Документ взят из кэша поисковой машины. Адрес оригинального документа : http://zmmu.msu.ru/rjt/articles/ther6_1%20051_062%20Mishta.pdf Дата изменения: Sat Feb 2 18:26:50 2008 Дата индексирования: Tue Oct 2 01:52:11 2012 Кодировка: Windows-1251 Поисковые слова: coma

Russian J. Theriol. 6 (1): 5162

ї RUSSIAN JOURNAL OF THERIOLOGY, 2007

Morphometric variation of the common shrew Sorex araneus in Ukraine, in relation to geoclimatic factors and karyotype
Alina V. Mishta
ABSTRACT. Comparisons were made of chromosomal and macro-morphological variation in two widespread chromosome races of the common shrew Sorex araneus Linnaeus, 1758 in the Ukraine using multivariate statistics. Representatives of the Kiev and Neroosa races cannot reliably be identified from craniometric characteristics. Within the two chromosome races, however, we observed a relationship between morphology and geography. About 60% of morphological variance could be explained by the influence of geoclimatic factors. It was confirmed that ecogeography is more important than karyotype as a morphological determinant in the common shrew. Highly differentiated southern Ukrainian forms of S. araneus may be regarded as ecotypes. KEY WORDS: Sorex araneus, morphometrics, geoclimatic factors, chromosome races.
Alina V. Mishta [amishta@izan.kiev.ua], Schmalhausen Institute of Zoology, National Academy of Sciences of Ukraine, Bogdan Khmelnitskii str. 15, Kiev30 01601, Ukraine.

Морфометрическая изменчивость обыкновенной бурозубки Sorex araneus на территории Украины в связи с геоклиматическими факторами и кариотипом
А.В. Мишта
РЕЗЮМЕ. С помощью методов многомерной статистики выполнено сопоставление хромосомного и макро-морфологического уровней изменчивости у двух широко распространенных хромосомных рас обыкновенной бурозубки Sorex araneus Linnaeus, 1758. Оказалось, что представители хромосомных рас Киев и Нерусса почти неразличимы морфологически по комплексу краниометрических характеристик. В то же время, в пределах обеих хромосомных рас были отмечены общие тенденции географической изменчивости комплекса краниометрических признаков. Около 60% морфологической изменчивости может быть объяснено влиянием геоклиматических факторов. Подтверждено, что эко-географическая принадлежность является более значимым морфологическим детерминантом у обыкновенной бурозубки, нежели различия кариотипа. Высоко дифференцированные южные формы S. araneus рассматриваются как экотипы. КЛЮЧЕВЫЕ СЛОВА: Sorex araneus, морфометрия, геоклиматические факторы, хромосомные расы.

Introduction
The common shrew Sorex araneus Linnaeus, 1758 is known as one of the most variable species of the genus Sorex with respect to morphology and karyology. Intensive investigation of S. araneus karyology has revealed about 70 chromosome races in Europe and Siberia (Wуjcik et al., 2002). The extent to which karyological and morphological differentiation coincide has been studied in different regions over the vast species range (Hausser, 1984; Zima & Krбl, 1985; Searle & Thorpe, 1987; Hausser et al., 1991; Meyer & Searle, 1994; Chкtnicki et al., 1996; Wуjcik et al., 2000; Banaszek et al, 2002; Polyakov et al., 2002; Okulova et al., 2004). It has been demonstrated that differences in morphology do not conform to the geographical distribution of chromosome races of S. araneus (Zima & Krбl, 1985), and could be explained

rather by differences in environmental conditions (Sulkava et al., 1985; Wуjcik et al., 2000). In some cases however, significant morphological differences have been detected between neighbouring chromosome races (Chкtnicki et al., 1996; Polyakov et al., 2002). In Siberia, chromosome differentiation of the common shrew has been found to coincide with subspecies division (Polyakov et al., 2002; Okulova et al., 2004). In Ukraine, three chromosome races of S. araneus have been identified (Mishta et al., 2000). Two of them, the Kiev (XX/XY1Y2, af, bc, hi, g/m, k/o, n, p, q, r) and the Neroosa (XX/XY1Y2, af, bc, hi, g/o, k/r, m/n, p/q), occur respectively on the right and left banks of the Dnieper River. In 1913 Thomas described a subspecies of the common shrew from the Danube delta: S. araneus peucinius Thomas. This form was distinguished from other S. araneus subspecies in its extremely large hind leg,


#

A.V. Mishta To examine morphological differentiation in the common shrew over the southern part of Eastern Europe, I investigated a complex of craniometric features using multivariate statistics and placed the results in the context of geography and karyotype.

Material and methods
A set of 20 measurements of immature common shrews were taken on the right side of the skull (Fig. 1). In total 223 skulls from different localities in Ukraine, Moldova, Belarus and Romania were analysed (Tab. 1, Fig. 2). Measurements were made using a micrometer eyepiece scale (0.01 mm) on a binocular microscope. Statistical treatment. For the analysis of mandible measurements, I used univariate and multivariate statistics. The influence of external factors on morphological variability (geographic origin, affiliation to the chromosome race if known and sex) was determined with univariate ANOVA. All external factors were considered as fixed. To account for correlated characters, trends of variation and differentiation in S. araneus were explored using principal component analysis (PCA). Eigenvectors and eigenvalues were calculated on the basis of the variance-covariance matrix. In order to test whether the intraspecific variance was distributed randomly for all the principal components obtained, the part of this variance accounted for

Figure 1. Measurements performed on skull and mandible.

1 condylobasal length; 2 breadth between lacrimal foramina; 3 maxillary breadth; 4 interorbital breadth; 5 cranial breadth; 6 palatal length; 7 postglenoid width; 8 length of the upper incisor; 9 length of the upper antemolars; 10 length of the upper molariform teeth; 11 length of the lower incisor; 12 length of the mandibular toothrow (except incisor); 13 length of m1m3; 14 length of m1; 15 length of the mandible (to the end of mandibular condyle); 16 distance from m1 to mental foramen; 17 distance from i1 to mental foramen; 18 length of the horizontal branch of the mandible; 19 height of the mandible; 20 breadth of the coronoid process.

peculiarities of its dentition and its monochromatic winter fur colour (Thomas, 1913, cited in Ognev, 1928). Later, another southern subspecies of S. araneus was described from the Dnieper delta S. araneus averini (Zubko, 1936, cited in Abelentsev et al., 1956). However, since then it has been regarded as pointless to try to classify shrews from the lower reaches of the Dnieper River into subspecies until proper taxonomic studies are conducted, given that all of the animals are generally the same size as S. a. peucinius (Migulin, 1938; Abelentsev et al., 1956). Thus, investigators have not reached a consensus about the variability and differentiation of S. araneus over the territory of the European part of the Former Soviet Union, and the Ukraine in particular. Some authors (Ognev, 1928; Migulin, 1938; Abelentsev et al., 1956; Mezhzherin et al., 1984) accept the existence of different subspecies in Ukraine, others (Gureev, 1957; Dolgov, 1985) do not attach importance to any differences observed, regarding variability in S. araneus as continuous.

Figure 2. Geographical origin of the material (localities numbered as in Tab. 1). Kiev chromosome race black circles; Neroosa chromosome race white circles; uncertain race affiliation grey circles.


Morphometrics of common shrews in Ukraine

#!

Table 1. Geographical origin and other details of samples (samples listed from north to south).
No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Sam ple name Belaru s C hern yg iv Voly n Sum y Z hy tomir Kiev Kan iv Kh ark iv Khm elnitsk y Kir ovo gr ad Lu gans k Karp aty Don etsk M oldo va Kh er so n Golaya Pr istan' Odessa Ro mania Geog raph ical o rig in Belar us : Gom el Pr ov ince, Abak umy v ic., Lo ev Ukr aine: C her nyg iv Pro vin ce, Ko zeletsk Dis trict Ukr aine: Voly n' Pro vin ce, Shatsk vic. Ukr aine: Sumy Pr ov ince, Yamp il Distr ict, L eb edy n Dis trict Uk rain e: Z hyto mir Pro vince Uk rain e: Kyiv v ic., Ter emk y Ukr ain e: Ch erk assy Pr ovin ce, Kan iv State Reser ve Uk rain e: Khar kiv Pr ov ince, Z mijiv vic. Uk raine: Khmeln ytsk y Pro vin ce, Volo chyn sk y Dis trict Uk rain e: Kiro vo gr ad Pro vince, Z nam en sk y Dis trict Ukr aine: L ug an'sk Pr ov ince, Krem en ets Dis trict Uk r ain e: Tr anscar pathian s, Rak hiv Distr ict, Petro s M t. ; T iach evsk y Distr ict; I vano -F rank ivsk Pr ov ince, Nadv or na Dis trict Uk raine: Donets'k Pro vince, Atamans'ka Village M oldo va: Kag ul vic. , M anta L ake Ukr aine: Kh erso n Pr ovin ce, Gola Pr ys tan ' Distr ict, I vano - Ry balche Ukr aine: Kherson Pr ovin ce, Gola P ry stan' Dis trict, Velyk yi Potiomkin sk y Island Uk rain e: Odessa Pr ovin ce, Vilko vo v ic. , islands o f Kilijskay a delta Romania: Dan ube delta

N
10 14 4 9 11 20 20 9 5 4 20 20 4 20 12 22 25 20

R ace NE KI KI NE KI NE KI NE KI NE NE KI

Co llectio n DE B K SU KSU K SU NHM DE B, KSU M SU, DE B M SU NHM NHM K SU DE B, USU, KSU M SU CМ KSU, M SU K SU CF CF

Race: NE Neroosa; KI Kiev DEB collection of the Department of Ecology and Biogeography of Schmalhausen Institute of Zoology NAScU; KSU Zoological Museum of Kiev State University; MSU Zoological Museum of Moscow State University; USU Zoological Museum of Uzhgorod State University; MNH National Museum of Natural History of Ukrainian Academy of Sciences Ukraine; CM private collection of Mikhailenko A.G.; CF private collection of Fedorchenko A.A.

by each component was compared to a broken stick distribution using the Kolmogorov-Smirnov goodness of fit test. This theoretical distribution specifies the distribution of a finite quantity partitioned at random (see Hausser, 1984). To test whether adaptations to local environmental conditions play an important role in morphological differentiation, a multiple regression analysis of each morphological variable against nine geoclimatic variables was carried out. Residuals obtained were also analysed using principal component (PCA) and discriminant function (DFA) analyses. Interdependence between geoclimatic factors and individual morphological characters was studied using multiple regression analysis. The geoclimatic factors used in the study are presented in Tab. 2. Climatic data from the weather stations nearest each sample point were taken from the Geographical Encyclopaedia of Ukraine (1989, 1990, 1993). The interdependence of two multivariate data arrays of craniometric and geoclimatic variables was determined by use of canonical correlation analysis.

Statistical analysis was performed using the Statistica 6.0 package (StatSoft Inc., 1984-2001, USA). Chromosome races: 1. Kiev race (XX/XY1Y2, af, bc, hi, g/m, k/o, n, p, q , r)
Table 2. Geoclimatic factors used in the study.
Ch ar acter istic L atitu de Lo ngitu de Altitu de Av er age tem per atu re of J an uar y Aver age tem peratu re of Ju ne Av er age an nu al r an ge of temp eratur es Av erage annu al n umb er of d ays with temper atur e exceed 10њ C Av er age sno w dep th Averag e ann ual p recipitatio n Abb reviatio n LAT LONG ALT TE MP_ J TEMP_I TEMP_D W_ DAYS SNOW RAIN


#"

A.V. Mishta

2. Neroosa race (XX/XY1Y2, af, bc, hi, g/o, k/r, m/n, p /q ) The details of racial distribution and chromosome polymorphism in the study area have been described elsewhere (Mishta et al., 2000). The approximate geographic distribution of the two races in the region studied is shown in Fig. 2. Samples used to analyse the relationships between karyotypic and macro-morphological differentiation are presented in Tab. 1. I analysed material from localities where karyotyping took place or from localities situated in core of the known race distribution. Material of uncertain racial affiliation was excluded from the analysis. For convenience of analysis, all populations were conditionally separated into northern and southern subgroups. Populations of Moldova, Odessa, Romania, Golaya Pristan and Kherson were classed into a southern subgroup and the remaining populations were considered northern. Populations were also simplistically divided into western and eastern subgroups, which approximates with the chromosome races. In the western subgroup representatives of the Kiev race (Zhytomir, Kiev, Khmelnitsky, Moldova, Odessa) as well as populations of uncertain racial affiliation (Kirovograd, Kaniv, Romania) are included. The eastern subgroup includes representatives of the Neroosa race (Sumy, Kharkiv, Lugansk, Kherson, Golaya Pristan) and also the Belarus and Chernigiv populations of uncertain racial affiliation.

Results and discussion
Mean values and standard deviations of 20 craniometric variables of S. araneus are presented in Appendix 1. Univariate ANOVA of all craniometric characters did not reveal differences in mandible size between males and females, except for interorbital breadth (p<0.05). At the same time it was demonstrated that differentiation of populations in the set of characters studied was best associated with geographical origin (for 19 craniometric characters out of 20: p<0.001; except for the distance between i1 and the mental foramen). Trends in craniometric variation in the common shrew.To study more precisely the geographic differentiation of S. araneus populations, comparative analysis of shrews from different regions of Ukraine and neighbouring territories was carried out using principal components analysis. We analysed six principal components (PCs), which altogether describe 77.19% of variance in the dataset. The positive end of each PC has shrews with the largest skulls, and the negative values are associated with the smallest skulls. Component loadings for craniometric characters of S. araneus are presented in Appendix 2. The PC1 explains 51.54% of total variance. The most important characters for separating S. araneus populations on this component are condylobasal length, length of the mandible to the end of mandibular condyle, length of the horizontal branch of the mandible, palatal

Figure 3. Geographical variation of normalized values of principal components: original morphological variables (A); their residuals after removing the effect of nine geoclimatic variables (B). Samples are ranked from left to right according to their geographical location from north to south and marked with colour according to their racial affiliation (see Fig. 2). Kiev chromosome race black; Neroosa chromosome race white; uncertain race affiliation grey.

length, length of the upper molariform teeth, length of the mandibular toothrow (except incisor), length of m1m3 and height of the mandible. Maxillary and cranial breadth and the length of m1 contribute a smaller loading. PC1 had a positive loading on all characters, which means that positive component values correspond to larger variable values and that PC1 is essentially a size component. Fig. 3A shows mean PC1 scores for samples ordered from left to right according to their geographic location from north to south. One can see that within the northern subgroup of populations (Belarus-Lugansk), the difference between mean values of the component is not significant. Populations from islands in the Danube delta (Romania, Odessa) and the Dnieper delta (Kherson) are substantially differentiated from populations of northern Ukraine. The Moldova population and the continental population of the Kherson region (Golaya Pristan) have an intermediate position. Thus, the magnitude of the craniometric charac-


Morphometrics of common shrews in Ukraine ters of S. araneus (determined by PC1) increases in a southerly direction. In addition to this, differentiation between neighbouring populations is more noticeable in the south. It is evident that populations of S. araneus are to a great extent differentiated by PC1, which represents general proportions of the skull. In both the Neroosa (Sumy, Kharkiv, Lugansk, Golaya Pristan, Kherson) and Kiev (Zhytomir, Kiev, Khmelnitsky, Karpaty, Moldova, Odessa) chromosome races there are parallel trends of geographical variation (Fig. 3A). PC2 explains 6.94% of the total variance. The largest contributions to the second component are made by the length of the upper incisor, the distance from i1 to the mental foramen and the length of the lower incisor. Thus, this component reflects the relative length of incisors against the distance from i1 to the mental foramen. The mean values do not vary systematically by geographical location. Mean values of this PC demonstrate large differences among shrews from western Ukraine: Khmelnitsky-Karpaty. Shrews from the Karpaty population are characterized by relatively long incisors and a very small distance from i1 to the mental foramen and shrews from Khmelnitsky show the reverse. PC3 explains 5.54% of the total variance. The largest contribution to the third component is made by the distance from i1 to the mental foramen, the breadth between the lacrimal foramina and the interorbital breadth. On this component, populations from eastern Ukraine are slightly differentiated from populations of western Ukraine. The mean values of this component also suggest a slight differentiation of the Moldova population within the western group. PC4 explains 4.94% of the total variance. The largest contributions in the values of the fourth component are made by the distance from m1 to the mental foramen and the breadth of the coronoid process. In this case there is also a slight differentiation between representatives of western and eastern Ukraine. The mean values of this PC suggest a slight differentiation of the Kherson population within the eastern group. PC5 explains 4.48% of the total amount of morphological variance. The mean values do not vary systematically by geographical location. The largest contribution in the values of this component is made by the length of the upper incisor. PC6 accounts for only 3.76% of the variation. There was a substantial loading on the postglenoid width. The mean values of this component suggest a separation of the Khmelnitsky population. To consider again the best example of geographic variation in PC scores, the increasing scores from north to south in PC1 (Fig. 3A) suggest that the main cause of morphological differentiation in the Ukraine is adaptation to the local environment. To test this hypothesis, a multiple regression analysis of each morphological variable against the nine above mentioned geoclimatic characteristics was performed to obtain residual morphological variables uncorrelated with these geoclimatic

##

Figure 4. Distribution of the intraspecific variance of the principal components computed on the original morphological characters and their residuals after removing the effect of nine geoclimatic variables. Principal components are ranked following their decreasing contribution to intraspecific variance.

characteristics. In Fig. 4 the partitioning of the total intraspecific variance among the untransformed principal components is compared with their residuals after removing the effect of the nine geoclimatic factors. A Kolmogorov-Smirnov test indicates that the first of these two distributions shows departure from normality (p<0.01), while the second one does not differ significantly from expectation (p<0.1). Furthermore, the residuals no longer show the same geographical pattern as the untransformed principal component scores (Fig. 3B). So, the general geographic trend of the complex of correlated craniometric characters within Ukrainian common shrews has been revealed. Overall, with some exceptions, these characters show a latitudinal gradient. Contributions of geoclimatic characteristics in differentiation of the common shrew. Geographic variation in the complex of craniometric variables studied was greater in southern regions of Ukraine, where environmental pressure is greatest (Geographical Encyclopaedia of Ukraine, 1989, 1990, 1993). The morphological clines, which are expected from the gradual geoclimatic clines throughout the studied area, are best displayed by means of a canonical correlation analysis. Correlations between the 20 macromorphological variables and 9 geoclimatic characteristics were obtained. The relationships between the canonical functions of the craniometric and the geoclimatic variables are presented in Tab. 3 and the loadings for these functions are presented in Appendix 3. The part of the total morphological variance explained by the geoclimatic function is 71.4%. As shown in Tab. 3, the canonical correlation coefficients between the first, second, third and fourth pairs of canonical functions are highly significant. This implies that at least 60.9% of variance of the complex of craniometric


#$

A.V. Mishta
Table 3. Relationships between the canonical functions of craniometric and geoclimatic variates.
Pair of Can onical fu nction s 1 2 3 4 5 E igen value 0. 77 4 0.42 5 0.35 2 0.29 6 0.22 4 Canon ical cor r elatio n R 0. 87 9 0. 65 1 0. 59 3 0. 54 4 0. 47 3
?
2

Var ian ce exp lained (%)

df 18 0 15 2 12 6 10 2 80

p
<0 .0 00 1 <0 .0 00 1 <0 .0 00 1 <0 .0 00 1 < 0. 00 1

MORPH 48. 2 6 4.1 0 3.8 7 2.5 7 2.0 8

GEOCL 50. 8 6 7.6 9 13. 0 8 3.0 0 11. 0 2

7 15. 9 2 4 07.8 9 2 93.4 1 2 03.5 7 1 30.8 0

variables is highly correlated with geoclimatic variation. The first morphological variate accounted for 48.26% of the total morphological variance, 50.86% of which can be attributed to the geoclimatic variables. The remaining four functions explained only a lower percentage of the morphological variability (Tab. 3). The contributions of different geoclimatic variables (Appendix 3b) demonstrate that the average annual range of temperatures has the greatest influence on the craniometric complex factor MORPH1 (Tab. 3). Those most related to the temperature factor variables are condylobasal length, maxillary breadth, palatal length and overall length of mandible (Appendix 3a). The canonical variate analysis revealed that canonical functions of morphological and geoclimatic variates correlate significantly and obvious clinal variation of the complex of craniometric characteristics exists. The combination of larger dimensions of craniometric features towards the south together with separation of S. araneus populations in the south suggests that adaptation to local environmental conditions could explain macro-morphological differentiation in S. araneus. To verify this hypothesis, I carried out a multiple regression analysis of the set of variables (20 craniometric and 9 geoclimatic characteristics taken together). Affiliation to the southern or northern subgroup was used as the dependent variable. Separation into southern and northern subgroups is significantly associated with the whole complex of geoclimatic factors (for all characteristics excluding altitude p<0.0001) (Tab. 4). Among the craniometric characteristics included in the model we can see influences from postglenoid breadth (#7, p<0.001), condylobasal length (#1, p<0.05), maxillary breadth (#3, p<0.05), palatal length (#6, p<0.05), length of the mandibular toothrow (#12, p<0.05), distance from m1 to mental foramen (#16, p<0.05) and breadth of the coronoid process (#20, p<0.05). Macro-morphological differences between shrews of different chromosome races. The relationships between chromosome races of the common shrew in Europe as a whole and in Eastern Europe in particular is a subject of wide speculation (Searle, 1984; Wуjcik, 1993; Searle & Wуjcik, 1998; Orlov et al., 2004). No doubt, as our knowledge of the diversity of chromosome races increases, it will be necessary to

revise previous phylogenetic schemes. The chromosome races under study likely represent different subgroups within the previously described West European Karyotypic Group (WEKG); the Neroosa race can be placed in a new group, which would be closely united with chromosome races from East Europe. The relationships between chromosome races of the common shrew in the region studied and in adjacent territories will be the subject of a later publication (work in progress). Univariate ANOVA showed that six characters out of 20 differ significantly between chromosome races:
Table 4. Craniometric and geoclimatic characteristics which determine division to southern and northern subgroups in the common shrew (R=0.98, R2 =0.97, p<0.0001).
Var iable In ter cept TE M P_D 12 W_ DAYS 1 7 TE MP_ J TEMP_I RAIN LONG 3 16 20 SNOW 6 ALT 8 19 13 18 11 15 B 1.340 0.298 0.15 9 0.008 0.04 8 0.09 0 0.49 9 0.37 8 0.00 1 0.02 0 0.141 0.18 6 0.13 8 0.006 0.093 0.04 7 0.09 4 0.06 1 0.12 6 0.08 5 0.06 4 0.046

t

201

p
<0 .0 00 1 <0 .0 00 1 <0.0 5 <0 .0 00 1 <0.0 5 <0.00 1 <0 .0 00 1 <0 .0 00 1 <0 .0 00 1 <0 .0 00 1 <0 .0 00 1 <0.0 5 0.05 4 <0 .0 00 1 <0.0 5 <0.0 5 0.21 5 0.16 5 0.13 1 0.06 0 0.14 5 0.27 1

4.340 11.65 7 2.10 7 1 2.83 8 2.31 0 3.20 5 15.096 15.421 5.25 9 4.45 9 3.514 2.25 8 1.93 1 4.222 3.038 2.12 2 1.24 3 1.39 1 1.51 5 1.88 9 1.46 1 1.103


Morphometrics of common shrews in Ukraine
Table 5. Classification results of the discriminant analysis of the 20 morphological variables and the residuals remaining after regression of the morphological variables by the geoclimatic factors.
Pr ed icted r ace mem bership ( % and n um ber) Actu al r ace M or ph olog ical var iables % Kiev ( n=1 01) Nero osa ( n= 45) Total 92. 079 76. 363 8 6. 54 2 Kiev Nero osa 93 13 106 8 42 50 % 9 1. 08 9 40. 000 73. 077 Resid uals Kiev 92 33 12 5 Nero osa 9 22 31

#%

length of m1 (p<0.001), length of m1m3 (p<0.01), distance from m1 to mental foramen (p<0.01), interorbital breadth (p<0.01), maxillary breadth (p<0.05) and length of the upper antemolars (p<0.05). Only one of these characteristics distance from m1 to mental foramen does not have a longitudinal geographic trend, so it could truly reflect genetic differences between the races. At the level of chromosome differentiation here (at least three chromosome rearrangements), morphological differences between representatives of two chromosome races of S. araneus were found to be insignificant. Only one measurement could be linked to genetic differentiation; the rest are associated with geographical variation and have a high correlation with geoclimatic factors. This statement is confirmed by discriminant function analyses of 20 morphological variables and their residuals after regression by geoclimatic ones. In the first case analysis led to the correct classification in approximately 87% of cases (Tab. 5). Specimens of the Kiev race were correctly classified in 92% of cases, and those from the Neroosa race in 76% of cases. The discriminant function explained 4.43% of the total morphological variance between the two studied chromosome races of shrews. The value of ч2 test obtained by transformation of Wilks lambda was statistically significant (ч2=148.53, p<0.0001). Standardized coefficients of the discriminant function for morphological variables were established. Variables such as maxillary breadth, interorbital breadth, length of the upper incisor, length of the upper antemolars, length of m1-m3, length of m1, distance from m1 to mental foramen, length of the horizontal branch of mandible and breadth of the coronoid process had important contributions for the discriminant function. Thus, it can be suggested that these variables account for most differences in mandible morphology between the two chromosome races. At the same time, the discriminant function analysis of residuals enables us to obtain a clear classification of individuals according their racial affiliation: ч2=21.63, p=0.36 (Tab. 5). In this case, only 73% of individuals were assigned to their actual chromosome races correctly.

So, our results do not show clear relations between morphometric variation and racial affiliation in Ukrainian S. araneus. Differences observed in this study may represent variation among western and eastern groups of populations rather then between representatives of different chromosome races. Numerous previous investigations have demonstrated that the differentiation of karyotypes between populations from different geographical localities does not correlate with phenotypic variation (Zima & Krбl, 1985; Searle & Thorpe, 1987; Meyer & Searle, 1994; Wуjcik et al., 2000). Even shrews from the Scottish island of Islay, which could be distinguished by nonmetrical (karyological and genetic) techniques, could not be separated from other British S. araneus by metric features (Meyer & Searle, 1994). Our results confirm the statement of Wуjcik et al. (2000) that karyotypic divergence does not play an important role in differentiating skull morphology among chromosome races of S. araneus of a given region, and that geography is more important than karyotype as a morphological determinant. However, it has also been demonstrated (Chкtnicki et al., 1996; Polyakov et al., 2002; Okulova et al., 2004) that when the level of karyotype differentiation is higher, morphological differences between chromosome races of S. araneus become more visible. Southern populations of the common shrewand possible causes of their macro-morphological distinctiveness. In the last century, numerous subspecies of S. araneus have been distinguished in Europe, which Zalesky (1948) divided into two main groups: araneus in the north and tetragonurus in the south. The basis of this division was a difference in general body size and peculiarities of fur coloration. The distribution range of the southern forms roughly coincides with the area occupied by vineyards. Notwithstanding the fact that several of populations included in the tetragonurus group are now recognized as different species (S. coronatus, S. antinorii, S. granarius, S. arunchi and S. samniticus), the basic idea of the division of S. araneus into these two groups remains subject to scrutiny. According to Niethammer & Krapp (1990) the description of numerous subspecies within the tetragonurus group and in marginal regions of the global distribution of the araneus group could be explained both by altered ecological conditions and above all by isolation in mountain regions. It is worth noting that only one northern European subspecies has been described: S. araneus bergensis Miller from Norway, which was isolated during the last glaciation. Dehnel (1950) considered the central and northEuropean populations of S. araneus to be genetically connected to forms that persisted in the glacial periods near ice-sheet borders. Common shrews from southern Europe lost contact with the northern group probably in the last glacial period, owing to the zone of dry and cold steppe stretching across Europe from East to West and reaching almost to the Pyrenees. The southern group retreated to the southern edge of this belt, subsisting in


#&

A.V. Mishta Southern regions of Ukraine, in which the southern populations in this study were obtained, represent the Prichernomorie zoogeographical province of the Palearctic steppe subregion. The province is divided into two zones. The first zone includes river valleys, seacoasts and islands (hydrophylous communities); the second one is field-steppe with artificial and natural islands of forests, forests belts and orchards (xerophilous communities) (Gizenko, 1965). So, in these regions the common shrew occurs specifically in river estuaries reeds and marshes, surrounded by typical steppe vegetation. Due to this fact, and under the influence of specific climatic factors, southern shrews have acquired macro-morphological differences from northern shrews. Moreover, in this steppe southern limit of the species range, wet habitats to which the common shrew is confined are mainly isolated from each other. The morphological differentiation between neighbouring populations in the southern regions may be explained by this relative isolation of their habitats.

the Pyrenees Mountains and their foothills in southern France, in the southern Alps and in the Balkans. These conditions of isolation during last glaciations were considered to generate the two different varieties of S. araneus the northern araneus type and the southern tetragonurus type (Dehnel, 1950). However, our data do not support the idea that common shrews of northern and southern groups in the Ukraine represent different stocks with independent postglacial histories. It is known, that during the last glaciation maximum, a number of forest refugia with broad-leaved species remained in the central part of the Russian Plain (4951њ N), in the middle reaches of the Dniester and Don Rivers, in Moldova and near the Azov Sea (Simakova, 2001). Mishta (2005) assumed that colonization of northern and central parts of Eastern Europe could take place from different refugia located in the mid-continental part of the region rather than from the Mediterranean area or Crimea. The recent distribution of widespread chromosome races of the common shrew in Eastern Europe and analysis of environmental conditions during the last glacial period support the idea of the existence of such refugia (Markova et al., 2001; Orlov & Kozlovskii, 2002). In the case of the Kiev and Neroosa races, it is likely that they were segregated before the last glacial maximum and maintained during glacial periods somewhere in refugia on either side of the Dnieper River (for example in the middle reaches of the Dniester and Don Rivers). After the glacial retreat, representatives of both races managed to occupy vast territories of Eastern Europe. It is noteworthy that the Dnieper River served as a natural barrier in the way of further expansion of both races. The scenario of colonization of S. araneus in Eastern Europe and relationships between neighbouring races will be the subject of a future publication. In the course of our investigation, we observed parallel geographic variation within two chromosome races of S. araneus, with increasing size in a southerly direction, along which there is a steep geoclimatic gradient. Moreover, this pattern has been found in other representatives of Soricidae (Mishta, 1997; Ochociсska & Taylor, 2003). Modern analyses of geographic variation in a variety of species have demonstrated that subspecies designations are often dubious (Thorpe, 1975a, b, 1976, 1980, 1981a, b; Clover, 1979; Smith, 1979; Lуpez-Fuster & Ventura, 1987; Panteleev et al, 1991; Panteleev, 1996). Many subspecies have been defined according to differences in exterior features only, without application of genetic criteria (because such techniques were not available at time of description). If such subspecies actually only differ in some features (such as body size) because of eco-geographical selection pressures, such taxa should be treated as ecotypes (Panteleev et al., 2000). As already demonstrated by comparison of condylobasal length in common shrews from many localities over its wide range, individuals are smaller in cold areas and, in general, larger in environments with high evapotranspiration (Ochociсska & Taylor, 2003).

Conclusion
No substantial differences in craniometric characters were revealed among the two main chromosome races of common shrews in the Ukraine. At the same time, within each chromosome race parallel clines of geographic variation can be seen. The extent of morphological differentiation of the common shrew populations is greater in the south. In other words, neighbouring groups of animals from northern, central, eastern and western Ukraine are less different than those in the south. Intraspecific differentiation in the common shrew is significantly determined by high ecological plasticity and represents the response of the species to geoclimatic factors. Of top priority are temperature (average annual temperature of January and June, annual range of temperatures and annual number of days with temperature exceed 10њC) and precipitation (average snow depth, average annual precipitation). I consider it unhelpful to recognise separate subspecies in southern Ukraine and regard the different southern forms as ecotypes. To avoid mixing concepts of geographical and racial differentiation, I recommend further investigations of macro-morphological variation and differentiation of S. araneus belonging to different chromosome races taking into consideration geoclimatic variation (especially by comparison of chromosome races which have an extensive range or with varied geographical origins). On the other hand, investigations on a smaller geographical scale may be more appropriate in studies of macromorphological difference between chromosome races.
ACKNOWLEDGMENTS. I am very grateful to the staff of Zoological Museum of Kiev State University, Zoological Museum of Moscow State University, Zoological Museum of Uzhgorod State University, National Museum of Natural History of Ukrainian Academy of Sciences, Department of


Morphometrics of common shrews in Ukraine
Ecology and Biogeography; and personally to Vitalii Mezhzherin, Sergei Mezhzherin, Igor Zagorodniuk, Alexander Fedorchenko and Alexander Mikhailenko for the opportunity to work with their collections. I wish to thank David Polly and Jeremy B. Searle for checking English, Igor Dzeverin for giving valuable suggestions and two anonymous reviewers for critical comments on the manuscript.

#'

References
Abelentsev V.I., Pidoplichko G.G. & Popov B.M. 1956. [Order Insectivora] // [Mammals. Fauna of Ukraine]. Kyiv: Vydavnytstvo Akademii Nauk URSR. P.79256 [in Ukrainian]. Banaszek A., Smakulska J., Fedyk S., Jadwiszczak K.A. & Chкtnicki W. 2002. Morphometric differentiation of shrews (Sorex araneus L., 1758) from the hybrid zone between the Guzowy M?yn and ?egucki M?yn chromosome races in Poland // Mammalia. Vol.67. P.217224. Brьnner H., Lugon-Moulin N., Balloux F., Fumagalli L. & Hausser J. 2002. A taxonomical re-evaluation of the Valais chromosome race of the common shrew Sorex araneus (Insectivora: Soricidae) // Acta Theriologica. Vol.47. P.245275. Chкtnicki W., Fedyk S., Banaszek A., Sza?aj K.A. & Ratkiewicz M. 1996. Morphometrical characteristics of the common shrew (Sorex araneus L.) from interracial hybrid zones // Hereditas. Vol.125. P.201207. Clover R.C. 1979. Phenetic relationships among populations of Podarcis sicula and P. melisellensis (Sauria: Lacertidae) from islands in the Adriatic Sea // Systematic Zoology. Vol.28. P.284-298. Dehnel A. 1950. [Studies on the genus Neomys Kaup] // Annales Universitatis Marie Curie-Sk?odowska. Sectio C. Lublin - Polonia. Vol.5. P.163 [in Polish, with English summary]. Dolgov V.A. 1985. [Red-toothed Shrews of the Old World]. Moskva: Izdatelsvo Moskovskogo Gosudarstvennogo Universiteta [in Russian]. [Geographical Encyclopedia of Ukraine]. 1989. T.1: AZH. Kyiv: Vydavnytstvo Ukrainska Radianska Encyklopediya imeni P. Bazhana [in Russian]. [Geographical Encyclopedia of Ukraine]. 1990. T.2: ZO. Kyiv: Vydavnytstvo Ukrainska Radianska Encyklopediya imeni P. Bazhana [in Russian]. [Geographical Encyclopedia of Ukraine]. 1993. T.3: LYA. Kyiv: Vydavnytstvo Ukrainska Radianska Encyklopediya imeni P. Bazhana [in Russian]. Gizenko A.I. 1965. [Ecology of mammals of steppe regions of Ukrainian Prichernomorie] // Materialy Zoologicheskogo Soveshchaniya po Probleme Biologicheskiye osnovy rekonstruktsii, ratsionalnogo ispolzovaniya i ohrany fauny yuzhnoi zony Evropeiskoi chasti SSSR. Kishinev. P.4145 [in Russian]. Gureev A.A. 1971. [Shrews (Soricidae) of the World Fauna]. Leningrad: Nauka [in Russian]. Hausser J. 1984. Genetic drift and selection: their respective weights in the morphological and genetic differentiation of four species of shrews in Southern Europe (Insectivora, Soricidae) // Zeitschrift fьr Zoologische Systematic und Evolutionsforschung. Vol.22. P.302320.

Hausser J., Bosshard F., Taberlet P. & Wуjcik J. 1991. Relationships between chromosome races and species of Sorex of the araneus group in the western Alps // Mйmoires de la Sociйtй Vaudoise des Sciences Naturelles. Vol.19. P.7995. Lуpez-Fuster M.J. & Ventura J. 1987. Estudio morfometrico le Sorex coronatus Millet, 1828 (Mammalia, Insectivora) en el notre de la Peninsula Iberica y sur de Francia // Sans-Coma V., Mas-Coma S. & Gosбlbez J. (eds.). Mamiferos y Helmintos. Volumen homenaje al Prof. Dr. H. Kahmann en su 81 aniversario. Barcelona: Editorial Ketres. P.5564. Markova A.K., Simakova A.N., Puzachenko A.Yu. & Kitaev L.M. 2001. Eastern European mammoth distribution and environments during the Middle Valdai Briansk Interstade (33,000-24,000 BP) // Cavarretta G., Gioia P., Mussi M. & Palombo M.R. (eds.). Proceedings of the 1st International Congress The World of Elephants. Rome. P.299304. Meyer A. & Searle J.B. 1994. Morphological studies on British common shrews // Folia Zoologica. Vol.43. Suppl.1. P.115. Mezhzherin V.A., Kuruts N.V., Mikhalevich O.A. & Revenko N.G. 1984. [The population levels of morphological differences as steps of evolutionary transformation] // Zhurnal Obschei Biologii. T.45. P.306317 [in Russian, with English summary]. Migulin O.O. 1938. [Mammals of URSR]. Kyiv: Vydavnytstvo AN URSR [in Ukrainian]. Mishta A. 1997. [Karyotypic and morphological differentiation of shrews genus Sorex L., 1758 (Insectivora, Soricidae) of Ukrainian fauna]. PhD thesis. Kiev: Schmalhausen Institute of Zoology. 22 p. [in Ukrainian, with English summary]. Mishta A.V., Searle J.B. & Wуjcik J.M. 2000. Karyotypic variation of the common shrew Sorex araneus in Belarus, Estonia, Latvia, Lithuania and Ukraine // Acta Theriologica. Vol.45. Suppl.1. P.4758. Mishta A. 2005. Late Pleistocene glacial refugia in Eastern Europe and postglacial recolonization of the region by shrews // Abstracts of Plenary, Symposium, Poster and Oral Papers presented at Ninth International Mammalogical Congress (IMC9). July 31August 5, 2005. Sapporo, Hokkaido, Japan. P.7071. Niethammer J. & Krapp F. (ed.) 1990. Handbuch der Sдugetiere Europas. Bd 3/1. Insektenfresser, Primaten. Wiesbaden: AULA Verlag. Ochociсska D. & Taylor J.R.E. 2003. Bergmanns rule in shrews: geographical variation of body size in Palearctic Sorex species // Biological Journal of the Linnean Society. Vol.78. P.365381. Ognev S.I. 1928. [The Mammals of the Eastern Europe and the Northern Asia]. Vol.1. Moskva-Leningrad: Izdatelstvo Glavnauka. [in Russian]. Okulova N.M., Balakirev A.E. & Orlov V.N. 2004. [Craniometrical characteristics of some Sorex araneus (Insectivora) chromosomal races] // Zoologicheskii Zhurnal. Vol.83. P.14761487 [in Russian, with English summary]. Orlov V.N., Bulatova N.Sh., Kozlovskii A.I. & Balakirev A.E. 2004. [Hierarchy of intraspecific taxa of the common shrew, Sorex araneus (Insectivora), and taxonomic struc-


$

A.V. Mishta
nal of Mammalogy. Vol.60. P.705722. Sulkava S., Vahtola M. & Fredga K. 1985. Structure of the upper tooth-row of Sorex araneus in Scandinavia // Acta Zoologica Fennica. Vol.173. P.237239. Thorpe R.S. 1975a. Quantitative handling of characters useful in snake systematics with particular reference to intraspecific variation in the ringed snake Natrix natrix L. // Biological Journal of Linnean Society. Vol.7. P.2743. Thorpe R.S. 1975b. Biometric analysis of incipient speciation in the ringed snake Natrix natrix L. // Experientia. Vol.31. P.180182. Thorpe R.S. 1976. Biometric analysis of geographic variation and racial affinities // Biological Review. Vol.51. P.407452. Thorpe R.S. 1980. Microevolution and the taxonomy of European reptiles with particular reference to the grass snake Natrix natrix L. and the wall lizards Podarcis sicula and P. melisellensis // Biological Journal of the Linnean Society. Vol.14. P.215233. Thorpe R.S. 1981a. The morphometrics of the mouse: a review // Symposia of the Zoological Society of London. Vol.47. P.85125. Thorpe R.S. 1981b. Racial divergence and subspecific status of the Gotland grass snake: a comment on Nilson and Andrens paper // Zoological Journal of the Linnean Society. Vol.72. P.369370. Wуjcik J.M. 1993. Chromosome races of the common shrew Sorex araneus in Poland: a model of karyotype evolution // Acta Theriologica. Vol.38. P.315338. Wуjcik J.M., Bogdanowicz W., Pucek Z., Wуjcik A. M. & Zalewska H. 2000. Morphometric variation of the common shrew Sorex araneus in Poland, in relation to karyotype // Acta Theriologica. Vol.45. Suppl.1. P.161172. Zalesky K. 1948. Die Waldspitzmaus (Sorex araneus L.) in ihrer Beziehung zur Form tetragonurus Herm. in Nordund Mitteleurope // Sitzungsberichten der Цsterreichischen Akademie der Wissenschaften, Mathematischenaturwissenschaftlische Klasse. Abt.I. Vol.157. P.9185. Zima J. & Krбl B. 1985. Karyotype variability in Sorex araneus in Central Europe (Soricidae, Insectivora) // Folia Zoologica. Vol.34. P.235243.

ture of species in mammals] // Zoologicheskii Zhurnal. Vol.83. P.199212 [in Russian, with English summary]. Orlov V.N. & Kozlovskii A.I. 2002. [On role of glacial period in formation of chromosomal polymorphism of the common shrew Sorex araneus L. (Insectivora, Mammalia)] // Doklady Akademii Nauk. Vol.386. P.423426 [in Russian]. Panteleev P.A. 1996. [About intraspecific systematics and taxonomic value of exterior and craniometric characteristics in Arvicola terrestris] // Vestnik Zoologii. Vol.30. P.2125 [in Russian, with English summary]. Panteleev P.A. 2000. [Role of temperature in geographic speciation (an example of rodents)] // Ekologiya. Vol.2. P.8388 [in Russian]. Panteleev P.A., Gerasimov S. & Khristov L. 1991. [Influence of different approaches to variation investigation on its results] // Ekologiya. Vol.5. P.4655 [in Russian]. Polyakov A.V., Onischenko S.S., Ilyashenko V.B., Searle J.B. & Borodin P.M. 2002. Morphometric difference between the Novosibirsk and Tomsk chromosome races of Sorex araneus in a zone of parapatry // Acta Theriologica. Vol.47. P.381387. Searle J.B. 1984. Three new karyotypic races of the common shrew Sorex araneus (Mammalia: Insectivora) and a phylogeny // Systematic Zoology. Vol.33. P.184194. Searle J.B. & Thorpe R.S. 1987. Morphometric variation of the common shrew (Sorex araneus) in Britain, in relation to karyotype and geography // Journal of Zoology. Vol.212. P.373377. Searle J.B. & Wуjcik J.M. 1998. Chromosomal evolution: The case of Sorex araneus // Wуjcik J.M. & Wolsan M. (eds.). Evolution of Shrews. Bia?owie?a: Mammal Research Institute, Polish Academy of Sciences. P.219268. Simakova A.A. 2001. The vegetation and mammoth distribution during the second half of the Late Pleistocene on the Russian Plain (33-17 ka) // Cavarretta G., Gioia P., Mussi M. & Palombo M.R. (eds.). Proceedings of the 1st International Congress The World of Elephants. Rome. P. 255258. Smith M.F. 1979. Geographic variation in genic and morphological characters in Peromyscus californicus // Jour-

Appendix 1. Mean values (upper row) and standard deviations (lower row) of 20 morphological variables(samples listed from north to south). Variables numbered as in Fig. 1.
Sam ple Belaru s (n= 10) Ch ern yg iv (n= 14) Sum y (n =9) Zhy tomir (n =11) Kiev (n= 20) Kan iv (n= 20) Var iables 1 19.3 7 0. 28 6 19.4 7 0. 35 3 19.1 9 0. 30 4 19.1 9 0. 30 4 18.9 8 0. 31 4 18.9 9 0. 33 8 2 2.6 1 0. 04 7 2.5 5 0. 08 3 2.6 0 0. 08 8 2.5 7 0. 06 8 2.5 7 0. 09 5 2.6 1 0. 07 7 3 5.3 9 0. 12 1 5.3 9 0. 12 9 5.3 1 0. 12 1 5.3 7 0. 16 8 5.4 7 0. 13 9 5.4 0 0. 13 1 4 3.9 9 0. 08 3 3.9 2 0. 09 1 3.8 4 0. 07 6 3.9 6 0. 12 0 3.8 9 0. 12 2 3.8 7 0. 13 3 5 9.7 5 0. 21 6 9.7 5 0. 11 4 9.5 5 0. 21 8 9.4 9 0. 19 9 9.6 6 0. 20 5 9.5 2 0. 24 3 6 8.0 6 0. 21 3 8.2 7 0. 14 1 8.2 0 0. 14 1 8.1 5 0. 18 0 8.0 0 0. 15 9 7.9 0 0. 20 9 7 5.5 9 0. 14 2 5.8 0 0. 06 4 5.7 2 0. 08 7 5.6 8 0. 13 4 5.7 7 0. 22 8 5.6 8 0. 16 9 8 1.3 8 0. 05 9 1.3 7 0. 04 9 1.3 7 0. 07 9 1.3 8 0. 07 1 1.3 4 0. 06 2 1.3 8 0. 07 4 9 2.5 7 0. 08 8 2.8 3 0. 10 2 2.7 1 0. 14 5 2.5 6 0. 09 2 2.6 1 0. 10 9 2.6 6 0. 10 8 10 4.6 3 0. 13 9 4.6 3 0. 07 2 4.6 5 0. 10 4 4.6 0 0. 12 5 4.6 8 0. 12 4 4.6 0 0. 13 8


Morphometrics of common shrews in Ukraine

$
Appendix 1 (continued).

Sam ple Kh ark iv (n =9) Lu gansk (n= 20) Karp aty (n= 20) Moldo va (n= 20) Kh er so n (n= 22) Golay a Pr istan' (n= 12) Odessa (n= 25) Ro mania (n= 20) Belaru s (n= 10) Ch ern yg iv (n= 14) Sum y (n =9) Zhy tomir (n =11) Kiev (n= 20) Kan iv (n= 20) Kh ark iv (n =9) Lu gansk (n= 20) Karp aty (n= 20) Moldo va (n= 20) Kh er so n (n= 22) Golay a Pr istan' (n= 12) Odessa (n= 25) Ro mania (n= 20)

Var iables 1 19.0 6 0. 35 0 19.2 0 0. 37 6 19.4 8 0. 23 7 19.5 2 0. 29 3 20.4 3 0. 32 1 19.9 2 0.3 3 20.4 4 0. 31 9 20.6 9 0. 31 8 11 3.9 1 0. 13 9 3.9 3 0. 07 4 3.8 3 0. 17 8 3.8 9 0. 11 7 3.8 6 0. 17 9 3.8 2 0. 11 6 3.7 6 0. 13 5 3.9 1 0. 08 7 4.0 0 0. 11 7 3.9 5 0. 14 5 4.1 5 0. 11 5 3.9 6 0.1 0 4.0 8 0. 14 1 4.0 4 0. 11 2 2 2.5 6 0. 09 9 2.5 8 0. 06 1 2.6 1 0. 05 9 2.5 9 0. 11 0 2.6 5 0. 14 7 2.6 0 0.0 9 2.7 7 0. 111 2.7 5 0. 09 1 12 5.5 0 0. 12 3 5.5 1 0. 08 0 5.5 6 0. 10 1 5.4 5 0. 08 4 5.4 5 0. 14 6 5.4 4 0. 12 2 5.4 2 0. 08 3 5.5 4 0. 09 8 5.5 9 0. 09 1 5.7 2 0. 07 3 5.8 2 0. 06 9 5.7 1 0.0 7 5.8 7 0. 13 6 5.8 7 0. 09 6 3 5.3 8 0. 15 1 5.3 6 0. 15 1 5.3 2 0. 12 2 5.5 6 0. 14 0 5.5 6 0. 15 7 5.5 7 0.0 7 5.7 3 0. 16 4 5.7 5 0. 13 8 13 3.8 3 0. 07 4 3.8 8 0. 06 2 3.8 8 0. 07 6 3.8 3 0. 07 5 3.8 9 0. 10 8 3.8 2 0. 11 2 3.7 5 0. 09 7 3.8 4 0. 08 9 3.9 1 0. 08 7 4.0 3 0. 06 8 4.0 7 0. 06 3 4.0 1 0.0 5 4.11 0. 08 7 4.0 5 0. 09 3 4 3.9 3 0. 10 5 3.9 4 0. 09 5 3.8 8 0. 07 8 3.9 3 0. 09 6 4.0 9 0. 14 1 4.1 2 0.0 9 4.1 0 0. 12 6 4.1 4 0. 11 0 14 1.7 2 0. 04 0 1.7 8 0. 05 6 1.7 3 0. 04 8 1.7 1 0. 05 4 1.7 5 0. 09 4 1.7 3 0. 07 1 1.6 8 0. 07 8 1.7 1 0. 03 8 1.7 8 0. 05 6 1.8 3 0. 04 1 1.8 4 0. 04 8 1.7 8 0.0 5 1.8 8 0. 05 9 1.8 2 0. 04 2 5 9.6 7 0. 16 4 9.7 4 0. 23 2 9.6 9 0. 23 3 9.6 5 0. 28 7 9.9 6 0. 17 6 9.9 8 0.11 10.0 1 0. 20 0 10.1 2 0. 17 1 15 7.4 1 0. 23 4 7.4 2 0. 12 3 7.4 2 0. 18 8 7.2 6 0. 12 3 7.3 4 0. 18 2 7.3 0 0. 19 2 7.2 5 0. 22 8 7.3 8 0. 13 7 7.4 1 0. 18 3 7.5 9 0. 14 0 7.8 6 0. 13 9 7.5 6 0.1 5 7.8 3 0. 18 8 7.8 7 0. 18 8 6 8.1 9 0. 21 5 8.2 2 0. 23 9 8.2 1 0. 17 3 8.4 5 0. 16 9 8.6 7 0. 21 3 8.4 2 0.1 7 8.8 8 0. 33 2 8.8 9 0. 23 0 16 0.5 6 0. 05 6 0.5 9 0. 03 7 0.5 8 0. 06 9 0.5 4 0. 05 8 0.5 5 0. 07 9 0.5 2 0. 04 8 0.6 0 0. 05 5 0.5 6 0. 05 5 0.5 1 0. 05 3 0.5 3 0. 05 4 0.7 0 0. 06 5 0.6 1 0.0 4 0.6 5 0. 05 8 0.6 0 0. 05 8 7 5.5 3 0. 19 3 5.7 2 0. 21 0 5.7 5 0. 14 6 5.9 4 0. 16 6 6.1 5 0. 18 3 6.0 3 0.3 0 5.9 7 0. 17 7 5.9 2 0. 12 9 17 1.0 2 0. 17 5 1.0 2 0. 12 3 1.0 9 0. 09 5 1.1 0 0. 06 7 1.0 7 0. 08 5 1.0 3 0. 09 9 1.0 7 0. 10 3 1.0 7 0. 13 7 1.0 0 0. 14 7 1.1 2 0. 11 9 1.1 0 0. 09 7 1.0 7 0.0 8 1.0 7 0. 15 2 1.0 8 0. 15 1 8 1.4 1 0. 08 6 1.4 0 0. 07 3 1.3 8 0. 06 8 1.4 3 0. 05 7 1.4 7 0. 06 3 1.4 5 0.0 5 1.4 3 0. 07 9 1.4 4 0. 05 4 18 6.9 3 0. 21 7 6.9 3 0. 21 6 6.9 8 0. 14 5 6.7 9 0. 15 2 6.8 2 0. 16 6 6.8 1 0. 17 2 6.9 3 0. 15 7 6.9 9 0. 13 7 6.9 5 0. 15 2 7.0 6 0. 11 4 7.4 3 0. 15 4 7.2 8 0.1 4 7.3 5 0. 16 9 7.4 8 0. 19 6 9 2.6 3 0. 10 1 2.7 1 0. 08 1 2.7 7 0. 10 1 2.7 0 0. 08 6 2.9 0 0. 06 8 2.7 3 0.1 2 2.7 5 0. 10 0 2.7 8 0. 11 9 19 4.8 2 0. 15 3 4.7 7 0. 10 0 4.7 3 0. 04 8 4.6 6 0. 11 4 4.7 5 0. 12 4 4.6 8 0. 13 9 4.6 5 0. 18 7 4.6 8 0. 13 9 4.7 9 0. 08 9 4.8 6 0. 09 4 5.1 2 0. 15 1 4.9 4 0. 15 0 4.9 8 0. 12 9 5.0 5 0. 10 2 10 4.5 6 0. 06 3 4.6 4 0. 10 6 4.6 6 0. 06 9 4.8 6 0. 10 0 4.9 0 0. 08 8 4.9 1 0.1 0 5.0 0 0. 12 1 4.8 9 0. 11 8 20 1.0 8 0. 06 8 1.0 5 0. 04 5 1.1 2 0. 06 1 1.0 8 0. 05 6 1.0 8 0. 08 6 1.0 5 0. 08 1 1.1 0 0. 09 5 1.0 9 0. 05 6 1.11 0. 08 3 1.1 5 0. 06 9 1.2 4 0. 06 2 1.2 2 0.0 8 1.1 4 0. 06 5 1.1 6 0. 06 4


$

A.V. Mishta
Appendix 2. Loadings for principal components based on craniometric variables. Appendix 3. Loadings for canonical functions based on (a) craniometric and (b) geoclimatic variables. a)
Var iable MORPH 1 MORPH 2 MORPH 3 MORPH 4 MORPH 5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0.376 0.005 0.139 0.037 0.09 1 0.217 0.11 0 0.087 0.12 0 0.11 2 0.03 9 0.085 0.03 3 0.102 0.22 4 0.036 0.02 7 0.273 0.007 0.127 0.14 0 0.451 0.49 9 0.20 4 0.205 0.407 0.12 6 0.215 0.017 0.07 4 0.01 8 0.091 0.25 9 0.29 5 0.38 7 0.558 0.229 0.413 0.27 6 0.146 0.135 0.329 0.466 0.16 0 0.63 7 0.062 0.045 0.148 0.47 5 0.338 0.31 5 0.80 5 0.649 0.05 7 0.088 0.395 0.051 0.25 2 0.045 0.19 1 0.60 8 0.06 1 0.49 1 0.002 0.12 6 1.336 0.11 3 0.131 0.16 0 0.034 0.19 3 0.839 0.10 4 0.01 7 0.142 0.30 5 0.253 0.60 2 0.25 4 0.226 1.775 0.197 0.56 5 0.331 0.48 8 0.25 8 0.20 0 0.40 8 0.14 5 0.37 6 0.11 5 0.035 0.026 0.224 0.315 0.284 0.017 0.82 6 0.083 0.30 8

Var iable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

PC 1 0. 91 6 0. 63 5 0. 77 6 0. 58 9 0. 73 1 0. 87 5 0. 66 4 0. 44 4 0. 55 5 0. 860 0. 635 0. 912 0. 841 0. 699 0. 89 3 0. 50 4 0. 14 7 0. 85 4 0. 81 2 0. 491

PC 2 0. 04 4 0. 33 5 0. 23 6 0. 14 1 0. 16 1 0. 03 0 0. 04 4 0. 45 5 0. 43 2 0. 010 0. 515 0. 096 0. 106 0. 192 0. 07 1 0. 21 3 0. 57 0 0. 11 9 0. 06 8 0. 24 3

PC 3 0. 06 7 0. 45 4 0. 17 6 0. 41 5 0. 29 3 0. 01 8 0. 11 9 0. 03 0 0. 24 0 0. 004 0. 05 9 0. 058 0. 082 0. 179 0. 19 7 0. 05 7 0. 61 9 0. 11 7 0. 10 4 0. 20 3

PC 4 0. 05 9 0. 03 3 0. 14 9 0. 22 5 0. 13 2 0. 07 8 0. 01 0 0. 00 6 0. 29 1 0. 193 0. 028 0. 144 0. 264 0. 192 0. 00 1 0. 48 7 0. 27 5 0. 05 6 0. 10 1 0. 558

PC 5 0. 12 3 0. 19 5 0. 11 9 0. 24 3 0. 00 9 0. 08 3 0. 25 7 0. 47 9 0. 20 6 0. 003

PC 6 0. 08 9 0. 17 4 0. 17 4 0. 11 0 0 . 111 0. 21 6 0. 41 0 0. 18 5 0. 09 7 0. 16 9

0. 165 0. 029 0. 067 0. 080 0. 254 0. 067 0. 375 0. 003 0. 06 3 0. 31 5 0. 16 3 0. 08 6 0. 064 0. 13 4 0. 04 9 0. 23 8 0. 16 9 0. 29 1 0. 21 5 0. 41 2

b)
Var iable LONG LAT ALT TE MP_ J TEMP_I TE M P_D W_ DAYS SNOW RAIN GEOCL 1 GEOCL 2 GEOCL 3 GEOCL 4 GEOCL 5 0. 18 5 0. 09 6 0. 19 0 1. 56 9 1. 92 3 2. 58 6 0. 32 9 0. 33 2 0. 43 1 0. 65 4 2. 07 1 2. 41 3 1. 27 4 2. 30 5 2. 12 0 0. 44 7 1. 58 1 0. 01 5 0. 53 0 1. 14 0 0. 60 4 3. 37 5 0. 12 7 2. 96 3 0. 35 6 1. 28 2 0. 47 7 1. 29 7 0. 70 1 0. 81 9 2. 98 7 0. 07 5 0. 82 9 0. 78 1 2. 19 6 0. 67 4 1. 61 7 0. 82 1 0. 82 4 0. 99 9 1. 33 0 1. 23 9 1. 21 6 0. 88 6 0. 03 9