Geographic variation of the large-eared field mouse (Apodemus latronum Thomas, 1911) (Rodentia: Muridae) with one new subspecies description verified via cranial morphometric variables and pelage characteristics

Abstract

Background

The large-eared field mouse (Apodemus latronum Thomas, 1911), a common rodent, mainly inhabits southwestern China. Since its first description nearly a century ago, there have been numerous taxonomical and morphological arguments as to its validity, but relatively little work was done on mapping out the geographic variations observed in samples of the species. In this study, we used 142 specimens of A. latronum from Xizang, Sichuan, and Yunnan to conduct a multivariate analysis, coefficient of difference (CD) analysis of cranial measurements, and a comparison of some pelage characteristics.

Results

The results of the analysis on 15 measurable cranial characters indicated that the specimens from Lijiang, Weixi, and Binchuan areas of the Yunnan province are apparently different from all the other specimens of A. latronum described so far and are allopatrically distributed.

Conclusions

These samples form the core of our new description for A. latronum lijiangensis subsp. nov. as a new subspecies, and a detailed discussion on the relationships between the differentiation of A. latronum and its evolvement in southwestern China is provided.

Background

The large-eared field mouse (Apodemus latronum Thomas, [1911]) mainly inhabits southwestern China, including both Sichuan and Yunnan provinces as well as southeastern Tibet and northern Burma (Corbet and Hill [1992]; Musser and Carleton [2005]). The species was first named Apodemus speciosus latronum (Thomas [1911]) based on an adult male specimen from Tatsienlu, Szechwan (modern-day Kangding, Sichuan province), and due to its large body size and big blackish ears, Osgood ([1932]) recognized it as a species that might have been related to A. flavicollis (Melchior, 1834), and Allen ([1940]) also accepted the naming as A. latronum in subgenus Sylvaemus. Ellerman ([1941]), however, thought it should be A. speciosus latronum in the speciosus group of Apodemus, and later, Ellerman and Morrison-Scott ([1950]) regarded it as A. flavicollis latronum. Zimmermann ([1962]) then identified it as a valid species, and Corbet ([1978]) accepted this result. Nearly a decade later, Feng et al. ([1986]) rearranged it as A. draco latronum, followed by arguments from Corbet and Hill ([1992]) and Musser and Carleton ([1993]), all of whom insisted on the earlier result of Zimmermann ([1962]). This finding was later supported by Huang et al. ([1995]) and Musser et al. ([1996]) who went on to expatiate on the validity of A. latronum with morphological or published genetic evidence (Suzuki et al. [2003]). Finally, both Wang ([2003]) and Musser and Carleton ([2005]) listed it as A. latronum.

After so many arguments on its taxonomic status and the subsequent genetic studies to clarify any inconsistencies, there has been no recent disagreement as to the status of A. latronum as a valid species. Interestingly though, there have been no further follow-up morphological studies conducted to obtain a more well-rounded view of this species, especially in regard to measureable cranial variables or comparisons of pelage characteristics as it pertains to the geographic variation of A. latronum. In this study, we opted to conduct a typical morphometric analysis of both morphometrics and related statistical analyses on the skull of A. latronum, and comparison on their pelage characteristics is performed in order to study their geographic variation. Additionally, we also discussed the relationships between the differentiation of A. latronum and its environment in southwestern China.

Methods

Data collection

The specimens used in the study were obtained from Kunming Natural History Museum of Zoology, Kunming Institute of Zoology, Chinese Academy of Sciences (KIZ, CAS) (Kunming, China), Institute of Zoology, Chinese Academy of Sciences (IOZ, CAS) (Beijing, China), and Sichuan Academy of Forestry (SAF) (Chengdu, China). The numbers and collection localities of specimens examined in the study are listed in detail in the Appendix.

A total of 142 specimens were studied. Excluding four specimens that have no noted sexual record, the remaining 138 specimens include 64 male samples and 74 female samples. All specimens used in this study are considered to have been adults at the time of their collection, because it was evident from examination that molar eruption had finished, and the grinding surface of M² appears as a longitudinal worn dentine link (Lu et al. [1987]). Totally, 15 cranial measurements were taken with a digital caliper at the greatest possible accuracy (0.01 mm) as follows: greatest length of skull (GLS), condylobasal length (CBL), basal length (BL), occipitonasal length (ONL), palatal length (PL), diastema length (DL), upper tooth row (UTR), breadth across molars (BAM), breadth of zygomatic plate (BZP), breadth of occipital condyles (BOO), zygomatic width (ZW), interorbital breadth (IB), breadth of braincase (BB), length of lower diastema (LLD), and lower tooth row (LTR) (Figure 1). In addition, four external measurements were taken from the initial accession labels attached to the skin: head and body length (HB), tail length (TL), hind foot length (HFL), and ear length (EL). Since these measurements were taken by a variety of different collectors who likely used different tools and methodologies, these four metrics were not included either in the multivariate or coefficient of difference (CD) analyses.

Figure 1

Sketch map indicating 15 cranial measurements in the study. GLS, greatest length of skull; CBL, condylobasal length; BL, basal length; ONL, occipitonasal length; PL, palatal length; DL, diastema length; UTR, upper tooth row; BAM, breadth across molars; BZP, breadth of zygomatic plate; BOO, breadth of occipital condyles; ZW, zygomatic width; IB, interorbital breadth; BB, breadth of braincase; LLD, length of lower diastema; LTR, lower tooth row.

Data analysis

Following the aforementioned data collection, the 15 cranial measurements we took were log-transformed and analyzed for sexual dimorphism (for the 138 samples with known sex records). After which, principal component analyses (PCA) without assuming a prior group was used to identify the various groupings of the samples. The groups identified via PCA were then assigned with names; then, multiple comparisons between the groups were performed for all 15 cranial variables, and the coefficient of difference analysis (CD) (Mayr [1969]) between the groups was calculated using the following equation: CD = (M b − M a) / (SDa + SDb), where M b is the mean of population b, M a is the mean of population a, SDa is the standard deviation of population a, and SDb is the standard deviation of population b. The principal component analyses and multiple comparisons were performed using SPSS version 11.0 (SPSS Inc., Chicago, IL, USA).

Pelage comparisons

Pelage characteristics are well known to change during different seasons, and as such, it can be difficult to make comparisons between samples taken at different times of the year. In order to eliminate the influence of seasonal pelage characteristic variation, we selected samples with the same collection date of topotype of A. latronum latronum and the type of A. latronum lijiangensis to compare the pelage characteristics.

Results

Sexual dimorphism and principal component analyses

Of the 142 total samples that were used for cranial measurements, 138 contained sexual information on their labels, with 64 male specimens and 74 female specimens (one without skull). The results from the Tests of Equality of Group Means indicated that 8 of the measured 15 cranial variables exhibited significant differences (p< 0.05) between males and females (Table 1). Since 8 out of 15 cranial variables show significant differences and some marked sexual dimorphism, we conducted principal component analyses and found that among the male group (N = 64), the eigenvalues for the first three principal components were 10.24, 1.23, and 0.86, respectively, thereby accounting for 82.22% of the total variance. Most of the measured characteristics had high positive loadings on the first principal component, suggesting that this component (68.26% of the total variance) represents size variation among the samples. The second principal component (8.21% of variance) was strongly correlated with IB and BB (factor loadings > 0.70), while the third principal component (5.76% of variance) is correlated primarily with LLD (factor loadings > 0.70) (Table 2). Figure 2A,B shows the plots of A. latronum male samples on principal component factors 1 vs. 2 and 1 vs. 3, respectively.

Table 1 Tests of equality of group means by gender

Table 2 Factor loadings and percentage of variance explained for principal component analysis on A. latronum male samples

Figure 2

Regional distribution of A. latronum male samples by plotting on principal components factors. (A) PC1 by PC2. (B) PC1 by PC3.

Similar results were found among the female group. For the 73 skull specimens in the female group, The PCA analysis showed that the eigenvalues for the first three principal components were 11.49, 0.89, and 0.67, respectively, accounting for 86.93% of the total variance. Most characteristics had high positive loadings on the first principal component, suggesting that this component (76.57% of the total variance) likewise represents size variation, while the second principal component (5.92% of variance) was also strongly correlated with LLD (factor loadings > 0.70); the third principal component (4.44% of variance) was correlated primarily with IB (factor loadings > 0.70) (Table 3). Scatter plots of A. latronum female samples on PC1 by PC2 and PC1 by PC3 are shown in Figure 3A,B, respectively.

Table 3 Factor loadings and percentage of variance explained for principal component analysis on A. latronum female samples

Figure 3

Regional distribution of A. latronum female samples by plotting on principal components factors. (A) PC1 by PC2. (B) PC1 by PC3.

The collective results from both the male and female groups together (Figures 2 and 3) indicate that the samples are actually composed of two different groups, and the results of multiple comparisons show significant difference between A. latronum latronum and A. latronum lijiangensis (Table 4); also, sustaining the samples could be distinguished into two different groups by skull variables. Paired with the data on the geographic distributions of the collection localities (Figure 4), we found that the samples from the Lijiang, Weixi, and Binchuan areas of the Yunnan province form a distinct group which we have named (A. latronum lijiangensis, subsp. nov.). The other samples from the Sichuan province (including Markam, Baoxing, Yajiang, Luhuo, Baiyu, Batang, Danba, Daocheng, and Muli areas), northwestern Yunnan province (including Xiaozhongdian and Deqin areas), and southeastern Tibet (including Mangkang, Zuogong, and Bomi areas) form the other absolute group (A. latronum latronum).

Table 4 Multiple comparisons on all 15 cranial variables between A. latronum latronum and A. latronum lijiangensis

Figure 4

Geographic distribution of the samples examined in this study. The black dots mean the places where A. latronum latronum were collected, and the black rectangles mean the places where A. latronum lijiangensis where collected.

Coefficient of difference analysis

The comparisons on the CD for the 15 cranial measurable variables between the two distinct groups, A. latronum lijiangensis subsp. nov. and A. latronum latronum (see Table 5) showed that the CDs of GLS, ONL, DL, and UTR are larger than 1.28 between A. latronum latronum and A. latronum lijiangensis. External and cranial measurements of A. latronum latronum and A. latronum lijiangensis subsp. nov. are given in Table 6.

Table 5 Comparison of coefficient of difference (CD) between A. latronum latronum and A. latronum lijiangensis , subsp. nov.

Table 6 Measurements (mm) of external and skull variable measurements (mm) of subspecies of Apodemus latronum

Description of one new subspecies, Apodemus latronum lijiangensis Li et Liu, subsp. nov.

Holotype: KIZ 016518, ♂, adult, collected 19 October 2007, from Yulong Snow mountain, Lijiang, Yunnan Prov., China, elevation 3,305 m.

Paratype: KIZ 016549, ♀, adult, collected 19 October 2007, from Yulong Snow mountain, Lijiang, Yunnan Prov., China, elevation 3,305 m.

Specimens examined: 17♂♂, 41♀♀, and one (no sexual recording) from Lijiang, Weixi, and Binchuan areas of the Yunnan province (Appendix).

Etymology: This new subspecies was named according to the type locality.

Diagnosis: General dorsal color is brown, mixed with black more or less, and from the back to the flank, it changes from fawn brown to clay brown. The muzzle and its sides are mainly grayish mixed black, while the forehead, the central area of the crown, and nape fawn brown are thinner, but largely the same as that at the back.

Description: A. latronum lijiangensis subsp. nov. exhibits several distinct and apparently fairly stable characteristics: (1) general dorsal color is brown, mixed with black more or less, and from the back center to the flank, it changes from fawn brown to clay brown, apart from A. latronum latronum, which is ochraceous or ochraceous buff, clearly; (2) underparts gray at base and tipped with dull white; (3) the muzzle and its sides mainly grayish mixed black; the forehead, the central area of the crown, and nape fawn brown, just the same as the back, but thinner more or less; apart from A. latronum latronum, which the muzzle mainly black brown, and its sides mainly yellow brown; (4) flanks and sides of cheeks and neck are clay brown; (5) orbits dark brownish; (6) the upper surfaces of the front and hind feet dull white; (7) the tail dark brown above and the undersurface brown in the base but more and more thinner to the twig, apart from A. latronum latronum, which the whole tail undersurface deep brown and mixed with black in its twig more or less.

Synonyms: Apodemus latronum latronum, Thomas [1911], Abstr. Proc.Zool. Soc. Lond., 100: 49.

Measurements (see Table 7 for external and skull of type specimens of A. latronum lijiangensis subsp. nov).

Table 7 Measurements (mm) of external and skull of type specimens of Apodemus latronum lijiangensis subsp. nov.

Remarks: This subspecies is known to presently dwell in Lijiang, Weixi, and Binchuan, Yunnan province, and its distribution apparently does not overlap with the other known subspecies. A. latronum latronum occurs both in Xiaozhongdian and Deqin of the Yunnan province as well as in southeastern Tibet and Sichuan province. According to the theory of subspecies differentiation (Mayr [1969]), when the CD of any parameter of variables within two samples exceeds 1.28, the samples can be regarded as different subspecies. The results of CD analysis indicate that among these 15 variables, 4 of the 15 CDs exceeded 1.28 between A. latronum lijiangensis subsp. nov. and A. latronum latronum (Table 5). On the pelage characteristics, A. latronum lijiangensis subsp. nov. can clearly be distinguished from A. latronum latronum by the following combinations of characteristics described above: (1), (3), and (7), further confirming the validity of this new subspecies. Figure 5 indicates the pelage differences between A. latronum lijiangensis and A. latronum latronum.

Figure 5

Representative pictures of A. latronum latronum and A. latronum lijiangensis . Dorsal (upper left), ventral (upper right), and lateral (lower left). (A) A. latronum lijiangensis (Holotype). (B) A. latronum lijiangensis (paratype). (C) A. latronum latronum (topotype).

Discussion

A. latronum has previously been regarded as a valid species (Allen [1940]; Zimmermann [1962]; Corbet and Hill [1992]; Musser and Carleton [1993]; Musser et al. [1996]; Musser and Carleton [2005]), and in 2010, based on molecular data, Sakka et al. ([2010]) found that ‘the phylogeographic pattern observed in A. latronum demonstrated strong differentiated populations in the Sichuan and Yunnan regions’ (Sakka et al. [2010]). In the present study, the results of principal component analyses (Figures 2 and 3) showed that the specimens of A. latronum (obtained from 18 sampling areas in southwestern China) cluster into two distinct geographic groups: A. latronum lijiangensis subsp. nov. which includes the specimens from Weixi, Lijiang, and Binchuan areas of Yunnan Province and A. latronum latronum which includes the specimens from Sichuan (Markam, Baoxing, Yajiang, Luhuo, Baiyu, Batang, Danba, Daocheng, Kangding, and Muli areas), southeastern Tibet (Mangkang, Zuogong, and Bomi areas), and northwestern Yunnan (including Deqin and Xiaozhongdian areas) (Figure 4). The multiple comparisons of all 15 cranial variables also indicate significant difference between A. latronum latronum and A. latronum lijiangensis (Table 4). The results of our analyses indicate that aside from the difference in pelage characteristics, apparent geographic variations also exist in cranial morphology.

The results of the CD analyses indicate that four CDs exceeded 1.28 between A. latronum lijiangensis subsp. nov. and A. latronum latronum among these 15 cranial measurements (Table 5). According to the results of Mayr ([1969]), the CD on subspecific differentiation should be equal to or larger than 1.28 (Mayr [1969]); thus, we accept that A. latronum lijiangensis subsp. nov. and A. latronum latronum should be two valid subspecies within A. latronum. Furthermore, pelage comparison between A. latronum lijiangensis subsp. nov. and A. latronum latronum also sustains their valid subspecies status (Figure 5). Most interestingly thought is that the geographic distribution of the specimens examined in this study (Figure 4) indicates that the A. latronum lijiangensis subsp. nov. and A. latronum latronum are allopatric populations. Given this finding, we strongly suggest that A. latronum lijiangensis subsp. nov. should be considered as a valid subspecies of A. latronum.

The locality of the specimens examined in the study nearly covers the distribution areas of A. latronum (Figure 4). In terms of biogeography, this area is often ringed by high mountains and deep gorges as a result of the collision of the Indian and Eurasian plates. This area is thought to have the highest species diversity not only in China (Chen [2002]) but also in the world (Myers et al. [2000]). For endemic species in southwestern China, such as A. latronum, it is certainly plausible that the complicated and varied environments have promoted intra-specific geographic variation, and potentially, that the high mountains and deep gorges act as the main geographic barriers to insulate different geographic populations and influence their gene flow. Though an interesting possibility, any potential connection between the genetic makeup of varying subspecies and their biographic conditions is far outside the realm of a morphometric analysis such as this one. All the same, more studies, especially molecular data analysis, should be performed to reveal the phylogenies of different geographic populations and to probe into the relationships between subspecies differentiation of A. latronum and geographical evolution in these areas. Such a case study may strengthen other works on trying to elucidate the actual biogeography of the Yunnan Plateau and its surrounding areas.

Conclusion

In closing, there was one intriguing finding that bears some exploration. The mean values of the 15 measured skull variables (Table 6) show that on average, the variables of A. latronum latronum are larger than those of A. latronum lijiangensis subsp. nov., suggesting that A. latronum living in lower latitudes have smaller skulls compared to those who inhabit higher latitudes, a supposition fitting Bergmann's rule (Bergmann [1847]). However, an entirely different Apodemus species, A. chevrieri, displays an entirely converse size that Bergmann's rule would predict (Li et al. [2008]). Exactly why the subspecies we encountered in this study follow the general understanding we have of species at altitude but another species of Apodemus show a contradictory image of skull variable size is a fascinating question. Perhaps further genetic and ecological studies on A. latronum as well as its subspecies and other Apodemus species may offer some future insights.

Key to the subspecies of A. latronum

• 1. General dorsal brown, mixed with black more or less; the muzzle mainly grayish; GLS 26.56 (24.51~28.08 mm), UTR 12.73 (11.82~13.52 mm) ------------------------------------------------------------------------------------A. latronum lijiangensis subsp. nov.

• 2. General dorsal ochraceous or ochraceous buff, mixed with black more or less; the muzzle mainly ochraceous but slightly grayer; GLS 28.82 (27.26~30.67 mm), UTR 13.85 (13.20~14.71 mm) ----------------------------------------------------------A. latronum latronum

Appendix

A. latronum lijiangensis, N = 61

Yunnan Province: Lijiang (KIZ 840♂, 79817♀, 79822♀, 79844♀, 79860♀, 79861♀, 016460♂, 016476♂, 016479♀, 016490♀, 016491♀, 016511♀, 016513♀, 016514♂, 016518♂ (holotype), 016519♀, 016520♀, 016521♀, 016525♀, 016527♀, 016528♂, 016530♀, 016532♀, 016533♀, 016535♂, 016537♀, 016538♀, 016541♂, 016543♀, 016544♀, 016545♀, 016546♀, 016549♀ (paratype), 016552♂, 016553♂, 016554♂, 016556♂, 016558♀, 016560♀, 016562♂, 016564♂, 016581♂, 016582♀, 016584♀, 016585♀, 016596♂, 016831♀, 016832♀, 016833♀); Weixi (KIZ 1115♂,1116♀, 810557♀, 810558♀, 810560♀, 810575♀, 810576♀, 810596♀, 810598♀, 810617♀); Binchuan (KIZ 810672♂, 810675).

A. latronum latronum, N = 81

Yunnan Province: Xiaozhongdian (KIZ 810018♂, 810021♂, 810029♂, 810036♀, 810037♂); Deqin (KIZ 79510♂, 79511♂, 79512♂, 79513♀, 79521♂, 79529♂, 79553♀, 79554♂, 79619♀, 79647♂, 79670♂, 79671♂, 79685♀, 79686♂, 79696♂, 79708♂, 79713♀, 79738♂, 79767♀, 79784♂, 79803♀, 79804♂).

Sichuan Province: Muli (KIZ 855♂, 820845♂, 820847♀, 820879♀); (SAF 003). (IOZ 17063♂); Daocheng (KIZ 1050♀); Batang (KIZ 1060♀); Yajiang (SAF 6-06♂, 3-13♀); Baoxing (SAF JJSB209♀, JJSB189♂); Markam (IOZ 20074♀, 20076♂, 20079♂, 20090♀, 20097♂, 20098♂, 20101♂, 20112♂, 20115♂, 20116♂, 20118♂, 20119♂, 20121♀, 20136♀, 20161♀, 20165♀, 20167♀, 20166♀, 20169♀, 20170♀, 20172♀, 20174♀, 20175♀); Baiyu (SAF BY003); Luhuo (SAF0404♂); Danba (SAF 01-03♂, B-4-06); Kangding (KIZ 820404♀, no skull).

Xizang Province: Mangkang (KIZ 1074♀, 1083♂); (IOZ 31382♂, 31384♀); Zuogong (IOZ 31366♂, 31367♀, 31368♂, 31369♂, 31370♀, 31371♂, 31373♀, 31374♂, 31375♂); Bomi (IOZ 27047♂).