UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Geographical variation in wolves (Canis lupis L.) of northwestern North America Jolicoeur, Pierre 1958

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1958_A8 J6 G3.pdf [ 1.66MB ]
JSON: 831-1.0106083.json
JSON-LD: 831-1.0106083-ld.json
RDF/XML (Pretty): 831-1.0106083-rdf.xml
RDF/JSON: 831-1.0106083-rdf.json
Turtle: 831-1.0106083-turtle.txt
N-Triples: 831-1.0106083-rdf-ntriples.txt
Original Record: 831-1.0106083-source.json
Full Text

Full Text

GEOGRAPHICAL VARIATION IN WOLVES (Canis lupus L. OF NORTHWESTERN NORTH AMERICA by PIERRE JOLICOEUR B.A., Uniyersite de MDntreal, 1953 B.Sc, Universite' de Montreal, 1956 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of ZOOLOGY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA Ap r i l , 1958 i ABSTRACT Five hundred wolf specimens were studied. They represent populations from Alaska to Keewatin and from Vancouver Island to Manitoba. Pelage color varies nearly from black to white. There are no discrete color phases. Pale wolves are more numerous and dark wolves less numerous toward the tundra (northeastward) between Great Slave Lake and Great Bear Lake. Judging from color variation, wolf populations intermingle by associating with caribou at migration. Male wolves are larger than females (approximately K$> i n linear skull dimensions). Northeastern individuals have a shorter and relatively broader skul l than southwestern ones. Multivariate divergence i n twelve skull dimensions i s approximately proportional to geographical separation. This may express genetic differentiation "by incomplete isolation. But the pronounced northeastward zonation of the environment may have direct influences upon growth processes. Interpretations i n terms of genetic a f f i n i t i e s are hypothetical and taxonomic conclusions are postponed. Simultaneous analysis of biometrical data appears indispensable to disclose major trends of geographic variation. In presenting t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree that permission f o r extensive copying of t h i s t h e s i s f o r s c h o l a r l y purposes' may be granted by the Head of my Department or by h i s r e p r e s e n t a t i v e . I t i s understood tha t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l gain s h a l l not be allowed without my w r i t t e n permission. Department of The U n i v e r s i t y of B r i t i s h Columbia, Vancouver^S, Canada. i i TABLE OF CONTENTS Text Abstract i Table of contents . i i Introduction 1 Material and data 3 Techniques of analysis . . . . . . . . 8 Variation i n pelage coloration 10 Variation i n skull size and form : bivariate analysis 13 Variation i n skull dimensions : multivariate analysis 18 Interpretations and conclusions . . . . . . . 28 References . . . • 32 Figures Fig. 1 Geographical origin of the samples h Fig. 2 Skull dimensions and coded designations 7 Fig. 3 Relative; frequency of pale wolves 11 Fig. h Variation i n overall skull size and proportions . . . . . . lU Fig. 5 Variation i n a relative growth rate 16 Fig. 6 Interbullar breadth and upper carnassial length 17 Fig. 7 K 1 : K 2 multivariate configuration 23 Fig. 8 K 1 : K 2 biometrical overlapping 2k Fig» 9 K 3 : K h multivariate configuration 25 Tables Table 1 Sample sizes and sex-compositions k Table 2 Discriminant functions 20 Table 3 Group means i n discriminant analysis . . . . , 21 1 INIRCDUCTION Bie last comprehensive taxonomic study of North American wolves (Canis lupus L.) was that of Goldman (Young and Goldman, 19kk). I t consisted largely of qualitative skull and pelage descriptions and such procedures fai l e d to show clearly geographic variation i n the species as a whole. Large collections of wolves have been made during recent control operations i n northwestern Canada by the Canadian Wildlife Service and the Manitoba Game Department. Qhis material has been deposited i n the Museum of Zoology of the University of Br i t i s h Columbia. I t forms the main object of the present study. Comparisons were made with B r i t i s h Columbia, Alaska and High Arctic material, some of which was borrowed from the Br i t i s h Columbia Provincial Museum, the National Museum of Canada and Dr. R. Rausch, Anchorage, Alaska. Much of the limitations of previous studies appears to be due to inefficient methods of analysis. Finding optimum biometrical techniques has therefore been a major aim of this investigation. Hie author worked under the guidance of Dr. I. McT. Cowan, Department of Zoology, University of Br i t i s h Columbia. Field aspects of the problem were discussed with several members of the Canadian Wildlife Service. Dr. S. W. Nash, Department of Mathematics, gave numerous explanations on multivariate analysis. Help i n mathematics was received from Marcel Banville, Dept. of Ehysics, W. R. Knight and Bomshik Chang, Dept. of Mathematics and many others. Most calculations were done at the Computing Centre, 2 with much assistance from the personnel. The author i s indebted to Dr. I. McT. Cowan for a c r i t i c a l reading of the manuscript and to other members of the Department of Zoology for advice on illustr a t i o n s . Discussions with fellow graduate students lead to cl a r i f i c a t i o n of several ideas. Financial support came from the Wildlife Conservation Fund of the Canadian Industries Limited. A l l of this i s gratefully acknowledged. •3 MATERIAL and DATA Specimens vere grouped by l o c a l i t i e s of origin ( f i g . l ) . The number of specimens at hand and their most obvious characteristics were considered i n delimiting the groups. Sample size and sex composition were taken into account throughout the analysis (table l ) . There was only half a dozen juvenile specimens (estimated younger than six months) and they were excluded. Only four areas are represented by large samples : Bri t i s h Columbia (group K), Manitoba ( I ), and the Northwest Territories between Great Slave Lake and Great Bear Lake ( groups D, E and G). Two arrows have been lined up on these large samples i n the map ( f i g . l ) and i n some subsequent graphs. They point approximat-ely northeastward and northwestward and help to refer biometrical differences to their geographical context. Skulls were available for most specimens while there were pelage and body data for only part of the collection. The analysis of geographic variation was therefore based primarily on skull dimensions. Photographic transparencies of the carcasses were available for four samples of the Northwest Territories and the frequencies of types of pelage coloration were compared. Twelve skull dimensions were measured. They were chosen for their descriptive value and for the ease with which they could be measured consistently. They are refered to by coded designations ( L i , Wi, Ci and Ti ) and defined as follows : If Figure 1 : Geographical origin of the samples. Sable 1 : Sample sizes and sex-compositions. Locality group males females undetermined tot a l High Arctic A 11 8 . _ 19 Alaska B 3 3 3 9 Keewatin C •5 3 6 lk D Ul 39 • - • 80 Northwest E hi 1+0. 81 territories F 12 8 20 G 33 33 - 66 . H - - • 9 ' 9 Manitoba I 73 - 137 Vancouver Island J 5 5 10 Interior B. C. K 15 12 18 5^ Rocky Mountains L 6 . 3 • 9 H99 5 ( L i ) : MEASUREMENTS OF LENGTH ( L 1 ) : Condylobasal length : Distance from the anterior t i p of the premaxillae to the plane of the posterior border of the occipital condyles. ( L 2 ) : Palatal length : Distance from the alveolus of the median upper incisor on one side to the notch of the posterior. edge of the palatal shelf on the same side. ( L 3 ) : Post-palatal length : Distance from the notch of the posterior edge of the palatal shelf on one side to the posterior face of the ventral l i p of the foramen magnum on the median li n e . ( ¥ i ) : MEASUREMENTS OF WIDTH ( W 1 ) : Zygomatic width : Greatest distance across the zygomatic arches. ( W 2 ) : Palatal width at M 1 : Greatest distance between the outer edges of the alveoli of the f i r s t upper molars. ( W 3 ) : Palatal width at Pm 2 : Least distance between the inner edges of the alveoli of the second upper premolars. ( W h ) : Interglenoid width : Least distance between the postglenoid foramina. ( W 5 ) : Interorbital width : Least distance across the frontal bones between the orbits. 6 ( C i ) : MEASUREMENTS EXPRESSING BRAINCASE DEVELOPMENT ( C 1 ) : Least width of the cranium : Least distance across the frontal hones behind the postorbital processes. ( C 2 ) : Interbullar "breadth : Distance between the auditory bullae where they angle with the basioccipital bone. ( T i ) : TOOTH MEASUREMENTS ( T 1 ) : Length of the upper carnassial : Distance from the anterior surface of the upper carnassial to i t s posterior surface at the level of emergence from the alveolus. ( T 2 ) : Length of the f i r s t upper molar : Greatest distance from the anterior surface of the f i r s t upper molar to i t s posterior surface at the level of the crown and i n the axis of the two outer cusps. figure These twelve skul l measurements are il l u s t r a t e d i n number two. 7 Figure 2 : Skull dimensions measured and coded designations. TECHNIQUES OF ANALYSIS 8 In the physico-chemical sciences variation arises mostly from errors of measurements. Biological variation on the other hand results largely from objective factors. In biological st a t i s t i c s therefore, describing variation concisely i s more important than assessing the probability that sets of observations f i t a single hypothesis. Associating biometrical data between themselves and/or with age data i s generally necessary to bring out their f u l l meaning. In the analysis of animal form large use has been made of arbitrary age estimates and of ratios of dimensions. Age estimates of wild mammals are generally far less precise than bone measurements except for a few species exhibiting "growth-rings" or other definite c r i t e r i a of age. Ratios express a proportion by a single figure but they dissociate form from size and they are inefficient for more than two dimensions. Bivariate scatter diagrams or their multivariate version (Anderson, 195*0 are the best simple analytical tool. However multiassociated data usually yield more information through multi-variate analysis (Hotelling, 1954; Quenouille, 1952; Yates, 1950). The latter takes into account a l l intercorrelations of the variables. Animal form can thus be analysed without age estimates save for a broad preliminary classification of the material. Multivariate techniques permit the analyst to express information with maximum conciseness. Most recent applications have unfortunately featured too abstract a presentation. Expressing the results of a study in terms of the original variables is preferable in practise. Multivariate analysis is now within the reach of biologists thanks to Murdoch's excellent introduction (1957) to linear algebra and analytic geometry. 10 VARIATION IN PELAGE COLORATION Pelage coloration of wolves is highly variable in inten-sity, in hue and in pattern. There are no obviously discrete color phases as in some polymorphic species. Detailed verbal descriptions are clearly unsuitable for large samples. The photographic transparencies examined for pelage coloration were classified into four arbitrary types according to the general darkness of pigmentation : dark, darkish, whitish and white. Such arbitrary types do not correspond in the wolf to actually discrete color phases. Such a classification is also only approximate and fits adequately only the present material. It does disclose however a gradual change in color-type frequencies analogous to the clines in color-phase frequencies of the red fox and the black bear (Cowan, 1938). The relative frequency of pale wolves increases in a northeastward direction (toward the tundra) between Great Slave Lake and Great Bear Lake in the Northwest Territories (fig. 3)» There are gradually more white and whitish and fewer dark and darkish indivi-duals in samples F, D, E, and G successively. Samples D and E differ l i t t l e from each other but differ significantly from the two extreme samples (95$ chi squared). Recent barren-ground caribou studies (Banfield, 195^J Kelsall, 1957) have shown caribou to migrate more through areas D and E than through areas F and G.Differences of pelage coloration between wolf populations appear therefore to be inversely proportional to the local importance of caribou migrations. But wolves are often observed 1 fTTTTTS 34 10 4 5 17 ni i i in i i 12 i » 4 3 8 5 7 6 : 11111 • 1111 d a r k w h i t e d a r k i s h w h i t i s h Great Bear Lake Great Slave Lake North West East South A b s o l u t e f r e q u e n c i e s G e o g r a p h i c a l l o c a l i t i e s Figure 3 : Northeastward increase i n the relative frequency of pale wolves between Great Slave Lake and Great Bear Lake ( toward the tundra ) . 12 with caribou herds (Banfield, 1951)* This suggests that wolf popula-tions intermingle by associating with caribou at migration. A relatively higher frequency of dark individuals has been reported for the Rocky Mountains (Cowan, 19^7)• The short-distance cline exhibited by the present material may therefore be part of a long-distance cline going at least from the Rockies to the Northwest Territories. More data on the pelage coloration of wolves may eventually show analogy with the pattern of geographical variation of color-phase frequencies of the red fox and the black bear (Cowan, 1938j Butler, 19^7). 13 VARIATION IN SKULL SIZE AND FORM : BIVARIATE ANALYSIS Overall skull size can be satisfactorily described by condylobasal length (L l ) and zygomatic width (W l ) . Bivariate dot diagrams of these two dimensions were made and 95$ equal-frequency ellipses were calculated following the procedure discussed by Defrise-Gussenhoven (1955)* Figure h summarizes the most important information : males reach a skull size approximately b$> greater than females (in linear dimensions). This agrees with Hildebrand's (1952) conclusions regarding the body size of Canidae. Other facts brought out are the lesser maximum skull size (they are closer to the l e f t lower corner of the graph) and the greater relative breadth (they are closer to the l e f t upper corner of the graph) of northeastern wolves. Groups L, • I, D + E + G, and A are successively closer to the l e f t side of the graph. This ordering of samples according to skull size and relative breadth i s strikingly similar to the ordering of the l o c a l i t i e s of geographical origin projected upon a line of northeastward direction. Such gradual geographic variation was termed "clines" by Huxley (1938). The same shortness and greater relative breadth of skull of northeastern wolves shows i n a scatter diagram (figure 5) of interglenoid width (\l k) on post-palatal length (L 3 ) . The skulls of wolves from the Northwest Territories ( G ) are shorter and broader than those of wolves from Br i t i s h Columbia ( K ) with respect to these two dimensions. But here the difference of proportion increases with size. This i s a difference of relative growth rate. mm. 220 230 240 250 260 L| Figure h : Sexual and geographic variation i n overall skull size and proportions as illustrated by condylobasal length (L 1) and zygomatic width (W l) Equal-frequency ellipses f i t the data satisfactorily; there i s no obvious trend curvature and no need for a logarithmic transformation. Rates of relative growth are of considerable biological interest (Huxley, 1932) and a multivariate analysis of growth i n wolf skulls i s planned for the near future. A third bivariate association shows geographical variation (figure 6) : interbullar breadth ( C 2 ) against carnassial length ( T 1 ). The wolves from Manitoba ( I ) and the Northwest Territories ( D + E ) are at the center of this graph and constitute the average. The wolves from Brit i s h Columbia ( K ) have a shorter carnassial than the average and those from Vancouver Island ( J ) a narrower interbullar space. Simple examination of the skulls confirms what the graphical analysis summarizes. Distinct spaces show i n between the small teeth of Brit i s h Columbia wolves and the ten Vancouver Island specimens have markedly "inflated" bullae with a narrow interval. Surprisingly i n this graph the Vancouver Island wolves differ the most from those to which they are the closest geographically. Further discussion of this w i l l follow the joint multivariate analysis of a l l twelve skull dimensions. Figure 5 : Geographical variation i n a relative growth rate. Interglenoid width (W h) against post-palatal length (L 3). Figure 6 : Geographical variation i n interbullar breadth (C 2) and upper carnassial length (T l ) . £j 18 VARIATION IN SKULL DIMENSIONS : MULTIVARIATE ANALYSIS Several multivariate techniques are available for joint biometrical variation. Some lead to overall estimates of between-sample differences ("distance functions"); others lead to combina-tions of measurements revealing the pattern of divergence or "configuration" of groups ("discriminant functions"). Distance functions express variation as a whole. Discriminant functions disclose the principal components of variation underlying the .intercorrelations of the variables. Discriminant analysis was carried on here following Rao's (1952 :370-378) procedure. Sexual skull differences having shown to be mostly size differences ( f i g . h), sexes were kept together to emphasize geographical variation i n sku l l proportions. The within-group product matrix W, generated by the individuals around their group means, came from the U09 specimens of the four largest samples (K, G, D + E, and I ). The between-group product matrix B, generated by the group means around the grand mean came from eleven geographical groups totalling U99 specimens. The B and W matrices were therefore divided by ^99 and k09 respectively^efore calculation of the discriminant functions. Inspecting the means of the twelve s k u l l dimensions i n the eleven geographical groups (table 3) permits a rapid check upon the reality of the trends of joint variation disclosed by discriminant functions. Tabulating other stat i s t i c s or the original data would consume too much space without making anything e x p l i c i t . Discriminant functions K (also called characteristic, 19 canonical, latent or eigen-vectors) and their variance components D (characteristic roots or eigen-values) are defined by the following matrix equation : KB = DKW . They were calculated on an electronic d i g i t a l computer by matrix operations (Murdoch, 1957 :l65-l66) corresponding to the transformations suggested by Rao (1952 :357>367). Matrices were diagonalized following Jacobi's method. The within-group variances and covariances of the discriminant functions checked ( KWK' = I ) to two or three significant digits, which i s acceptable. A l l of these mathematical manipulations correspond to the analysis of between-group variation taking within-group variation as a unit of measurement. This standardization should minimize the effects of differences i n age-composition of the samples. A l l components of standardized between-group variation add up to I.5U6U . The f i r s t five add up to 1.4533 and account for 9*$ of the t o t a l . Tb each of these five components correspond twelve coefficients for the original variables i n the discriminant functions (table 2)« She s t a t i s t i c a l significance of these variance components was tested as prescribed by Rao (1952 :372) for large samples taking k09 as total number of observations. The probability of such large components under a nu l l hypothesis i s less than 1$ for the f i r s t four and less than 5$ for the f i f t h . She configuration of groups i n the two f i r s t discriminant functions, (figure 7) i s recognizably similar to the disposition of the lo c a l i t i e s of origin on a geographical map. Northern samples congre-gate i n the- l e f t upper corner of the graph, eastern samples i n the Table 2 : Discriminant functions. Variance components and coefficients of the skull dimensions. Function K 1 K 2 K 3 Kl+ K 5 variance component. .80kQ .2761 .181+6 .1221 .0657 $ of total variance 52$ 18$ 12$ 8$ 1+$ L 1 -.1557 .06^5 .231+5 .0728 -.1623 L 2 -.0198 -.161+0 -.31+80 -.0302 . 3 7 ^ L 3 -.0097 .0755 -.2077 -.11+02 .1201+ ¥ 1 .0538 -.0^98 -.0786 -.0137 -.I65O Coefficients of W 2 .0172 -.01+1+2 -.1125 -.0182 -.0332 the skul l W 3 -.0080 -.0080 .6301 .3135 .01+72 dimensions W k .2271 .0993 .01+28 -.1227 .11*59 W 5 .1712 -.11+20 .0910 -.1803 .1783 C 1 -.1261 .001+1+ -.0952 .1221 '-.0366 C 2 -.1970 .132^ .0908 -.271+1 -.1258 T 1 .5036 .7033 -.2729 .1+297 .2522 T 2 .1781* -.31+66 .3857 -.1+070 .181+8 Table 3 : Group means of the skull dimensions i n discriminant analysis. GROUP N L 1 L 2 L 3 W 1 W 2 W 3 W 4 W 5 c 1 C 2 T 1 T 2 A 19 231.63 113.84 98.84 139-37 80.32 32.90 65.84 45.44 40.08 19.20 25.77 17.54 B 9 245.67 i23.ll 102.22 140.45 81.33 35.36 65.65 46.86 43.06 19.46 24.40 17.70 C 14 234.14 U6.79 98.50 135.29 78.72 33.67 64.01 45.71 40.52 18.70 24.50 17.33 D+E 161 234.73 117.30 98.49 138.56 78.37 33.58 63.78 46.37 41.28 18.45 23.99 17.28 F 20 242.15 119.55 103.15 140.75 79.63 33.79 65.07 45.64 41.64 19.67 25.13 17.42 G 66 235.98 118.35 99.00 137.76 78.13 33.18 64.05 46.08 41.03 18.59 24.42 17.36 H 9 243.33 120.89 102.89 140.22 79.59 34.39 64.34 46.27 40.49 20.54 23.91 17.19 I 137 237.20 117.75 99.88 136.52 78.63 34.01 63.82 45.11 41.15 19.03 24.61 17.34 J 10 236.30 119.60 98.30 136.70 77.73 31.85 61.15 44.13 41.94 17.03 24.82 16.70 K 45 240.18 119.36 101.40 135.27 76.92 32.97 62.49 44.07 42.16 19.50 23.28 16.80 L 9 251.00 123.45 106.33 139.67 79.91 32.73 66.10 47.51 42.61 22.23 25.23 17.76 right upper corner and inversely for southern and western samples. The two arrows of northeastward and northwestward directions correspond to those of the map ( f i g . l ) and help to evaluate the similarity of the pattern of biometrical divergence with the pattern of geographical origin. Discrepancies come mostly from small samples. The major one i s the respective position of Alaska ( B ) and Vancouver Island ( J ) wolves. But the group configuration of the third and fourth discriminant functions (figure 9) compensates largely that discrepancy : Vancouver Island wolves contrast sharply with a l l others and Alaska wolves are further from the southern ones than a l l other northern ones. The f i r s t component of multivariate variance ( D 1 = 52$ of t o t a l ) corresponds very closely to a northeastward direction and i s markedly greater than the next largest one ( D 2 = 18$ of total ) .> Sets of vectors ("arrows") bearing the coded designations of the skull dimensions indicate their contributions to the discriminant functions. Each vector shows the change i n the discriminant plane that the corresponding dimension would generate i f i t varied independently (by 1 standard deviation i n f i g . 9 and by 2 i n f i g . 7)-» Such i s not the case of course and these vectors must be considered j o i n t l y rather than separately. Northeastern wolves differ generally from southwestern ones ( f i g . 7) by a decrease i n skull length ( L I and L 3 ) and i n braincase development ( C 1 and C 2 ) opposed to an increase i n skul l breadth ( W 1, W h} and W 5 ), Eastern wolves have a longer upper i r 5 6 7 K 2 Figure 7 : Group configuration (left) i n the f i r s t two discriminant functions ( K 1 and K 2 ) and variation of the s k u l l dimensions (right) j N-W and N-E arrows correspond to those of the map ( f i g . 1); see text for explanations. i i i - r r r r~ f 4 5 6 7 8 9 K 2 10 Figure 8 : Biometrical overlapping i n discriminant functions K 1 and K 2 illustrated by 95> equal-frequency ellipses; crosses and dots represent group means and individuals respectively; letters refer to closest symbols. Figure 9 : Group configuration (left) i n discriminant functions K 3 and K h and variation of the skull dimensions (right); 95$ equal-frequency-ellipse of group K ; see text for explanations. K> 26 carnassial ( T 1 ) and a shorter palate ( L 2 ) than western ones. Such East-West variation had not shown up with simpler analytical techniques. Vancouver Island wolves ( J ) di f f e r very much from others (figure 9) "by six skull dimensions (greater T 1, C 1; lesser T 2, C 2, W k and W 5) and very l i t t l e with respect to the six others. The role of these two groups of dimensions i s contrasted not only by the directions but also by the lengths of their vectors. Vancouver Island wolves are much further from the grand mean than the arrows ( 1 standard deviation each ) of their discriminators are long. The amount of biometrical overlapping can be shown satisfactorily by the individual observations of small samples and by 95$ equal-frequency ellipses of large samples. Brit i s h Columbia wolves ( K ) overlap by approximately 50$ (figure 8) with Manitoba ( I ) and Northwest Territories wolves ( D + E ). Ihe wolves from the Rocky Mountains ( L ) are intermediary and overlap largely both with those from Brit i s h Columbia and those from Manitoba. High Arctic wolves ( A ) overlap by approximately 50$ with those from the mainland. Ihe lowermost point of sample A represents a subadult female from Coronation Gulf which should have been grouped with mainland specimens and i s relatively narrow-skulled. Save for this exception, High Arctic wolves do not overlap with those from the Rockies. Larger samples would probably do to some extent however. Vancouver Island wolves overlap ( f i g . 9) by approximately 50$ with others. To sum up, this material shows northeastern wolves to • have generally shorter and relatively broader skulls than southwestern 27 ones and eastern wolves to have a shorter palate and a longer carnassial tooth than western ones. Such a generalization i s approximate however : the correspondence "between the patterns of biometrical divergence and of geographical separation i s imperfect and the f i r s t two discriminant functions account for only of total variance. More variance.is associated with a northeastward direction than with any other one. Vancouver Island wolves differ markedly from others by six skull dimensions but very l i t t l e with respect to the six others. The amount of biometrical divergence and overlapping between a l l groups i s approximately proportional to the degree of geographical separation by distance, insularity, etc. 28 INTERPRETATIONS AND CONCLUSIONS The proportionality of biometrical divergence to geographical separation could readily be interpreted i n terms of .population genetics. Genetic differentiation within an incompletely panmictic population should theoretically be proportional to geographical distance and other factors of isolation (Male'cot, l^hQ; Wright, 1951). Tne high mobility of wolves would compensate for the extent of their area of distribution and tend to erase the amount of differentiation probably induced by isolation during recent glaciations. Between-group variation i s most pronounced northeastward. Sampling has perhaps much to do with the predominance of northeast-ward variation i n this study. But genetical differentiation may be actually greater i n that direction. The genetical interpretation of geographical variation i s not the only one available however. The marked northeastward zonation of the environment may have direct influences upon the growth processes involved i n skull development. The peripheral dimensions of length and breadth of the skul l of Canidae reach f u l l development at a later age than the posterior central region (Huxley, 1880). This appears to be indeed a general pattern of mammalian skull development (Baer, 195*0 • Particular growth processes could be especially affected i f they were i n progress during temporary physiological disturbances. Juvenile sheep with thyroid deficiencies grow skulls with normal braincase and teeth but with underdeveloped f a c i a l region (Todd and Wharton, 193*+) • Their descriptions would f i t 2 9 surprisingly well the skulls of northeastern wolves with large teeth cramped i n a short palate. Stockard and others ( 1 9 ^ 1 ) found pituitary and thyroid abnormalities more frequently i n domestic dog breeds with short-broad skulls than i n those with long-narrow skulls. The f a c i a l development of arctic wolves may therefore possibly be hindered by a low activity of the pituitary and thyroid glands. Seasonal periodicity of the environment (light, temperature, food, etc.) may have effects upon growth just as on other physiological a c t i v i t i e s . Molts and coat-color changes of weasels were controlled photoperiodically by Bissonnette and Bailey ( 19kh ); the pituitary gland was considered to be involved. Seasonal periodicity i s also known to act through endocrine glands and metabolic factors upon bird migrations, on the reproductive cycles of various vertebrates, etc. Large mammals should be especially affected by seasonal, periodicity i n prairies and tundra where climatic and ecological conditions are so homogeneous. The northward increase in seasonal periodicity of the environment may therefore have something to do with the skul l dimensions reached by wolves. Studies of seasonal variations i n wolf behavior may give valuable clues on the effect of arctic winters on the endocrine balance and the metabolism of young wolves. Such studies should also lead to a more integrated view of wolf and dog behavior than either Scott ( 1 9 5 0 ) or Stockard ( 1 9 4 1 ).have reached. Inasmuch as geographical variation expresses genetic differentiation, this analysis may improve our knowledge of genetic a f f i n i t i e s . Manitoba wolves are quite closely similar to the ones of 30 the Northwest Terri t o r i e s from which Alaska wolves also show l i t t l e difference. Vancouver Island wolves have features of t h e i r own but i n other respects they resemble northern wolves more than those presently inhabiting the In t e r i o r of B r i t i s h Columbia. I t i s perhaps ' with northern populations that Vancouver Island had i t s l a s t free b i o t i c contact. As for the High A r c t i c wolves, t h e i r biometrical characteristics are i n good accordance with t h e i r geographical position and they give no clear indications of unsuspected genetic a f f i n i t i e s . Taxonomical interpretations of geographical v a r i a t i o n can only be accepted when the l a t t e r i s known to express mostly genetic d i f f e r e n t i a t i o n . More research i s necessary to evaluate d i r e c t environmental effects i n the present problem. The wolves of Vancouver Island and those of the I n t e r i o r o f . B r i t i s h Columbia exhibit pronounced ch a r a c t e r i s t i c s . Such characteristics f i t quite w e l l into the general pattern of v a r i a t i o n however and there seems to be no point i n thinking of subspecific units unless further studies show va r i a t i o n between populations to be somewhat abrupt. Ascertaining the relationships of western wolves requires more material from Vancouver Island, Alaska, Alberta and the regions i n between. On the other hand, the analysis of v a r i a t i o n should be extended to the species as a whole or at least to a l l i t s North American representatives. There are quite c e r t a i n l y too many subspecific designations in. use ( M i l l e r and Kellog, 1 9 5 5 ) . Goldman's (Young and Goldman, lykk) f a i l u r e to detect the major trends of geographical v a r i a t i o n seems largely due to his 31 approach. He compared specimens i n detail only with those from neighbouring l o c a l i t i e s . Gradual variation cannot show up clearly unless a l l samples are compared simultaneously. Joint trends of variation constitute a "multidimensional f i e l d " of variation rather than just "clines" (Huxley, 1938). Multi-variate analysis i s optimum for multiassociated biometrical data. I t should eventually bring out relationships of growth phenomena and geographic variation with physiology and population genetics. 32 REFERENCES ANDERSON, E. 1954 Efficient and inefficient methods of measuring specific differences. •-93-106 i n KEMPIHORNE and others, 1954. BAER, M.J. 1954 Patterns of growth of the skull as revealed by v i t a l staining. Human Biology 26:80-126. BANFIELD, A.W.F. 1951 Notes on the mammals of the Mackenzie D i s t r i c t , Northwest „ "territories. Arctic 4:112-121. 1954 Preliminary investigation of the barren ground caribou. Wildl.Man.Bull. 1, 10 A and 10 B. Canadian Wildlife Service, Ottawa. BISSONNETTE, T.H. and E.E. BAILEY 1944 Experimental modification and control of molts and changes of coat-color i n weasels by controlled lighting. Ann. N.Y. Acad. Sci. 45:221-260. BUTLER, L. 1947 The genetics of the colour phases of the red fox i n the Mackenzie River l o c a l i t y . Can.Jour.Res. D 25:190-215. COWAN, I.McT. 1938 Geographic distribution of color phases of the red fox and black bear i n the Pacific Northwest. Jour.Mamm. 19 :202-206. 1947 The timber wolf i n the Rocky Mountain National Parks of Canada. Can.Jour.Res. D 25:139-174. DEFRISE-GUSSENHOVEN, E. 1955 Ellipses equiprobables et taux d'eloigneraent en biometrie. Bull.Inst.Royal Sci.nat. Belgique XXXI ( 2 6 ) . Bruxelles. HILDEBRAND, M. 1952 An analysis of body proportions i n the Canidae. Amer.Jour.Anat. 90:217-256. HOTELLING, H. 1954 Multivariate analysis. :67-80 i n KEMPIHORNE and others, 1954. 33 HUXLEY, J.S. 1932 Problems of relative growth. xix- 276. Methuen and Co., London. 1938 Clines : an auxiliary taxonomic principle. Nature 11+2:219. HUXLEY, T.H. 1880 Cranial and dental characters of the Canidae. Proc.Zool.Soc. London :238-287. KELSALL, J.P. 1957 Continued barren-ground caribou studies. Wildl.Man.Bull. 1, 12 . Canadian Wildlife Service, Ottawa. KEMPTHORNE, 0 . and others (editors) 195k Statistics and mathematics i n biology. v i i - 6 3 2 . Iowa State College Press, Ames. MALECOT, G. 19k& Les mathematiques de l'he'redite. 6 3 . Masson et Cie, Paris. MILLER, G.S. and R. KELLOG 1955 List of North American Recent mammals. U.S,Nat.Mus.Bull. . 2 0 5 : 9 5 1 + . Smithsonian Institution, Washington. MURDOCH, D.C. 1957 Linear algebra for undergraduates. xi - 2 3 9 . John Wiley and Sons, Inc. New York. QUENOUILLE, M.H. 1952 Associated measurements. X - 2 U 2 . Butterworths Scientific Publications, London. RAO, CR. 1952 Advanced s t a t i s t i c a l methods i n biometric research. x v i i - 3 9 0 . John Wiley and Sons, Inc. New York. SCOTT, J.P. 1950 The social behavior of dogs and wolves : an i l l u s t r a t i o n of sociobiological systematics. Ann.N.Y.Acad.Sci. 5r:1009-1021. STOCKARD, CR. and c o l l . 19hl The genetic and endocrinic basis for differences i n form and behavior as elucidated by studies of contrasted pure-line dog breeds and their hybrids. xx- 775» Amer.Anat.Mem. 19 . Wistar Inst.Anat.Biol., Philadelphia. TODD, T.W. and R.E. WHARTON 193k The effect of thyroid deficiency upon skull growth and form i n the sheep. Amer.Jour.Anat. 55:97-115. WRIGHT, S. 1951 The genetical structure of populations. Ann.Eugenics 15:323-35^ . YATES, F. 1950 The place of statistics i n the study of growth and form. Proc.Royal Soc. B, 137:^ 79-^ 89. YOUNG, S.P. and E.A. GOLDMAN I9I4.I4. The wolves of North America. xx-636. American Wildlife Institute, Washington. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items