[Pg 1]

University of Kansas Publications

Museum of Natural History

Volume 9, No. 1, pp. 1-68, figures 1-18

  December 10, 1955  

Speciation of the Wandering Shrew

BY

JAMES S. FINDLEY

University of Kansas

Lawrence

1955

[Pg 2]

University of Kansas Publications, Museum of Natural History



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








Volume 9, No. 1, pp. 1-68, figures 1-18

Published December 10, 1955













University of Kansas

Lawrence, Kansas

PRINTED BY

FERD VOILAND, JR., STATE PRINTER

TOPEKA, KANSAS

1955



25-7903

[Pg 3]

[Pg 4]

The purpose of this report is to make clear the biological relationships
between the shrews of the Sorex vagrans-obscurus "species
group." This group as defined by H. H. T. Jackson (1928:101)
included the species Sorex vagrans, S. obscurus, S. pacificus, S.
yaquinae, and S. durangae. The last mentioned species has been
shown (Findley, 1955:617) to belong to another species group.
Sorex milleri, also assigned to this group by Jackson (1947:131),
seems to have its affinities with the cinereus group as will be explained
beyond. The position of the vagrans group in relationship
to other members of the genus will be discussed.

Of this group, the species that was named first was Sorex vagrans
Baird, 1858. Subsequently many other names were based on members
of the group and these names were excellently organized by
Jackson in his 1928 revision of the genus. Subsequent students of
western mammals, nevertheless, have been puzzled by such problems
as the relationship of (1) Sorex vagrans monticola to Sorex
obscurus obscurus in the Rocky Mountains, (2) Sorex pacificus, S.
yaquinae, and S. obscurus to one another on the Pacific Coast, and
(3) S. o. obscurus to S. v. amoenus in California. Few studies have
been made of these relationships. Clothier (1950) studied S. v.
monticola and S. o. obscurus in western Montana and concluded that
the two supposed kinds actually were not separable in that area.
Durrant (1952:33) was able to separate the two kinds in Utah as
was Hall (1946:119, 122) in Nevada. Other mammalogists who
worked within the range of the vagrans-obscurus groups have
avoided the problems in one way or another. Recently Rudd (1953)
has examined the relationships of S. vagrans to S. ornatus.

Approximately 3,465 museum study skins and skulls were studied. Most
of these were assembled at the University of Kansas Museum of Natural History,
but some were examined in other institutions.

Specimens were grouped by geographic origin, age, and sex. Studies of
the role of age and sex in variation were made. Because it was discovered
that secondary sexual variation was negligible, both males and females, if of
like age and pelage, were used in comparisons designed to reveal geographic
variation.

External measurements used were total length, length of tail, and length of
hind foot. After studying a number of cranial dimensions I chose those listed
below as the most useful in showing differences in size and proportions of the
skull. Figures 1 and 2 show the points between which those measurements
were taken.

[Pg 5]
Condylobasal length.—From anteriormost projection of the premaxillae to posteriormost
projection of the occipital condyles (a to a´).

Maxillary tooth-row.—From posteriormost extension of M3 to anteriormost extension
of first unicuspid (b to b´).

Palatal length.—From anteriormost projection of premaxillae to posteriormost
part of bony palate (c to c´).

Cranial breadth.—Greatest lateral diameter of braincase (d to d´).

Least interorbital breadth.—Distance between medialmost superior edges of
orbital fossae, measured between points immediately above and behind
posterior openings of infraorbital foramina (e to e´).

Maxillary breadth.—Distance between lateral tips of maxillary processes
(f to f´).

Figs. 1 and 2. Showing where certain cranial
measurements were taken. × 3½. (Based on
Sorex vagrans obscurus, from Stonehouse Creek,
5½ mi., W junction of Stonehouse Creek and
Kelsall River, British Columbia, , 28545 KU.)

In descriptions of color, capitalized terms refer to those in Ridgway (1912).
In addition the numerical and alphabetical designations of these terms are given
since a knowledge of the arrangements of these designations enables one quickly
to evaluate differences between stated colors. Color terms which are not
capitalized do not refer to any precise standard of color nomenclature.

In the accounts of subspecies, descriptions, unless otherwise noted, are of
first year animals as herein defined. Descriptions of color are based on fresh
pelages.

[Pg 6]
Unless otherwise indicated, specimens are in the University of Kansas
Museum of Natural History. Those in other collections are identified by the
following abbreviations:

In nature, the subspecies of Sorex vagrans form a cline and are distributed
geographically in a chain which is bent back upon itself. The subspecies in
the following accounts are listed in order from the southwestern end of the
chain clockwise back to the zone of overlap.

The synonymy of each subspecies includes the earliest available name and
other names in chronological order. These include the first usage of the name
combination employed by me and other name combinations that have been
applied to the subspecies concerned.

In the lists of specimens examined, localities are arranged first by state or
province. These are listed in tiers from north to south and in any given tier
from west to east. Within a given state, localities are grouped by counties,
which are listed in the same geographic sequence as were the states and
provinces (N to S and W to E). Within a given county, localities are arranged
from north to south. If two or more localities are at the same latitude the
westernmost is listed first. Marginal localities are listed in a separate paragraph
at the end of each account. The northernmost marginal locality is listed first
and the rest follow in clockwise order. Those records followed by a citation to
an authority are of specimens which I have not personally examined. Marginal
records are shown by dots on the range maps. Marginal records which cannot
be shown on the maps because of undue crowding are listed in Italic type.

To persons in charge of the collections listed above I am deeply indebted.
Without their generous cooperation in allowing me to examine specimens in
their care this study would not have been possible. Appreciated suggestions
in the course of the work have been received from Professors Rollin H. Baker,
A. Byron Leonard, R. C. Moore, Robert W. Wilson, and H. B. Tordoff, and
many of my fellow students. Mr. Victor Hogg gave helpful suggestions on the
preparation of the illustrations. My wife, Muriel Findley, devoted many hours
[Pg 7]
to secretarial work and typing of manuscript. Finally I am grateful to Professor
E. Raymond Hall for guidance in the study and for assistance in preparing the
manuscript. During the course of the study I received support from the
University of Kansas Endowment Association, from the Office of Naval Research,
and from the National Science Foundation.

Non-geographic variation, that is to say, variation within a single
population of shrews, consists of variation owing to age and normal
individual variation. In Sorex I have detected no significant secondary
sexual differences between males and females; accordingly
the two sexes are here considered together.

Variation with age must be considered in order to assemble
comparable samples of these shrews. Increased age results in wear
on all teeth and in particularly striking changes in the size and shape
of the first incisors. Skulls of older shrews develop sagittal and
lambdoidal ridges, and further differ from skulls of young animals
in being slightly broader and shorter, and in developing thicker
bone, particularly on the rostrum which thus seems to be, but is
not always in fact, more robust. Pruitt has recently (1954) noted
these same cranial differences in specimens of Sorex cinereus of
different ages.

Several students of American shrews, notably Pearson (1945)
on Blarina, Hamilton (1940) on Sorex fumeus, and Conaway (1952)
on Sorex palustris, have shown that young are born in spring and
summer, usually reach sexual maturity the following spring, and
rarely survive through, or even to, a second winter. The result is
that collections made, as most of them are, in spring and summer,
contain two age classes, first year and second year animals. These
two age classes are readily separable on the basis of differences in
the skull as well as on the decreased pubescence of the tail and the
increased weight of second year animals. My own examination of
hundreds of museum specimens confirms this for the Sorex vagrans
group. Separation of the two age classes in an August-taken series
of Sorex vagrans from coastal Washington is shown in figure 3, in
which two tooth-measurements that are dependent upon wear are
plotted against one another.

First year animals are more abundant in collections than are
second year animals. Within the first year, that is to say from spring
to late fall, animals vary but little. Dental characters are best
studied in first year shews. For this reason I have used them as the
basis for the study of geographic variation, and descriptions are
based on first year animals unless otherwise noted.

[Pg 8]

Within the Sorex vagrans complex, the only characters of taxonomic
significance that I have detected are in size and color. It is
true that cranial proportions, such as relative size of rostrum, may
change from population to population, but these proportions seem
to me to be dependent upon actual size of the individual shrew as
I shall elsewhere point out. Of the cranial measurements here
employed, palatal length and least interorbital breadth are the most
significant and useful. Color in the S. vagrans group seems to be
in Orange and Cadmium Yellow, colors 15 and 17 of Ridgway
(1912). No specimens actually possess these pure colors, but most
colors in these shrews are seen to be derived from the two mentioned
by admixture of black and/or neutral gray. In color designations
an increase in neutral gray is indicated by an increased number of
prime signs ( ´ ), whereas increase in black is indicated by progressive
characters of the Roman alphabet (i, k, m). Thus, 17´´k is grayer
than 17´k and 17´´m is blacker than 17´´k. In subspecific diagnoses in
this report, color and size, and sometimes relative size, are the
characters usually mentioned.

Fig. 3. Two measurements (in millimeters)
reflecting tooth-wear plotted against one another. First year and second
year individuals of Sorex vagrans vagrans, all taken in August at
Willapa Bay, Washington, are completely separated. Open circles represent
teeth of second year shrews; solid circles represent teeth of first year
shrews.

[Pg 9]

In general, winter pelage is darker than summer pelage in these
shrews. Winter pelage comes in first on the rump and spreads
caudad and ventrad. The growth line of incoming hair is easily
detected on the fur side of the skin. Throughout the winter the
color of the pelage changes, often becoming somewhat browner,
although no actual molt takes place. This was noted by Dalquest
(1944) who assumed that the color change resulted from molt
although he was unable to detect actual replacement of hairs.
Summer pelage usually comes in first on the back or head and moves
posteriorly and laterally. Time of molt depends on latitude and
altitude. Summer pelage may appear fairly late in the season and
may account for the anomalous midsummer molt noted by Dalquest.
Fresh pelages of summer and winter are best seen in first year
animals and are less variable than are worn pelages and hence are
used as the basis of color descriptions.

The largest shrews of the vagrans group (large in all dimensions)
occur in the coastal forests of northern California and of Oregon.
Those shrews are reddish, large-skulled, large-toothed, and have
rostra that are large in proportion to the size of the skull as a whole.
The very largest of these shrews live along the coast of northwestern
California. To the southward they are somewhat smaller, and at successively
more northern localities, to as far as southwestern British
Columbia, they are likewise progressively smaller and also somewhat
less reddish. The relative size of the rostrum decreases with
the decrease in size of the skull; consequently smaller shrews have
relatively smaller rostra (see fig. 4). In addition the zygomatic
ridge of the squamosal decreases in relative size with decrease in
actual size of the skull. Thus, these features change in a clinal
fashion as one proceeds from, say, Humboldt County, California,
northward to Astoria, Oregon.

Turning our attention now farther inland to the Cascade Mountains
of northern Oregon, the shrews there also are smaller and less
reddish (more brownish) than in northwestern California, and the
trend to smaller and darker shrews culminates in the northern
Cascades of Washington. Shrews from there, and from the southwestern
coast of British Columbia, compared with those from
northwestern California, are much smaller and have so great a suffusion
of black that they appear brown rather than red. At places
[Pg 10]
along the coast successively farther north of southwestern British
Columbia the shrews become larger again, the largest individuals
being those from near Wrangell, Alaska. From that place northwesterly
along the coast of Alaska, size decreases again.

Fig. 4. Condylobasal length (in millimeters)
plotted against palatal index (palatal length/condylobasal length × 100)
in several subspecies of Sorex vagrans to show relative increase
in size of rostrum with actual increase in size of skull.

The shrews so far discussed inhabit forests in a region of high
rainfall and a minimum of seasonal fluctuation in temperature. Such
a habitat seems to be the optimum for shrews of the vagrans group
since the largest individuals are found there. In addition, shrews
seem to be as common, or commoner, in this coastal belt, than they
are in other places.

The large shrews of the vagrans group on the Pacific coast were
divided into three species by H. H. T. Jackson in his revision of the
North American Sorex in 1928. The large reddish shrews of the
coast of California and southern Oregon were called S. pacificus.
The somewhat smaller ones from the coast of central Oregon were
called S. yaquinae. Still smaller shrews from northwestern Oregon
and from the rest of the Pacific coast north into Alaska were called
S. obscurus. I find these kinds to intergrade continuously one with
the next in the manner described and conclude that all are of a
single species.

[Pg 11]

Inland from the coasts of British Columbia and Alaska the size
of the vagrans shrew decreases rapidly. Specimens from western
Alaska, central Alaska, and the interior of British Columbia are
uniformly smaller than coastal specimens. In addition the red of
the hair is masked more by neutral gray than by black with the
result that the pelage is grayish rather than brownish or reddish.
Shrews of this general appearance are found southward through the
Rocky Mountain chain to Colorado and New Mexico. On the more
or less isolated mountain ranges of Montana east of the continental
divide the vagrans shrew is somewhat smaller still. On the Sacramento
Mountains of southeastern New Mexico the shrew is somewhat
larger and slightly darker. Southwestward from the Colorado
Rockies this shrew becomes smaller and slightly more reddish (less
grayish).

All of these montane populations of the vagrans shrew are commonest
in hydrosere communities, that is to say, streamsides and
marshy areas where the predominant vegetation is grass, sedges,
willows, and alders. Since these animals are less common within
the montane forests, hydrosere communities, rather than the actual
forest, seem to be the positive feature important for the shrews.

The shrews of the montane region just described were regarded
by Jackson as belonging to two species: Sorex obscurus, occupying
all the Rocky Mountains south to, and including, the Sacramento
Mountains; S. vagrans, made up of small individuals from various
places in Wyoming, Montana, and Colorado, and all the shrews of
western New Mexico and all of Arizona. My study of these animals
has led me to conclude that the smaller shrews of Arizona and New
Mexico intergrade in a clinal fashion with the shrews of Colorado
and in fact represent but one species. Since some individuals from
Colorado are as small as larger individuals from this southwestern
population of small animals, I conclude that such specimens are
the basis for reports of S. vagrans from Colorado. The shrews of
the Sacramento Mountains resemble those of the Colorado Rockies
more than they do the smaller shrews of western New Mexico and
Arizona, possibly because the climate is similar in the Sacramento
Mountains and the higher Colorado Rockies. There is less precipitation
in the more western mountain ranges in New Mexico and in
Arizona in April, May, and June than in the Colorado Rockies.
These months are critical for the reproduction and growth of shrews.

[Pg 12]
As mentioned above, the shrews from east of the continental divide
in Montana are smaller than those of the other mountains of the
state, and it is upon such small animals that the name Sorex vagrans
has been based in this area. It is clear, however, that these smaller
animals intergrade with the larger shrews of the more western
mountains. The small size might be an adaptation to the lesser
precipitation and harsher continental climate east of the continental
divide in Montana.

The vagrant shrews of the Great Basin and adjoining Columbia
Plateau and Snake River Plains are smaller than their relatives in
the Rocky Mountains and, by virtue of less gray in their pelage,
are reddish in summer and blackish rather than grayish in winter.
There is little significant geographic variation in shrews throughout
this region, although owing to their restriction to the vicinity of
water, the populations of shrews are more or less isolated from one
another and each is somewhat different from the next. Those from
nearest the Rockies are sometimes slightly larger and those from
some places in Nevada are slightly paler than the average. This small
reddish shrew is found all the way to the Pacific coast of California,
Oregon, and Washington. In these coastal areas it is somewhat
darker and sometimes a trifle larger than elsewhere. It intergrades
with a somewhat larger, grayer shrew in the Sierra Nevada of
California. Along the Wasatch front in Utah, this Great Basin shrew
intergrades with the larger, grayer shrew of the Rockies. Owing
to the abrupt change in elevation, the zone of intergradation is
rather narrow horizontally. In the latitude of Salt Lake City,
populations of intergrades occur at between 8,700 and 9,000 feet
elevation. The lowland shrew occurs in the eastern part of the Snake
River Plains, and along the valleys of the Bear and Salt rivers into
Wyoming. Along the northern edge of the Snake River Plains and
on the western edge of the mountains of central Idaho the transition
from lowland to montane habitats is abrupt and in consequence the
zone of contact between small and large shrews is narrow. In
northern Idaho and northwestern Montana the transition from lowland
to highland is more gradual. Tributaries of the Columbia
River system, especially the Clark Fork, provide a path for movement
of lowland forms into intermontane basins of western Montana.
In addition, the vegetational zones are found at lower elevations,
and there are boreal forests in the lowlands rather than only in
[Pg 13]
the mountains as is the case in Utah and Colorado. In this area,
therefore, the zone of intergradation between the smaller lowland
shrew and the larger montane shrew is more gradual and gradually
intergrading populations are found over a relatively large area.
This has been well demonstrated for northwestern Montana by
Clothier (1950). In southern British Columbia and northern Washington
this shrew in the mountains is large and in the intermontane
valleys is small. There is extensive interdigitation of valleys and
mountain ranges, and, consequently, of life-zones in this region.
In a few places, recognizably distinct populations of the vagrant
shrew occur within a few miles of one another, but in other places
there are populations of intergrades. West of the Cascades no
evidence of intergradation has been found and the two kinds occur
almost side by side and maintain their distinctness.

These Great Basin shrews dwell in hydrosere communities as do
their Rocky Mountain counterparts. In this arid region such a
habitat obviously is the only one habitable for a shrew of the vagrans
group. These shrews often maintain their predilection for such
habitats when they reach the Pacific coast, and are commonly found
in such places as coastal marshes, marshy meadows, and streamsides,
while the woodlands are inhabited by other species.

These small shrews of the Great Basin and the small vagrant
shrews of the Pacific Coast were called Sorex vagrans by Jackson.

Large reddish shrews of the coast of California and southwestern
Oregon become smaller and darker to the north. From southwestern
British Columbia they again become larger as one proceeds
northward along the coast to Wrangell, Alaska, and north of that
they again become smaller. Moving inland from the coast the
shrews become markedly smaller in Alaska and British Columbia.
The smaller inland and montane form occurs south through the
Rocky Mountains, becoming slightly smaller in central Montana,
slightly larger in southeastern New Mexico, and slightly smaller in
western New Mexico and in Arizona. This montane form intergrades
with a smaller more reddish Great Basin shrew, the zone of
intergradation roughly following the western slope of the Rocky
Mountains. The Great Basin shrew occurs westward to the Pacific
Coast; there the Great Basin shrew occurs with, although in part it
is ecologically separated from, the large reddish coastal shrews.

There seems to be an intergrading chain of subspecies of one
[Pg 14]
species, the end members of which (the small Great Basin form
and the large coastal form) are so different in size and ecological
niche that they are able to coexist without interbreeding. In southern
British Columbia the morphological differences are not so
marked as farther south along the Pacific Coast. There, in British
Columbia, reproductive isolation is not complete and occasional
populations of intergrades occur. In Montana extensive intergradation
occurs in a broad zone of transitional habitat. Along the western
edge of the Rockies from Idaho south to Utah the zone of
transition from montane to basin habitat is sharp and the zone of
intergradation, although present, is fairly narrow, perhaps because
there is little intermediate habitat which logically might be expected
to be most suitable for intergrading populations.

The oldest name applied to a shrew of the group under consideration
is Sorex vagrans Baird, 1858, the type locality of which is
Willapa Bay, Pacific County, Washington. The name applies to
the small vagrant shrew of this area, rather than to the larger forest
dweller which has been known as Sorex obscurus. The name S.
vagrans, in the specific sense, must therefore apply to all the shrews
discussed which have heretofore been known by the names S. pacificus,
S. yaquinae, S. obscurus, and S. vagrans.

A situation such as the one here described where well differentiated
end members of a chain of subspecies overlap over an extensive
geographic range throughout the year without interbreeding—thus
reacting toward one another as do full species—so far as I know
has not previously been found to exist in mammals. The overlapping
end-members of the chain of subspecies of Sorex vagrans
really do coexist; specimens of the overlapping subspecies have
been taken together at the same localities from California to British
Columbia. I have taken a specimen of S. v. vagrans and several of
S. v. setosus in the same woodlot at Fort Lewis, Pierce County,
Washington. Two subspecies of deer, Odocoileus hemionus, in the
Sierra Nevada of California, occur together over a sizeable area but
for only a part of each year that does not include the breeding
season (Cowan, 1936:156-157). In the deer mouse, Peromyscus
maniculatus, the geographic ranges of several pairs of subspecies
meet at certain places without intergradation of the two kinds. In
these instances well marked ecological differences exist between the
subspecies involved. In western Washington, for example, the
geographic range of the lowland subspecies, P. m. austerus, interdigitates
to the east and west with the range of the montane and
coniferous forest-inhabiting subspecies, P. m. oreas, and the two
kinds have not been shown to intergrade. Peromyscus maniculatus
artemesiae and P. m. osgoodi come together without interbreeding
in Glacier National Park, Montana. P. m. artemesiae is almost entirely
a forest-dwelling subspecies, whereas osgoodi is an inhabitant
of open country. The two kinds do not actually occur together
ecologically although they occur together in buildings at the edge
of the woods (A. Murie, 1933:4-5).

[Pg 15]

Fig. 5. Probable present geographic
distribution of Sorex vagrans. The range of S. v. vagrans
and its derivatives S. v. vancouverensis, S. v. halicoetes,
and S. v. paludivagus, is shown by lines slanting in a different
direction than those which mark the range of all the other subspecies of
S. vagrans. The region in which S. v. vagrans occurs
together with other subspecies of S. vagrans is shown by the
superposition of one pattern upon the other.

[Pg 16]

Cases of sympatric existence of two subspecies of one species are
known in birds and in reptiles. Notable examples are in the gull,
Larus argentatus (Mayr, 1940), in the Old World warbler, Phylloscopus
trochiloides (Ticehurst, 1938), and in the great titmouse,
Parus major (Rensch, 1933), of the Old World. In the first species
the two end-members, the herring gull and the lesser black-backed
gull, occur together over an extensive region from northern Europe
and the British Isles throughout Fennoscandia. Fitch (1940) described
a rassenkreis with overlapping subspecies in the garter
snake Thamnophis ordinoides.

The geographic distribution of the species Sorex vagrans is shown
in figure 5. The geographic range of the Great Basin subspecies
is shown by a different pattern of lines than the other subspecies of
S. vagrans. In the region in which the geographic range of the
Great Basin subspecies overlaps those of the subspecies of the
Pacific Coast, the pattern of shading for the Great Basin subspecies
is superimposed on the patterns for the other subspecies.

The distribution of the species Sorex vagrans and that of its immediate
ancestors obviously has not always been the same; during
glacial ages much of the present range of the species in Canada and
in some of the higher mountains of the United States was covered
with ice and not available to the shrew. Furthermore, large areas
that are now too hot and dry to permit the existence of S. vagrans
were at one time habitable. If we are to speculate on the manner
in which the Sorex vagrans rassenkreis originated we must inquire
into the nature and extent of these climatic changes.

The most recent epoch of geological time, the Pleistocene, is
known to have been divided into a series of alternating glacial and
interglacial ages. During the glacial ages continental and montane
glaciers are judged to have covered much of Canada and the northern
United States. Concurrently the major storm tracks of the west
probably were shifted southward; in any event much of the now
arid intermontane west was much better watered than it is today.
[Pg 17]
The increased precipitation, and probably glacial meltwater, formed
large lakes in the closed basins of the Great Basin. There were
boreal forests at lower elevations than there are today in comparable
latitudes and continuous boreal habitat probably connected many
of the isolated mountain ranges of the southwest. That probability
is supported by the presence of boreal animals and plants on many
[Pg 18]
of these isolated ranges today. A boreal tree squirrel, such as
Tamiasciurus, could hardly be suspected of crossing a treeless, intermontane
desert valley, miles wide.

Figs. 6a-6f. Fig. 6a.
Sorex vagrans pacificus, 1 mi. N Trinidad, Humboldt Co.,
California, FC 1442. Fig. 6b. S. v. yaquinae, Newport,
Lincoln Co., Oregon, AW 707. Fig. 6c. S. v. yaquinae
(near bairdi), McKenzie Bridge, Lane Co., Oregon, AW 82.
Fig. 6d. S. v. setosus, Reflection Lake, Jefferson Co.,
Washington, CMNH 4275. Fig. 6e. S. v. obscurus, 10 mi. SSW
Leadore, Lemhi Co., Idaho, FC 1499. Fig. 6f. S. v. vagrans,
Baker Creek, White Pine Co., Nevada, 88042 (after Hall, 1946:113).

Interglacial ages were characterized by warmth and aridity as
compared to the glacial ages. Glaciers retreated or disappeared,
boreal forests became montane in much of the United States, and
the lakes in the Great Basin were reduced or disappeared. One can
envision that during such times boreal mammals were isolated,
their geographic ranges were restricted, and Sonoran mammals expanded
their ranges.

Evidence is more extensive concerning the number and extent of
glacial ages in the eastern than in the western part of North America.
This evidence suggests a division of the Pleistocene into four glacial
ages and four interglacial ages, the fourth interglacial age corresponding
to the present time. More information is available about
the Wisconsinan, or last, glacial age, than about the earlier ones,
because the last glaciation in many montane areas destroyed evidence
of earlier glaciations. The names of currently recognized
glacial and interglacial ages of the Pleistocene are listed below. The
names of interglacial ages are in Italic type.

We may think of these ages as an alternating series of cool moist
and warm dry periods during which boreal mammals, and other
organisms, alternately moved southward (disappearing in the glaciated
regions) and northward into previously glaciated areas (while
disappearing from southern areas except on isolated mountain
ranges). Sorex vagrans probably followed this pattern of movement
and now is restricted to forested or well-watered places.

One possible series of events culminating in the formation of the
Sorex vagrans rassenkreis may be thought of as having begun during
the Illinoian age. With much of Canada, and perhaps also many
areas in the Rockies, Cascades, and the Sierra Nevada covered with
glacial ice, the shrew-stock ancestral to Sorex vagrans may well have
occupied a more or less continuous range over the Colorado Plateau,
the Columbian Plateau, the Great Basin, and in the forests of the
Pacific Coast (as well as over part of eastern United States, as will
[Pg 19]
be explained beyond; see fig. 7). At that time the species probably
was a continuously interbreeding unit.

Fig. 7. Possible distribution in Illinoian
(inset) and Sangamonian times of the ancestor of the Sorex
vagrans-ornatus-longirostris-veraepacis complex. Approximate
southern boundary of Illinoian glaciation marked by heavy line.

In the ensuing Sangamonian interglacial age all glaciers retreated
or disappeared thereby opening up extensive areas in the north and
in the higher mountains which were occupied by a boreal fauna,
including S. vagrans. Concurrently the Great Basin, and probably
also much of the Columbian Plateau, became dry, and desert conditions
developed, perhaps much as they are today. Increasing
aridity eliminated shrew habitat in most places between the Rocky
Mountains and the Sierra Nevada-Cascade mountain chain with the
result that the geographic range of the species resembled an inverted
"U", one arm lying along the Rocky Mountains and the other along
the Cascade-Sierra Nevada axis; the connection between the two
arms was in British Columbia (see fig. 7). At present Sorex vagrans
[Pg 20]
does occur in isolated places in the Great Basin, but its existence
there is tenuous and seemingly dependent upon the occurrence of
permanent water such as Ruby Lake and Reese River. With such
an arrangement as this it can readily be seen that gene flow between
the eastern and western arms of the "U" would be greatly reduced
by distance; consequently differentiation between the two might
be expected.

Fig. 8. Possible distribution of Sorex
vagrans at two different times in the Wisconsinan Age. Left, early
Wisconsinan; right, mid-Wisconsinan.

Wisconsinan glaciation again rendered Canada uninhabitable,
and it is quite possible that extensive areas in the Rocky Mountains,
the Cascades and the Sierra Nevada were heavily glaciated. With
the elimination of the northern part of the "U", the eastern and
[Pg 21]
western arms became isolated, if not by the width of the Columbian
Plateau at least by the glaciated Cascade Mountains. At the same
time extensive areas on the Colorado Plateau and much of the area
south to the Mexican highlands were again occupied by the species.
Finally the Great Basin, again being well-watered, provided suitable
habitat for, and was reoccupied by, Sorex vagrans (see fig. 8).
This reoccupation of the Great Basin took place probably from
the Colorado Plateau and mountains of Arizona and Utah, since the
present day shrews of the species S. vagrans in the Great Basin
closely resemble Rocky Mountain shrews but differ markedly from
the large endemic subspecies of the Pacific Coast.

Finally, with the waning of Wisconsinan ice, the species again
was able to occupy northern and montane areas as it had during
Sangamonian times. Again dessication of the Great Basin caused
drastic restriction of shrew habitat. The small, marsh-dwelling kind
of wandering shrew which had developed there around the lakes
of Wisconsinan time occupied suitable habitat all the way to the
Pacific coast where its range came into contact with that of the western
arm of the Sangamonian "U."-pattern of shrew distribution (see
fig. 9). The animals of this western segment and the new arrivals
from the east were by this time so different from one another that
the two kinds lived in the same areas without interbreeding. The
descendants of the original western arm now are known as Sorex
vagrans sonomae, S. v. pacificus, S. v. yaquinae, and S. v. bairdi. The
newcomers from the east are known as S. v. vagrans, S. v. halicoetes,
S. v. paludivagus and S. v. vancouverensis.

In addition to occupying the Pacific Coast from San Francisco
Bay north to the Fraser Delta, the Great Basin subspecies populated
the Columbia Plateau and the western foothills of the central and
northern Rockies. By so doing that subspecies came into secondary
contact with its own parent stock with which it was still in reproductive
continuity in Utah. In some places in British Columbia differentiation
between the two kinds had proceeded to such an extent
that some reproductive isolation was effected, but in many other
places the two interbred. The Rocky Mountain form spread north
and west and occupied the Cascades and coastal lowlands in southwestern
British Columbia and in Washington. Here the differentiation
between the Rocky Mountain subspecies and the Great Basin
subspecies was great enough to cause complete reproductive isolation.

[Pg 22]

Fig. 9. Probable changes in the distribution
of Sorex vagrans concurrent with and following the dissipation
of Wisconsinan ice. Dark arrows in Washington, Idaho, Oregon, and
California, shows S. v. vagrans.

[Pg 23]

Deglaciation of the Sierra Nevada opened it up for reoccupation
from the east by Sorex vagranss of the Great Basin. In response to
the montane environment the subspecies obscuroides, resembling
the subspecies obscurus of the Rockies, developed.

Desiccation of the intermontane parts of New Mexico, Arizona,
and Chihuahua, left "marooned" populations of Sorex vagrans on
suitable mountain ranges. In this way Sorex vagrans orizabae may
have been isolated in southern Mexico. The isolated populations
of Arizona and New Mexico differentiated in situ into the subspecies
monticola and neomexicanus.

Western Canada and Alaska were populated by shrews which
originated in the habitable parts of the Rocky Mountains and
Colorado Plateau during Wisconsinan time (as opposed to shrews
originating, as subspecies, in the Great Basin or on the Pacific
Coast). These shrews differentiated into the currently recognized
subspecies of the west coast and coastal islands of British Columbia
and Alaska in response to the different environments in these places,
many of which were isolated; the subspecies isolatus, mixtus, setosus,
longicauda, elassodon, prevostensis, malitiosus, and alaskensis are
thought to have originated in this fashion after the areas now occupied
by them were freed of Wisconsinan ice.

This group of shrews from the Rocky Mountains probably came
into contact with the Pacific coastal segment of the species somewhere
in northwestern Oregon. The clinal decrease in size from
S. v. pacificus to S. v. setosus seems steepest in this area. Upon the
establishment of this contact reproductive continuity was resumed,
probably because the temporal separation of the two stocks involved
was not so great as, say, that between S. v. vagrans and S. v.
pacificus, and in addition the morphological differentiation was not
so great.

On the eastern side of the Rockies the montane stock moved
northeastward, occupying suitable territory opened up by the dissolution
of the Laurentide ice sheet. Still later changes in the
character of the northern plains owing to desiccation divided the
range of the species and isolated S. v. soperi in Manitoba and central
Saskatchewan and a population of S. v. obscurus, in the Cypress
Hills. A number of semi-isolated stocks in central Montana became
differentiated as a recognizable subspecies there.

A number of other boreal mammals have geographic ranges
which resemble that of Sorex vagrans, except that the geographic
ranges of subspecies do not overlap. Because of the general similarities
[Pg 24]
of these geographic ranges, it is pertinent to examine the
reasons suggested by students to account for the present geographic
distributions of some of these other boreal species.

The red squirrel genus, Tamiasciurus, has a Rocky Mountain (and
northern coniferous forest) species, T. hudsonicus, that occurs all
along the Rocky Mountain chain and northward into Alaska. In the
Cascade Mountains of Washington and British Columbia this
species meets the range of a well marked western species, T. douglasii,
with no evidence of intergradation. Dalquest (1948:86)
attributes the divergence of the two species to separation in a
glacial age but feels that the degree of difference between the two
is too great to have all taken place during the Wisconsinan. Perhaps
he has overemphasized the importance of the differences between
the two, but, be that as it may, it seems that the two kinds differentiated
during a glacial age when they were isolated, perhaps by ice
on the Cascades into a coastal population and an inland population.
One difference between the distribution of the red squirrels and
vagrant shrew is that the squirrel of the Sierra Nevada is the species
of the Pacific Coast, whereas the vagrant shrew of the Sierra Nevada
was derived from the Great Basin population, which in turn was
derived from the Rocky Mountain kind. Red squirrels do not occur
on any of the boreal montane "islands" of Nevada. During the
pluvial periods when hydrosere-loving shrews populated the Great
Basin, that region may have been a treeless grassland. Vagrant
shrews, then as now, probably depended on hydrosere communities,
while red squirrels required trees. Therefore the shrews were able
to traverse the Great Basin, while the Sierran red squirrels were of
necessity derived from the coastal population.

The ecological requirements of jumping mice, genus Zapus, and
the subspecies of Sorex vagrans that dwell in hydroseres are essentially
similar. The species Zapus princeps lives in the Rocky
Mountains, the Great Basin, the Sierra Nevada, and north to Yukon
(Krutzsch, 1954:395). Its geographic range is similar to that of
the montane and basin segments of S. vagrans. The species Z. trinotatus
occurs along the Pacific coast and in the Cascades north to
southwestern British Columbia. Its distribution thus coincides in
general with that of the large red coastal subspecies of S. vagrans.
Krutzsch (1954:368-369) thought that these two kinds of jumping
mice were first separated by the formation of the Cascade Mountains
and the Sierra Nevada and finally by Pleistocene glaciation.
The Sierran jumping mouse (Zapus princeps), as is the Sierran
vagrant shrew, is more closely related to the jumping mouse of the
[Pg 25]
Great Basin and of the Rocky Mountains than it is to the jumping
mouse (Z. trinotatus) of the Pacific Coast, just as the Sierran vagrant
shrew is related to the shrew of the Great Basin and Rocky Mountains.
The jumping mouse also is limited in its distribution by
hydrosere communities, not by forests.

In western North America there are two species of water or marsh
shrews: Sorex palustris and S. bendiri. They have been placed in
separate subgenera, but, as pointed out beyond, are closely related
and here are placed in the same subgenus. The species palustris
is found throughout the Rocky Mountains, north into Alaska, across
the Great Basin into the Sierra Nevada, and west to the Pacific
coast in Washington. The species bendiri is found from northwestern
California north along the Pacific coast to southwestern British
Columbia and east to the Cascades. Where the ranges of the two
species overlap in western Washington they do not interbreed so far
as is known, and are somewhat different in their ecology, bendiri
being a lowland, and palustris being a montane, species. The two
species probably were separated in a glacial period as seems to have
been the case with the wandering shrews. Also, the water shrew of
the Sierra Nevada is derived from that of the Great Basin and Rocky
Mountains. Sorex palustris is tied closely in its distribution to hydrosere
communities and is not dependent upon the presence of forests.

Red-backed mice, genus Clethrionomys, occur throughout the
Rocky Mountains and west to the Cascades in Washington as the
species C. gapperi. The species C. californicus is found along the
Pacific Coast from California north to the Olympic Peninsula. Where
the ranges of the two species meet in Washington they seem not to
intergrade. In some glacial interval these two species may have
evolved in the same manner as has been described for the species of
Zapus and those of Tamiasciurus. No Clethrionomys are found in
the Sierra Nevada, nor are red-backed mice found in the boreal
islands of the Great Basin. It is not known why Clethrionomys
californicus does not occur in the Sierra Nevada. Some boreal birds
have distributional patterns similar to those of the mammalian
examples cited above. One kind of sapsucker, Sphyrapicus varius
nuchalis, occurs in the Rocky Mountains north into British Columbia
and west to the Cascades and Sierra Nevada. A related kind, S.
varius ruber, occurs along the Pacific Coast from California north
into British Columbia. Recently Howell (1952) has shown that
some intergradation takes place between ruber and nuchalis in
Washington and British Columbia, although they do not intergrade
[Pg 26]
freely. Previously the two kinds were thought not to intergrade
and were regarded as two species. The two kinds intergrade also
in northeastern California, although in that state S. v. daggeti, rather
than S. v. ruber, is involved in the intergradation. Howell considered
the two kinds to be conspecific with one another as well as with the
eastern S. varius. He attributed a measure of the distinctness of
nuchalis and ruber to their separation during a glacial period, but
felt that the separation was much older than Wisconsinan. Whatever
the time of separation, the pattern seems clear: nuchalis and
ruber (as well as varius) were separated into montane, coastal, and
eastern segments respectively, probably by glaciation (it seems to
me in the Pleistocene), and have since re-established contact with
one another.

The grouse genus Dendrogapus is divided into a Great Basin
species, D. obscurus, which extends northward into British Columbia,
and a Rocky Mountain species, D. fuliginosus, that is found
in the Sierra Nevada and northward along the coast and Cascades
into British Columbia. Although the two kinds have at times been
considered conspecific, they differ in voice, hooting mechanism,
and characters of the downy young, and so far no actual
intergradation between the two has been shown (Grinnell and
Miller, 1944:113). These grouse thus seem to offer additional evidence
for a Pleistocene, possibly Wisconsinan, separation of the
boreal fauna into a Rocky Mountain and a Pacific coastal segment.

A notable sidelight on these data is the frequency with which
species in the Sierra Nevada have their closest relatives in the Rocky
Mountains, rather than in the geographically nearer Cascades or
coastal areas. This similarity in fauna of the Sierra Nevada and the
Rockies was noted long ago by Merriam (1899:86).

During the Sangamonian interval, isolated segments of the once
widespread ancestral Sorex vagrans quite possibly persisted in such
places as the Sierra Nevada, coastal southern California, the mountains
of Arizona, New Mexico, and southern Mexico, and in the
Black Hills (see fig. 6). One might expect that by Wisconsinan
time these populations would have become reproductively isolated
from their parent stock. They would therefore have remained
specifically distinct when Wisconsinan Sorex vagrans, reoccupied
these outlying areas, and may still be found isolated in places
peripheral to the range of the ancestral species.

[Pg 27]

Fig. 10. Probable distribution of S.
veraepacis, S. longirostris, and the S. ornatus group
(stipple) and of their Wisconsinan ancestors (lines). Heavy line
indicates limits of Wisconsinan glaciation.

In fact, we do find species closely related to Sorex vagrans in just
such places today (fig. 10). Probably Sorex ornatus, including
members of the ornatus group such as S. trigonirostris, S. sinuosus,
S. willeti, S. tenellus, and S. nanus, and also S. veraepacis, arose by
separation from the ancestral vagrans stock in Sangamonian time.
Probably the eastern S. longirostris arose in a like manner. The
ancestor of S. ornatus may have been isolated in southwestern California
during Sangamonian time, spread north and south during the
Wisconsinan age, and afterward given rise to S. trigonirostris and
the modern S. ornatus complex of California and Baja California.
In at least one place reproductive isolation between ornatus and the
invading S. vagrans has broken down (Rudd, 1953); the place is a
salt marsh along San Pablo Bay, where a hybrid population between
S. vagrans and S. sinuosus, an ornatus derivative, has formed. Sorex
tenellus may have been isolated in the Sierra Nevada in the Sangamonian
[Pg 28]
interval, moved into the valleys east of the mountains during
the Wisconsinan age, and become restricted to its present range
since the retreat of the last ice. Sorex nanus may have occurred in
the Black Hills and isolated mountains of Arizona and New Mexico
during the Sangamonian interval and remained in these general
areas during the Wisconsinan age. Its present range is peripheral
to the main body of the Rockies and the Colorado Plateau.

The eastern species Sorex longirostris has many similarities with
shrews of the ornatus-vagrans stock. S. l. longirostris is close in
many ways to S. nanus. Indeed, the differences between the species
S. nanus, S. ornatus, and S. longirostris seem to me to be of the
same magnitude and indicate a similar period of differentiation from
a common ancestor. The ancestor of S. longirostris may have gained
access to the eastern United States in the Illinoian Age via the northern
Great Plains south of the glacial boundary (fig. 7). The ancestor
of Sorex veraepacis of southern Mexico probably reached that
area in Illinoian time as part of the ancestral vagrans stock and probably
attained its differentiation during the Sangamonian interval.

All the kinds of shrews so far discussed, including the S. vagrans
complex, might thus be thought of as having had a common ancestor
in the Illinoian Age. This entire group of shrews has the third unicuspid
smaller than the fourth, a pigmented ridge from the apex to
the cingulum of each upper unicuspid, and, in most individuals,
lacks a post-mandibular foramen in the lower jaw (Findley, 1953:636-637).
The pigment is not always prominent in S. longirostris.

Two other species of North American shrews, Sorex palustris, the
water shrew, and Sorex bendiri, the marsh shrew, show these three
characters to a greater or lesser degree, and it seems that these two
species and the vagrans-ornatus-veraepacis group had a common
ancestor, probably before Illinoian time for reasons stated beyond.
I judge, however, that far from being subgenerically distinct as they
have been considered to be, S. palustris and S. bendiri are actually
closely related species of the same subgenus and may have differentiated
from one another because of separation into eastern (palustris)
and western (bendiri) segments in the Sangamonian interval,
much as has been postulated concerning the eastern and western
stocks of Sorex vagrans. Indeed, Jackson (1928:192) has noted that
in the Pacific northwest the characters of the two kinds approach
one another and become differences of degree only.

The widespread species Sorex cinereus resembles all the foregoing
species in the ridges on the unicuspid teeth and in the lack of a
post-mandibular foramen, but differs from those other species in
[Pg 29]
having the third upper unicuspid larger than the fourth. The subspecies
S. cinereus ohionensis, however, often has the sizes of these teeth
reversed. With S. cinereus I include S. preblei (eastern Oregon) and
S. lyelli (Sierra Nevada), both obviously closely related to
cinereus as Jackson (1928:37) recognized when he included them in the
cinereus group. Sorex milleri (Coahuila and central western Nuevo
Leon) seems to me to resemble S. cinereus more than it does other
species of North American Sorex, and I judge that it also belongs to
the cinereus group. Sorex cinereus and its close relatives seem more
closely related to the species which have thus far been discussed than
they do to such other North American species as S. arcticus, S.
fumeus, S. trowbridgii, S. merriami, and the members of the S.
saussurei group; most of these five species last mentioned possess a
post-mandibular foramen, lack pigmented unicuspid ridges, and have the
third unicuspid larger than the fourth. Because of the morphological
resemblances mentioned above, it seems likely to me that S. cinereus
and the vagrans-ornatus-veraepacis-palustris complex had a common
ancestor in early Pleistocene time. Sorex cinereus has recently been
considered to be conspecific with the Old World S. caecutiens Laxmann
(Van den Brink, 1953) which name, being the older, would apply to the
circumpolar species.

Hibbard (1944:719) recovered S. cinereus and a species of
Neosorex (a name formerly applied to the water shrew) from the
Pleistocene (late Kansan) Cudahy Fauna. This indicates that the
ancestors of the modern S. cinereus and of the water shrew had
diverged from one another before that time. Brown (1908:172)
recorded S. cinereus and S. obscurus from the Conard Fissure in
Arkansas. These materials were deposited probably at a later
time than was the Cudahy Fauna. The S. obscurus from Conard
Fissure probably represents the ancestral S. vagrans stock which I
think reached eastern United States in Illinoian time and gave rise
to S. longirostris. The Conard Fissure material was deposited at a
time (Illinoian?) when northern faunas extended farther south than
they do today.

All of the species mentioned as having structural characters in
common with S. vagrans seem to have arisen from a common ancestor
which had already differentiated from the ancestor of such
species as S. arcticus, S. saussurei, and others. Consequently all
are here included in a single subgenus. The oldest generic name
applied to a shrew of this group, other than the name Sorex, is
Otisorex DeKay, 1842, type species Otisorex platyrhinus DeKay, a
[Pg 30]
synonym of Sorex cinereus. The subgenus can be characterized as
follows.

Third unicuspid usually smaller than fourth; upper unicuspids
usually with pigmented ridge extending from apices medially to
cingula, uninterrupted by antero-posterior groove; post-mandibular
foramen usually absent. Includes the species S. cinereus,
S. longirostris, S. vagrans, S. ornatus, S.
tenellus, S. trigonirostris, S. nanus, S. juncensis,
S. willeti, S. sinuosus, S. veraepacis, S.
palustris, S. bendiri, S. alaskanus, and S.
pribilofensis.

Figs. 11-14.	Characters of the subgenera Sorex and Otisorex.
Fig. 11.	Medial view of right ramus of Sorex (Otisorex) vagrans. × 14.
Fig. 12.	Medial view of right ramus of Sorex (Sorex) arcticus. × 14.
Fig. 13.	Anterior view of left second upper unicuspid of Sorex (Otisorex) vagrans. × 45.
Fig. 14.	Anterior view of left second upper unicuspid of Sorex (Sorex) arcticus. × 45.

[Pg 31]

Other species of Sorex now occurring in North America differ from
Otisorex in having the 3rd unicuspid usually larger than 4th, in lacking
a pigmented ridge from the apices to the cingula of the upper
unicuspids, and in usually possessing a well-developed post-mandibular
foramen. Exceptions to the last mentioned character are
S. fumeus and S. dispar. The subgenus Sorex in North America
should include only the following species: S. jacksoni, S. tundrensis,
S. arcticus, S. gaspensis, S. dispar, S. fumeus, S. trowbridgii, S.
merriami, and all the members of the Mexican S. saussurei group.

The subgenera Otisorex and Sorex probably separated in early
Pleistocene or late Pliocene. Sorex is unknown in North America
earlier than the late Pliocene (Simpson, 1945:51).

In the genus Microsorex the characters of the subgenus Otisorex
are carried to an extreme; the unicuspid ridges are prominent and
end in distinct cusplets, and the 3rd unicuspid is not merely smaller
than the 4th, but is reduced almost to the vanishing point. In addition,
the post-mandibular foramen is absent. Although it is closer
structurally to Otisorex than to Sorex, the recognition of Microsorex
as a distinct genus seems warranted.

Figure 15 is intended to represent graphically some of the relationships
discussed above. It must be re-emphasized that much of
it is purely speculative, especially as regards actual time when
various separations took place. It will be noted that I have indicated
most separations as having taken place in interglacial ages.
They are generally regarded as periods of warmth and aridity and,
therefore, probably are times of segmentation of the ranges of boreal
mammals and hence times exceptionally favorable to the process of
speciation. Glacial ages, characterized by extensive and continuous
areas of boreal habitat, probably were times of relatively unrestricted
gene flow between many populations of boreal mammals
and hence not favorable to rapid speciation.

The size of the wandering shrew varies from small in the subspecies
monticola and vagrans to large in the subspecies pacificus. The tail
makes up from a little more than a third to almost half of the total
length. The color pattern ranges from tricolored through bicolored
to almost monocolored. Color ranges from reddish (Sayal or Snuff
Brown) to grayish in summer pelage and from black to light gray
in winter. Diagnostic dental characters include: 3rd upper unicuspid
smaller than 4th, and unicuspids, except 5th, with a pigmented ridge
extending from near apex of each tooth medially to cingulum and
sometimes ending as internal cusplet. S. vagrans differs from members
[Pg 32]
of the ornatus group in less flattened skull, and in more ventrally
situated foramen magnum that encroaches more on the basioccipital
and less on the supraoccipital. The wandering shrew differs
from S. trowbridgii and S. saussurei in the dental characters
mentioned above. These dental characters also serve to distinguish
S. vagrans readily from S. cinereus, S. merriami, and S. arcticus
which may occur with vagrans. The large marsh shrew and water
shrew, S. palustris and S. bendiri, can be distinguished at a glance
from S. vagrans by larger size and darker color.

Fig. 15. Diagrammatic representation of the
probable phylogeny of Sorex vagrans and its near relatives.

In the following treatment of the 29 subspecies of Sorex vagrans,
the subspecies are arranged in geographic sequence, beginning with
the southernmost large subspecies on the California coast and proceeding
clockwise, north, east, south, and then west back to the
starting point.

Type.—Adult female, skin and skull; No. 19658, Mus. Vert. Zool.; obtained
on July 2, 1913, by Alfred C. Shelton, from Gualala, on the Sonoma County
side of the Gualala River, Sonoma Co., California.

Range.—Coastal California from Point Reyes north to Point Arena.

Diagnosis.—Size large; average and extreme measurements of 3 topotypes
are: total length, 141.7 (141-143); tail, 59 (54-63); hind foot, 17 (17-17).
Color reddish in summer, somewhat grayer in winter.

[Pg 33]

Fig. 16. Probable geographic ranges of 16
subspecies of Sorex vagrans.

Guide to subspecies 1. S. v. shumaginensis 2. S. v. obscurus 3. S. v. alascensis 4. S. v. soperi 5. S. v. isolatus

6. S. v. setosus  7. S. v. bairdi  8. S. v. permiliensis  9. S. v. yaquinae 10. S. v. pacificus 11. S. v. sonomae

12. S. v. longiquus 13. S. v. parvidens 14. S. v. monticola 15. S. v. neomexicanus 16. S. v. orizabae

[Pg 34]

Comparisons.—Differs from S. v. pacificus, with which it intergrades to the
north, in average smaller size and somewhat darker color; differs from the
sympatric S. v. vagrans in much larger size and more reddish color in both
summer and winter.

Remarks.—This subspecies inhabits the Transition Life-zone below 300 feet,
and occurs on moist ground in forests and beneath dense vegetation.

Marginal records.—California: Point Arena (Grinnell, 1933:82); Monte
Rio (Jackson, 1928:144); Inverness (Grinnell, 1933:82).

Type.—Adult, sex unknown, skin and skull; No. 3266 U. S. Nat. Mus.; date
of capture unknown; received from E. P. Vollum and catalogued on March 8,
1858; obtained at Ft. Umpqua, mouth of Umpqua River, Douglas Co., Oregon.

Range.—Coast of California and Oregon from Mendocino north to Gardiner.

Diagnosis.—Size large, largest of the species; average and extreme measurements
of 8 specimens from Orick, Humboldt Co., California, are: total length,
143.1 (134-154); tail, 65.5 (59-72); hind foot, 17.5 (16-19). Color reddish
in summer, browner or grayer in winter.

Comparisons.—See account of S. v. sonomae for comparison with that subspecies;
averaging larger in all dimensions than S. v. yaquinae with which it
intergrades to the north; much larger and has more reddish than the sympatric
S. v. vagrans.

Remarks.—This subspecies occurs in the Canadian and Transition life-zones
below 1500 ft. where there is found moist ground in or adjacent to heavy
forests.

Specimens examined.—Total number, 76.

Oregon: Douglas Co.: Umpqua, 1 BS. Coos Co.: Marshfield, 1 BS;
Myrtle Point, 1 BS. Josephine Co.: Bolan Lake, 1 SGJ.

California: Del Norte Co.: Smith River, 2 BS; Gasquet, 4 BS; Crescent
City, 17 BS. Humboldt Co.: Orick, 13 BS; 1 mi. N Trinidad, 18 FC; Trinidad
Head, 1 BS; Carson's Camp, Mad River, Humboldt Bay, 5 BS; Arcata, 3 BS;
Cape Mendocino, 2 BS; 5 mi. S Dyerville, 1 BS. Mendocino Co.: Mendocino,
6 BS.

Marginal Records.—Oregon: Marshfield; Umpqua. California: Gasquet;
5 mi. S Dyerville; Mendocino, thence up coast to point of beginning.

Type.—Adult female, skin and skull; No. 73051 U. S. Biol. Surv. Coll.,
obtained on July 18, 1895, by B. J. Bretherton, from Yaquina Bay, Lincoln Co.,
Oregon.

Diagnosis.—Size large for the species; average and extreme external measurements
[Pg 35]
of 11 specimens from Oakridge, Lane Co., Oregon, are: total length,
125.3 (11-136); tail, 55.1 (49-61); hind foot, 14.9 (14-16). Color reddish in
summer, browner or grayer in winter.

Comparisons.—See account of S. v. pacificus for comparison with that
subspecies. Larger and more reddish than S. v. bairdi with which it intergrades
to the north and east. Much larger and more reddish than the sympatric
S. v. vagrans.

Remarks.—The name yaquinae actually applies to a population of intergrades
between pacificus and bairdi. There is much variation over the range
of the subspecies, and individuals from the western and southern parts are
larger than those from the west slope of the Cascades. Specimens from Vida
and McKenzie Bridge are smaller than those from Mapleton, Mercer, and the
type locality but still seem closer to yaquinae than to topotypes of bairdi.
Between Marshfield and Umpqua on the one hand, and the Columbia River
and the Cascade Mountains on the other, the size of Sorex vagrans decreases
quite rapidly from the large pacificus to the smaller permiliensis. Size decreases
less rapidly northward along the coast than it does eastward toward the mountains;
consequently, at any given latitude, coastal shrews are larger than
mountain shrews. In this area of rapid change in size it is difficult to draw
subspecific boundaries between pacificus, yaquinae, and bairdi, and this must
be done somewhat arbitrarily.

Jackson (1928:141) remarked upon the possibility that intergradation
between pacificus and yaquinae took place. He noted also the close resemblance
between yaquinae and bairdi, and stated (loc. cit.) that specific affinity between
the two might be demonstrated with more specimens. He had a series
of eight specimens from Vida, Oregon, seven of which he assigned to S. o. bairdi
and one to yaquinae. I have examined these specimens and find no more
variation between the largest and the smallest than would be expected in any
normally variable series of shrews. Vernon Bailey (1936:364) arranged
yaquinae as a subspecies of pacificus without giving his reasons for so doing.

Specimens examined.—Total number, 65. Oregon: Lincoln Co.: type locality,
2 AW. Benton Co.: Philomath, 2 BS. Lane Co.: Mable, 1 OU; Vida,
4 BS, 1 OSC, 3 OU; McKenzie Bridge, 8 OSC, 3 AW, 17 OU, 2 SGJ; Mercer,
1 OSC, 1 OU; Mapleton, 3 BS; Oakridge, 11 OU. Douglas Co.: Gardiner,
2 BS; Elkhead, 1 BS. Klamath Co.: Crescent Lake, 3 OU.

Marginal Records.—Oregon: Yaquina Bay; Philomath; McKenzie Bridge;
Prospect (Jackson, 1928:140); Crescent Lake; Gardiner.

Type.—Adult female, skin and skull; No. 17414/24318, U. S. Biol. Surv.
Coll.; obtained on August 2, 1889, by T. S. Palmer, from Astoria, Clatsop Co.,
Oregon.

Range.—Northwestern Oregon, south to Otis and east to Portland.

Diagnosis.—Size medium for the species; average and extreme external
measurements of 6 specimens from the type locality are: total length, 126.3
(124-130); tail, 55.0 (52-57); hind foot, 15.0 (14-15). Color Fuscous to Sepia
in summer, darker in winter, underparts buffy.

[Pg 36]
Comparisons.—For comparisons with yaquinae see account of that subspecies.
More reddish and larger than permiliensis with which bairdi intergrades to the
east; specimens from Portland show evidence of such intergradation. Some
specimens from southern Tillamook County show an approach to yaquinae.

Remarks.—S. v. bairdi lives primarily in forests as do yaquinae and pacificus.

Specimens examined.—Total number, 39. Oregon: Clatsop Co.: type locality,
12 BS; Seaside, 3 BS. Tillamook Co.: Netarts, 1 OU; Tillamook, 2 OSC;
Blaine, 1 AW; Hebo Lake, 1 SGJ; 5 mi. SW Cloverdale, 1 AW. Multnomah
Co.: Portland, 6 USNM. Lincoln Co.: Otis, 7 USNM; Delake, 1 KU. Lane
Co.: north slope Three Sisters, 6000 ft., 4 BS.

Marginal Records.—Oregon: type locality; Portland; north slope Three
Sisters; Taft (Macnab and Dirks, 1941:178).

Type.—Adult male, skin and skull; No. 91048, U. S. Biol. Surv. Coll.;
obtained on October 2, 1897, by J. A. Loring from Permilia Lake, W base Mt.
Jefferson, Cascade Range, Marion Co., Oregon.

Range.—The Cascade Mountains of Oregon from Mt. Jefferson north to the
Columbia River.

Diagnosis.—Size medium for the species; average and extreme measurements
of 14 specimens from the type locality are: total length, 117.7 (110-124);
tail, 51.9 (45-58); hind foot, 14.0 (14-15). Pale reddish in summer, darker
and brownish in winter.

Comparisons.—For comparison with S. v. bairdi see account of that subspecies.
Larger than S. v. setosus except tail relatively shorter. More reddish
in summer pelage than setosus.

Remarks.—S. v. bairdi is larger in the southern part of its range than elsewhere.
Specimens from McKenzie Bridge, herein referred to yaquinae, are
intermediate in character between yaquinae and bairdi or between yaquinae
and permiliensis. The transition between yaquinae and bairdi is much more
gradual than between yaquinae and permiliensis.

Specimens examined.—Total number, 21. Oregon: Hood River Co.: Mt.
Hood, 2 BS. Wasco Co.: Camas Prairie, E base Cascade Mts., SE Mt. Hood,
1 BS. Marion Co.: Detroit, 1 BS; type locality, 17 BS.

Marginal Records.—Oregon: Mt. Hood; type locality; Detroit.

Type.—Adult male, skin and skull; No. 6213/238, Chicago Nat. Hist. Mus.;
obtained on August 18, 1898, by D. G. Elliott from Happy Lake, Olympic
Mts., Clallam Co., Washington.

Range.—Washington from the Cascades west; southwestern British Columbia
west of 120° W Longitude north to Lund.

Diagnosis.—Size medium for the species; average and extreme measurements
of 20 specimens from the Olympic Mountains, Washington, are: total length,
[Pg 37]
117.3 (107-125); tail, 49.8 (41-54); hind foot, 13.4 (12-14). Color dark in
both summer and winter.

Comparisons.—For comparison with permiliensis see account of that subspecies.
Darker, longer-tailed, and somewhat larger cranially than S. v. obscurus
with which it intergrades in southwestern British Columbia. Smaller in
all dimensions, but much the same color as S. v. longicauda with which it
intergrades along the British Columbian coast north of Lund. Larger, darker,
less reddish, and longer-tailed than the sympatric S. v. vagrans.

Remarks.—S. v. setosus lives mostly in forests. According to Dalquest
(1948:139) it is commonest at high altitudes in western Washington. In the
Hudsonian Life-zone where shrew habitat is more restricted and marginal than
it is at lower altitudes in the humid part of Washington, setosus might be expected
to compete with S. v. vagrans and to supplant it. Records of occurrence
in the Olympic Mountains suggest a degree of such separation there.

Specimens examined.—Total number, 135.

British Columbia: Lund, Malaspina Inlet, 4 BS; Gibson's Landing, 10 BS;
Port Moody, 19 BS; Langley, 2 BS; Chilliwack, 1 BS; Manning Park, 2 PMBC.

Washington: Whatcom Co.: Mt. Baker, 6 WSC; Barron, 1 BS. Chelan
Co.: Clovay Pass, 1 WSC; Stehekin, 6 (4 WSC, 2 BS); Cascade Tunnel,
1 WSC. King Co.: Scenic, 1 WSC. Kittitas Co.: Lake Kachess, 1 WSC;
Easton, 10 BS. Clallam Co.: 8 mi. W Sekin River, 1 WSC; mouth of Sekin
River, 1 WSC; Clallam Bay, 1 CMNH; 7 mi. W Port Angeles, 1 WSC; Ozette
Lake, 1 CMNH; 12 mi. S Port Angeles, 4 WSC; Forks, 1 CMNH; Deer Lake,
7 CMNH; Hoh Lake, 1 CMNH; Bogachiel Peak, 1 CMNH; Sol Duc Hot
Springs, 3 CMNH; Sol Duc Park, 1 CMNH; Canyon Creek, 1 WSC; Sol Duc
Divide, 2 WSC; Cat Creek, 2 WSC. Jefferson Co.: Jackson Ranger Station,
1 CMNH; Mt. Kimta, 2 CMNH; Reflection Lake, 6 CMNH; Blue Glacier,
3 CMNH. Gray's Harbor Co.: Westport, 1 WSC. Pierce Co.: Fort Lewis,
1 FC; Mt. Rainier, 19 (16 BS, 3 WSC). Pacific Co.: Tokeland, 2 BS. Yakima
Co.: Gotchen Creek, 3 WSC; Mt. Adams, 1 WSC. Skamania Co.: Mt. St.
Helens, 1.

Oregon: Hood River Co.: 2 mi. W Parkdale, 2 BS.

Marginal Records.—British Columbia: Rivers Inlet (Anderson, 1947:20);
Agassiz (Jackson, 1928:136); Chilliwack Lake. Washington: Barron; Lyman
Lake (Jackson, 1928:137); Mt. Stuart (Dalquest, 1948:141); Mt. Adams.
Oregon: 2 mi. W Parkdale. Washington: Ilwaco (Jackson, 1928:137);
Lund, Malaspina Inlet.

Type.—Adult male, skin and skull; No. 74711, U. S. Biol. Surv. Coll.; obtained
on September 9, 1895 by C. P. Streator, from Wrangell, Alaska.

Range.—The British Columbian and Alaskan coasts from Rivers Inlet north
to near Juneau and also certain islands including Etolin, Gravina, Revillagigedo,
Sergeif, and Wrangell.

Diagnosis.—Size medium for the species, tail relatively long; average and
extreme measurements of 17 specimens from the type locality are: total length,
128.4 (122-138); tail, 57.8 (53-66); hind foot, 15.1 (14-16). Color dark in
summer and winter.

Comparisons.—For comparison with S. v. setosus see account of that subspecies.
Larger and darker than S. v. obscurus with which it intergrades east
of the humid coastal region; larger and darker than S. v. alascensis with which
[Pg 38]
it intergrades in the Lynn Canal area; larger and darker than S. v. calvertensis
which occurs on Calvert Island and Banks Island, British Columbia; differs from
S. v. insularis of Smythe, Townsend, and Reginald islands in larger size and
blackish rather than brown winter pelage; larger and relatively longer-tailed
than S. v. elassodon which occurs on most of the islands west of the range of
longicauda; larger and relatively longer-tailed than S. v. isolatus.

Specimens examined.—Total number, 151.

Alaska: Wrangell, 54 BS; 8 AMNH; Crittenden Creek, 1 BS; Ketchikan,
2 BS; Loring, 11 BS.

British Columbia: Port Simpson, 25 BS; Inverness, 15 BS; head of Rivers
Inlet, 35 BS.

Marginal Records.—British Columbia: Great Glacier, Stikine River (Jackson,
1928:133). Alaska: Burroughs Bay (ibid.). British Columbia: Bella
Coola region (Anderson, 1947:19); head of Rivers Inlet; Spider Island (Cowan,
1941:101); Goose Island (Cowan, 1941:99); Princess Royal Island (Cowan,
1941:98); Pitt Island (ibid.); Metlakatla (Jackson, 1928:133); Port Simpson.
Alaska: Gravina Island (ibid.); Helm Bay (ibid.); Etolin Island (ibid.);
Sergeif Island, mouth of Stikine River (ibid.); Sumdum Village (ibid.); Port
Snettisham (ibid.).

Type.—Adult male, skin and skull; No. 70376, Mus. Vert. Zool.; obtained
on May 4, 1936, by R. A. Cumming, from Vanada, Texada Island, Georgia
Strait, British Columbia.

Range.—Known only from the type locality.

Diagnosis.—Size medium; average and extreme measurements of 5 specimens
from the type locality are: total length, 111 (108-117); tail, 48 (44-49);
hind foot, 12 (12-13) (Hall, 1938:463).

Comparisons.—Color much as in S. v. setosus or S. v. isolatus; palate longer
than that of isolatus or setosus; hind foot shorter than either; smaller than
S. v. longicauda.

Type.—Adult male, skin and skull; No. 177719, U. S. Biol. Surv. Coll.;
obtained on May 21, 1911, by Alexander Wetmore from mouth of Millstone
Creek, Nanaimo, Vancouver Island, British Columbia.

Range.—Vancouver Island.

Diagnosis.—Size medium; measurements of two from the type locality are:
total length, 113, 118; tail, 48, 49; hind foot, 14, 14. Dark in summer and
winter, underparts brownish.

Comparisons.—Smaller than S. v. setosus but color much the same; resembles
S. v. obscurus in size and cranial characters but darker in all pelages; similar
in color to S. v. vancouverensis with which isolatus is sympatric but with longer
tail, longer hind feet, broader rostrum and larger teeth. For comparison with
S. v. mixtus see account of that subspecies.

Remarks.—S. v. isolatus and S. v. vancouverensis seemingly approach one
[Pg 39]
another morphologically more closely than do any other pair of sympatric subspecies
of Sorex vagrans. The exceptions may be S. v. vagrans and S. v.
obscurus which are geographically sympatric in a few places although they may
be ecologically separated.

Specimens examined.—Total number, 9. British Columbia, Vancouver
Island: Nanaimo, 3 BS; Barclay Sound, 1 AMNH; Goldstream, 5 BS.

Marginal Records.—British Columbia, Vancouver Island. (Anderson,
1947:19): Cape Scott; Victoria.

Type.—Adult female, skin and skull; No. 3110, Prov. Mus. British Columbia;
obtained on August 24, 1938, by T. T. and E. B. McCabe from Smythe Island,
Bardswell Group, British Columbia.

Range.—Smythe, Townsend, and Reginald islands, British Columbia.

Diagnosis.—Size medium; average and extreme measurements of 50 specimens
from within the range of the subspecies are: total length, 122.3 (111-134);
tail 52.6 (46-58); hind foot, 14.6 (13-15) (Cowan, 1941:107).

Comparisons.—Smaller externally and cranially than S. v. longicauda and
brown instead of blackish or grayish in winter pelage. Skull broader than that
of S. v. calvertensis and color brown rather than blackish or grayish in winter
pelage.

Remarks.—S. v. insularis occurs together with S. cinereus on Townsend and
Smythe islands. S. vagrans far outnumbered the cinereus shrew (Cowan,
1941:96).

Records of occurrence.—British Columbia (Cowan, 1941:104): Smythe
Island, Townsend Island, Reginald Island.

Type.—Adult male, skin and skull; No. 1947, Prov. Mus. British Columbia;
obtained on July 14, 1937, by T. T. and E. T. McCabe from Safety Cove,
Calvert Island, British Columbia.

Diagnosis.—Size medium; average and extreme measurements of 13 specimens
from Calvert Island are: total length, 121.6 (109-129); tail, 54.0 (52-58);
hind foot, 14.7 (13-15) (Cowan, 1941:106). Blackish or grayish in winter
pelage.

Comparisons.—Smaller externally and cranially and paler in winter and
summer than S. v. longicauda; for comparisons with S. v. insularis see account
of that subspecies.

Remarks.—S. v. calvertensis seems to be the only shrew on Calvert and
Banks islands.

Records of occurrence.—British Columbia (Cowan, 1941:103): Safety
Cove, Calvert Island; Larson Harbor, Banks Island.

Marginal Records.—British Columbia: Larson Harbor, Banks Island; type
locality.

[Pg 40]

Sorex obscurus malitiosus Jackson, Proc. Biol. Soc. Washington, 32:23,
April 11, 1919.

Type.—Adult female, skin and skull; No. 8401; Mus. Vert. Zool.;
obtained on May 21, 1909, by H. S. Swarth from east side of Warren
Island, Alaska.

Range.—Warren and Coronation islands, Alaska.

Diagnosis.—Size medium; average and extreme measurements of 5
topotypes are: total length, 129.8 (126-135); tail, 56.4 (53-61); hind
foot, 15.4 (15-16). Color brownish in summer, brownish rather than
blackish in winter.

Comparisons.—Somewhat more brownish than S. v. longicauda but
resembling it in size; skull slightly more flattened and rostrum
broader. Larger than S. v. elassodon. Larger and relatively
longer-tailed than S. v. alascensis.

Records of occurrence.—Alaska
(Jackson, 1928:130): Warren Island; Coronation Island.

Type.—Adult male, skin and skull; No. 100597, U. S. Biol. Surv.
Coll.; obtained on June 13, 1900, by W. H. Osgood from Cumshewa Inlet,
near old Indian village of Clew, Moresby Island, Queen Charlotte
Islands, British Columbia.

Range.—Alaskan and British Columbian islands from Admiralty Island
south to Moresby Island.

Diagnosis.—Size medium; average and extreme measurements of 4 from
the type locality are: total length, 126, (119-131); tail, 53.5
(52-55); hind foot, 13.8 (13-14). Color dark.

[Pg 41]

Fig. 17. Probable geographic ranges
of the subspecies of Sorex vagrans on the coast of British
Columbia and southeastern Alaska.

1. Sorex vagrans malitiosus

2. Sorex vagrans elassodon

3. Sorex vagrans prevostensis

4. Sorex vagrans calvertensis

5. Sorex vagrans insularis

6. Sorex vagrans longicauda

7. Sorex vagrans obscurus

Comparisons.—Smaller with relatively smaller tail and hind feet than
S. v. longicauda, but resembling it in color. Smaller and paler than S. v. prevostensis
with relatively narrower rostrum. Larger, darker, and with relatively
longer tail than S. v. obscurus. Resembles S. v. alascensis but hind foot smaller
and skull relatively narrower. Smaller than S. v. malitiosus.

Remarks.—In the northern part of its range S. v. elassodon occurs with
Sorex cinereus. In the southern part it is the only shrew present.

Specimens examined.—Total number 93.

Alaska: near Killisnoo, Admiralty Island, 2 BS; Kupreanof Island, 15 BS;
Petersburg, Mitkof Island, 10 BS; Woewodski Island, 4 AMNH; Kasaan Bay,
Prince of Wales Island, 18 BS.

British Columbia: Cumshewa Inlet, Moresby Island, 25 BS; Massett,
Graham Island, 6 BS; Queen Charlotte Islands, 13 AMNH.

Marginal Records.—Alaska: Hawk Inlet, Admiralty Island (Jackson,
1928:131); Kupreanof Island; Mitkof Island; St. John Harbor, Zarembo Island
(Jackson, 1928:131); Kasaan Bay, Prince of Wales Island; Duke Island (Jackson,
1928:131). British Columbia: Massett, Graham Island, Queen Charlotte
Islands; type locality; Langara Island, Queen Charlotte Islands (Jackson,
1928:131). Alaska: Forrester Island (ibid.); Rocky Bay, Dall Island (ibid.);
Shakan (really on Kosciusko Island) (ibid.); Point Baker (ibid.); Kuiu Island
(ibid.); Port Conclusion, Baranof Island (ibid.).

Type.—Adult male, skin and skull; No. 100618, U. S. Biol. Surv. Coll.;
obtained on July 3, 1900, by W. H. Osgood from north end of Prevost Island
(Kunghit Island on some maps) on coast of Houston Stewart Channel, Queen
Charlotte Islands, British Columbia.

Range.—Known only from the type locality.

Diagnosis.—Size medium; measurements of two specimens from the type
locality are: total length, 132, 142; tail, 53, 59; hind foot, 14, 15. Color dark.

Comparisons.—Larger and darker than S. v. elassodon. Resembles S. v.
longicauda but darker, tail relatively somewhat shorter on the average and
rostrum relatively slightly broader.

Specimens examined.—Total number, 14. British Columbia: Prevost Island,
Queen Charlotte Group, 14 BS.

Type.—Adult female, skin and skull; No. 73539, U. S. Biol. Surv. Coll.;
obtained on July 10, 1895, by C. P. Streator from Yakutat, Alaska.

Range.—The coast of southern Alaska from the vicinity of Juneau west to
include eastern part of the Kenai Peninsula.

[Pg 42]
Diagnosis.—Size medium for the species; average and extreme measurements
of 9 specimens from 9 mi. W and 4 mi. N of Haines, Alaska, are: total
length, 110 (104-128); tail, 45.4 (41-52); hind foot, 14 (14-14). Color grayish
brown.

Comparisons.—For comparison with S. v. longicauda and S. v. elassodon
see accounts of those subspecies. Resembles S. v. obscurus in color but differs
in larger skull, longer hind foot and in somewhat darker color. Larger and
darker than S. v. shumaginensis; the two intergrade near the base of the Kenai
Peninsula.

Remarks.—This subspecies is transitional between the large, usually dark
subspecies of the southeastern Alaskan and British Columbian coast and islands,
and the smaller, paler subspecies of western and interior Alaska. There seem
to be no sharp breaks between alascensis and shumaginesis. North of Haines,
Alaska, size of shrews decreases in a short distance across a narrow intergradational
zone between alascensis and obscurus. Throughout most of its range
S. v. alascensis occurs with Sorex cinereus.

Specimens examined.—Total number, 88.

Alaska: Orca, 1 BS; Montague Island, Prince William Sound, 2 BS; Yakutat,
8 BS; north shore Yakutat Bay, 2 BS; Yakutat Bay, 1 BS; E side Chilkat River,
100 ft., 9 mi. W and 4 mi. N Haines, 12 KU; 1 mi. S Haines, 5 ft., 10 KU; 7 mi.
SSE Haines, 10 ft., 2 KU; N end Sullivan Island, 10 ft., 6 KU; SE end Sullivan
Island, 10 ft., 2 KU; Glacier Bay, 3 BS; Mendenhall River, 1 BS; Juneau, 36 BS.

British Columbia: Sheslay River, 1 AMNH; headwaters Sheslay River,
1 AMNH.

Marginal Records.—Alaska: Valdez Narrows, Prince William Sound (Jackson,
1928:128); north shore Yakutat Bay; east side Chilkat River, 100 ft., 9 mi.
W and 4 mi. N Haines. British Columbia: Sheslay River (Jackson, 1928:128).
ALASKA: Juneau; Glacier Bay; Montague Island, Prince William Sound
(ibid.); Port Nell Juan (ibid.).

Type.—Adult male, skin and skull; No. 97993, U. S. Biol. Surv. Coll.; obtained
on July 17, 1899, by De Alton Saunders from Popof Island, Shumagin
Islands, Alaska. (Measured by C. Hart Merriam and numbered 2210 in A. K.
Fisher's catalog.)

Range.—Southwestern Alaska from Seward Peninsula southeasterly to western
part of Kenai Peninsula and southwesterly to the southwestern end of the
Alaskan Peninsula.

Diagnosis.—Size medium to small for the species; average and extreme
measurements of 6 specimens from King Cove, Alaska, are: total length, 112.7
(107-118); tail, 48.3 (45-52); hind foot, 13.8 (13-14). Tending toward the
development of a tricolor pattern, the back darkest, the sides buffy, and the
venter paler.

Comparisons.—Paler and more definitely tricolored than S. v. obscurus; also
with relatively shorter palate, narrower rostrum and smaller teeth. For comparison
with S. v. alascensis see account of that subspecies.

[Pg 43]
Remarks.—S. v. shumaginensis occurs together with Sorex cinereus over
much of southwestern Alaska. Part of the range of shumaginensis falls within
the tundra of the Arctic Life-zone. This may be a partial explanation of the
tricolored pattern of the animal. Sorex tundrensis, S. cinereus ugyunak, and
S. cinereus haydeni, shrews which dwell mostly in treeless areas, are markedly
tricolored, or bicolored. Sorex arcticus, however, although tricolored, is found
in forested areas.

Specimens examined.—Total number, 340. Alaska: Sawtooth Mts., Nome
River, 2 AMNH; Nulato, 5 BS; St. Michaels, 1 BS; Bethel, 7 BS; Aniak, 1 BS;
Skeventna River, 1 BS; 6 mi. WSW Snowshoe Lake, 1 KU; 1 mi. NE Anchorage,
1 KU; Tyonek River, 48 BS; Hope, 15 BS; Hope, Mts. near, 13 BS; Moose
Camp, 3 AMNH; Kenai Peninsula, 24 AMNH; Kakwok River, 80 mi. up, 1 BS;
Kakhtul River, 5 BS; Kakwok, 3 BS; Goodnews Bay, 1 BS; Lake Aleknagik,
6 BS; Nushagak River, 25 mi. above Nushagak, 1 BS; Dillingham, 1 BS; Nushagak
Village, 15 BS; Homer, 1 AMNH; Kenai Mts., 37 AMNH; Seldovia, 24
AMNH; Barabor, 1 AMNH; Caribou Camp, 7 AMNH; Ugagik River, 3 BS;
Becharof Lake, 8 BS; Cold Bay, 14 BS; Kanatak, Portgage Bay, 4 BS; Chignik,
6 BS; Moller Bay, 1 BS; Alaska Peninsula, near Popof Island, 6 AMNH; Frosty
Peak, 15 BS; Morzhovoi Bay, 7 BS; Ungu Island, 3 BS; Sand Point, Popof
Island, 45 AMNH; Popof Island, 3 BS.

Marginal Records.—Alaska: Nome River; Nulato; Kuskokwim River, 200
mi. above Bethel, Crooked Creek (Jackson, 1928:126); 6 mi. WSW Snowshoe
Lake; Seldovia; mts. near Hope; Morhzovoi Bay; thence along coast to St.
Michael.

Type.—Adult female, skin and skull; No. 23525/30943, U. S. Biol. Surv.
Coll.; obtained on August 26, 1890, by Vernon Bailey and B. H. Dutcher from
near Timber Creek, 8200 ft., Lemhi Mts., 10 mi. SSW Junction (now Leadore),
Lemhi Co., Idaho.

Range.—Mountainous interior of western North America from central Alaska
east across Yukon and southwestern Northwest Territories to northeastern
Alberta, south in the mountains through north-central and western Washington,
Idaho, western Montana, Wyoming, Utah, and Colorado, into northern New
Mexico.

Diagnosis.—Size medium to small for the species; average and extreme measurements
of 9 topotypes are: total length, 110.3 (105-117); tail, 46.4 (42-50);
hind foot, 13.1 (12.5-13.5). Color grayish or brownish gray in summer, light
grayish in winter.

Comparisons.—For comparisons with S. v. setosus, S. v. longicauda, S. v.
alascensis and S. v. shumaginensis see accounts of those subspecies. Paler and
slightly larger than S. v. soperi. Larger than the subspecies from central Montana
herein described as new. Smaller than S. v. neomexicanus. Averaging
larger in all dimensions than S. v. monticola with which obscurus intergrades in
northern New Mexico and northern Arizona. Larger than S. v. vagrans with
more grayish rather than reddish fresh summer pelage and light gray rather
than dark grayish-black fresh winter pelage.

[Pg 44]
Remarks.—Intergradation of S. v. obscurus with S. v. setosus, S. v.
longicauda, S. v. alascensis, and the new subspecies from Montana
takes place in the usual way with specimens from intermediate localities
being intermediate in size and color. However the relationship
of S. v. obscurus and S. v. vagrans (as the latter subspecies is
defined in this study) is rather complicated. In southern British
Columbia where the two subspecies come together a situation of
remarkable complexity prevails. Series from some localities seem
to represent intergrades between obscurus and vagrans; from other
localities some specimens seem to be referable to one and some to
the other subspecies; from other localities all specimens seem referable
to one subspecies. A similar situation is seen in specimens from
northeastern Washington, northern and central Idaho, and extreme
western Montana. The region mentioned is one of extensive interfingering
of life-zones. In southern British Columbia the main axes
of the rivers, valleys and mountain ranges are north and south.
Most of the valleys are in the Transition Life-zone; the forests are
rather dry and of pine with more or less isolated hydrosere communities
about streams and ponds. These hydrosere situations are
the habitat of Sorex vagrans. Shrews from these situations are
usually referable to vagrans. The high ridges and mountain ranges
are usually in the Canadian Life-zone or higher and most of the
shrews referable to obscurus come from such places. Marginal
localities with regard to life-zone produce most of the populations
which seem to represent intergrades between the two subspecies.
Isolated areas of Canadian Life-zone, even though surrounded
with Transition Life-zone, often harbor a population of obscurus,
whereas the streams in the nearby dry valleys harbor populations
of vagrans. Farther south in the Rocky Mountain chain, obscurus
seemingly intergrades regularly with vagrans. This intergradation
is seen in populations from several localities in Utah. There the
lower elevations west of the Wasatch and Uinta mountains are
inhabited by S. v. vagrans, the higher elevations by obscurus and
where the ranges of the two abut intergrading populations occur.
In these series of intergrades there are specimens which, using size
as a subspecific criterion, would unhesitatingly be assigned, as individuals,
to obscurus, and others would be assigned to vagrans, but
these individuals represent extremes of a normally variable population.
At Cuddy Mountain, Idaho, the two subspecies seemingly
abut without intergradation; anyhow the available specimens from
this locality are referable to one or the other subspecies and none
[Pg 45]
is intermediate. The situation just described understandably has
been the source of much anguish to students who sought to identify
shrews from the Rocky Mountains. The reason for the relationship
just described has been discussed at length in a previous section.

In the Rocky Mountains of Wyoming and Colorado the subspecies
S. v. obscurus ranges almost uninterruptedly over relatively large
areas, but southward in New Mexico and southwestward into Utah
and Arizona, suitable boreal habitat becomes insular in nature and
obscurus there is confined to the higher mountains. With one exception,
once the shrew populations become 'insular' in this region
they become smaller and show intergradation with Sorex vagrans
monticola. The exception is the population in the Sacramento
Mountains of southeastern New Mexico which is larger than obscurus
and has been rightly recognized as a distinct subspecies,
neomexicanus.

Almost without exception the range of typical Sorex vagrans
obscurus is sympatric with that of Sorex cinereus, usually the subspecies
S. c. cinereus. So close is this correspondence that the presence
of S. cinereus comes near to being a useful aid in identifying
S. v. obscurus. In areas where individuals of obscurus show intergradation
with vagrans, Sorex cinereus is absent or rare. The implication
is that as the species S. vagrans approaches the size of the
species S. cinereus, competition between the two increases with
resultant displacement of cinereus.

Specimens examined.—Total number, 982.

Alaska: Wahoo Lake, 69° 08' N, 146° 58' W, 2350 ft., 2 KU; Chandler Lake,
68° 12' N, 152° 45' W, 2900 ft., 1 KU; Bettles, 1 KU, 5 BS; Alatna, 1 BS; Yukon
River, 20 mi. above Circle, 1 BS; Tanana, 1 BS; Mountains near Eagle, 18 BS;
Richardson, 8 BS; head of Toklat River, 11 BS; Savage River, 8 BS.

Yukon: MacMillan Pass, Mile 282, Canol Road, 1 NMC; MacMillan River,
Mile 249, Canol Road, 1 NMC; S. fork MacMillan River, Mile 249, Canol Road,
2 NMC; Sheldon Lake, Mile 222, Canol Road, 5 NMC; Rose River, Mile 95,
Canol Road, 1 NMC; McIntyre Creek, 3 mi. NW Whitehorse, 2250 ft. 1 KU;
Nisutlin River, Mile 40, Canol Road, 6 NMC; SW end Dezadeash Lake, 2 KU;
3 mi. E and 1½ mi. S Dalton Post, 2500 ft., 1 KU.

Mackenzie: Nahanni River Mtns., Mackenzie River, 1 BS; Fort Simpson,
3 BS; Fort Resolution, Mission Island, 1 BS.

British Columbia: W. side Mt. Glave, 4000 ft., 14 mi. S and 2 mi. E
Kelsall Lake, 1 KU; Stonehouse Creek, 5½ mi. W jct. Stonehouse Creek and
Kelsall River, 4 KU; Bennett City, 6 BS; Wilson Creek, Atlin, 1 PMBC; McDame
Post, Dease River, 6 BS; McDame Creek, 3 BS; Hot Springs, 3 mi. WNW
jct. Trout River and Liard River, 1 KU; NW side Muncho Lake, 1 KU; Little
Tahtlan River, 1 AMNH; Junction (4 mi. N Telegraph Creek), 7 BS; Raspberry
Creek, 16 AMNH; Klappan River Valley, 1 BS; Chapa-atan River, 4 BS;
Fort Grahame, 3 BS; Kispiox Valley, 23 mi. N Hazleton, 1 BS; Bear Lake,
site of Fort Connully, 2 BS; Tetana Lake, 1 PMBC; Hudson Hope, 2 BS;
Charlie Lake, 3 PMBC; Babine Mts., 6 mi. N Babine Trail, 5200 ft., 1 BS;
Big Salmon River (S branch near Canyon), 1 BS; Ootsa Lake, 2 PMBC;
Indianpoint Lake, 4 PMBC; Barkerville, 7 BS; Yellowhead Lake, 2 NMC, 1
[Pg 46]
PMBC; N. fork Moose River, 1 BS; Moose Lake, 2 BS; Moose Pass, 1 BS;
Glacier, 7 AMNH, 12 BS; Golden, 1 BS; Field, 2 BS; Caribou Lake, near
Kamloops, 2 BS; Sicamous, 1 BS; Monashee Pass, 4 PMBC; Paradise Mine,
3 PMBC; Level Mtn., 4 AMNH; 6 mi. S Nelson, 6 BS; Morrissey, 5 NMC;
Wall Lake, 1 BS.

Alberta: Hays Camp, Slave River, Wood Buffalo Park, 1 NMC; Kinuso,
Assineau River, 1920 ft., 2 KU; Athabaska River, 30 mi. above Athabaska
Landing, 8 BS; Smokey Valley, 50 mi. N Jasper House, 1 BS; Sulfur Prairie,
Grande Cache River, 3 BS; Stoney River, 35 mi. N Jasper House, 1 BS; Moose
Mtn., 1 NMC; Rodent Valley, 25 mi. W Henry House, 1 BS; Henry House,
3 BS; Jasper, 2 NMC; Shovel Pass, 4 NMC; mouth of Cavell Creek, Jasper
Park, 1 NMC; 11 mi. S Henry House, 2 BS; 15 mi. S Henry House, 1 BS;
Red Deer River, 1 AMNH; 27 mi. W Banff, 3 NMC; 12 mi. WNW Banff,
4500 ft., 1 NMC; N. Fork Saskatchewan River, 5000 ft., 1 NMC; Cypress
Hills, 1 NMC; Waterton Lakes Park, 53 NMC.

Saskatchewan: Cypress Hills, 21 NMC.

Washington: Okanogan Co.: Pasayten River, 1 BS; Bauerman Ridge, 1 BS;
Conconully, 2 BS. Pend Oreille Co.: 2 mi. N Gypsy Meadows, 2 WSC; Round
Top Mtn., 1 WSC; head Pass Creek, 1. Chelan Co.: Stehekin, 4 BS; head
Lake Chelan, 4 BS; Wenatchee, 1 BS. Kittitas Co.: Easton, 10 BS.

Idaho: Boundary Co.: Cabinet Mtns., E Priest Lake, 2 BS. Adams Co.:
½ mi. E Black Lake, 1 KU; 1 mi. N Bear Creek R. S., SW slope Smith Mtn.,
2 KU. Washington Co.: 1 mi. NE Heath, SW slope Cuddy Mtn., 4000 ft.,
4 KU. Lemhi Co.: 10 mi. SSW Leadore (type locality), 4 BS; 5 FC. Fremont
Co.: 7 mi. W West Yellowstone, 4 KU. Custer Co.: head Pahsimeroi River,
Pahsimeroi Mtns., 1 BS. Blaine Co.: Perkins Lake, 1 KU. Bear Lake Co.:—Caribou
Co. line: Preuss Mts., 1 BS.

Montana: Glacier Co.: Sherburne Lake, 3 UM; 2½ mi. W and 1½ mi. S
Babb, 1 KU; St. Mary's, 6 UM; St. Mary Lakes, 9 BS; Fish Creek, 2 BS; Gunsight
Lake, 2 BS. Flathead Co.: Nyack, 3 UM, 1 BS; 1 mi. W and 2 mi.
S Summit, 1 KU. Ravalli Co.: 8 mi. NE Stevensville, 3 BS; Sula, 1 BS.
Meagher Co.: Big Belt Mtns., Camas Creek, 4 mi. S Fort Logan, 7 BS. Gallatin
Co.: West Gallatin River, 4 BS. Park Co.: Emmigrant Gulch, 3 mi. SE Chico,
2 BS; Beartooth Mtns., 2 BS; Carbon Co.: Pryor Mtns., 2 BS.

Wyoming: Yellowstone Nat'l Park: Mammoth Hot Springs, 11 BS; Tower
Falls, 1 BS; Astringent Creek, 1 BS; Flat Mtn., 1 BS; Yellowstone Park, 1 UM.
Park Co.: Beartooth Lake, 15 BS; SW slope Whirlwind Peak, 1 KU; Pahaska
Tepee, 6300 ft., 8 BS; Pahaska, mouth Grinnell Creek, 15 BS; Pahaska, Grinnell
Creek, 7000-7500 ft., 18 BS; 25 mi. S and 28 mi. W Cody, 1 KU; Valley,
Absaroka Mts., 14 BS; Needle Mtn., 2 BS. Big Horn Co.: 28 mi. E Lovell,
9000 ft., 12 KU; head Trapper's Creek, W slope Bighorn Mtns., 7 BS; 17½ mi.
E and 4½ mi. S Shell, 1 KU. Teton Co.: Two Ocean Lake, 6 FC; Emma
Matilda Lake, 2 BS; 1 mi. N Moran, 1 FC; 2½ mi. E and ¼ mi. N Moran, 6230
ft., 2 KU; Moran, 7 FC, 1 KU; 2½ mi. E Moran, 6220 ft., 1 KU; 1 mi. S Moran,
1 FC; 3¾ mi. E and 1 mi. S Moran, 9 KU; 7 mi. S Moran, 3 FC; Timbered
Island, 14 mi. N Moose, 6750 ft., 3 KU; Bar BC Ranch, 2½ mi. NE Moose,
6500 ft., 1 KU; Beaver Dick Lake, 1 UM; Teton Mtns., Moose Creek, 6800 ft.,
9 BS; Teton Mtns., S Moose Creek, 10,000 ft., 3 BS; Teton Pass, above Fish
Creek, 7200 ft., 15 BS; Whetstone Creek, 4 UM; Flat Creek-Gravel Creek
Divide, 2 UM; Flat Creek-Granite Creek Divide, 1 UM; Jackson, 3 KU, 2 UM.
Fremont Co.: Togwotee Pass, 5 FC; Jackey's Creek, 3 mi. S Dubois, 1 BS;
Milford, 5400 ft., 2 KU; Mosquito Park R. S. 17½ mi. W and 2½ mi. N Lander,
1 KU; 17 mi. S and 6½ mi. W Lander, 9300 ft., 1 KU; Mocassin Lake, 19 mi.
W and 4 mi. N Lander, 10,000 ft., 1 KU; 23½ mi. S and 5 mi. W Lander,
8600 ft., 1 KU; Green Mts., 8 mi. E Rongis, 8000 ft., 4 BS. Washakie Co.:
9 mi. E and 5 mi. N Tensleep, 7400 ft., 2 KU; 9 mi. E and 4 mi. N Tensleep,
7000 ft., 2 KU. Lincoln Co.: Salt River Mtns., 10 mi. SE Afton, 5 BS; Labarge
Creek, 9000 ft., 1 BS. Sublette Co.: 31 mi. N Pinedale, 8025 ft., 3 KU; Surveyor's
Park, 12 mi. NE Pinedale, 8000 ft., 2 BS; N. side Half Moon Lake,
7900 ft., 1 KU; 2½ mi. NE Pinedale, 7500 ft., 2 KU. Natrona Co.: Rattlesnake
Mts., 7000-7500 ft., 18 BS; Casper Mts., 7 mi. S Casper, 6 BS. Converse Co.:
21½ mi. S and 24½ mi. W Douglas, 7600 ft., 7 KU; 22 mi. S and 24½ mi. W
[Pg 47]
Douglas, 7600 ft., 4 KU; 22½ mi. S and 24½ mi. W Douglas, 7600 ft., 2 KU.
Uinta Co.: 1 mi. N Fort Bridger, 6650 ft., 1 KU; Fort Bridger, 3 KU; Evanston,
1 BS; 9 mi. S Robertson, 8000 ft., 6 KU; 9 mi. S and 2½ mi. E Robertson, 8600
ft., 1 KU; 10 mi. S and 1 mi. W Robertson, 8700 ft., 3 KU; 10½ mi. S and
2 mi. E Robertson, 8900 ft., 1 KU; 13 mi. S and 1 mi. E Robertson, 9000 ft.,
1 KU; 13 mi. S and 2 mi E Robertson, 9200 ft., 1 KU. Carbon Co.: Ferris
Mts., 7800 to 8500 ft., 13 BS; Shirley Mts., 7600 ft., 7 BS; Bridget's Pass, 18 mi.
SW Rawlins, 7500 ft., 2 KU; 10 mi. N and 12 mi. E Encampment, 7200 ft.,
1 KU; 10 mi. N and 14 mi. E Encampment, 8000 ft., 6 KU; 9½ mi. N and 11½
mi. E Encampment, 7200 ft., 2 KU; 9 mi. N and 3 mi. E Encampment, 6500 ft.,
1 KU; 9 mi. N and 8 mi. E Encampment, 7000 ft., 1 KU; 8 mi. N and 14 mi.
E Encampment, 8400 ft., 3 KU; 8 mi. N and 14½ mi. E Encampment, 8100
ft., 2 KU; 8 mi. N and 16 mi. E Encampment, 4 KU; 8 mi. N and 21½ mi. E
Encampment, 9400 ft., 2 KU; S. base Bridger's Peak, 8800 ft., Sierra Madre
Mts., 3 BS; 8 mi. N and 19½ mi. E Savery, 8800 ft., 2 KU; 7 mi. N and 17 mi.
E Savery, 8300 ft., 1 KU; 6½ mi. N and 16 mi. E Savery, 8300 ft., 1 KU; 6 mi.
N and 15 mi. E Savery, 8500 ft., 1 KU; 5 mi. N and 10½ mi. E Savery, 8000
ft., 2 KU; 14 mi. E and 6 mi. S Saratoga, 8800 ft., 1 KU. Albany Co.: Springhill,
12 mi. N Laramie Peak, 6300 ft., 10 BS; Laramie Peak, N. slope, 8000 to
8800 ft., 7 BS; Bear Creek, 3 mi. SW Laramie Peak, 7500 ft., 6 BS; 2½ mi.
ESE Brown's Peak, 10,500 ft., 2 KU; 3 mi. ESE Brown's Peak, 10,000 ft., 1 KU;
27 mi. N and 5 mi. E Laramie, 6960 ft., 2 KU; 1 mi. SSE Pole Mtn., 8350 ft.,
3 KU; 2 mi. SW Pole Mtn., 3 KU; 3 mi. S Pole Mtn., 8100 ft., 2 KU; 8¾, mi. E
and 6½ mi. S Laramie, 8200 ft., 2 KU; Woods P. O., 1 BS. Laramie Co.: 5 mi. W
and 1 mi. N Horse Creek P. O., 7200 ft., 2 KU.

Utah: Weber Co.: Mt. Willard, Weber-Box Elder Co. line, 9768 ft., 2 UU.
Salt Lake Co.: Butterfield Canyon, 7000 ft., 1 UU; Brighton, Silver Lake P. O.,
8700 ft., 2 UU; Brighton, Silver Lake P. O., 8750 ft., 8 UU; Brighton, Silver
Lake P. O., 9000 ft., 2 UU; Brighton, Silver Lake P. O., 9500 ft., 1 UU.
Summit Co.: Jct. Bear River and East Fork, 2 CM; Smith and Morehouse
Canyon, 7000 ft., 1 UU; Mirror Lake, 10,000 ft., 1 UU. Daggett Co.: Jct. Deep
and Carter creeks, 7900 ft., 1 UU. Utah Co.: Nebo Mtn., 1 mi. E Payson
Lake, 8300 ft., 1 UU; Nebo Mts., 12 mi. SE Payson Lake, 1 UU. Wasatch Co.:
Current Creek, Uinta Mts., 1 BS; Wasatch Mts., 1 BS. Uintah Co.: Paradise
Park, 21 mi. W and 15 mi. N Vernal, Uinta Mts., 10,050 ft., 2 CM, 3 KU;
Paradise Park, Uinta Mts., 10,100 ft., 6 UU. Sanpete Co.: Manti, 3 BS. Sevier
Co.: 7 mi. Creek, 20 mi. SE Salina, 5 CM; Fish Lake Plateau, 2 BS. Emery
Co.: Lake Creek, 11 mi. E Mt. Pleasant, 4 CM. Grand Co.: Warner R. S., La
Sal Mts., 9750 ft., 2 UU; La Sal Mts., 11,000 ft., 1 BS. Beaver Co.: Puffer
Lake, Beaver Mts., 2 BS. Wayne Co.: Elkhorn G. S., Fish Lake Plateau, 14
mi. N Torrey, 9400 ft., 3 UU. Garfield Co.: Wildcat R. S., Boulder Mtn.,
8700 ft., 6 UU; 18 mi. N Escalante, 9500 ft., 1 UU. Washington Co.: Pine
Valley Mts., 7 BS. San Juan Co.: Geyer Pass, 18 mi. SSE Moab, 3 CM; Cooley,
8 mi. W Monticello, 3 CM.

Colorado: Larimer Co.: Poudre River, 1 KU. Rio Blanco Co.: 9½ mi. SW
Pagoda Peak, 2 KU. Grand Co.: Arapaho Pass, Rabbit Ears Mts., 2 BS.
Boulder Co.: Willow Park, Rocky Mtn. Nat'l Park, 8 UM; Longs Peak, 1 BS;
¾ mi. N and 2 mi. W Allenspark, 8400 ft., 5 KU; Ward, 9500 ft., 1 BS; Buchanan
Pass, 1 BS; 3 mi. S Ward, 1 KU; 7 mi. NW Nederland, 1 KU; 5 mi.
W Boulder, 3 BS; Boulder, 3 BS, 1 ChM; Nederland, 6 BS, 4 ChM; Eldora,
1 BS. Garfield Co.: Baxter Pass, 8500 ft., 2 BS. Eagle Co.: Gores Range,
1 BS. Gilpen Co.: Black Hawk, 1 BS. Lake Co.: 3 mi. W Twin Lakes, 2 KU;
12 mi. S and 1 mi. W Leadville, 1 KU. Gunnison Co.: 2 mi. W Gothic, 2 FC;
Copper Lake, 2 FC; Gothic, 1 FC. Chaffee Co.: St. Elmo, 10,100 ft., 2 BS;
E side Monarch Pass, 7 mi. W Salida, 2 ChM. Teller Co.: Glen Core, Pikes
Peak, 2 UM. El Paso Co.: Hunters Creek, a tributary of Bear Creek, 7250-7400
ft., 1 AMNH. Montrose Co.: Uncomphagre Plateau, 8500 ft., 3 BS.
Saguache Co.: 3 mi. N and 16 mi. W Saguache, 8500 ft., 2 KU; Cochetopa
Pass, 10,000 ft., 4 KU; Monshower Meadows, 27 mi. W Saguache, 2 BS. San
Juan Co.: Silverton, 4 BS. Mineral Co.: 23 mi. S and 11 mi. E Creede, 1 KU.
Costilla Co.: Fort Garland, 2 BS. Huerfano Co.: 5 mi. S and 1 mi. W Cuchara
Camps, 8 KU.

[Pg 48]

New Mexico: Taos Co.: 3 mi. N Red River, 2 BS; Taos, 1 BS. Colfax Co.:
1 mi. S and 2 mi. E Eagle Nest, 8100 ft., 2 KU. Sandoval Co.: Jemez Mts.,
3 BS. Santa Fe Co.: Hyde Park, 5 mi. NE Santa Fe, 2 HC; Santa Fe Field
Station, 1 HC; Santa Fe Ski Basin, 1 KU; Pecos Baldy, 4 BS. Torrance Co.:
Manzano Mts., 2 BS.

Marginal Records.—Alaska: Chandler Lake, 68° 12' N, 152° 45' W; Yukon
River, 20 mi. above Circle; Mountains near Eagle. Mackenzie: Nahanni
River Mts.; Fort Simpson; Fort Resolution, Mission Island. Alberta: Wood
Buffalo Park; Athabaska River, 30 mi. above Athabaska Landing. Saskatchewan:
Cypress Hills. Montana: St. Mary; 4 mi. S Fort Logan; Pryor Mts.
Wyoming: 1 mi. W and 1 mi. S Buffalo, 27424 KU; Springhill, 12 mi. N
Laramie Peak; 5 mi. W and 1 mi. N Horse Creek PO. Colorado: Boulder;
Hunters Creek; 5 mi. S and 1 mi. W Cuchara Camps. New Mexico: 3 mi. N
Red River, 10,700 ft.; Pecos Baldy; Manzano Mts.; Jemez Mts. Colorado:
Navajo River (Jackson, 1928:120); Silverton. Utah: La Sal Mts., 11,000 ft.
Colorado: Baxter Pass. Utah: junction Trout and Ashley Creeks, 9700 ft.
(Durrant, 1952:35); Mirror Lake, 10,000 ft.; Mt. Baldy R. S. (Durrant, 1952:53);
Wildcat R. S.; Pine Valley Mts.; Puffer Lake; Butterfield Canyon. Idaho:
Preuss Mts.; 4 mi. S Trude (Davis, 1939:104); head Pahsimeroi River, Pahsimeroi
Mts.; Perkins Lake; 1 mi. NE Heath; ½ mi. E Black Lake. Montana:
Sula; 8 mi. NE Stevensville. Washington: head Pass Creek; Conconully;
Wenatchee; Easton; Stehekin; Pasayten River. British Columbia: Second
Summit, Skagit River, 5000 ft., (Jackson, 1928:120); Babine Mts., 6 mi. N
Babine Trail, 5200 ft.; Hazleton (Jackson, 1928:120); 23 mi. N Hazleton;
Flood Glacier, Stikine River (Jackson, 1928:120); Cheonee Mts. (ibid.); Level
Mtn.; west side Mt. Glave, 4000 ft., 14 mi. S and 2 mi. E Kelsall Lake.
Alaska: head Toklat River; Tanana; Alatna; Bettles.

Type.—Adult male, skin and skull; No. 18249, Nat. Mus. Canada; obtained
on September 21, 1940, by J. Dewey Soper, from 2½ mi. NW Lake Audy,
Riding Mtn. Nat'l Park, Manitoba.

Range.—Southwestern Manitoba to central Saskatchewan.

Diagnosis.—Size medium to small for the species; measurements of type and
two topotypes are: total length, 107, 108, 117; tail, 45, 45, 45; hind foot, 12.1,
12.3, 12.5. Color dark brownish or fuscous in summer pelage; winter pelage
unknown.

Comparison.—Resembles S. v. obscurus in size; darker than obscurus in
summer pelage; cranium slightly higher and top more nearly flat; larger and
darker in summer pelage than the new subspecies from central Montana.

Remarks.—In their description of this subspecies Anderson and Rand
pointed out that specimens from the type locality and from central Saskatchewan
represent the dark extreme in a color cline which begins in south-central British
Columbia with "pale, brownish-tinged animals." These authors referred shrews
from Cypress Hills, southwestern Saskatchewan and southeastern Alberta to
S. o. soperi, although they noted that these specimens, taken by themselves, are
not strikingly different from S. o. obscurus from the Rocky Mountains. The
specimens from the Cypress Hills were included in soperi because the authors
felt that the subspecific boundary should be drawn "where specimens average
about half way between the extremes (of the cline) in characters."

It is true, as Anderson and Rand say, that the shrews from Cypress Hills
are hardly separable from those from, say, Waterton Lakes Park. The specimens
[Pg 49]
from the Cypress Hills are noticeably different from specimens from the
Okanagan area, but some of the latter, in my opinion may represent intergrades
between S. v. obscurus and the more reddish S. v. vagrans and are not, at any
rate, typical obscurus. In view of the similarity of shrews from Cypress Hills
to typical S. v. obscurus and since the Cypress Hills are much nearer to the
range of S. v. obscurus than to the record-stations of occurrence in central
Saskatchewan and Manitoba, I have chosen to restrict the name soperi to shrews
from these latter two localities. Seemingly S. vagrans is absent from the plains
separating the Cypress Hills from the Rocky Mountains and from Riding
Mountain National Park.

Specimens examined.—none.

Marginal records.—Saskatchewan: Prince Albert National Park, 1700 ft.
(Anderson and Rand, 1945:48). Manitoba: Riding Mountain National Park,
2½ mi. NW Audy Lake (ibid.).

Type.—First year male, skin and skull; No. 87332, Univ. Michigan Mus.
Zool.; obtained on July 21, 1942, by Emmet T. Hooper from 25 mi. ESE Big
Sandy, Eagle Creek, Chouteau Co., Montana, original no. 2184.

Range.—Central Montana; marginal localities are: Bearpaw Mts., Zortman,
Big Snowy Mts., Buffalo, Little Belt Mts.

Diagnosis.—Size small for the species; measurements of three topotypes are:
total length, 101, 105, 108; tail, 39, 40, 42; hind foot, 11.5, 11.5, 12. Color
pale; summer pelage: back near (17´´´k) Olive Brown but hairs of dorsum with
a pale, buffy band proximal from the tips which imparts a pale over-all appearance;
flanks near Wood Brown; underparts Pale Smoke Gray, usually not with
a buffy wash; color of underparts often extending along margin of upper lip.
Skull small for species; rostrum relatively broad and heavy; relatively broad
interorbitally.

Comparisons.—From S. v. obscurus, S. v. longiquus differs as follows: size
smaller; skull smaller in all dimensions although similar in proportion. From
S. v. soperi, S. v. longiquus differs in: size smaller; color paler in summer
pelage. From S. v. vagrans, S. v. longiquus differs in: color paler in summer
pelage, less brownish; color of venter extending higher on flanks; venter Pale
Smoke Gray, rarely tinged with buffy rather than usually tinged with buffy.
From S. v. monticola, S. v. longiquus differs in: summer pelage slightly paler,
venter Pale Smoke Gray rather than suffused with buffy.

Remarks.—The subspecies longiquus is obviously derived from the neighboring
S. v. obscurus and differs from it mainly in size. Some specimens of
obscurus from western Montana show evidences of intergradation with S. v.
vagrans in possessing a somewhat buffy belly and these are thus more strikingly
different from longiquus than are other specimens of obscurus. Many specimens
of obscurus from the eastern slope of the Lewis and Clark Range in Montana
show the tricolored pattern seen in many specimens of longiquus. The
smallest individuals of longiquus are found on the Big Snowy Mountains. Intergradation
with obscurus is seen in specimens here referred to S. v. obscurus
from the Big Belt Mountains.

Specimens examined.—Total number, 45. Montana: Hill Co.: Bearpaw
Mts., 5 UM, 2 BS. Phillips Co.: Zortman, 1 BS. Chouteau Co.: type locality,
[Pg 50]
3 UM; Highwood Mts., 13 BS. Cascade Co.: Neihart, Little Belt Mts., 1 BS.
Judith Basin Co.: 3 mi. W Geyser, 4100 ft., 1 KU; Otter Creek, 10 mi. SW
Geyser, 1 BS; Dry Wolf Creek, 20 mi. SW Stanford, 1 BS. Buffalo, 13 mi. W
Buffalo Canyon, 2 BS. Fergus Co.: Moccasin Mts., 15 mi. NW Hilger, 3 BS;
Judith Mts., 17 mi. NE Lewiston, 1 BS; 15 mi. S Heath, N. fork Flat Willow
Creek, Big Snowy Mts., 1 BS; Timber Creek, Big Snowy Mts., 1 BS; Crystal
Lake, 6000 ft., Big Snowy Mts., 2 UM; Rocky Creek, 5600 ft., Big Snowy Mts.,
3 UM; Big Snowy Mts., 3 BS. Meagher Co.: Sheep Creek, 16 mi. N White
Sulphur Springs, Little Belt Mts., 1 BS.

Marginal records.—Montana: Bearpaw Mts.; Zortman; Big Snowy Mts.;
16 mi. N White Sulphur Springs; Highwood Mts.

Type.—Adult male, skin and skull; No. 100440, U. S. Biol. Surv. Coll.;
obtained on May 29, 1900, by Vernon Bailey, from Cloudcroft, 9000 ft., Otero
Co., New Mexico.

Range.—Sacramento and Capitan Mountains of New Mexico.

Diagnosis.—Size medium for the species; average and extreme measurements
of 4 topotypes are: total length, 105.2 (103-107); tail, 41.0 (39-42); hind
foot, 13.1 (12.5-14). Color near Olive Brown in summer; winter pelage
unknown; skull large and relatively broad; teeth relatively large.

Comparisons.—Skull larger than that of S. v. obscurus and relatively somewhat
broader; much larger in all cranial dimensions than S. v. monticola.

Remarks.—S. v. neomexicanus is a well-marked subspecies seemingly limited
to the mountains of southeastern New Mexico. It is the only species of Sorex
thus far recorded from that area.

Specimens examined.—Total number, 12. New Mexico: Otero Co.: SW
slope Capitan Mts., 2 BS; 10 mi. NE Cloudcroft, 2 BS; Cloudcroft, 7 BS, 1 UM.

Marginal records.—New Mexico: SW slope Capitan Mts.; 10 mi. NE
Cloudcroft; type locality.

Type.—Adult male, skin and skull; No. 17599/24535, U. S. Biol. Surv. Coll.;
obtained on August 28, 1899, by C. Hart Merriam and Vernon Bailey from
San Francisco Mtn., 11,500 ft., Coconino Co., Arizona.

Range.—Mountains of western New Mexico, eastern Arizona, and the northern
Sierra Madre Occidental of Mexico.

Diagnosis.—Size small for the species; average and extreme measurements
of 12 specimens from the White Mountains, Arizona, are: total length, 104.3
(98-112); tail, 41.2 (37-45); hind foot, 12.0 (11-13). Summer pelage between
(15´m) Proutts Brown and (15´´m) Bister, venter tinged with (15´f) Pale
Ochraceous Buff; winter pelage near (17´´´k) Olive Brown; skull relatively
broad.

[Pg 51]
Comparisons.—For comparisons with S. v. obscurus and S. v. neomexicanus
see accounts of those subspecies. Skull slightly larger and relatively broader
than that of S. v. orizabae, and color slightly paler. Differs from S. v. vagrans
in: winter pelage grayish (near 17´´´k Olive Brown) rather than blackish
(17´´´´k or 17´´´´m Chaetura Drab or Chaetura Black); summer pelage slightly
grayer; skull relatively slightly broader rostrally and interorbitally.

Remarks.—S. v. monticola intergrades gradually with S. v. obscurus to the
north and east; indeed the type locality is actually in this area of intergradation.
So far as I know, monticola is not in reproductive continuity with any
other subspecies of Sorex vagrans. Specimens from southeastern Arizona are
the smallest and seem to be the most "typical" in the sense that they are most
different from S. v. obscurus. Some specimens from the whole length of the
Rocky Mountain chain in the United States have for years been referred to
monticola. Some of these, as I have pointed out, belong to S. v. longiquus, and
others are intergrades between S. v. obscurus and S. v. vagrans. Since vagrans
and monticola resemble one another somewhat, and since topotypes of S. v.
monticola actually show the influence of intergradation with obscurus, it is
easy to understand how intergrades between obscurus and vagrans could have
been assigned to monticola.

Throughout most of its range, S. v. monticola is the only Sorex present. In
some places monticola may occur with S. nanus or S. merriami. S. v. monticola
occurs with the water shrew in southeastern Arizona.

Specimens examined.—Total number, 80.

Arizona: Coconino Co.: San Francisco Mtn., 3 BS, 6 CMNH. Apache Co.:
Spruce Creek, Tunitcha Mts., 7 BS; Springerville, 1 BS; North Fork White
River, White Mts., 12 SD; White River, Horseshoe Cienega, 8300 ft., White
Mts., 5 BS; Mt. Thomas, 9500 to 11,000 ft., White Mts., 12 BS; Little Colorado
River, White Mts., 4 BS; near head Burro Creek, 9000 ft., White Mts., 1 BS.
Graham Co.: Graham Mts., 9200 ft., 2 BS. Greenlee Co.: Prieto Plateau,
9000 ft., S. end Blue Range, 1 BS. Pima Co.: Summerhaven, 7500 ft., Santa
Catalina Mts., 3 BS, 1 SD. Cochise Co.: Fly Park, Chiricahua Mts., 4 BS;
Rustler Park, Chiricahua Mts., 1 SD; Long Park, Chiricahua Mts., 1 UM;
Huachuca Mts., 1 BS. Santa Cruz Co.: Stone Cabin Canyon, 8500 ft., Santa
Rita Mts., 1 BS.

New Mexico: San Juan Co.: Chusca Mts., 1 BS. Catron Co.: Mogollon
Mts., 3 BS; 10 mi. E Mogollon, 1 KU. Socorro Co.: Copper Canyon, Magdalena
Mts., 3 BS. Sierra Co.: Mimbres Mts., near Kingston, 1 BS.

Chihuahua: Sierra Madre, near Guadalupe y Calvo, 5 BS.

Marginal records.—Arizona: Tunitcha Mts. New Mexico: Chusca Mts.;
Copper Canyon, Magdalena Mts.; Mimbres Mts., near Kingston. Chihuahua:
Guadalupe y Calvo. Arizona: Huachuca Mts.; Santa Catalina Mts.; White
River, Horseshoe Cienega, 8300 ft., White Mts.; San Francisco Mtn.

Type.—Adult female, skin and skull; No. 53633, U. S. Biol. Surv. Coll.;
obtained on April 24, 1893, by E. W. Nelson from W slope of Mt. Orizaba,
9,500 ft., Puebla.

Range.—Transverse volcanic belt of mountains at the southern end of the
Mexican Plateau.

[Pg 52]
Diagnosis.—Size small for the species; measurements of 3 specimens from
Volcan Toluca, Mexico, are: total length, 98, 100, 108; tail, 35, 39, 40; hind
foot, 13, 13, 14. Summer pelage Mummy Brown tending toward Olive Brown;
Fuscous to Fuscous-Black in winter; skull and teeth relatively narrow.

Comparisons.—For comparison with S. v. monticola see account of that
subspecies.

Remarks.—The range of S. v. orizabae probably is not now in contact with
that of any other subspecies of S. vagrans, although judging by the slight degree
of difference between orizabae and monticola the separation
between the two has not been of great duration.

Sorex vagrans orizabae occurs with S. saussurei saussurei throughout the
transverse volcanic belt.

Specimens examined.—Total number, 23.

Michoacán: Patambán, 1 BS; Nahuatzín, 3 BS; Mt. Tancítaro, 4 BS.

Mexico: Salazar, 2 BS, 1 KU; Volcan de Toluca, 3 BS.

Tlaxcala: Mt. Malinche, 2 BS.

Puebla: Mt. Orizaba, 6 BS.

Veracruz: Cofre de Perote, 1 BS.

Marginal records.—Michoacán: Patambán. Veracruz: Cofre de Perote.
Puebla: Mt. Orizaba. Michoacán: Mt. Tancítaro.

Type.—Adult male, alcoholic; No. 1675, U. S. Nat. Mus.; obtained at
Shoalwater (Willapa) Bay, Pacific Co., Washington; received from J. G. Cooper,
and entered in Museum catalog on October 23, 1856.

Range.—The Great Basin, and Columbia Plateau west across the mountains
to the Pacific coast of northern California, Oregon, Washington and southwestern
British Columbia.

Diagnosis.—Size small for the species; average and extreme measurements
of 8 topotypes are: total length, 104.1 (99-109); tail, 43.3 (42-45); hind foot,
12.9 (12-14). Summer pelage ranging from (15´k) Cinnamon Brown through
(15´m) Proutt's Brown to (17´m) Mummy Brown. Winter pelage (13´´´´m)
Fuscous Black to (17´´´´m) Chaetura Black.

Comparisons.—For comparison with S. v. monticola see account of that
subspecies. Differs from S. v. halicoetes in relatively narrower and more attenuate
rostrum and in less brownish underparts in winter pelage; smaller and
more brownish (less grayish) than Sorex vagrans from the southern Sierra
Nevada.

[Pg 53]

Fig. 18. Probable geographic ranges of
Sorex vagrans vagrans, its derivative subspecies, and
S. v. mixtus.

1. S. v. vancouverensis 2. S. v. vagrans 3. S. v. halicoetes 4. S. v. paludivagus 5. S. v. obscuroides 6. S. v. mixtus

Remarks.—Restriction of the range of S. v. monticola to Arizona
and New Mexico leaves shrews that were formerly assigned to this
subspecies from Utah, Idaho, Washington and southern British
Columbia unassigned. Several names are available for consideration.
The name Sorex vagrans dobsoni Merriam, 1891, type locality
Alturas Lake, Blaine Co., Idaho, was once applied to small shrews
from Idaho, Montana, Wyoming, and Utah, but was considered by
Jackson to be synonymous with S. v. monticola. The name Sorex
vagrans amoenus Merriam, 1895, type locality near Mammoth, Mono
Co., California, has been applied to wandering shrews from western
Nevada, northeastern California and southern Oregon. Sorex vagrans
nevadensis Merriam, 1895, type locality Reese River on Nye-Lander
Co. line, Nevada was considered by Hall (1946:119) to be
synonymous with S. v. amoenus. Specimens of Sorex vagrans west
[Pg 54]
of the Cascade Mountains have long been referred to the nominate
subspecies which has its type locality at Willapa Bay, Pacific Co.,
Washington. Over so wide an area it is only to be expected that
some geographic variation is to be found. Thus specimens from
central Nevada average slightly paler in summer pelage than those
from the Pacific Coast or from the foothills of the Rocky Mountains.
In addition there are slight average differences in size from place
to place. Topotypes of S. v. vagrans, however, show a fair degree
of variability and some are nearly as pale as the paler Great Basin
stocks. Furthermore topotypical individuals of vagrans can be lost
in series of S. v. amoenus, although amoenus is shorter-tailed on the
average. Specimens from the western foot-hills of the Rocky Mountains
show an amazing series of relationships with the montane S. v.
obscurus. In Utah, as previously pointed out, complete intergradation
occurs. At 1 mi. N Heath, Washington Co., Idaho, the lowland
and the highland forms approach each other within a short distance
and still maintain a degree of distinctness, especially in size. In
northwestern Montana intergradation is extensive (Clothier, 1950).
In northeastern Washington distinctly separable populations occur
within a few miles of one another. In southern British Columbia
some populations are clearly intergrades while at 6 mi. S Yahk
intergradation seemingly has not taken place. Where some intergradation
has occurred the result often has been increased size of the
lowland shrews, although they usually retain the reddish summer
pelage rather than acquiring the more grayish pelage of obscurus.
The name dobsoni was based upon shrews from a place where lowland
and highland forms occur almost together with only a slight
amount of intergradation. Examples of "dobsoni" may not with
certainty be distinguished from typical vagrans except that they
are, as Merriam (1895:68-69) points out, somewhat larger. Merriam
(loc. cit.) further notes that dobsoni is "intermediate in size
and cranial characters between S. vagrans and obscurus;" a statement
which hits very close to the heart of the matter. I consider
the name dobsoni to apply to intergrades. To attempt to apply the
name to the highly variable populations of intergrades from British
Columbia to southern Idaho seems inadvisable. I have examined
the possibility of using the name amoenus for the animals from this
region. The characters which set amoenus apart from vagrans,
slightly shorter tail and slightly darker summer pelage, however,
are not universally found in shrews from the Columbian Plateau
and eastern Great Basin and furthermore these differences between
amoenus and vagrans do not seem to me to be of great enough
[Pg 55]
magnitude to warrant subspecific recognition of the former. Thus
the name S. v. vagrans may apply to shrews in the region under
consideration. The subspecies, as thus thought of, embraces several
incipient subspecies, namely (1) the populations on the isolated
mountain ranges of Nevada, (2) the coastal rain forest population
and possibly (3) the population on the Columbian Plateau.

In western British Columbia, Washington, and Oregon no evidences
of intergradation between S. v. vagrans and the races setosus,
permiliensis, bairdi, yaquinae, or pacificus are seen. In this region
S. v. vagrans occurs sympatrically with one or the other of these subspecies.
Different degrees of differentiation thus obtain between
the subspecies vagrans as here defined and the surrounding subspecies
of Sorex vagrans to wit: complete intergradation and
allopatry in Utah with S. v. obscurus; partial intergradation and
partial sympatry with S. v. obscurus in the foothill region from Idaho
to British Columbia; no intergradation and complete sympatry with
all the other races of Sorex vagrans from the Cascades to the coast
and south to San Francisco Bay. The relationship of S. v. vagrans to
the wandering shrews of the high Sierra is discussed on page 58.

Throughout most of the Great Basin and Columbian Plateau Sorex
vagrans is, with the exception of the rare S. merriami and S. preblei,
the only small shrew. In the Cascades and in the coastal lowlands
it is the only small shrew except for S. cinereus and S. trigonirostris,
both extremely rare and local in this region. S. vagrans seemingly
competes to a certain extent with the larger S. trowbridgii in western
Washington and seems to be partially dominant to trowbridgii, at
least in marshy habitats (Dalquest, 1941:171).

Specimens examined.—Total number, 1197.

British Columbia: Osoyoos District: Okanagan, 20 PMBC; Okanagan
Landing, 2 PMBC; Nahun Plateau, 2 PMBC. Vancouver District: Vancouver,
2 PMBC. New Westminister District: Port Moody, 16 BS; Westminster Jct.,
4 AMNH; Langley, 1 BS; Vedder Crossing, 1 PMBC; Huntingdon, 69 NMC;
Sumas, 16 BS; Cultus Lake, 1 NMC. Similkameen District: Princeton, 6 Mile
Creek, 1 NMC. Hedley, Stirling Creek, 7 NMC; Fairview-Keremeos Summit,
5 NMC; Oliver, 1 NMC; Westbridge, 6 NMC; Osoyoos, 1 PMBC; Osoyoos-Bridesville
Summit, 4 NMC; Cascade, 7 NMC. Nelson District: Kuskonook,
1 PMBC; Rossland, 14 NMC; Trail, 2 NMC; Creston, 4 PMBC, 4 NMC; near
Creston, 7 NMC. Cranbrook District: Cranbrook, 5 BS; Yahk, 2 NMC; Yahk
Camp 6, 2 NMC; Goatfell, 2 NMC. Fernie District: Newgate, 3 NMC.

Washington: Whatcom Co.: Blaine, 1 BS; Beaver Creek, 5 WSC; Glacier,
1 BS; Mt. Baker Lodge, 1 WSC; Lake Whatcom, 1 BS; Barron, 2 BS. Okanogan
Co.: Sheep Mtn., 3 BS; E. end Bauerman Ridge, 1 BS; Oroville, 1 BS; Hidden
Lakes, 1 BS; Loomis, 1 BS; Conconully, 1 BS; Twisp, 1 BS. Ferry Co.: 5 mi.
W Curlew, 2 BS. Stevens Co.: Marcus, 1 BS. Pend Oreille Co.: Canyon, 1
WSC; Metaline, 2 BS; Sullivan Lake, 1 BS. San Juan Co.: East Sound, Orcas
Island, 3 BS; Friday Harbor, San Juan Island, 1 BS; San Juan Park, 2 WSC;
Blakely Island, 1 KU; Richardson, 6 BS. Skagit Co.: Cypress Island, 1 KU;
Hamilton, 1 BS; Sauk, 1 BS; Avon, 3 BS; Mt. Vernon, 2 BS; La Conner, 5 BS.
[Pg 56]
Island Co.: San de Fuca, Whidby Island, 3 BS; Greenbank, Whidby Island,
2 BS; 3 mi. N Clinton, Whidby Island, 1 BS. Snohomish Co.: Oso, 2 BS;
Hermosa Point, Tulalip Indian Reservation, 7 mi. W and ½ mi. N Marysville,
3 KU. Chelan Co.: Entiat, 2 BS. Lincoln Co.: 6 mi. E Odessa, 4 BS.
Spokane Co.: Marshall, 7 BS. Clallam Co.: Neah Bay, 29 BS; 8 mi. W Sekin
River, 1 WSC; mouth Sekin River, 1 WSC; Dungeness, 1 BS; Port Townsend,
3 BS; Ozette Indian Reservation, 1 CMNH; Sequim, 4 BS; Tivoli Island, Ozette
Lake, 1 CMNH; Garden Island, Ozette Lake, 3 CMNH; Elwah, 1 WSC; Blyn,
1 BS; Soleduck River, 1 BS; 12 mi. S Port Angeles, 1 WSC; Forks, 9 CMNH,
1 BS; Cat Creek, 1 WSC; Lapush, 5 BS. Jefferson Co.: Jefferson Ranger Station,
N Fork Hoh River, 5 CMNH; Duckabush, 6 BS. Kitsap Co.: Vashon Island,
2 BS. King Co.: Redmont, 2 BS; Kirkland, 20 BS; Seattle, 1 WSC, 3 KU; Northbend,
2 BS: Lake Washington, near Renton, 2 BS; Kent, 1 BS; Enumclaw, 1 BS.
Grays Harbor Co.: Lake Quinault, 9 BS; Aberdeen, 20 BS; Westport, 5 BS,
2 WSC; Oakville, 1 BS. Mason Co.: Lake Cushman, 11 BS; Hoodsport, 1 BS;
North Fork Skokomish River, 1 BS; Shelton, 2 BS. Pierce Co.: Puyallup, 6
BS; Steilacoom, 1 BS; 6 mi. S Tacoma, 2 BS; Roy, 3 BS; Bear Prairie, Mt.
Rainier, 1 BS; Reflection Lake, Mt. Rainier, 1 WSC. Kittitas Co.: Blewett Pass,
3 BS; Easton, 3 BS; 2 mi. E Cle Elum, 4 FC; Ellensburg, 2 BS. Grant Co.:
Moses Lake, 1 BS; 9 mi. S and 1 mi. W Neppel, 1 UM. Whitman Co.: Hangman
Creek, Tekoa, 1 WSC; 4 mi. ENE Pullman, 1 KU; 2 mi. N Pullman, 2
WSC; 2 mi. NW Pullman, 1 WSC; 2 mi. W Pullman, 1 WSC; Pullman, 5 WSC;
Armstrong, 1 WSC; 5 mi. NE Wawawai, 1 BS; Wawawai, 5 WSC. Thurston Co.:
Nisqually Flats, 2 BS; Nisqually, 1 BS; 4 mi. S Olympia, 1 BS; Tenino, 4 BS.
Pacific Co.: Tokeland, 4 BS; 1 mi. S Nemah, 2 FC; 1 mi. N Bear River, Willapa
Bay, 8 FC; ¼ mi. N Bear River, 3 FC; 3½ mi. E Seaview, 6 FC; Ilwaco, 1 BS.
Lewis Co.: 8 mi. W Chehallis, 2 BS; Chehallis, 2 BS; Toledo, 1 BS. Yakima
Co.: Selah, 7 KU; Wiley City, 4 BS. Wahkiakum Co.: Cathlamet, 1 BS.
Skamania Co.: 45 mi. SE Toledo, 2 BS; Carson, 1 BS; Stevenson, 1 BS; 15 mi.
NW White Salmon, 1 BS. Klickitat Co.: Trout Lake, 15 mi. S Mt. Adams,
2 BS; 15 mi. N Goldendale, 1 WSC; Goldendale, 1 BS. Walla Walla Co.:
College Place, 1 KU. Columbia Co.: Starbuck, 3 BS. Garfield Co.: 1 mi. E
Pomeroy, 1 SGJ. Asotin Co.: 21 mi. SE Dayton, 1 BS; Rogersburg, 1 BS.

Idaho: Bonner Co.: 4 mi. S Sandpoint, 1 UM. Kootenai Co.: Coeur
d'Alene, 2 BS. Shoshone Co.: Osburry, 1 BS; Mullan, 2 BS. Latah Co.:
Felton's Mills, 1 WSC; Cedar Mtn., 5 WSC. Lewis Co.: Nezperce, 2 BS.
Idaho Co.: Seven Devils Mtn., 1 BS. Adams Co.: Summit of Smith Mtn.,
7500 ft., 5 KU; New Meadows, 1 BS; Tamarack, 1 BS. Washington Co.: 1 mi.
NE Heath, SW slope Cuddy Mtn., 4000 ft., 7 KU. Boise Co.: Bald Mtn. R. S.,
10 mi. S Idaho City, 1 BS. Elmore Co.: Cayuse Creek, 10 mi. N Featherville,
1 BS. Canyon Co.: Nampa, 5 BS. Blaine Co.: Sawtooth City, 5 BS; Alturas
Lake, 1 BS. Bonneville Co.: 10 mi. SE Irwin, 5 BS. Bannock Co.: Pocatello,
1 BS, 1 KU; 1 mi. W Bancroft, 1 KU; Swan Lake, 1 BS. Owyhee Co.: Grasmere,
1 SGJ. Cassia Co.: 10 mi. S Albion, Mt. Harrison, 1 BS.

Montana: Sanders Co.: Prospect Creek, near Thomson Falls, 4 BS. Lake
Co.: Flathead Lake, 5 BS. Ravalli Co.: Bass Creek, NW Stevensville, 2 BS;
2 mi. NE Stevensville, 1 UM; Corvallis, 4 BS; 6 mi. E Hamilton, 1 KU.

Oregon: Clatsop Co.: Seaside, 1 BS. Washington Co.: 5 mi. SE Hillsboro,
1 BS; Beaverton, 1 BS. Multnomah Co.: Portland, 20 BS; Portland, Switzler
Lake, 5 BS. Hood River Co.: 2 mi. W Parkdale, 1 BS; north slope Mt. Hood,
2 BS. Umatilla Co.: 10 mi. W Meacham, 2 BS; Meacham, 3 BS. Union Co.:
Elgin, 2 BS; Kamela, 2 BS; Hot Lake, 2 BS. Wallowa Co.: 25 mi. N. Enterprise,
4 BS; Wallowa Lake, 23 BS; S Wallowa Lake, 1 BS. Clackamas Co.:
Estacada, 1 KU. Marion Co.: Salem, 8 BS; Permilia Lake, 2 BS. Benton Co.:
Corvallis, 2 BS; 5 mi. SW Philomath, 5 BS. Linn Co.: Shelburn, 1 BS. Jefferson
Co.: 20 mi. W Warm Springs, 2 BS. Grant Co.: Beech Creek, 6 BS;
Austin, 1 BS; Strawberry Butte, 1 BS; Strawberry Mts., 12 BS. Baker Co.:
Homestead, 1 BS; Cornucopia, 11 BS; Rock Creek, 1 BS; Bourne, 7 BS; McEwen,
1 BS; Huntington, 1 BS; Anthony, 42 AMNH. Lane Co.: north slope
Three Sisters, 3 BS; Vida, 1 BS; Mapleton, 1 BS; Eugene, 2 BS; 10 mi. S
[Pg 57]
McKenzie Bridge, 1 BS; Florence, 1 BS. Deschutes Co.: Paulina Lake, 7
BS; Lapine, 8 BS. Crook Co.: 1 SGJ. Douglas Co.: Winchester Bay, 1
SGJ; Scottsburg, 3 BS; Drain, 5 BS; Lookingglass, 1 BS; Diamond Lake, 6
BS. Coos Co.: Empire, 5 BS. Curry Co.: Port Orford, 1 BS; Gold
Beach, 4 BS. Klamath Co.: Anna Creek, Mt. Mazama, 1 BS; Crater Lake,
14 BS; Upper Klamath Marsh, 2 BS; Ft. Klamath, 35 BS; Klamath Falls, 6
BS. Lake Co.: 10 mi. SW Silver Lake, 3 BS; west fork Silver Creek,
Yamsay Mts., 4 BS; Plush, 1 BS; Warner Creek, Warner Mts., 1 BS; Warner
Mts., 3 BS; Gearhart Mts., 17 SGJ; Harney Co.: Diamond, 2 BS; Keiger
Gorge, Steens Mts., 3 BS. Malheur Co.: 8 mi. W Jordon Valley, 1 BS.

Wyoming: Lincoln Co.: 13 mi. N and 2 mi. W Afton, 6 KU; 10 mi. N
Afton, Salt River, 2 BS; 9 mi. N and 2 mi. W Afton, 6 KU; 7 mi. N and 1 mi.
W Afton, 4 KU; Cokeville, 1 BS; 12 mi. N and 2 mi. E Sage, 2 KU; 6 mi. N
and 2 mi. E Sage, 1 KU.

California: Del Norte Co.: Smith River, 2 BS; Crescent City, 20 BS.
Siskiyou Co.: Beswick, 1 BS; Hornbrook, 3 BS; Brownell, Klamath Lake, 1 BS;
Mayten, 2 BS; Squaw Creek, Mt. Shasta, 5 BS; Upper Ash Creek, Mt. Shasta,
1 BS; upper Mud Creek, Mt. Shasta, 8 BS; Wagon Camp, Mt. Shasta, 5 BS;
Warmcastle Soda Springs, Squaw Creek Valley, 2 BS; Sisson, 7 BS. Modoc Co.:
Davis Creek, Goose Lake, 1 BS. Humboldt Co.: Humboldt Bay, 10 BS.
Trinity Co.: Canyon Creek, 2 BS. Shasta Co.: Fort Crook, 11 BS; Dana, 17
BS; Fall Lake, Fall River Valley, 3 BS; Cassel, 2 BS; 12 mi. E Burney, 1 BS;
Lassen Peak, 13 BS; Kellys, Warner Creek, 1 KU; Drakes Hot Springs, Warner
Creek, 2 BS. Mendocino Co.: Russian Gulch State Park, 2 FC. Plumas Co.:
12 mi. NE Prattville, 2 BS; Spring Garden Ranch, Grizzly Mts., 3 BS; Sierra
Valley, 1 BS. Sierra Co.: Lincoln Creek, 1 BS. Sonoma Co.: Petaluma, 3 BS;
Point Reyes, 7 BS. Placer Co.: Donner, 3 BS. El Dorado Co.: Tallac, 3 BS.
Mono Co.: Mt. Conness, 1 BS; Mono Lake, 1 BS; near Mammoth, 8000 ft.,
head of Owens River, 2 BS. Inyo Co.: Alvord, 1 BS.

Nevada: Elko Co.: Mountain City, 1 BS; Three Lakes, 1 KU; west side
Ruby Lake, 3 mi. N White Pine Co. line, 8 KU; Ruby Mts., 9 BS; W side Ruby
Lake, 3 BS. White Pine Co.: W side Ruby Lake, 3 mi. S Elko Co. line, 1
KU. Nye Co.: Cloverdale, Reese River, 3 BS.

Utah: Weber Co.: Beaver Creek, S Fork Ogden River, 2 UU; Huntsville,
10 mi. E Ogden, 1 UU; Hooper Bay Refuge, 4200 ft., 1 UU; Riverdale, 4200
ft., 3 UU; Riverdale, 4250 ft., 1 UU; 3 mi. SE Ogden, 2 UU; Snow Basin,
2 UU; Snow Basin, S part Wheeler Canyon, 1 UU; Uinta, 2 mi. W Weber
Canyon entrance, 4 UU; 2 mi. W Uinta, 1 UU. Salt Lake Co.: City Creek
Canyon, 6 mi. NE Salt Lake City, 4700 ft., 2 UU; 1 mi. up City Creek Canyon,
4600 ft., 1 UU; ¾ mi. above Forks, City Creek Canyon, 1 UU; The Firs, Millcreek
Canyon, 1 UU; Olympus Water Box, 1 UU; Salamander Lake, Lamb's
Canyon, 9000 ft., 3 UU (near obscurus); Salt Lake City, 7500 ft., 1 UU; 1 mi.
W Draper, 4500 ft., 6 UU; Draper, 4500 ft., 5 UU; 1½ mi. SW Draper, 4500
ft., 1 UU; 3 mi. SW Draper, 4400 ft., 2 UU; 3 mi. S Draper, 4400 ft., 2 UU;
1 mi. S Draper, 4500 ft., 1 UU. Juab Co.: W side Deep Creek Mts., Queen
of Sheba Canyon, 8000 ft., 3 UU. Wasatch Co.: Midway Fish Hatchery,
5450 ft., 1 UU.

Marginal records.—British Columbia: Okanagan; Westbridge; Kuskonook;
Cranbrook. Montana: Flathead Lake; 6 mi. E Hamilton; Prospect Creek.
IDAHO: Cedar Mtn.; New Meadows; Alturas Lake; 10 mi. SE Irwin. Wyoming:
13 mi. N and 2 mi. W Afton; 6 mi. N and 2 mi. E Sage. Idaho: 1 mi. W
Bancroft; Swan Lake. Utah: Beaver Creek, South Fork, Ogden River; Midway
Fish Hatchery; west side Deep Creek Mts., Queen of Sheba Canyon, 8000
ft. Nevada: Baker Creek (Hall, 1946:120); Reese River (ibid.); 2 mi. S Hinds
Hot Springs (ibid.). California: Mono Lake (Jackson, 1928:110); near
Mammoth; Alvord; Mount Conness; Donner; Buck Ranch (Jackson, 1928:110);
Warner Creek, Drake Hot Springs (ibid.); Canyon Creek; Cuddeback (Jackson,
1928:105); Novato Point (ibid.), thence northward along the coast to Washington:
Friday Harbor, San Juan Island. British Columbia: Port Moody.

[Pg 58]

Type.—First year female, skin and skull; No. 30064/42074, U. S. Biol. Surv.
Coll.; obtained on August 9, 1891, by Frank Stephens from Bishop Creek, 6600
ft., Inyo Co., California, original no. 811.

Range.—The Sierra Nevada of California, north at least to El Dorado
County, intergrading northerly with S. v. vagrans.

Diagnosis.—Size medium for the species; average and extreme measurements
of 5 topotypes are: total length, 107 (103-112); tail, 47 (45-50); hind foot,
12.8 (12-13.5). Skull relatively broad interorbitally; color of dorsum in summer
pelage nearest (17´´´k) Olive Brown.

Comparisons.—Differs from S. v. vagrans, with which it intergrades to the
north, in: longer tail and total length; skull larger and relatively broader
interorbitally; color in summer grayer (less reddish), the lighter subterminal
color bands of the hair often showing through the darker tips and imparting
a grizzled appearance to the dorsum. Differs from S. v. parvidens to the south
in: skull relatively broader interorbitally and less flattened; teeth slightly larger.

Remarks.—S. v. obscuroides has long been called S. v. obscurus. In fact,
obscuroides is separated from the range of obscurus by the intervening, smaller
subspecies S. v. vagrans. S. v. obscuroides resembles S. v. obscurus in color
and size but the skull is smaller, although relatively slightly broader. The resemblance
in color is possibly due to the fact that obscuroides, like obscurus,
is a high mountain form. S. v. obscuroides intergrades with S. v. vagrans
along the crest of the Sierra between Yosemite National Park and Lassen Peak
and on the eastern slope of the Sierra from approximately Mammoth northward.
Specimens from Donner are intergrades but are closest to S. v. vagrans.
Although all specimens from Lassen Peak are referable to S. v. vagrans, some
show cranial characters of obscuroides.

Specimens examined.—Total number, 76. California: Mono Co.: Mt.
Dana, 6 BS; Mt. Lyell, 11 BS. Mariposa Co.: Tuolumne Meadows, Muir
Meadow, 9300 ft., 1 BS; Tuolumne Meadows, Mt. Unicorn, 1 BS; Tuolumne
Meadows, N base Mt. Lyell, 8 BS; Tuolumne Meadows, Soda Springs, 4 BS;
Lake Tenaya, 5 BS. Madera Co.: San Joaquin River, 8000 ft., 4 BS. Fresno
Co.: Horse Corral Meadows, 3 BS. Mono Co.: head of Owens River near
Mammoth, 2 BS. Inyo Co.: Bishop Creek, 5 BS; Round Valley, 1 BS. Tulare
Co.: E. Fork Kaweah River, 7 BS; Mt. Whitney, 5 BS; Whitney Creek, Mt.
Whitney, 4 ChM; Whitney Meadows, 9700 ft., 1 BS; Mineralking, 2 BS; N.
Fork Kern River, 9600 ft., 1 BS; S. Fork Kern River, 4 BS; Kern Lakes, 1 BS.

Marginal records.—California: Pyramid Peak; near Mammoth; Round
Valley; Bishop Creek; Mt. Whitney; Kern Lakes; Halstead Meadows; Horse
Corral Meadows; east fork Indian Canyon (Jackson, 1928:121).

Sorex obscurus parvidens Jackson, Jour. Mamm., 2:161, August 19, 1921.

Type.—Adult male, skin and skull; No. 56561, U. S. Biol. Surv. Coll.; obtained
on October 3, 1893, by J. E. McLellan from Thurmans Camp, Bluff Lake,
7500 ft., San Bernardino Mts., California.

Range.—Confined, so far as known, to the San Bernardino and San Gabriel
mountains, San Bernardino Co., California.

Diagnosis.—Size medium for the species; measurements of two specimens
from the San Bernardino Mountains are: total length, 105, 106; tail, 41, 48;
hind foot, 12, 14. Upper parts in summer Olive-Brown to Buffy-Brown;
[Pg 59]
cranium flattened and relatively narrow; unicuspids and incisors relatively
small.

Comparisons.—For comparison with S. v. obscuroides, the only adjacent
subspecies, see the account of that subspecies.

Remarks.—S. v. parvidens is seemingly an uncommon mammal. I have been
informed by Terry Vaughan that repeated attempts by him to obtain it in
suitable habitat in the San Gabriel Mountains failed. This shrew is probably
no longer in reproductive continuity with Sorex vagrans of the Sierra Nevada.

Specimens examined.—Total number, 4. California: San Bernardino Co.:
type locality, 4 BS.

Marginal records.—California: Camp Baldy, San Antonio Canyon (Jackson,
1928:124); type locality.

Type.—Young adult male, skin and skull; No. 3638, Mus. Vert. Zool.; obtained
on May 6, 1908, by Joseph Dixon from salt marsh near Palo Alto, Santa
Clara Co., California.

Range.—Marshes in the southern part of San Francisco Bay, California.

Diagnosis.—Size small for the species; measurements of two topotypes are:
total length, 105, 106; tail, 39, 40; hind foot, 12, 13. Upper parts in winter
Chaetura Black or near Fuscous-Black; underparts brownish; upper parts in
summer near (17´m) Mummy Brown; underparts with a decided buffy wash,
near (15´d) Light Ochraceous Buff; rostrum relatively large; maxillary tooth-row
relatively long; teeth relatively large.

Comparisons.—Darker ventrally, both summer and winter, than S. v. vagrans;
slightly more reddish dorsally in summer pelage than S. v. vagrans, rostrum and
teeth relatively larger; smaller externally than S. v. paludivagus, paler; skull
longer, narrower cranially and broader rostrally.

Remarks.—This subspecies seems to be restricted to salt marshes where
it occurs with Sorex ornatus.

Specimens examined.—Total number, 12. California: San Francisco Co.:
San Francisco, 4 BS. Alameda Co.: West Berkeley, 1 BS; Berkeley, 1 BS;
Dumbarton Point, 1 KU. San Mateo Co.: San Mateo, 2 BS. Santa Clara Co.:
Palo Alto, 3 BS.

Marginal records.—California:
Berkley, Elmhurst; Palo Alto; San Mateo.

Type.—Adult male, skin and skull; No. 5053, Los Angeles Museum of
History, Science and Art, obtained on November 3, 1938, by Jack C. von
Bloeker, Jr., from salt marsh at mouth of Elkhorn Slough, Moss Landing,
Monterey Co., California, original no. 9456.

Diagnosis.—Size medium for the species; average and extreme measurements
of 6 topotypes are: total length, 115 (113-118); tail, 46.5 (42-48); hind foot,
14.5 (14-15) (von Bloeker, 1939:94). In winter nearly black dorsally, deep
mouse gray ventrally; in summer nearly as dark dorsally as in winter, hairs of
[Pg 60]
venter tipped with Clove Brown; skull short, relatively broad cranially and
relatively narrow rostrally.

Comparisons.—For comparison with S. v. halicoetes see account of that
subspecies.

Remarks.—This subspecies, occurring at the limits of the range of the species,
is uncommon in most collections. Seven specimens were available for the
original description. The summer pelage is not completely described in the
original description, but is stated to be darker than the winter pelage of S. v.
vagrans, and must thus be considerably darker than the summer pelage of
S. v. halicoetes. Two specimens in the Museum of Vertebrate Zoology, from
San Gregario, referred by Jackson to S. v. halicoetes, were included in the
present subspecies by von Bloeker.

Specimens examined.—None.

Records of occurrence (von Bloeker, 1939:94).—California: San Mateo
Co.: San Gregario. Monterey Co.: Seaside; mouth of Salinas River; Moss
Landing.

Marginal records.—California: San Gregario; Seaside.

Sorex vancouverensis Merriam, N. Amer. Fauna, 10:70, December 31, 1895.

Sorex vagrans vancouverensis, Jackson, N. Amer. Fauna, 51:106, July, 1928.

Type.—Adult male, skin and skull; No. 71913, U. S. Biol. Surv. Coll.;
obtained on May 10, 1895, by Clark P. Streator, from Goldstream, Vancouver
Island, British Columbia.

Range.—Vancouver Island from Sayward south, and Bowen Island.

Diagnosis.—Size small for the species; average and extreme measurements of
6 specimens from Alberni Valley, Vancouver Island, are: total length, 106.5
(97-115); tail, 41.7 (40-43); hind foot, 12 (11-13) (Jackson, 1928:107).
Ventral parts brownish, winter pelage reddish brown rather than grayish.

Comparisons.—Differs from S. v. vagrans in more brownish ventral parts
and more brownish, rather than grayish, winter pelage; differs from the sympatric
S. v. isolatus in shorter tail, shorter hind foot, more narrow skull, and
smaller teeth.

Remarks.—This is a poorly differentiated subspecies which is closely related
to S. v. vagrans. The differences in color noted are average ones. Some individuals
of this shrew might be difficult to separate from S. v. isolatus. The
slight degree of morphological divergence is such that intergrades might be
expected to occur. Possibly some habitat separation occurs, but such has not
been reported.

Specimens examined.—Total number, 3. British Columbia: Vancouver Island:
Mt. Washington, 1 KU; Nanaimo, 1 BS; type locality, 1 BS.

Marginal records.—British Columbia: Sayward (Anderson, 1947:18);
Bowen Island (Hall, 1938:463); Alberni (Jackson, 1928:107).

[Pg 62]

26-7903

[Pg i]

Institutional libraries interested in publications exchange may
obtain this series by addressing the Exchange Librarian, University
of Kansas Library, Lawrence, Kansas. Copies for individuals, persons
working in a particular field of study, may be obtained by
addressing instead the Museum of Natural History, University of
Kansas, Lawrence, Kansas. There is no provision for sale of this
series by the University Library, which meets institutional
requests, or by the Museum of Natural History, which meets the
requests of individuals. Nevertheless, when individuals request
copies from the Museum, 25 cents should be included, for each
separate number that is 100 pages or more in length, for the purpose
of defraying the costs of wrapping and mailing.

* An asterisk designates those numbers of which the Museum's
supply (not the Library's supply) is exhausted. Numbers published to
date, in this series, are as follows: