In comparing the ventral views of skulls note especially the size and location of incisive foramina and posterior palatine foramina as well as the breadth of mesopterygoid fossae. In comparing the left maxillary tooth-rows note especially the size of M3 and the reduced posterointernal cusp on Ml in A. peninsulae.]
More recently, Ellerman (1949:32) and Ellerman and Morrison-Scott (1951:566) have arranged peninsulae as a subspecies of Apodemus flavicollis under the assumption that all the members of the subgenus Sylvaemus on the eastern Asiatic mainland are subspecies of one or another of the species of western Europe, A. flavicollis or A. sylvaticus. Ellerman (in Ellerman and Morrison-Scott, 1951:564) states: "The majority of the forms I distribute in a somewhat arbitrary manner between sylvaticus, average smaller skull, and flavicollis, average larger skull; occurring together nearly throughout the Palaearctic. I feel fairly sure that there are some errors of judgment in my arrangement, and equally sure that there is no other way to define species in this very large and difficult group." I have compared the specimens of peninsulae available to me from central and southern Korea with specimens of A. f. flavicollis from Denmark, Germany and Sweden and find, although the two are similar in many ways, that peninsulae differs from flavicollis in several important characters: Mammae 1-2=6 in flavicollis, and 2-2=8 in peninsulae; incisive foramina reaching level of alveoli of M1, or nearly so, in flavicollis, but ending conspicuously short of that level in peninsulae; posterior palatine foramina large in flavicollis and opposite a point where M1 and M2 meet, but small in peninsulae and situated farther back on the palate, opposite M2. Moreover, peninsulae lacks the characteristic buffy throat patch of flavicollis, has a much reduced posterointernal cusp on the M1, a relatively (frequently actually) larger M3 and, on the average, a broader mesopterygoid fossa. In view of these differences, all of which appear to be constant, I consider peninsulae specifically distinct from flavicollis. Throughout its known geographic range (see below) peninsulae is evidently confined to wooded terrain, either scrub or brush types or forested areas, and the vernacular name wood mouse, therefore, seems appropriate for this species.
The type specimens of Apodemus praetor Miller (type from Sungari River, 60 mi. SW Kirin, Manchuria) and Apodemus nigritalus Hollister (type from Tapucha, Altai Mountains, Siberia) agree with peninsulae as concerns the above characters and differ from it only in minor external and cranial features. They are, therefore, here considered as subspecies of the latter.
Ellerman (1949:32) and Ellerman and Morrison-Scott (1951:567) regarded nigritalus, like peninsulae, as a subspecies of flavicollis. The subspecies praetor, on the other hand, has generally been regarded as a synonym of peninsulae by recent authors. Howell (1929:58) noted that the holotype was, "... a phenomenally large specimen such as is encountered occasionally in almost all groups of rodents." He ascribed the color differences noted by Miller to "seasonal" variation. The holotype of praetor is undeniably larger than the other adult specimens listed in the original description. These paratypes and other specimens of praetor available to me are approximately the same size externally and average only slightly larger cranially than specimens of peninsulae from central and southern Korea. However, the dorsal coloration of praetor is somewhat darker and duller than that of peninsulae, especially in summer pelage when praetor lacks the conspicuous bright ochraceous tinge of the Korean specimens. In addition, praetor has broader zygomatic plates with correspondingly deeper zygomatic notches and the color on the face of the upper incisors averages much more orange than in peninsulae.
In the north then, wood mice range from Korea and Manchuria westward at least as far as the Altai Mountains. For mice from the intervening Siberian areas Russian workers have used the name major which, as noted above, is unavailable. The exact relationships of the mice of these areas to previously named subspecies is unknown to me and I have not seen specimens of "Mus (Alsomys) major rufulus" of Dukelsky, 1928, the type locality of which is 75 versts (approximately 50 miles) SE Vladivostok, Siberia. It appears to be of the same species as peninsulae and judging from the original description it closely resembles praetor. Neither have I seen specimens of the Sakhalin Island mouse, giliacus, which Ellerman (1949:32) regards as a subspecies of Apodemus sylvaticus. I feel reasonably sure, however, that it will prove to be a subspecies of peninsulae. In the original description giliacus was referred to as, "Most closely allied to the Korean subspecies..." (Thomas, 1907:411).
In China the extent of the distribution of Apodemus peninsulae is also uncertain. Allen (1940:949-50) reported its occurrence from Jehol and Hopeh in the northeast, southwestward through Shansi, Shensi and eastern Kansu to Szechuan and northwestern Yunnan. Throughout most of this region it occurs with another mouse, currently regarded as conspecific with Apodemus sylvaticus, and the two kinds have been confused by some previous authors. Howell (1929:58), for instance, reported twelve specimens of peninsulae from 65-75 mi. NE Peking but my examination of these mice indicates that only four are peninsulae while the others are referrable to what is currently regarded as Apodemus sylvaticus draco. Another subspecies of sylvaticus, A. s. orestes, occurs in Szechuan and Yunnan and it is certain that some records of distribution ascribed to peninsulae from those provinces actually represent orestes (see Allen, 1940:949-50). A. sylvaticus is distinguishable from peninsulae by darker ears, blackish preauricular patches, dark eye rings, a noticeably smaller skull, incisive foramina that reach the level of Ml (or nearly so), much larger auditory bullae, and a more fully developed posterointernal cusp on M1. Too, sylvaticus typically has 1-2=6 mammae although Allen reports finding a 2-2=8 formula in some specimens. Apodemus latronum, regarded as a full species by Osgood (1932:318) and G. M. Allen (1940:950) but as a subspecies of flavicollis by Ellerman (1949:32) and Ellerman and Morrison-Scott (1951:567), also occurs in Szechuan and Yunnan. Its relatively dark color, large feet and large ears, flavicollis-like skull and large molar teeth immediately separate it from peninsulae although the two possibly have been confused in the earlier literature. Until a complete revisionary study of the Asiatic members of the subgenus Sylvaemus can be undertaken the presence of peninsulae in southwestern China must remain in question.
The western limits of the geographic range of Apodemus peninsulae are unknown. Apodemus gurkha Thomas, 1924, from Nepal is said to have 2-2=8 mammae but the description is not otherwise suggestive of close relationship to peninsulae. Farther to the west, Apodemus flavicollis rusiges Miller, 1913, from Kashmir seems to have been properly assigned as a subspecies of flavicollis (cotypes and large series in USNM).
Wood mice almost certainly do not occur in the Gobi Desert. They are known as far west as the Altai Mountains to the north of the Gobi and at least as far west as Kansu (see below) to the south of it. Whether the geographic range of the species skirts the western edge of the arid regions of northern China is at present unknown; perhaps it does not. At any rate, mice available to me from the North Chinese provinces of Jehol, Shansi, Shensi and Kansu are notably different in certain external and cranial features from other known races of Apodemus peninsulae and are here given subspecific recognition. All measurements are in millimeters. Capitalized color terms are from Ridgway (1912).
Apodemus peninsulae sowerbyi, new subspecies
Type.—Adult female molting from winter to summer pelage, skin and skull, U. S. National Museum no. 175523, from 30 miles west of Kuei-hua-cheng, 7000 ft., northern Shansi, China; obtained on 23 May 1912 by Arthur de Carle Sowerby, original no. 456.