the braincase. These features serve to set off from northern races of frenata all those subspecies of frenata which occur from southern México southward to the northern and western limits of the Amazon drainage of South America. The Amazon Basin is inhabited by another species, Mustela africana, having more primitive characters.

In the species frenata, the explanation for this abrupt change in characters between the animals of the temperate highlands and those of the tropical lowlands may be this: In the early Pleistocene, after the emergence of much or all of Central America took place, weasels distributed themselves over the Isthmus and into South America. These weasels were more generalized in structure than those now inhabiting the uplands of México. Failure of this stock of weasels often to cross some still-persisting water barrier, or failure of this stock to cross some water barrier that was widened or reformed because of a rise in sea level in some one of the interglacial periods of the Pleistocene cut the frenata stock into two or more parts. After the land connection was established or re-established and when the necessary precedent plants and rodents again had established themselves, the two groups of weasels, one from the northern tableland of México, and the other from the southern area of tropical complexion, met. The weasels of the frenata stock that reinvaded the area from the north probably did so by following along the chain of high volcanic cones and narrow uplifts. If and when a subsequent inundation occurred in some part of Central America, weasels were stranded on the adjacent mountains—converted into islands—only the higher parts of which were above water. Mustela frenata costaricensis and Mustela frenata goldmani may be examples of a northern stock of weasel that pushed southward in the highlands and became stranded for a short time. Following the latest emergence of land to provide a continuous highway between the two continents, weasels from the south and the insular populations, as for example, M. f. costaricensis, were the first to invade the low tropical areas most recently under water. When the Pleistocene history of Central America is better known, the facts will provide a useful means of testing the hypothesis that has been outlined immediately above.

As explained above, fossil specimens of M. frenata from deposits of the last half of Pleistocene time show that no appreciable change occurred in some areas, for example, in the vicinity of Hawver Cave and Samwel Cave of California, and that but slight change occurred in other areas, for example, in southern California (fossils from Rancho La Brea) and probably in the central United States (fossil from Conard Fissure). It is possible to imagine, therefore, that the two groups of weasels, one occurring southward only as far as the highlands of Central America and the other occurring in northern South America, had not differentiated sufficiently in the period of their isolation to prevent crossbreeding when they last came into contact. If the separation of the two groups had been maintained for a longer period, the two groups, tropical weasels and austral weasels, probably would have been so different when the two met as to prevent crossbreeding and they would have constituted two full species instead of only one.

Mustela africana is the most primitive of the American weasels. Some of the most important structural features that mark it as such are in the basicranial region. The tympanic bullae are less inflated than in other weasels, are pointed anteriorly and posteriorly, and do not have the lateral margins carried outward to the outer margins of the braincase. The mastoid sinus is not involved, by inflation or marked modification in the production of the auditory complex. Between the alisphenoid and the squamosal there is a clear demarcation posteriorly from a point directly lateral to the foramen ovale. This demarcation permits a transverse rounding of the alisphenoid to form a longitudinal ridge between the anterior margin of each bulla and the base of the pterygoid of the same side. Nevertheless, there is no such specialization of this primitive, structural feature such as occurs in some African and Asiatic mustelids in which the tympano-pterygoid part of the alisphenoid fuses with the tip of the hamulus of the pterygoid. However, the tympano-pterygoid eminence has not been obliterated in M. africana as it has in the other American weasels. Another primitive feature in the basicranial region of M. africana is the tendency toward separation of the paroccipital processes from the tympanic bullae. The thenar pad of the foot probably is an inheritance from a primitive ancestor since the pad is present in the viverrids and in a majority of mustelids judged to be more primitive than Mustela.

Some specializations are obvious in Mustela africana. One is the reduction in number of premolars; p2 is absent whereas it is normally present in the other weasels; P2 has one instead of two roots; and, in relation to the other teeth, m2 is smaller. The shortness of the preorbital part of the skull in relation to the length of the skull as a whole may reflect the mentioned reduction of the premolars or retention of a primitive shape of skull, or both. Also, certain features which denote immaturity in other weasels are retained in adults of this species, as for example, sutures on the dorsal face of the preorbital region of the skull.

Figs. 17-22. Views of the feet of American weasels (subgenus Mustela) to show differences in number and arrangement of the pads and variation in degree of hairiness of the soles. × 1-1/2. In each figure, left-forefoot on left, and left hind foot on right.

Fig. 17. M. rixosa rixosa, Halifax, N.S.; juv., ♀, 7425 U.S.N.M.

Fig. 18. M. erminea richardsonii, Ft. Chimo; ad. ♀, 14866 U.S.N.M.

Fig. 19. M. frenata noveboracensis, Mich., July 7, 1913; ad. ♂, 44689 M.Z.

Fig. 20. M. f. frenata, Brownsville, June 1, 1892; yg. ♂, 34043 U.S.N.M.

Fig. 21. M. frenata panamensis, Panamá, February 17, 1911; sad. ♀, type.

Fig. 22. M. a. africana, Pará, Brazil, Sept., 1908; yg. ♂, 37475 A.M.N.H.

Figs. 17, 18 and 19. Drawn from specimens preserved in alcohol.

Figs. 20, 21 and 22. Drawn from relaxed feet of dried skins.

Mustela africana, all characters considered, is the most aberrant of the American weasels. That is to say, greater difference prevails between M. africana and any other American weasel than exists between any other two American weasels. The distinctive cranial and dental characters, excepting the reduction in number of premolars, are of a primitive nature. For example, the relatively wide postorbital region, the large braincase that is inflated anteriorly, and the flattened tympanic bullae are points of resemblance to the holarctic Mustela erminea, the species which is regarded as most closely resembling the stem form. Also, the mentioned characters in adults of M. africana resemble ontogenic stages passed through by other weasels. Consequently, it is thought that M. africana crossed the filter-barrier from North America to South America, remained isolated from the original stock for a length of time sufficient to permit africana to differentiate from North American weasels and vice versa to such a degree that crossbreeding with the frenata stock was prevented when frenata, at a later time, pushed southward over the, then zoölogically less-effective, water barrier, or continental bridge if it was by this time in existence.

Fig. 23. Diagram indicating probable relationships of the species of American weasels.

The four full species of American weasels may well be thought of as having the same stem form of which erminea is the most nearly direct descendant. Geographic and climatic changes may have operated to isolate, and then to foster morphologic differentiation of, first rixosa in Eurasia, next africana, third the tropicalis section of M. frenata, and finally M. frenata itself, leaving M. erminea as a modern version, somewhat altered to be sure, of the stem form. Some of these ideas are expressed in figure [16]. The climate is different in the ranges of the several species and the climate has changed through time in the ranges of at least many subspecies. Natural selection of morphological features best adapted to a particular kind of climate probably has altered some species more than others. M. erminea in almost every one of its characteristics is generalized and potentially progressive whereas africana retains more characters which are truly primitive along with a few which are specializations. M. africana is potentially the least progressive of any of the American weasels. The most specialized weasels are the North American races of Mustela frenata. A progressive series of increasing specialization is comprised in (1) M. africana, (2) the M. tropicalis (Central American, lowland) section of M. frenata, and (3) the races of M. frenata in North America.

Considering now features of the environment which have obviously influenced the distribution and speciation of weasels, water barriers are important. Bering Strait, Carquinez Strait (along with San Francisco Bay) which opens through the Golden Gate, and the channels between the islands of southeastern Alaska, have contributed to the formation of subspecies. The difference is really slight on the two sides of Bering Strait and San Francisco Bay and is slightly more on two sides of each of several of the channels between the islands of southeastern Alaska. The differences between the weasels on the two sides of one of these water barriers supposedly result from the preservation in animals on one side, or on one island, of small mutations, which would be swamped by crossbreeding if the water barrier were not present. The effect of this isolation is easily seen if ermines from the Queen Charlotte Islands are compared with those of the opposite mainland. The degree of morphological difference is great. Isolationwise, the Queen Charlotte Islands are the seaward end of a chain, beginning with Admiralty Island in southeastern Alaska, and are farther from the mainland, zoölogically, than the distance in actual miles across the water channel would suggest. Between any two islands that are geographically consecutive, however, and between the mainland and the first island of the chain, the difference in the ermines is small. In other places, water barriers of equal or greater width have contributed little if anything to the differentiation from one another of weasels on the two sides of the water barrier. The strait between eastern Canada and Newfoundland is an example.

The absence of water, or scarcity of it to a degree that closely approaches absence, in any large area appears to prevent weasels from living there. At any rate, the one sizeable region of North America from which weasels are unknown is the desert of the southwestern United States and adjoining part of northwestern México. More precisely, in western Arizona, the Mohave Desert and the desert of northwestern Sonora, collectors of mammals have repeatedly sought small carnivores without ever finding any weasels.

Degree of moisture is closely correlated with color in weasels. Humidity and cloudiness as well as actual precipitation seem to be involved. Even if we take into account average annual rainfall alone, the darkest-colored weasels are found in the areas of heaviest rainfall and the lightest-colored weasels in areas of lightest rainfall (extreme type of desert where no weasels occur being excepted). In any large region where there is a geographic gradient in rainfall, the transition from light to dark color almost exactly parallels the increase in amount of rainfall. Within a given species the same color reappears in widely separated areas that have the same amount and seasonal distribution of rainfall. This correlation is repeated so often that one can almost certainly say that heavy rainfall, or the associated phenomena of high humidity and cloudiness, acting separately or together, causes an increase in intensity of color. Relative extent of the color of the upper parts and underparts and presence and absence of light facial markings seem also to be correlated, in a more general way, with differences in rainfall. A fuller discussion of the nature and amount of the variation in color is given on page [51].

Temperature seems not to be an important factor in directly limiting the distribution of weasels, since M. frenata occurs from the hottest to some of the coldest parts of the Americas. Do M. erminea and M. rixosa range no farther south, than they do at present, because high temperatures constitute a barrier? No evidence is known to me which provides an answer, one way or the other, to this question. Granting that temperature is unimportant in limiting the distribution of weasels, it seems to cause geographic variation. Increase in mean annual temperature is correlated with decreased size in M. erminea and with increased size in M. rixosa. Temperature, it seems, causes the hair to vary; the pelage is harsher and sparser in weasels from tropical regions than in those from boreal regions. Difference in number of hairs is especially well shown on the soles of the feet. In the weasels from the far north, the pads are concealed by hair and in the weasels from the tropical regions the soles are mostly bare. Also, the hair on the soles of the feet is longer in northern than in southern weasels. Furthermore there is seasonal change in length of the hair on the soles of the feet; at a given locality in southern Canada the hair of the white winter coat is so long on the soles of the feet as to obscure completely the palmar and plantar pads whereas the hair of the brown summer coat is shorter and leaves these pads boldly exposed to view. This seasonal change, as would be expected, is most marked in animals of northern regions and is not perceptible in those from the tropics; it is correlated with increase in seasonal change as the distance from the equator increases.

Temperature and moisture acting together may cause extensive white facial markings, that neither alone would cause. In Mustela frenata these markings occur where there is heavy rainfall and high mean annual temperature. Where there is heavy rainfall and a low mean annual temperature they do not occur and where there is high mean annual temperature and light rainfall the markings are not pure white but are of the same color as the underparts. [Plate 1] and the description of color on page [51] may be consulted in this connection. Extremely high mean annual temperature together with extremely heavy rainfall may inhibit the development of light facial markings. M. f. meridana, panamensis and costaricensis are cases in point. In either direction, north or south, from the territory inhabited by these three subspecies a similar combination of temperature and rainfall is found and similar light facial markings appear there.

Considering the delicate response of structure to climate, a person naturally questions whether or not natural selection accounts for all of the differences between subspecies. To show that natural selection determines the color of Mustela frenata, it would be necessary to assume that climate, color, and utility of color are positively correlated. Although climate (rainfall) and color are correlated in such a manner that three subspecies of weasel in places as far apart as New England, Perú, and the state of Washington are colored alike, other features of the three environments are unlike. Kinds of animals which the weasel catches for food, and flora in which the weasel finds concealment, are dissimiliar. If natural selection alone determined the color, some difference in color would be expected between the weasel which needed to be obliteratively colored, that is camouflaged, the better to catch a Phyllotis in Perú and the weasel in Washington which needed nature's aid in catching Microtus. Mustela frenata goldmani of the highlands of southern México, which is known to attack the huge pocket gophers, Orthogeomys and Cratogeomys, has a weaker dental armature than Mustela frenata texensis which does not have to overcome prey so formidable as does goldmani. Equally formidable enemies endanger M. f. goldmani and texensis. Examples of this nature could be multiplied. Without actually proving anything concerning selection, these examples give reason for us to suppose that some characters are not determined by natural selection.

Another question upon which data obtained from a study of Mustela has some bearing, is this: Where the geographic ranges of two subspecies meet, why does not the swamping effect of crossbreeding cause one subspecies to disappear? Although swamping may have occurred in some instances, it does not occur in the majority of instances. Witness the long-continued existence of the living subspecies Mustela frenata nevadensis of which skulls are available from Pleistocene deposits. Therefore, its distinctive characters, cranially at least, have been maintained for a long time. Furthermore, these characters are maintained over a large geographic region more than a thousand miles across. On the eastern margin of its range, at the eastern base of the Rocky Mountains in Colorado, M. f. nevadensis intergrades in a relatively narrow belt with the lighter-colored, longer-tailed and cranially different Mustela frenata longicauda, which has a geographic range almost equally extensive. M. f. longicauda also is uniform in its characters over a large area but at approximately 400 miles east of the base of the Rocky Mountains, it begins to intergrade with the darker-colored, shorter-tailed and cranially different Mustela frenata primulina and does so over a belt of 100 miles or more in width. At any given locality within this wide belt of intergradation the range of individual variation ordinarily does not exceed that in animals from a given locality well within the geographic range of M. f. longicauda. In the narrow belt of intergradation along the eastern base of the Rocky Mountains, the range of individual variation at several places is greater than in animals from a given locality well within the geographic range of M. f. longicauda or for that matter from well within the geographic range of M. f. nevadensis.

Considering the dominance and recessiveness of genes and the genetic mechanism in general by which characteristics of offspring are inherited from their parents, it would seem that M. f. longicauda and for that matter M. f. nevadensis and M. f. primulina would lose their distinctive characteristics because of the crossbreeding that is every year going on between longicauda and nevadensis on the one hand and between longicauda and primulina on the other hand.

Sumner (1932:84) suggests that homogeneity is prevented by population pressure. Applying his suggestion to the species Mustela frenata we could say that the subspecies longicauda pressing westward meets strong pressure from the subspecies nevadensis pressing eastward and that the width of the zone of intergradation between the two subspecies varies inversely with the strength of the population pressure from the two sides. Sumner recognizes that according to his hypothesis the two contiguous races would remain distinct only so long as there was a preponderance of centrifugal movement from both of the centers of dispersal. Sumner (op. cit.:85) recognizes that an abrupt change of environmental conditions could account in part for the boundaries of the ranges of the two subspecies and finally that his hypothesis does not certainly answer the question of why crossbreeding does not result in homogeneity between two subspecies with contiguous geographic ranges.

The hypothesis of harmoniously stabilized complexes of genes was offered by Timofeeff-Ressovsky (1940:124) to explain why the swamping effect of crossbreeding does not obliterate subspecies. The hypothesis takes into account that any one of several characters of a subspecies may be caused by several genes. Some characters of this kind may be favored by natural selection more than others. In the belt of intergradation between two subspecies, where two of these favored characters meet, a "biological tension" as Huxley (1939:415) terms it "will result, which will produce partial discontinuity between the two groups. Each group will evolve a gene-complex which is not only broadly adapted to the external environment of the central area of its range, but is also harmoniously stabilized, in adaptation to the internal genetic environment, by the selection of modifiers." Crosses, that is to say intergrades, between the two subspecies will lack this stabilization and will therefore be at a selective disadvantage. The zone of intergradation will therefore remain narrow; intermediates are constantly being brought into existence there by crossing but are as constantly being extinguished by selection.

These two hypotheses are the best that geneticists yet have offered. Neither has been tested and both, as originally proposed, would hardly apply everywhere because there are some contradictions.