UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The genus Martes (Mustelidae) in North America: |b its distribution, variation, classification, phylogeny… Hagmeier, Edwin Moyer 1955

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

Item Metadata


831-UBC_1955_A1 H2 G3.pdf [ 26.89MB ]
JSON: 831-1.0106267.json
JSON-LD: 831-1.0106267-ld.json
RDF/XML (Pretty): 831-1.0106267-rdf.xml
RDF/JSON: 831-1.0106267-rdf.json
Turtle: 831-1.0106267-turtle.txt
N-Triples: 831-1.0106267-rdf-ntriples.txt
Original Record: 831-1.0106267-source.json
Full Text

Full Text

THE GENUS MARTES (MUSTELIDAE) IN NORTH AMERICA: ITS DISTRIBUTION, VARIATION, CLASSIFICATION, PHYLOGENY AND RELATIONSHIP TO OLD WORLD FORMS by Edwin M. Hagmeier A Thesis Submitted i n Partial Fulfilment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY i n the Department of Zoology We accept this thesis as conforming to the standard required from candidates for the degree of Doctor of Philosophy Members of the Department of Zoology THE UNIVERSITY OF BRITISH COLUMBIA September 1955 F a c u l t y o f G r a d u a t e S t u d i e s P R O G R A M M E O F T H E Jffinai ^xxatttinatimt for ttye ]^t$xtz of Jioxtor of ^Irilrfsoijliy EDWIN HAGMEIER B . A . ( Q u e e n ' s ) 1 9 5 0 M . A . ( B r i t i s h C o l u m b i a ) 1 9 5 2 A U G U S T 8 t h , 1 9 5 5 , a t 3 : 0 0 p . m . I N T H E B I O L O G I C A L S C I E N C E S B U I L D I N G R o o m 1 0 7 A C O M M I T T E E I N C H A R G E H . F . A N G U S , Chairman I . M c T . C O W A N V . J . O K U L I T C H ," H . C . G U N N I N G P . FORD K . G R A H A M J . G . SPAULDING G . J . S P E N C E R V . K R A J I N A External Examiner — R . J . R U S S E L L N a t i o n a l M u s e u m o l C a n a d a ABSTRACT T h e G e n u s Martes ( M u s t e l i d a e ) i n N o r t h A m e r i c a : i t s D i s t r i b u t i o n , V a r i a t i o n , C l a s s i f i c a t i o n , P h y l o g e n y a n d R e l a t i o n s h i p t o O l d W o r l d F o r m s T h r e e s u b g e n e r a o f t h e g e n u s Martes e x i s t i n t h e w o r l d t o d a y . T h e f i r s t , Martes, c o n s i s t s o f t h e f o l l o w i n g s p e c i e s : M. americana, M. foina, M. martes, M. melampus a n d M. zibellina; t h e s e c o n d , Pekania, o f o n l y o n e , M. pennanti; a n d t h e t h i r d , Charronia, o f o n e , M. flavigula, w i t h s o m e t i m e s a s e c o n d , M. guatkinsi. W i t h i n t h e s u b g e n u s Martes, M. americana, M. martes, M. zibellina a n d p o s s i b l y M. melampus a r e s o c l o s e l y r e l a t e d m o r p h o l o g i c a l l y t h a t t h e r e a p p e a r s g o o d r e a s o n t o b e l i e v e t h a t t h e y a l l b e l o n g t o o n e s p e c i e s . T w o s p e c i e s o c c u r i n N o r t h A m e r i c a , n a m e l y M. americana a n d M. pennanti- U n t i l r e c e n t l y M. americana w a s c o n s i d e r e d t o c o n s i s t o f t w o s p e c i e s , M. americana a n d M. caurina, a n d t h i r t e e n o r s o s u b s p e c i e s . T h e r e c e n t w o r k o f P . L . W r i g h t i n d i c a t e s t h a t w h i l e t h e t w o " s p e c i e s " a r e d i s t i n c t i v e m o r p h o l o g i c a l l y , t h e y i n t e r g r a d e a t t h e p o i n t w h e r e t h e i r r a n g e s m e e t a n d m u s t b e c o n s i d e r e d a s i n g l e s p e c i e s . Martes pennanti h a s b e e n c o n s i d e r e d o n e s p e c i e s , c o n s i s t i n g o f t h r e e s u b s p e c i e s . T h e c o n c e p t o f t h e s u b s p e c i e s p r o v e s i n m a n y r e s p e c t s t o b e u n -s a t i s f a c t o r y . I t l a c k s r e a l i t y , i t i n v o l v e s t h e a r b i t r a r y p a r t i t i o n i n g o f c o n t i n u a , i t p o s s e s s e s n o l o w e r l i m i t , a n d i t i s d e t e r m i n e d d e d u c t i v e l y . T h i s , t o g e t h e r w i t h t h e c l i n a l n a t u r e o f v a r i a t i o n i n m a r t e n a n d f i s h e r l e a d s t o t h e c o n c l u s i o n t h a t m a r t e n o f t h e N e w W o r l d s h o u l d b e c o n -s i d e r e d a s r e p r e s e n t e d b y o n l y t w o s u b s p e c i e s ( M . a. americana a n d M . a. caurina), t h e f i s h e r b y o n e s p e c i e s , a n d n o n a m e d s u b s p e c i e s . T h e d i s t r i b u t i o n o f m a r t e n a n d f i s h e r c o r r e s p o n d s c l o s e l y t o t h e d i s t r i b u t i o n o f t h e n o r t h e r n e v e r g r e e n f o r e s t s . T h e d i s t r i b u t i o n i s l e s s p r e c i s e i n f i s h e r t h a n i n m a r t e n . F o s s i l s r e f e r a b l e t o t h e g e n u s Martes a r e f i r s t r e c o r d e d f r o m t h e M i o c e n e o f b o t h t h e O l d a n d N e w W o r l d s . T w e n t y - e i g h t f o s s i l s p e c i e s a r e k n o w n ( w h e n s y n o n y m s a r e d i s p o s e d o f ) , o f w h i c h f i v e a r e s t i l l l i v -i n g , t w o o f t h e m i n N o r t h A m e r i c a . I t a p p e a r s t h a t m o d e r n m a r t e n s a n d f i s h e r s a r r i v e d i n t h e N e w W o r l d ( o r e v o l v e d t h e r e ) l a t e i n t h e P l i o c e n e o r e a r l y i n t h e P l e i s t o c e n e . D u r i n g t h e P l e i s t o c e n e m a r t e n f o u n d h a b i t a b l e e n v i r o n m e n t s i n t h e f o r e s t r e f u g i a o f s o u t h e a s t e r n U n i t e d S t a t e s , t h e R o c k y M o u n t a i n s s o u t h o f t h e i c e s h e e t , t h e C o a s t a n d C a s c a d e M o u n t a i n s s o u t h o f t h e i c e s h e e t , a n d A l a s k a a n d Y u k o n . F i s h e r p r e s u m a b l y o c c u r r e d i n a l l o f t h e s e r e f u g i a e x c e p t t h e A l a s k a - Y u k o n o n e . W i t h p o s t - g l a c i a l c l i m a t i c a m e l i o r a t i o n t h e y m i g r a t e d t o t h e r e g i o n s o f t h e i r p r e s e n t o c c u r r e n c e . GRADUATE STUDIES F i e l d o f S t u d y : Z o o l o g y V e r t e b r a t e Z o o l o g y — I . M c T . C o w a n I n v e r t e b r a t e Z o o l o g y — W . A . C l e m e n s O t h e r S t u d i e s : V e r t e b r a t e P a l a e o n t o l o g y — J . T . G r e g o r y i i TABLE OF CONTENTS Page LIST OF ILLUSTRATIONS V INTRODUCTION 1 CLASSIFICATION 6 THE GENUS MARTES 6 HISTORY OF CLASSIFICATION 6 TAXONOMIC POSITION 12 THE SUBGENERA OF MARTES 14 HISTORY OF CLASSIFICATION 14 TAXONOMIC CHARACTERS 16 THE SPECIES OF MARTES 18 HISTORY OF CLASSIFICATION 18 In Eurasia 18 In North America 22 Marten 22 Fisher 32 SYNOPSIS OF THE SPECIES AND SUBSPECIES OF MARTEN .. 37 Martes martes 41 Martes foina 46 Martes zibellina . 53 Martes melampus 62 Martes americana (including caurina) 65 Martes pennanti 69 Martes flavigula 70 RELATIONSHIPS OF THE MARTEN 79 /continued i i i CLASSIFICATION (continued) page THE MARTENS OF THE NEW WORLD 90 LIFE HISTORY 90 VARIATION 9 4 Nongeographic variation 9 4 Methods of study of geographic variation . 106 MARTENS 110 Distribution 110 Synoptic examination 129 "Martes americana" 130 "Martes americana americana 130 "Martes americana brumalis" 143 "Martes americana atrata" 148 "Martes americana abieticola" 153 "Martes americana actuosa" (including"boria") 159 "Martes americana kenaiensis" 173 "Martes americana abietinoides" 178 "Martes caurina" 187 "Martes caurina caurina" 188 "Martes caurina origenes" 203 "Martes caurina humboldtensis" 208 "Martes caurina sierrae" 212 "Martes caurina Vancouverensis" 219 "Martes caurina nesophlla" 223 Summary 229 "Species" of marten 230 /continued iv CLASSIFICATION (continued) Page FISHERS 242 Distribution 242 Synoptic examination 262 "Martes pennanti pennanti" 262 "Martes pennanti columbiana" 274 "Martes pennanti pacifica" 2 79 CLASSIFICATION OF MARTEN AND FISHER .. 289 PHYLOGENY .. . 298 PALAEONTOLOGY 298 BIOGEOGRAPHY 312 INTRODUCTION 312 REFUGIA OF THE WISCONSIN 327 POST-WISCONSIN MIGRATION 340 MARTEN AND FISHER 342 SUMMARY 348 LITERATURE CITED 354 APPENDICES 439 BRONGERSMA'S INDICES 439 SYNONYMY 441 CRANIAL MEASUREMENTS 463 SOURCES OF LOCALITY RECORDS 466 ABSTRACT 468 V ILLUSTRATIONS Figures Facing Page 1 and 2. The skulls of the species of the world's marten 42 2, 3» 4, The distribution of Eurasian marten 42 5 and 6. 7. Ratio of canine width to bulla length i n the skulls of the world's marten 86 8. Pie diagrams depicting certain characteristics of the skulls of the world's marten 88 9. Evidence of intergradation between "Martes  americana" and "Martes caurina" i n Idaho and Montana, after Wright 1953 235 10. Evidence of intergradation between "Martes  americana" and "Martes caurina" in Manning Park, Br i t i s h Columbia 235 11. Distribution of the boundary separating "Martes  americana" from "Martes caurina" 236 12. The distribution of the f o s s i l species of marten 308 13. A hypothesized phylogeny of the Recent species of marten 312 14. North America i n Mankato time 327 15. 16, North America during the Pleistocene 32? and 17. 18. The distribution of Pleistocene forest f o s s i l s . . 329 19. The relative densities of marten i n North America 333 20. Pdst-Mankato migration of marten and fisher i n North America 346 21. Variation in M. a. americana and i t s possible relation to climate 348 ENCLOSURE A. Geographic distribution of condylobasal length in marten. ENCLOSURE B. Distribution of marten i n North America ENCLOSURE C.' Distribution of fisher i n North America 1 INTRODUCTION The genus Martes constitutes the fishers and martens. Seven or eight species are considered to occur i n the world today, including the pine and stone marten, the sable, yellow bellied and Japanese marten in the old world, and the American marten and fisher i n the new. No revisionary work has been done on the new world forms in half a century, and i t i s the purpose of this paper to f i l l the need. Yeager (1941) points out that of the fur bearing mammals he considered, only badger, otter and raccoon have received less attention in the literature than the wolverine, fisher and marten together, so that i t i s apparent that much future work i s needed before these animals can be considered relatively well understood. Myers (1952) suggests that systematic papers should make the problems being solved clear, hence the following state-ment of the problem. This i s an attempt to define the limits of the genus concerned, to study i t s geographic variation i n North America, to devise a useable classification from this, to determine the relationship of the new world species to those of the old world, and to devise a plausible phylogeny for the whole genus, with emphasis here on the American forms. This study is unusual i n that i t attempts to bridge two points of view concerning marten classification which neces-sitates the use of certain arbitrary standards of usage, with which the reader must be familiar to understand what follows. 2 The older classification considers North American marten and fisher to be made up of three species containing between them sixteen subspecies, whereas I attempt to advance the view that a truer conception of re a l i t y i s to recognize only two species, and two subspecies. These two ideas are not by any means self-exclusive, but merely a different interpretation of the same thing. Classification i s always arbitrary and since whatever view accepted i s a matter of opinion, I am bound to present both views, the older scheme and my own. In the following, as a means by which the reader may recognize what classification i s • referred to at any given place in the text, the term species, subspecies (or form or race) and the trinomial or i t s components are always, when referred to i n the older sense, placed i n Inverted commas. The use of inverted commas for the older scheme i s not meant to imply derogatory treatment of i t . The scheme could have been reversed. Since, however, I am champion-ing the new plan, i t seems logical to employ the above method. To orientate the reader then, the old scheme of c l a s s i f i -cation i s as follows: "Martes americana" "M. a. americana" "M. a. atrata" "M. a. brumalis" "M. a. abieticola" "M. a. actuosa" "M. a. kenaiensis" "M. a. abietinoides" "Martes caurina" "M. c. caurina" "M. c. orieenes" HM. c. sierrae" "M. c. humboldtensis" "M. c. vancouverensis" "M. c. nesophila" "Martes pennanti" "M. p. pennanti" "M. p. columbiana" "M. p. p a c i f i c a " The new scheme I o f f e r , a s i m p l i f i c a t i o n of the above, i s as follows: Martes americana M. a. americana M. a. caurina  Martes pennanti This study i s based upon the examination of 3301 marten and f i s h e r s k u l l s of which measurements were taken on 1698. The s p e c i f i c d i s t r i b u t i o n of these i s as follows: Martes americana americana: 1803 examined, 1044 measured M. a. caurina: 936 examined, 390 measured M. pennanti: 352 examined, 206 measured M. f l a v i g u l a and guatkinsi: 32 examined, 2 measured M. fo i n a : 81 examined, 19 measured M. martes ; 12 examined, 7 measured M. z i b e l l i n a : 77 examined, 20 measured M. melampus: 8 examined, 3 measured 4 Besides these a representative number of skins of each were exam-ined. The specimens examined are i n the collections of the following: University of Br i t i s h Columbia, Bri t i s h Columbia Prov-i n c i a l Museum, National Museum of Canada, Royal Ontario Museum of Zoology, University of Michigan, Carnegie Museum, Museum of Vertebrate Zoology i n Berkeley, American Museum of Natural His-tory, United States National Museum, United States Biological Survey, University of Utah, Museum of Comparative Zoology i n Cambridge, and Peabody Museum, Yale University. Besides these, I have been allowed to examine specimens belonging to Messrs R. Y. Edwards, W. Cottle, R. Webb, J. Bryant, K. Racey and Doctors P. L. Wright and L. E. Yeager. Mr. D. Flook and Mr. C. Lensinck have donated specimens of their own to me, and Mr. E. McEwen provided me with measurements of specimens i n his possession. It i s pleasant to name the individuals who have provided assistance during the course of the study. They include the custodians of the above named collections, especially Drs. D. Johnson, and H. Tate, and Miss Viola Schantz. Dr. Johnson also loaned specimens and transcribed portions of his rare copy of Kerr's Animal Kingdom for my use. Dr. P. L. Wright, besides allowing me to examine his specimens, sent me a manuscript copy of his paper on marten conspecifity a year before i t s publication. Mr. K. Racey provided valuable advice on western martens. Dr. L. Yeager cleaned a large series of skulls so that I might examine them. Mr. Calvin Lensinck and Mr. Donald Flook both made gifts of marten skulls from Alaska and Mackenzie. Mr. Lensinck also provided material on marten distribution. Dr. G. G. Simpson 5 provided information on nomenclature. Dr. I. McT. Cowan is directly responsible for the instigation of this study i n that he f i r s t suggested the problem to me, and has guided me throughout, though he must not be held responsible for i t s errors. Both he and Dr. W. Clemens have been a source of ready information and encouragement; I am obliged to both too, for having assisted i n obtaining funds to assist me i n the work. These funds were provided by the B r i t i s h Columbia Sugar Refining Company and the B r i t i s h Columbia Research Council. Further assistance has been provided by the following: Drs. Peter Larkin, K. Graham, V. Krajina, J. Spencer, M. Udvardi, J. T. Gregory, N. Preble, G. Swanson, L. Butler, E. L. Coekrum, Messrs. R. Denney, J. Bryant, A. Cameron, R. W. Sutton, F. W. Fay, J. S. Tener, F. C. Kleinschnitz, B. S. Wright, D. Pimlott, and Miss Helenette Silver. Mrs. E. F. Bonoff helped i n the transla-tion of Russian literature. The academic year 1953-54 was spent i n study at Yale University. For this, obligation i s owed to Drs. E. S. Deevey, Jr., J. S. Nicholas and Mr. G. E. Hutchinson. It i s a pleasure to acknowledge the help given by my wife, Elizabeth Hagmeier, whose assistance during the preparation of this thesis has been incalculable. 6 CLASSIFICATION A. THE GENTJS MARTES HISTORY OF CLASSIFICATION The post-Linnaean synonymy of the generic term Martes as applied to martens and fishers, follows: 1758 1758 1765 1765 1771 1774 1775 1777 1792 1800 1816 1820 1829 1843 1847 1865 1865 1865 1865 1873 Mustela Linnaeus, vol. 1, p. 45. Zibelline Buffon and D'Aubenton, vol. 7, p. 309. Pekan Buffon and D'Aubenton, vol. 1, p. 304 and p i . 42. Fouine Buffon and D'Aubenton, vol. 13, p. l 6 l . Fisher.Pennant, p. Pekan Charlevoix, vol. 3 , p. 134. Martes Frisch, p. 11. Mustella Scopoli, p. 491. Martes, Pinel, vol. 1, p. 55. Type Martes domestica Pinel = Martes foina Erxleben. Viverra Shaw, vol. 1, p. 414. Tavra Oken , p. 1001. Martes, Nilsson, vol. 1, p. 38. Zibellina Kaup, vol. 1, p. 31. Galadicitis Smith, W. Jardines Nat. Library, vol. 35> p t . l , p. 167. Putorius Boitard, vol. 8, p. 12 (Diet. Univ. Hist. Nat. 8:) Charronia Gray, p. 108. Foina Gray, p. 108. Pekania Gray, p. 107. Gulo "H. Smith", according to Gray 1865: 108. Aelurodon Cope, p. 1. 7 1904, Palaeogale Trouessart, p. 201. 1907, Pustella Millard, p. 1040 (Journ. Bombay Wat. Hist. Soc. 1 7 0 . 1911, Martes. Thomas, p. 139. 1924, Plionictis Matthew, vol. 21, p. 135. 1928, Lamprogale Ognev, p. 26. 1937* Charionia Thorn, p. 317 (Journ. Bombay Nat. Hist. Soc. 39). Ray, i n 1693 was the f i r s t to use the name Martes in a generic sense, according to Trouessart (1897). Linnaeus, however, used the term Mustela for the genus including with i t the weasels. For this reason i t took a long while for agreement to be reached as to the proper designation for the genus of martens i n the st r i c t sense. During the 18th and 19th centuries, Mustela was held to be the proper generic term, the weasels being named P u t o r i u s b u t since the f i r s t decade of the 20th century, Martes has been used, Mustela being reserved for the weasels, minks and ermines, and Putorius as genus or subgenus for the black-footed ferret Mustela (Putorius) nigrepes. Alston, in 1879> was the f i r s t to interest himself i n the problem. He wrote that most writers used the term Mustela for the martens on the ground that Cuvier (1817) had done so. Alston concluded that the term was used not as Linnaean genus, but rather as "sous-genres" and hence was inapplicable. The f i r s t proper naming of martens was done by Nilsson i n 1820, who used the term Martes. Thus, concludes Alston, the proper term 8 should be Martes Nilsson 1820. He adds that certain authors have used the term Martes i n the past, not after Nilsson, but from Cuvier in 1797* This i s incorrect, he states, since the word used was merely the plural form of the French Martre. Flower and Lydekker i n 1891 state that while Cuvier (1817) used the terra Mustela as a sub-genus for the martens, usual practice allows them consideration as a f u l l generic term. They conclude that martens should be named Mustela Cuvier 1817. Trouessart (1898-99) followed Alston in refuting Mustela Linnaeus as the proper term, and retaining Martes Nilsson. Allen i n 1902 (a) gave consideration of Oken's (1816) use of Tavra for the genus name, and while considering him scarcely Linnaean, nowhere completely condemned him. Sherborn i n 1902-1933 gave authority for Martes to Pinel (1792). In 1904, Palmer, i n his Index Generum, attributed the name Martes to Frisch (1775)* Thomas and Miller i n 1905 con-demned Palmer's usage of Frisch's term. They state that while Frisch appears superficially binomal, he i s not, and often used phrases for his generic terms, and that his terminology was thus unacceptable. Thomas, i n 1911? reviewing the conflicting claims of the terms Mustela and Martes as applying to the martens, con-cluded that only Martes was valid. He said that Linnaeus chose many of his terms from Gesner (1606 and 1617-1621) and that by going to Gesner one could determine what "type" Linnaeus was thinking of i n his application of names. Concerning the genus Mustela (Systema Natura 10, p. 45), the type, he concluded, i s 9 by tautonymy Mustela erminea ("Mustela" Gesner, misquoted as Mustela vulgaris). In this instance alone of the Gesner quotations i n the Systema, he said, i s a second name attached to the primary one, for which tautonymy i s claimed. This proved to be a misquotation by Linnaeus, for while Linnaeus wrote "Mustela vulgaris" Gesner, Gesner actually wrote only "Mustela" i n the German edition quoted by Linnaeus, while in the Latin edition he wrote "Mustela proprie sic dicta" i n contradistinction to his "Mustelis dlversis", which included the marten and the polecat} Thomas concluded that the Mustela referred to by both Gesner and Linnaeus was clearly the ermine, and that i t must be considered the type species. Almost a l l writers since Thomas have used Martes, under various authorities, for the martens, restricting Mustela to the weasels and ermines. Heider, Kuhlgatz, Hesse, Schulze and Kukenthal (1926-35) assert that a reference to Martes was made by Brisson i n 1762, but assert that neither he nor Plnel can be considered legitimate authorities. Wagler's (1830) usage of the term was used for Herpestes (a vlverrid) and concluded that Nilsson (1820) must be considered the proper authority. Ognev i n 1931 credited the term to Pinel, as did Neave (1930-1940). Brongersma, i n 1941, reviewed the problem of the use of the term Mustela and concluded that four different species were involved as types. De Blaineville (1841), he asserts, employed Mustela foina as an anatomical type, even though i t was not men-tioned by Linnaeus (see above). Chenu and Desmarest (date uncertain, but probably 1852 or 1853> P« 263) named M. martes as 1G type, although not i n a nomenclature^ sense. Coues i n 1877 (p. 60) named Mustela martes. This type was also used by Cuvier (1817, pp. 142-149), Miller and Rehn (1901, p. 226), Pal-mer (1904, p. 436) and E l l i o t (1905b, p. 419). E l l i o t in 1901 (p. 33) mentioned Mustela lutra as the type, but this apparently was a lapsus, since on page 352 he mentioned Mustela lut r a as the type of the genus Lutra Erxleben. Thomas i n 1911 > as pointed out above, named Mustela erminea Linnaeus the type of the genus. Brongersma concluded that i f the International Regula-tions were to be closely followed, Coues* designation (M. martes) would have to be maintained over that of Thomas (M. erminea). Since, however, Thomas' usage had been employed by most workers for thirty years, much confusion would be created by a change and the author expressed the hope that the Commission would pass an opinion making Mustela erminea the type of the genus Mustela, so that the martens might retain the name Martes. Brongersma be-lieved the authority for the latter term should be attributed to Pinel. In 1945, Simpson i n his Classification of Mammals revert-ed to Palmer's usage of Frisch as the authority for Martes. In a letter dated 1951 he writes: "I took Martes Frisch 1775 from Palmer.... Since Frisch i s doubtfully or not binomial, the more usual and probable reference i s Martes Pinel.... There i s , however, also some question as to Pinel's absolute purity ... so perhaps i t should be Martes Nilsson 1820. If Nilsson should not pass, either, then the name is lost because the next reference, Wagler, 1830, uses Martes for a group of viverrids... 11 Hershkovitz (1948a) says that Palmer i n 1941 i n private conversation denied the v a l i d i t y of Frisch's names, which Hersh-kovitz did also. The same author (Hershkovitz 1948b) reasserted his opinion, saying that Frisch was non-Linnaean, and persisted only because Palmer and Allen (Bull. Amer. Mus. Nat. Hist. 1902, 16: 13-22) had employed them. In 1949 the same author denied the v a l i d i t y of Oken's (1816) use of Tavra for the martens, saying the author was non-Linnaean. Ellerraan and Morrison-Scott (1951) suggest that Brisson's names could not be considered Linnaean and that Oken's names f a l l into much the same category. Concerning Frisch's terminology they report that the International Commission in Paris in July, 1948, declared his work to be unavailable (Bull. Zool. Nomenclature 1950, 4:549). I have been unable to examine this journal. To summarize the above, i t appears that by s t r i c t application of International Regulation, Mustela should be used as the generic term for the martens but that common usage and con-sideration of types gives the term Martes priority (Brongersma 1941, Thomas 1911). Although Frisch was the f i r s t to use the term, his names have been declared Illegitimate (Ellerman and Mor-rison-Scott 195D and the authority for the name must l i e with any of Brisson (1?62), Pinel (1792) or Nilsson (1820). Brisson has generally been considered unavailable, and so i t appears li k e l y that Pinel w i l l retain permanent authority for the term. For these reasons I accept Martes Pinel as designating the genus. In 1918, Pocock, on the basis of the structure of the baculum of the Indian and Malaysian species M. flavigula and M. guatklnsi, separated them as the distinct genus Charronia. 12 Ogney i n 1928 showed that this name was preoccupied by Charonia Gistel 1848, p. 559, for a genus of Molluscs, and replaced i t with the name Lamprogale. Ognev i n 1931 and Pocock i n 1936(b) accepted the change but Howell 1929, Allen 1938 and 1940, Simpson 1945 and Tate 1947 retained Charronla. Allen and Simpson i n s i s t -ing that the change was not merited. Most recent authors, however, refuse to accept Pocock1s distinction as generic, and Ognev ( 1 9 3 D , Carter H i l l and Tate (1944), Bobrinskoy et. a l . (1944) and Ellerman and Morrison-Scott (195D consider Charronia or Lamprogale merely a subgenus of the genus Martes, which interpretation i s accepted i n this paper. TAXONOMIC POSITION The martens and fishers (Martes) have invariably been placed within the family Mustelidae, although how this family l i e s relative to the Arctoids and Aeluroids has never been precisely determined. For convenience i t has generally been considered to l i e nearer the former, however. Within the Mustelidae a great many subfamilies have been devised, especially by Pocock, although most authors e.g. Simpson and Ellerman, Morrison-Scott name only four or five. Martes i s generally placed i n the subfamily Mustelinae, although some,,for example Pocock and Allen (1938 and 1940), name both the Martinae and Mustelinae, the f i r s t for the martens, the last for the weas-els etc. For greater detail see Simpson (1945). For ordinal characters of the family Mustelidae, and the sub-family Mustelinae see Miller 1912a, Pocock 1921a and b, Allen 1938 and 1940, Stiles and Baker 1935. 13 About a dozen recent genera of Mustelinae are consid-ered to e x i s t today (Simpson 1945) but other authors have named more. About another twenty to twenty-five genera have been named as f o s s i l forms but again there i s great disagreement among authors. Keys to the recent genera are given by Gray 1865, M i l l e r 1912a, Pocock 1921b, A l l e n 1938-40, Dammerman 1940, Ognev 1931 and others. Keys to the f o s s i l forms (now outdated) as well as the recent ones are given by Matthew 1924 and Winge 1941. The characters of the genus Martes have been given by many authors, including the following, Flower and Lydekker 1891, M i l l e r 1912a, Pocock 1921a and 1921b, Ognev 1931» S t i l e s and Baker 1935> H a l l 1936, Dammerman 1940, and Winge 1941. There follows a summary of the characters of the genus, compiled from the above: are Mustelid carnivores with body long, slender and f l e x i b l e , though less so than i n Mustela; head some-what t r i a n g u l a r , muzzle pointed; limbs short, feet rounded, toes short, f i v e toes on each foot; t a i l moderately long, more or less bushy; s k u l l long, narrow and ranging i n b a s i l a r length from 60 to 115 mm.; f a c i a l angle s l i g h t ; tympanic bullae moder-ately i n f l a t e d , with t h i n w a l l s , not i n close contact with para-o c c i p i t a l processes; palate extending behind l a s t upper molars; the brain although large, i s not excessively so; the anterior edge of the o r b i t , although s l i g h t l y pushed forward, i s not greatly displaced; the jugular process i s d i s t i n c t ; dental formula ^ i l t ] ^ inner moiety of M 1 large, at least as large as outer with two i n t e r n a l cusps (protocone and an accessory cusp) and a cingulum; P4" with a single deuterocone, t h i s forming a 14 distinct internal lobe; length of whole tooth greater than i t s width; lower with evident though small metaconid and entoconid (thus with a total of 5 cusps); trigonid always longer than talonid, the latter semibasened. B. THE SUBGENERA OF MARTES HISTORY OF CLASSIFICATION Although Cuvier (1817) and certain other authors notably Lydekker (1885) used subgenera to separate the martens from the weasels, interest here must be restricted to subgeneric division within the martens themselves. The f i r s t to interest himself in the problem was Gray, who i n 1865 and I869 named four subgenera, which with a modern interpretation of the species constituting them, follows: Genus Martes: Subgenus Martes: Martes martes, M. melampus, M.' zib-e l l i n a and M. americana. Subgenus Pekania: Martes pennanti. Subgenus Foina: Martes foina. Subgenus Charronia: Martes flavigula and M. guatkinsi. Pocock i n 1918, as has been described earlier, reasserted the v a l i d i t y of Gray's subgenera Martes and Charronia, naming Charronia a f u l l genus and basing distinction on baculum structure. Pekania, as a subgenus, was not considered but in North America at least has continued to persist as a taxonomic unit. Foina was not upheld by Pocock, nor has i t persisted. Ognev, i n 1928 (as discussed earlier) changed Charronia to Lamprogale. Certain 15 authors, namely Ognev (1928), Ognev (193D and Pocock 1936b accept Lamprogale, but most (Howell 1929 > Allen 1938 and 1940, Simpson 1945» Tate 1947 and EllermariiMorrison-Scott 195D retain Charronia. Howell, Simpson and Tate give Charronia f u l l generic status, while Carter H i l l and Tate (1944) and Ellerman, Morrison-Scott consider i t a subgenus. Ognev 1931» Allen 1938-40 and Bob-rinskoy 1944 give Lamprogale subgeneric status only. In 1931 Hall named Martes gazini from the tertiary of North America and erected the new subgenus Tomictis for i t s use. Similarly i n 1946 Shikama named the subgenus Ten for Martes ten from the Pleistocene of Japan. We find then, that the following are the subgenera that have been named to constitute the genus Martes. Subgenus Martes Pinel 1792, vol. 1, p. 55 fide Gray 1865 and 1869. Subgenus Pekahia Gray 1865, p. 107. Subgenus Charronia Gray 1865, p. 108, Pocock 1918, p. 307, synonymous with Lamprogale Ognev 1928, p. 26 and 30, considered by some a genus. Subgenus Foina Gray 1865> p. 108. Subgenus Tomictis Hall 1931 > P« 156. Subgenus Ten Shikama 1946, p. Only three of these subgenera have been held valid by students of recent mammals (e.g. Ognev 1931» Bobrinskoy 1944, Miller 1924, Anderson 1946, Ellerman Morrison-Scott 195D• These with their constituent species are as follows: Subgenus Martes: Martes martes, M. foina, M. zibellina, M. melampus and M. americana. Subgenus Charronia = Lamprogale: Martes flavigula and M. guatkinsi. 16 Subgenus Pekania; Martes pennanti. Because t h i s c l a s s i f i c a t i o n i s currently accepted i t i s the one used i n the present paper. Although only two subgenera of f o s s i l martens have been named, i t appears l i k e l y that a c r i t i c a l examination of f o s s i l members of the genus would result i n the naming of several to many more. Nor i s i t known with a few exceptions what f o s s i l species are represented i n the subgenera of recent martens. TAXONOMIC CHARACTERS I t i s generally stated that nine species of marten exist i n the world today, these being "Martes americana", "caurina", martes, f o i n a , z i b e l l i n a , pennanti, f l a v i g u l a , guat-k i n s i , and melampus. For reasons given elsewhere, however, t h i s paper assumes the number to be only seven, by considering "M. americana" and "caurina" a single species (M. americana), and M. f l a v i g u l a and M. guatkinsi another. As stated e a r l i e r , these are grouped into the following subgenera: Subgenera Martes: Martes americana, martes, f o i n a , z i b e l l i n a , and melampus Subgenus: Pekania: Martes pennanti Subgenus Charronia ( F l a v i g u l a ) : Martes f l a v i g u l a The c h a r a c t e r i s t i c s of the subgenera Martes and Char-ronia have been c a r e f u l l y considered by Pocock (1918), Ognev (19330, Pocock (1936b) and Bobrinskoy (1944). The d i s t i n c t i v e nature of the subgenus Pekania has never been considered serious-l y and i n fact i s not so clear cut as i s that between the f i r s t 17 two named groups. A summary of the characteristics of the subgenera of martens follows. 1. Subgenus Charronia = Lamprogale. Colour varied black and yellow above; upper l i p not divided by a distinct vertical groove; baculum distinctive, the distal end curved somewhat abruptly upwards and backwards in the form of a hook, the tip slightly expanded to produce four processes, each with a rounded condyle-like head. Distribution: through the mountainous regions of northern India, northwards through China to Amur, and southward through Malaysia to Sumatra, Java and Borneo, also i n the h i l l s of southern India. 2. Subgenus Martes: Colouration generally monotone, brownish, but often with a lighter or yellowish spot on breast; upper l i p divided by a distinct vertical groove; baculum not as above, gradually inclined upwards and forwards at i t s distal end, where i t divides into two processes which remain separate, or which may unite. Distribution: a l l of holarctic Europe, Asia and North America. 3. Subgenus Pekania: Colouration usually monotone, brownish, with chest spot absent or reduced, usually whitish; upper l i p divided by a distinct v e r t i c a l groove; baculum as i n the sub-genus Martes,. It is so like the preceding subgenus that i t i s best distinguished by i t s relatively larger size, the length of i t s head and body being more than 570 mm. instead of less and i t s condylpbasal length being more than 95 nun* instead of less. Distribution: holarctic North America. 18 C. THE SPECIES OF MARTES HISTORY OF CLASSIFICATION In Eurasia Adequate histories of the development of the present concepts of the classification of the genus Martes are not avail-able, although attempts of varying degree have been made by de Blainville (1841), von Martens (1870) as summarized by Coues (1877) and Alston (1879)• Much of what follows i s collated from these. It appears that the ancient Greeks were the f i r s t to recognize and name what we now place i n the genera Martes, Mustela and Putorius. Three names were used for these, i n their Anglicized form, I c t i s , Gale and Mustela, but these terms were used rather indiscriminately. Aristotle compared Gale to Icti s , saying that the latter had a thicker pelage and a paler belly. Elsewhere he said that Gale of the forest waged war against rats. De Blainville believed that by Gale Aristotle meant weasel and by Ictis marten, and that the distinction of the two had been made by then. Gesner and later authors refer to the marten even more distinctly. It i s believed, however, that these writers s t i l l confused the polecat (Putorius) with the weasel (Mustela) because Aristophanes used the word Gale to refer to an unclean woman, and Herodotus referred to a Gale of Africa, which was l i k e l y a Viverrid. The Roman writers did not use the older Greek termin-ology, developing rather their own, e.g. Martes and Mustela, i n their present form. Thus the origin of these words i s nearly unknown, or at least obscure. In. spite of this these are the 19 terms used by modern zoologists. In Leviticus 11, 29 Moses listed a number of animals that were considered impure. Among these was the Choleb which many translators have interpreted as Gale or Mustela. De Bla i n v i l l e believes this was the stone marten (Martes foina) of today. It i s uncertain who f i r s t made the distinction between the two species of European marten (Martes foina and the pine marten Martes martes). De Blainville attributes the f i r s t dis-tinction to Pliny, who while transcribing Aristotle's statements referred to above, substituted the word Viverra for Gale and Mustela for I c t i s . Ever since, he said, Mustela has appertained primarily to the stone marten. Viverra, i t is assumed, was meant by Pliny to be the pine marten. Martial replaced the term Viverra with Martes, though providing no description to either. Most later authors, especially Agricola and Gesner, maintained the distinction, most applying Martes to the pine marten, Mustela to the stone marten, and Putorius and Ictis to the weasels and polecats. Von Martens and Alston (1879) however, considered neither Aristotle nor Pliny to have recognized the distinction seen by de Bl a i n v i l l e . It may have f i r s t been done by Albertus Magnus (who mentioned both Martarus or Martes fagorum and Martes  abietum). Just as l i k e l y though, they believed, the distinc-tion was not really made unt i l as late as the time of Buffon and Linnaeus. The cranial and dental characters by which the pine and stone martens could be distinguished were f i r s t recognized and pointed out by Hensel i n 1853 and further elaborated by Blasius i n 1857 (Saugethier Deutschlands, pp. 211-214), Gray 1865, Coues 1877, Alston 1879, Miller 1912a, Reynolds 1912, Ognev 1931, Bobrinskoy et al.1944, Rode and Didier 1944, among others. The origin of the name Mustela i s believed by von Martens to originate i n the word Mus. Sundevall believed further that i t came from the Greek word for hunt, since Palladlus referred to the mouse hunting a b i l i t y of the animal and since i t i s s t i l l called Moustelle i n Nice and Lorraine. Aubert and Wimmer, said von Martens, concluded, however, that a l l the older references are really to the stone marten i n the s t r i c t sense, since i t occurs throughout Greece (see also Rolleston 1868) and i s s t i l l called Ictis there. The origin of the word Martes i s s t i l l obscure. Kerr (1792) considered,the Satherius of Aristotle to have been the f i r s t reference to the sable, and Shitkov (1940) agrees. Pallas (1811) and Heptner (1934) attributed the name to Albertus Magnus, however, and de B l a i n v i l l e (1841) to Agricola. The f i r s t statement of the distinctive marks of the species was made by Aldrovandi and Ambrosini (according to de Blainville) and further elaborated by Brandt (1855), Gray (1865 and 1869) and others named above. The specific names of the three species of martens dis-cussed previously were considered by von Martens as follows: Zibellina has been known since the latter half of the middle ages under many variations such as sabelus, zibellina, zebel, sebell, sable, sobol, soboli and samur. Graff says the Germanic term zebel i s as old as the 9th century. 21 Martes now occurs i n nearly a l l Germanic and Romantic languages as marta, martre, mart, martora, martin, marten, marter, mard, maar, meardtr and martarus. The word was f i r s t used by Martial as Martes, probably being taken from some non-Romantic language. Foina i s present i n nearly a l l Romantic languages. Forms include fuina, fulntra, foulne, faina, faguino, fahino, f ayna, fagina, farveina, fierna, feb, feh-wamme and fagorum. The Celtic, Slavonic, Finnish, Russian, and Magyarish words for the animal appear to be of totally different origin than the Latin term foina. Martes martes has at times undergone various specific designations, these including sylvestris, svlvatiea, vulgaris and ableturn. M. foina has been named variously domestica and fagorum. It was not recognized by Linnaeus, who did not see i t near his native Upsalla. Ray introduced the concept of the genus and i t was he who grouped the several species of marten then known into a single genus. He gave to this the generic name Martes. Later species of Asiatic marten were described by Boddaert i n 1735 (Martes flavigula, the yellow bellied marten) and by Wagner i n 1841 (M. melampus, the Japanese marten). Buffon later described many discoveries from America "Pekan","Martre", "Vison", etc., the f i r s t two of which are dis-cussed subsequently. 22 In North America Marten: The classification of species of European marten was f a i r l y well understood by the time Buffon and d'Aubenton (1758 and 1765) described the New World marten and i t was with these that they were f i r s t compared. Buffon employed French names rather than Latin binomials and hence his terminology has never been preserved in use. Pennant in 1771 and again i n 1781 f i r s t used Linnaean terminology to describe the New World forms, saying that two species of marten occurred there; these being Mustela martes and Mustela zibellina. Pennant said he had been informed by Dr. Pallas that no sables (M. zibellina) occur northeast of the "Anadyr" River i n the country of the "Ichutki" Indians. In 1772 Forster named the American marten Mustela martes after that name of Linnaeus. Pallas (1776-1780) named i t Mustela zibellina. In 1792 Kerr renamed i t Mustela zibellina  americana being the f i r s t to apply the epithet americana to the New World form. He also named two other forms, Mustela martes and Mustela zibellina alba as being present. Thomas i n 1879 and Allen i n 1895 both believe Kerr should be given authority for the present usage of the current name Martes americana, but most subsequent authorities have attributed i t to Turton. Allen says that Turton copied most of his names from Kerr and that the latter's work i s "mere trash" compared to that of Kerr. For purposes of the present paper, I consider Kerr rather than Turton the authority for the epithet americana as applied to martens. 23 Turton i n 1802 named the marten as Mustela americana. Coues in 1877 considered Turton's 1806 edition the authority of the specific name americana and although Rhoads (1902) pointed out that Turton f i r s t used the combination i n his 1802 edition and although Sherborn does the same, most authorities refer the authorship to the 1806 volume. I have examined both editions of Turton's work, arid find Rhoad's designation and not Coues1 to be correct. Nevertheless for reason given previously, authority i s given to Kerr for the name and not to Turton i n either of his volumes. Gray 1865 and 1869 and Trouessart erroneously give authority for the name to a non-existent 1803 edition of Turton. Ord i n 1815 named three American species, namely Mustela martes, M. zibellina and M. americanus. Rafinesque (1819a and b) described a new species, Mustela vulpina from "the upper Missouri" which Coues (1877) concluded i s really a white-tailed semi-albinistic individual of the typical form. Kuhl i n 1820 named a form as Mustela leucopus which Coues (1877) considered a white-footed albinlstic form. In 1823 Cuvier named a form Mustela huro, which Coues again explains as merely a light coloured Individual of the typical eastern marten. Harlan (1825) considered two forms, Mustela zibellina and Mustela  martes, to occur and confused these martens with the mink, Mustela vlson, which he considered identical. Godman (1828) named two forms, Mustela martes and Mustela zibellina. American pine marten, he states, d i f f e r from their European equivalent in possessing a yellow rather than white breast. The sable is 24 much paler than most marten, he says, a specimen collected by Lewis and Clark being "bleached". G r i f f i t h Smith and Pidgeon (1827) apply the name Mustela leucotis, this l i k e l y being a corruption of Kuhl's term leucopus. Waterhouse (1$28) named only one form, Mustela zib-bellina, though ten years later Waterhouse (1838) considered zibellina to be absent from the continent and to be replaced by Mustela martes and M. foina. Richardson i n 1829 believed only Mustela martes to be present, as did Gapper i n 1830. The latter, however, pointed out that distinctions existed between the Old and New World forms. Godman (1831) named Mustela zibellina as the only form present, while Emmons i n 1840, De Kay i n 1842 and Linsley in the same year named Mustela martes. Gray (1843) named the marten Martes leucopus, using Frisch's generic name for the f i r s t time. Smith in the same year described three forms, Martes zibellina, Martes huro, and Martes leucotis, and Schinz (1845) named Mustela vulpine. Mustela leucopus and Mustela  huro. Gesner in 1847 grouped a l l as Mustela martis [ s i c ] . Audubon and Bachman (1845-48 and 1851-54) concluded that a l l martens they had examined were referrable to Mustela martes and denied the presence of the sable on the continent. Brandt i n 1855> however, compared skins of Asiatic and American marten and concluded that the two were almost identical. The American form he named Mustela zibellina var. americana. Billings (1857) said that three species of martens occurred i n the New World but did not name them. 25 Baird (1857 and 1859)» after considering Brandt's views, expressed disagreement with him, concluding that two forms existed i n America, one with a whitish head, lying near to the sable, and one with a darkish head near to the pine marten but both probably specifically distinct from the corresponding Old World forms. The name he used for,both was Mustela americana. Ross i n 1861 (a and b) concluded that two species occurred, these being Mustela americanus and M. zebillina [ s i c ] . He believed the Mackenzie River form was only a colour variety of the Asiatic sable. A l l workers up to this period worked primarily on colour and structure of pelage in their attempts to determine the spec-ies of marten present in the New World. For this reason i t i s not surprising that they referred American forms to either the pine marten (Martes martes) , to the sable (M. zibellina) or named albinistic variants as distinct species (M. huro, M. leucopus). Before a better understanding of the status of the New World forms could be determined, i t was necessary to examine cranial and dental characteristics. One of the f i r s t to do this was Gray, who i n 1865 and again i n 1869 concluded that the last upper molar i n New World forms was small and quadrate, whereas i n Old World forms i t was large, massive and nearly twice as long on the inner as on the outer side. So long as one remembers that Gray had no specimens from the Cordillera and west coast of North America and that to these his distinction does not hold, Gray's separation is valid. He further divided "Martes americana" into three races, which as subspecies have been held as valid to the present. 26 These are: variety huro from Fort Franklin (the equivalent of "Martes americana actuosa"); variety abietinoldes (now "Martes  americana abietinoides"); and variety leucopus (now "M. ameri-cana americana")* Gray's work may be considered to mark the end of the exploratory period and the beginning of the scientific one, insofar as American marten are concerned. Allen, in 1869, examined many Old and New World skins and skulls and concluded that a l l belonged to a single circum-polar species, with possibly one or two well marked continental races. Thus he named a l l marten, including the American mar-tens, Mustela martes. Ames i n 1874-, however, named two species as being present, Mustela americana and Mustela martinus. He gave no description of either. In I876 (a) Allen reviewed his older paper and retracted the views i t contained. He agrees to the distinctions made by Gray between the Old and New World forms and adds that while the character was inadequate for the separation of M. foina and M. americana the two could be distinguished by differences i n the second lower molar which i n a l l the Old World forms possesses an inner cusp lacking i n the American form. Rhoads (1902) points out, however, that this distinction i s invalid, or at least not constant, and my own studies have led me to the same conclusions. Allen also pointed out, quite correctly, that American marten increase i n size from south to north i n the fashion stipulated by Bergmann. A l l these he named Mustela americana. Coues i n 1877 monographed the Mustelid genera including the genus of martens. He agreed with Gray's separation of Old 27 and New World forms and through detailed comparison named other distinguishing characteristics. He concluded that M. americana lay nearest to the sable, of a l l Old World forms, a relationship not much doubted since. He disagreed with Gray's division of the species into three varieties, saying that they represented individual variation only. A l l forms he considered to be one species Mustela americana. In his 1884 catalogue, True li s t e d Coues one species, M. americana, without comment, as did T y r r e l l i n 1889. With Merriam, i n I890, began the third phase i n the history of marten classification. His impress upon mammalian taxonomy i n North America was extreme and the publication of his description of the species Mustela caurina resulted i n what may be termed the phase of taxonomic s p l i t t i n g . From 1890 until recently the martens have been divided into thirteen or fourteen forms, a l l but one s t i l l considered valid. Much tax-onomic splitting of the period i s to be deplored, but i t i s a fact that Merriam's description of caurina from the Cordillera of North America marks the most basic division of marten that can be made. Merriam believed M. americana to l i e nearest to M. zibellina of a l l Eurasian forms. Flower and Lydekker (189D regress to the conclusion that Mustela americana lay very near to M. martes and M. zibel-l i n a , and was d i f f i c u l t to distinguish from them. Herrlck (1892) likewise emphasizes this lack of distinctiveness, as does Lydekker I896. 28 Poland (1892) erroneously names the marten Mustela  martes, saying i t i s indistinguishable from the Old World form and very nearly inseparable from M. foina. In spite of this, he provides a correct view of geographic variation i n marten, showing extreme insight into the problem. Alaskan martens, he said, are large and pale; Californian animals are small and uniformly coloured; the Labrador marten i s large and richly furred; those from southern Canada and the northeastern United States are small and more l i g h t l y furred. These are apt des-criptions of the current forms "actuosa", "sierrae", "brumalls", and "americana". He concluded that "each d i s t r i c t of North America has i t s own peculiar type". In 1897 Bangs described as a distinct species Mustela  atrata, the Newfoundland marten. The following year, the same author described another from Labrador, Mustela brumalls. In 1900 Osgood named Mustela americana actuosa from Alaska, Yukon and Mackenzie. In 1901 he described Mustela nesophila from the Queen Charlotte Islands. E l l i o t (1901a) named five forms, Mustela americana, Mustela caurina, M. c. atrata, M. c. brumalis and M. c. actuosa. Miller and Rehn (1901) l i s t them as M. americana, M. a. actuosa, M. atrata, M. brumalis and M. caurina. In 1902 Preble described Mustela americana ableticola from Manitoba, and Rhoads i n 1902 Mustela caurina origenes from Colorado. Rhoads' work was revis-ionary i n nature and he named the following forms: M. americana (with the subspecies abietinoides, actuosa and brumalis), M. atrata and M. caurina with i t s new subspecies origenes. E l l i o t 29 i n 1903(a) described Mustela americana kenaiensis from the Kenai, Alaska, and i n 1905(a) M. boria from Mackenzie. E l l i o t (1905b) l i s t e d six species of martens: Mustela americana (with subspec-ies ableticola, abietinoides, actuosa and brumalis), M. kenai-ensis, M. atrata, M. boria and M. nesoohila. In 1908 Preble synonymized M. boria with M. a. actuosa. Grinnell and Storer, i n 1916, described Martes caurina sierrae from the Sierra Nevadas and Grinnell and Dixon i n 1926, Martes caurina humboldtensis from the Coast Mountains of California and M. c. vancouverensis from Vancouver Island. These latter mark the last forms of recent North American marten to be described. The systematic catalogues of the period 1912 to the present show a progressive attempt on the part of systematists to group the known forms of marten into r e a l i s t i c specific groups. Miller i n 1912(b) li s t e d six species: Martes americana (with subspecies americana, kenaiensis, abieticola, abietinoides and actuosa), M. atrata, M. boria, M. brumalis, M. caurina (with subspecies caurina and origenes) and M. nesophila. Anthony i n 1917 named four species, using Miller's classification but omit-ting M. boria, origenes and nesophila. Miller, i n his Catalogue of Mammals, 1924, again l i s t e d six species, they being now Martes americana (with subspecies americana, kenaiensis, abieticola, abietinoides and actuosa), M. atrata, M. boria, M. brumalis, M. caurina (with subspecies caurina, sierrae and origenes) and Martes nesophila. Anthony, in his Field-book (1928) used Miller's classification, with the exception of M. boria which he omitted. It would appear that 30 Grinnell and Dixon (1926) were the f i r s t to suspect that only two major types (species) of martens occurred i n North America. They wrote that "the material as yet available for study i n d i -cates the existence of two distinct species of marten i n north-western America — Mustela caurina, with short bullae and other minor characters, on the Pacific coast from Admiralty Island south to east-central California; and Mustela americana, with long bullae, etc., i n the interior of the continent and west at the north clear to the Alaska Peninsula. Each of these species contains several well marked geographic races". Allen (194-2) employed the most modern classification of a l l , the one with which most systematists now concur. He also named only two species, Martes americana and Martes caurina, the f i r s t with the subspecies,americana, kenalensis, abieticola, abietlnoides, actuosa, atrata, boria and brumalis; the second with the sub-species caurina, sierrae, origenes, nesophila, vancouverensis and humboldtensis. Anderson, in his Catalogue of Canadian forms (194-6) followed Allen except that he omitted M. a. boria and the American forms M. c. sierrae, origenes and humboldten-sis and raised M. atrata to specific status. Miller and Kellogg (1955) l i s t a l l of Allen's subspecies except "boria". and name three species, "M. americana", "M. caurina", and"M. atrata". The realization that only two major forms of marten exist on the continent was l i k e l y recognized by some systematists long before Allen put the idea into print. Just how these two forms (M. caurina and M. americana) were related to each other 31 (i.e. whether distinct species or subspecies) has only recently been made clear. In 1936 Hall wrote that " c r i t i c a l study of the described species and subspecies of true martens i n North America (subgenus Martes) almost certainly would show that the named forms belong to no more than two, and possibly only to one, f u l l species". In 1942 Allen published his heretofor discussed volume in which he grouped a l l as subspecies of the two species M. americana and M. caurina. In 1946 Hall again wrote "I do not know whether caurina i s a species distinct from americana and the species name should be regarded as tentative pending a revision of the genus". Rand in 1948(b) lis t e d these as the only two species of marten occurring in Canada. Dalquest, the same year, studied the problem and concluded that "east of the Rocky Mountains the ranges of the two species approach closely but each retains distinctive characters.... Gn the basis of the evidence at hand, the two should be considered f u l l species until positive proof of intergradation i s established." Durrant i n 1952 said that "my studies of martens cause me to doubt that M. caurina i s distinct from Martes americana", but added no detail. It remained for Wright (1950 and 1953) to provide the proof of conspecificity demanded by Dalquest. He showed that the two underwent complete intergradation i n western Montana and central Idaho. On the basis of this, Wright liste d a l l thirteen forms of marten as subspecies of the single species Martes ameri-cana. In view of the v i t a l nature of Wright's work, i t i s reviewed i n greater detail elsewhere i n this paper. 32 Fisher: Because fisher occur only i n the New World and were discovered after the European forms were relatively well known, the problems of their nomenclature and classification are not the same as were those of New World martens. Buffon, i n 1765> published the f i r s t description of the Fisher and later the name was used by Brisson. In 1771 Pennant named and described not only Buffon's Fisher but another form he named Pekan. Concerning the f i r s t of these, he expressed belief that i t might be identical to the sable of Asia, although i n 1781 he retracted this view. Concerning the second, he said that he could not separate i t from the Vison of Buffon. Commenting on this situation Coues (1877) fide Baird (1857 and 1859) writes that Buffon, Brisson and Pennant a l l "described the animal from the same specimen - one i n the cabinet of M. Aubry of Paris.... Pennant's account of his Fisher i s unmistakeable, but he describes i n addition, the Pekan of Buffon, not recognizing in i t the same species. These two accounts furnished for many years the bases of a l l the scie n t i f i c binomial names imposed by various authors." Erxleben, i n 1777 gave for the f i r s t time a Linnaean binomial to the fisher, naming i t Mustela pennantli, although very shortly later Schreber (1778) renamed i t M. canadensis. Thus not only were two animals named from the same specimen, but two names were applied at nearly the same time to the same species. This situation provided much of the nomenclatural problem assoc-iated with the history of fisher taxonomy. In 1785, Boddaert 33 used the name M. melanorhyncha for the fisher and Kerr i n 1792 named two forms, M. canadensis and M. zibellina nigra (both for the fisher). Shaw (1800) changed the generic name to Viverra, naming two forms, V. piscator and V. canadensis. Turton (1802) named Mustela nigra and M. canadensis and used the same nomen-clature again i n 1806. In 1815 Ord used the names M. nigra and M. hudsonius and used the term M. canadensis for the otter (Lutra canadensis) (Rhoads 1894-). Desmarest (1820-1822) intro-duced the name M. mfa. In 1825 Harlan named only one form, calling this M. canadensis, while the following year Godman (1828) named a single form M. pennanti [ s i c ] . Lesson (182 7) named three species, a l l referable to the fisher. These being M. pennanti, M. canadensis, and M. piscatoria. G r i f f i t h Smith and Pidgeon, in 1827, followed Godman1s nomenclature. Fischer (1829) used the term M. godmani (corrupted by Gray 1865 to M. goodmanii). Richardson (1829) named two forms, these being M. canadensis and a pale variant M. canadensis var. alba. Swainson 1835, for the f i r s t time employed the generic name Martes for the fisher, naming i t Martes canadensis. Emmons In 1840 remarked that earlier division of the fisher into two species was due to colour variation and grouped them together as one form, Mustela canadensis. Smith i n 1843 named three species, Martes canadensis, M. nigra and M. pennantii. Gray i n 1843 reduced a l l to the single species Martes canadensis. Schinz 1845 and Audubon and Bachman (1845-1848 and 1851-1854) * named one form Mustela canadensis. Baird i n 1857 and 1859 was the f i r s t to attempt to systematize the chaotic nomenclatural state of the fisher. He 34 makes clear that, as pointed out previously, Buffon, Brisson and Pennant a l l named t h e i r f i s h e r s from the same specimen. Buffon named his Pekan, as did Brisson, but Pennant named both Pekan and Fisher, not knowing that both were the r e s u l t of exam-in a t i o n of the same specimen. Baird then went on to the problem of p r i o r i t y of date between the two names, canadensis and pennant! applied to the species of Buffon and Pennant. "While Erxleben's work i s dated 1777 and volume 3 of Schreber's work i s dated 1778, Baird f e l t i t possible that parts of Schreber's work may have been issued p r i o r to 1777 and states as evidence that Erxleben quotes the plates of Schreber i n various places. Nonetheless, i n l i e u of better evidence, Baird accepted Erxleben as having p r i o r claim. On page 455 of Erxleben, i s described a Mustela cana- densis, the synonomy of which includes both the Vison and Pekan of Buffon, but which i n actual fact describes only the mink (Mustela vison). On page 470, however, Erxleben described under M. pennant11 the Fisher of Pennant and concluded that the proper name f o r the species i s Mustela pennantii Erxleben. He suggests elsewhere that two species, the f i s h e r and the black cat, might e x i s t , however. In i860 Godman described both Mustela pennantii and M. canadensis but expressed b e l i e f i n Pennant's idea that the l a t t e r was named f o r a mink skin. Gray i n 1865 used the binom-i a l Martes pennantii, employing modern terminology f o r the f i r s t time. Coues (1877) continued the c l a r i f i c a t i o n of nomenclature 35 begun by Baird. He concluded that Erxleben 1777 (Mustela  pennanti), Shaw 1800 (Viverra piscator), Turton 1802 and 1806 (Mustela nigra), Lesson 182 7 (Mustela piscatoria) and Boddaert 1785 (Mustela melanorhyncha) a l l named their forms after the Fisher of Pennant. Schreber 1778 (Mustela canadensis) named his after the Pekan of Buffon. Erxleben named a M. canadensis but applied i t to Mustela vison. Shaw named a Viverra piscator whose origin Coues could not determine. Turton named a Mustela  nigra for both Buffon's Pekan and the otter (Lutra) which he considered identical. Like Baird, he concluded that the fi s h -er's name should rightly be Mustela pennantii Erxleben. Else-where, however, Coues drops the double i ending of pennantii, reducing i t to pennanti. The current use of the shortened epithet may be said to date from Coues. True i n 1884, used Coues* term, M. pennanti, as did Tyrrell 1889. Rhoads (1894) considered the nomenclature! problem and concluded as did Baird and Coues, that the proper name was Mustela pennantii. In I898, however, i n his revision of the species, he employed the name M. canadensis Schreber. Contrary to earlier opinions he concluded that since Erxleben made refer-ence to Schreber's plates, and also on Sherborn's opinion, that Schreber was published (or the Illustrations at least) prior to the issuance of Erxleben's work i n 1777 and that he thus gained priority. Rhoads, in this paper, described a new 'subspecies from the Pacific coast, M. c. pacifica. This has been the f i r s t named subspecies to persist i n the literature. Allen, i n I898, replied to Rhoads decision, saying that since 36 Schreber's plate name M. canadensis i s preceded by the plate name M. lutra canadensis, i t preoccupies the epithet canadensis as applied to the fisher. Hence Allen concluded, the proper name for fisher was Erxleben's M. pennanti. In 1899 E l l i o t used the contracted epithet penanti, this probably being a lapsus, Miller 1900, Miller and Rehn 1901 and E l l i o t 1901, returning to the orthodox pennanti. Bangs i n 1910 used the term pennantii. In 1912(b) Miller using Thomas' (1911) precedence, made i t Martes pennanti, which form i t has held to the present, with few exceptions (see Jordan 1929, Warren 1942 for exceptions). In 1935 Goldman named the third and last subspecies to be accepted by modern workers. This was M. p. columbiana from the Rocky Mountains, which intergraded in the Prairie Provinces, he believed with typical M. pennanti. A l l three ^subspecies" are l i s t e d i n Miller and Remington's (1955) catalogue. Some criticism has been directed towards the division of fisher into subspecies. Grinnell i n 1933 and i n 1937 (Grin-n e l l , Dixon and Linsdale 1937) wrote that after c r i t i c a l exami-nation of the specimens in the United States National Museum, he was unable to separate M. pennanti pacifica from M. p. pen-nanti and concluded they were a l l to be named simply M. pennanti. Hall, i n 1936, working with fewer specimens, came to the same conclusion. None of them considered M. p. columbiana, of course, which was not described until 1935 (Grinnell's later manuscripts being completed before this date). In 1950 (Hemming 1950) redrafted Article 14 of the 37 International Rules such that t r i v i a l names based on a male person's surname should "consist of the exact surname ... to which should be added ... the termination i n the genitive case ... - i " . Thus the epithet pennanti i s given v a l i d i t y over the older form pennantii by the Regulations. The currently accepted c l a s s i f i c a t i o n of the fi s h e r s i s as follows: "Martes pennanti pennanti" Erxleben 1777. "M. p. p a c i f i c a " Rhoads 1898. "M. p. columbiana" Goldman 1935. SYSOPSIS OF THE SPECIES AND SUBSPECIES OF MARTEN Comparisons of the d i s t i n c t i v e q u a l i t i e s of the various species of marten have been made i n the past by the following authors: A l l e n 1876(A), A l l e n 1938-40, Alston 1879, Blanford 1888-91, Bobrinskoy et a l . 1944, Bonhote 1901, Brongersma 194-1, Coues 1877, de B l a i n v i l l e 1841, de Vos 1952, Didier 1947, Ellerman, Morrison-Scott 1951, Flower 1885, Gerrard 1862, Gray 1865 and 1869, H a l l 1926, Kneeland 1859, Merriam 1890, M i l l e r 1912(a), Ognev 1925, Ognev 1931, Owen,1853, Pocock 1918, Pocock 1921(a and b), Pocock 1936(b), Reynolds 1912, Rode and Didier 1944, Schmidt 1943, Tate 1947, and Bannikov 1953. At no time, however, have more than three or four species been compared simultaneously. The synopsis provided below, prepared through reference to the above named authors and by examination of speci-mens, attempts to help remedy t h i s lack. For those interested i n the general anatomy of the martens the reader i s referred to the writings of de B l a i n v i l l e 38 (1841), Reynolds (1912), Hall (1926 a and b), and Cooper ( 1 9 5 4 ) . Dental terminology i s given i n an i l l u s t r a t i o n by Gregory ( 1 9 5 D f i g . 20.08. I have no way of knowing how carefully the specific distinction of species of marten i n a biological sense has been determined. It i s l i k e l y that the species of the subgenera Pekania and Charronia are truly species. Within the remaining subgenus Martes, the situation is less l i k e l y . The sympatricity of the ranges of M. martes and M. foina suggests their distinct specificity. Streuli (1932) says that not a single hybrid or intermediate was found among several hundred specimens of the two forms examined although Schmidt (1943) says Severtzov reported hybrids i n the Tienschan Mountains. Rode and Didier (1944) report that 42 skulls of the two species were distinctive i n every case. Severtzov (I876) considered M. intermedia inter-mediate between M. foina and M. martes but he apparently worked only with skins. Ellerman, Morrison-Scott (195D conclude, however, that the two species are not always easily distinguish-able, which suggests the possibility that Intermediates may some-times occur. Martes zibellina has a range that barely overlaps that of M. martes; i t is uncertain that i t is i n contact with that of M. foina. although Schmidt (1943) cites Severtzov as saying that in the Tienschan Mountains M. foina has been crossed with the sable. The cranial characteristics of a l l three are marked i n the specimens I have examined, but Ponomarev i n 1946, recorded that he obtained hybrids from M. zi b e l l i n a and M. martes crosses. 39 Whether this situation occurs under natural conditions I do not know. Martes melampus has always been considered a distinct species. Ellerman, Morrison-Scott (1951) say that while i t s coloration i s near M. zibellina the skull is nearer M. foina and M. martes and conclude that i t is an "isolated and valid species". Its true status i s unknown to me, since older students have never considered i t s specificity i n a c r i t i c a l fashion. It has never been demonstrated that North American martens are specifically distinct from Old World forms. M. amer-icana americana i s morphologically distinct from M. zibellina, i t s nearest Old World neighbour, but this cannot be said to be marked i n the relationship of M. americana caurina to M» zibel-l i n a . These two can almost always be separated, but their dis-tinctions are small, far smaller i n fact than exist between the two true subspecies M. americana americana and M. a. caurina. It i s quite possible that the Asiatic and American forms w i l l eventually be shown to be potentially capable of interbreeding, as Rausch (1953) and Bobrinskoy (1944) believe to be the case i n many Siberian-Alaskan forms. If this i s found to be so, M. am-ericana w i l l be reduced nomenclaturally to a subspecies of M. zibellina or of M. martes. The relationship between Asiatic and American forms i s considered elsewhere i n greater d e t a i l . In order to f a c i l i t a t e the identification of unknown skulls, a key is provided below, which is based almost completely on dental characteristics. It makes no attempt to denote rela-tionship and i s completely a r t i f i c i a l . It i s based largely upon a series of indices devised by Brongersma (1941) which were 40 enlarged upon during the course of the present study. These indices are given in detail i n Appendix A. The key considers M. flavigula and M. guatkinsi to be conspecific, and further, takes M. americana to subspecies. 1. Width of inner lobe of P 4 x 100/width of trenchant part of P 4 more than 65 2 l b . Less than 65 7 2a. Greatest width of canines at base x 100/bulla length more than 100.5 3 2b. Less than 100.5 5 3a. Mesial length of M1 x 100/breadth of M1 more than 55 4 3b. Less than 55 M. flavigula (including M. guatkinsi) 4a. Width of inner lobe of P x 100/length of P more than 28 M. melampus 4b. Less than 28 M. americana caurina 5a. Lateral length of M1 x 100/mesial length of M1 more than 79 M. americana americana 5b. Less than 79 6 6a. Greatest width of canines at base x 100/bulla length more than 95.5 M. martes 6b. Less than 95.5 M. zib e l l i n a 7a. Breadth of M1 x 100/length pf P4 more than 8 9 . . M. foina 7b. Less than 89 M. pennanti Where the key proves inadequate for the identification of specimens, I suggest reference be made to the stati s t i c s on which i t i s based, given i n Appendix A; even when identification 41 appears f a i r l y certain, I would suggest that the reader make comparison to these. Other keys to a restricted number of species are given by Miller 1912(a), Ognev 1931» and Bobrinskoy et a l . , 1944. Diagrams of the skulls of the various species of marten are given on Figures 1 and 2. Maps showing the distribution of Eurasian forms are given on Figures 3, 4, 5 and 6} these have been prepared from many sources, but chiefly from the maps of Kozhantschikov (1930), Ognev (1931), Stirton (1939), Schmidt (1943) and Bobrinskoy et a l . (1944) and from the accounts of Sowerby (1922, 1930), Allen (1938-1940), Chasen (1940) Brongersma (1941), Harper (1945), and E l l e r -man, Morrison-Scott (1951). Synonymies for the species and subspecies considered'in this paper are, because of their lengthy nature, summarized alphabetically i n Appendix B. The measurements of condylobasal length provided are meant to give some idea of geographic varia-tion. They have been compiled from the specimens examined and the literature. Because the samples are small, only the means, observed extremes, and sample size are given, measurements being in centimeters. The classification that follows is that of Ellerman, Morrison-Scott (1951 and 1953) and Chasen (1940). Other recent classifications of importance are given by Miller 1912(a), Ognev 1925, Ognev 1931, Pocock 1936(b), Schmidt 1943, Bobrinskoy et al.., 1944, and Harper 1945. Martes martes (Linnaeus .1758) = M. sylvestris, svlvatica, vulgaris and ableturn of authors (see Appendix B for detail). Type l o c a l i t y : Upsala, Sweden. Range: B r i t i s h Islands including Ireland, Norway, Sweden, Figures 1 and 2. (To face page 42) Figures 1 and 2. The skulls of the species of the world's marten, i n ventral aspect. From sources referred to i n the text, and from original material. C A U R INA A M E R I C A N A M E L A M P U S P E N N A N T I F O I N A F L A V I G U L A M A R T E S Z I B E L L I N A Figures 3? 4, 5 and 6. (To face page 42) Figures 3? 4 , 5 The distribution of Eurasian marten. and 6. Collated from Kozhantschikov (1930), Ognev ( 1 9 3 D , Schmidt (1943), Bobrinskoy et al (1944) and many other sources. Type lo c a l i t i e s are given for the classifications of Ellerman, Morrison-Scott (1951 and 1953) and Chasen (1940). 42 Denmark, France, Belgium, Holland, Switzerland, Italy, northern Spain, Balearic Islands, Sardinia, Bohemia, Poland, to Russia and the very southwestern parts of Siberia, to the Obe and Eptischa Rivers, the White Sea and the Caucasus, also Asia Minor and Persia (Bobrinskoy 1944 and Ellerman, Morrison-Scott 1951). Sowerby (1922-30) l i s t s them from Heilungkiang and K i r i n i n Manchuria, and Sowerby (1916) from southwest Kansu and northwest Ssuchuan. Brongersma: (1941) records a specimen very near to M. martes from Nepal, possibly another from Tibet. Skull diagrams and photographs: Coues 1877, pl» 3; Alston 1879, f i g . 2} Reynolds 1912, p i . 2; Miller 1912(a) figs. 76, 77, 78, 79 and 80; Ognev 1931, vol. 2, pp. 593, 600, 601 and 602; Schmidt 1943, f i g s . 8 and 9} Rode and Didier 1944, figs. 357, 358, 359, and 360. Other sources are given by Ognev 1931, v o l . 2. Specimens examined: twelve from Europe This forest dweller i s a typical occupant of the old hardwood or coniferous second growth forests. It lives i n woods of up to 2000 meters i n the Caucasus. It feeds on nuts, insects and berries. The den i s built i n a cavity 5 to 30 meters high i n trees. Infrequently i t uses the dens of wild pigs. The period of heat begins i n February and reaches i t s peak i n July. Gestation lasts for eight months (270 to 285 days). The l i t t e r of two to six, i s born i n April or May (Bobrinskoy 1944, Schmidt 1943). Geographic variation, according to Bobrinskoy (1944) chiefly expressed i n size and colour. Jurgenson (1951) says much geographic variation i n skull and teeth i s associated with 4-3 the type of food eaten. Characteristics: More active than M. foina; smaller and lighter (males weighing 1200-1650 grams, females 800-1350 grams); body long and slim with short limbs; length of head and body less than 57 cm.; fur fine, soft, brown below and above, rarely slatey; outer fur rich brown, under fur red grey with red yellow tips; top of head same colour as spine; often with a light coloured throat patch of yellow, which extends behind rear of forelegs; upper l i p with distinct vertical groove; t a i l bushy, rather short, being considerably longer (without tip hairs) than one half the length of head and trunk (Ellerman, Morrison-Scott 1951 say only 49$ length of head and trunk); t a i l vertebrae said to be variously 20-22, 17, 18-19, 18, 10+, 18-33, 17-19, 18-19 (Bobrinskoy et a l . 1944-, Ognev 1931, Gray 1869, de Blain-v i l l e 1841, Flower 1885, Reynolds 1912, Schmidt 1943, Gerrard 1862); hair endings of t a i l long, being more than quarter length of t a i l ; t i p of t a i l rounded; soles of feet covered with thick hair; skull massive, zi b e l l i n a shaped, i . e. long, narrow, and rather produced, thin and deep when seen from behind; skull and braincase less massive, especially in the posterior part; condy-lobasal length 70 to 90 mm.; zygomatic width about 48 mm., the zygomatic width being rather more than half condylobasal length; zygomatic arches highest at rear, whence they slope suddenly down and forward, being longer and narrower; sagittal crest heavy* only rarely completely absent; rostrum with parallel sides, narrow and elongate; anterior nares oval i n shape; nasal bones with moderate central constriction; frontal profile only slightly sloping, concavity of profile not very pronounced; supraorbital 44 constriction moderate, pointed, rather deep and always present; post orbital process nearly midway between point of greatest constriction of cranium and the anterior root of zygoma, usually weak, occasionally absent; postorbltal region (in front of point of greatest constriction) short and convex; palate rela-tively narrow and long (about 4.3 cm. i n length), often with a distince azygos notch on rear surface; anteorbital foramen relatively large; bullae moderate in length and rather widely spaced, so that twice the distance between carotid canals i s greater than the greatest length of the bullae; the distance between the outer edges of the jugular foramina is shorter than the greatest length of the bullae; distance between the carotid canals equals about half the distance between the foramen lacerum posteriore and the eustachian foramen; bulla containing three or more chambers, with partitions strongly developed; mastoid process extends somewhat beyond the edge of the meatus; general dentition rather strong; upper premolars not crowded and straight; 3 PM strongly convex on inner surface, strongly concave on outer; 4 relatively large PM with thick inner cusp, i t s width being equal to i t s length and nearly equal to width of trenchant part, trenchant portion as long as M1 i s wide; M3" i s large, massive, inner moiety large - twice as large as outer moiety and one third longer, waist separating the two strongly constricted; M^  with a slightly developed inner tubercle at base of main cusp (i.e. hypoconid); M2 with inner cusp only slightly developed; bacu-lum of the Martes type gradually bending upward throughout i t s di s t a l third, 35 to 45 mm. long ending either In two v e r t i c a l 45 prongs, or these being closed to form a transverse perforation (according to Pocock 1918 always forked, which conclusion I find to be incorrect). Martes martes martes (Linnaeus 1758) Type l o c a l i t y : Upsala, Sweden. Range: Europe north of the Mediterranean east as far as the White Sea, Kiev and Vetebsk i n Asia. Large and dark i n colour (Bobrinskoy et a l . 1944). Condylobasal length: males 8.66 (8.41-9.00) n = 6; females 7.90 (7.78 - 8.16) n = 3 . Martes martes borealis Kutznetzov 1941 (according to Bobrinskoy et a l . 1944). Type l o c a l i t y : not known. Range: northern parts of European Russia, excluding the Kola Peninsula. Very like M. m. ruthena, but with winter pelage lighter (Bobrin-skoy et a l . 1944). Martes martes latinorum (Barrett-Hamilton 1904) Type l o c a l i t y : Nurri Mountains, Sardinia. Range: Italy, Sardinia, Balearic Islands (Ellerman, Morrison-Scott 1951)• As i n M. m. martes but with pelage lighter i n colour, throat patch yellower (Miller 1912a). Condylobasal length: males 8.66 (8.50 - 8.80) n s 3; females 7.89 (7.70 - 8.30) n = 6. Martes martes lorenzl Ognev 1926. Type l o c a l i t y : Storojevala, Kuban Di s t r i c t , Caucasia. 46 Range: the forested area of the whole Caucasus. Large, as i n M. m. martes. but with winter fur coarse and thick, dark brown with a characteristic red-olive cast (Ognev 1926, Bobrinskoy et a l . 1944). Condylobasal length: males 8.53 (8.41 - 8.66) n = 4; females 7.89 (7.70 - 8.30) n = 6. Martes martes notialis (Cavazza 1912) Type l o c a l i t y : South of Abruzzi, southern Italy. Range: Italy. Smaller, paler than M. m. martes. with less dense pelage (Cavazza 1912). Martes martes ruthena Ognev 1926. Type l o c a l i t y : Dmitrovsk Subdistrict, Moscow Government, Russia. Range: central part of European Russia. Smaller than M. m.  martes and lighter i n colour with a reddish sandy cast (Ognev 1926, Bobrinskoy et a l . 1944). Condylobasal length: males 8.20 (7.89 - 8.66) n = 22; females 7.57 (7.11 - 8.01) n = 11. Martes martes uralensis Kutznetzov 1944 (according to Bobrinskoy et a l . 1944). Type l o c a l i t y : unknown. Range: the whole of the Ural Range. Colour very much as i n M. m. ruthena and borealis but larger i n size than either (Bobrin-skoy et a l . 1944). Martes foina (Erxleben 1777) = domestica, alba, fagorum. Type l o c a l i t y : Germany. Range: Europe, except southern Spain and Italy and the B r i t i s h 47 Isles; present in northern Spain and Italy, Bosnia, France, Belgium, Holland, Germany, Denmark, Switzerland, Crete, Poland, Finland, Ukraine, Crimea, Caucasus, Transcaucasus, Russian and Eastern Turkestan, northwards to the A l t a i , Asia Minor, Persia, Afghanistan, Syria, Palestine, Baluchistan, Kashmir, Himalayas and the Tibetan Plateau, Punjab, Chinese Turkestan, Mongolia, Manchuria, and probably parts of northern China, northern C h i l i , Shansi and into West China, probably not much further south than Shansi and Szechwan. (Sowerby 1922-30, Sowerby 1916, Ognev 1931, Allen 1938-40, Brongersma 1941, Bobrinskoy et a l . 1944, Tate 1947, Ellerman, Morrison-Scott 1951, and Bannikov 1953). Specimens examined: 81 from Europe and Asia. Skull diagrams and photographs: de Blainville 1841, p i . G4 and G13; Bronn 1874-90; Klassen und Ordnungen vol. 1, pt. 5, P i . 15; Coues 1877, p i . 4; Alston 1879, f i g . l j Miller 1912a, f i g . 80; Festa 1914, p i . 1; Ognev 1931, vol. 2, pp. 623, 625, 626, 627, and 628; Migulin 1938, p. 167; Schmidt 1943, f i g s . 8 and 9; Bobrinskoy et a l . 1944, f i g . 77; Rode and Didier 1944, pp. 127 and 128; Baumann 1949, f i g s . 365, 366, 367 and 368; Bannikov 1953, P« 40. Other sources are named by Ognev 1931, vol. 2. This marten i s unlike M. martes i n that while i t i s found i n woods, i t also occurs i n barren rock slides. Its den is made i n crevices i n the rock and unlike M. martes i t does not avoid human habitation, often l i v i n g i n abandoned dwellings. It feeds on rodents, birds, birds' eggs, frogs, various insects, 48 berries, and grapes. It lives i n the h i l l s to an altitude of 3000 meters. Time and length of the period of reproduction i s as In H. martes (Bobrinskoy et a l . 1944). Asdell (1946) says that i n England i t mates i n July and August and that the gesta-tion period is 8.5 to 9 .6 months. In India, according to Blan-ford (1888 - 189D mating occurs i n February and gestation lasts nine weeks. This interpretation may be a result of lack of knowledge concerning delay in implantation of the blastocyst, but Asdell suggests that i f i t i s correct that delayed implanta-tion occurs only in the northern part of the animal's range and not in the southern, i t may be a function of seasonal daylength. According to Bobrinskoy individual variation i n M. foina is very great, while geographic variation i s only slight. Characteristics: A sluggish animal, being relatively large and heavy (male weighing 1700 - 2100 gms., females 1100 - 1500 gms.); body broad and compact with long limbs; length of head and body less than 57 cm.; fur coarse, less soft, brown above and below, often slaty; outer fur dull grey brown, under fur greyish white; top of the head same colour as spine; usually with a lighter coloured throat patch, usually white or creamy, not extending between legs but rather forking, each branch running down the front legs; upper l i p with distinct vertical groove; t a i l bushy, rather short, though somewhat longer than in M. martes, without hair tips, considerably longer than one half head and body length (Ellerman, Morrison-Scott say actually 49$ of head and trunk length); t a i l vertebrae various-stated to be 17 - 19, 18 and 20 - 22 (Schmidt 1943, de Blainville 1841, Bobrinskoy 1944 and 49 Ognev 193D; hair endings of t a i l long, being more than half length of t a i l ; tip of t a i l pointed; soles of feet lacking hair; skull light, flavigula shaped (i . e . swollen, flattened, with short nose); skull broad, more massive, especially i n posterior part; condylobasal length 70 - 80 mm. usually less than 83 mm., zygomatic width 48 mm., being much more than half condylobasal length; zygomatic arches regularly curved, broad-est and highest near their middle, broader and shorter; sagittal crest absent or very weak; rostrum wide and short, with con-verging sides; anterior nares heart shaped; nasal bones strongly constricted; frontal profile more sloping, concavity much more pronounced; supraorbital constriction slight, f l a t , blunt and often absent; postorbital process much nearer point of greatest constriction than anterior root of zygoma; always present, strong, coming to a distinct point; postorbital region (in front of point of greatest constriction) long and straight; palate relatively broad and short ( 35 nmu long), truncated at posterior margin; antorbital foramen relatively small; brain-case broader, less extended; bullae short and widely spaced so that the distance between carotid canals i s greater than the greatest length of the bullae; the distance between the outer edges of the jugular foramina i s the same or more than the great-est length of bulla; distance between carotid canals equals two-thirds the distance between foramen lacerum posterior and eustachian foramen; bulla with 3 or more chambers, the parti-tions less well developed than in M. martes; mastoid process extends somewhat beyond edge of meatus; general dentition 50 weaker, upper premolars crowded, often placed diagonally; PM2 4 evenly biconvex, somewhat small; PM with inner cusp small (total width less than length, and barely half width of trench-ant part); inner cusp usually placed diagonally to axis of 4 1 1 trenchant part; length of PM greater than width of M ; M small, inner lobe small, hardly larger than outer lobe; con-striction of waist weak; M^  with a well developed hypoconid; M£ likewise with an inner cusp present and well developed; bacu-lum of the Martes type, bending gradually upward throughout i t s distal third, 56 to 59 nun. long, occasionally with two vertical prongs at t i p , these most often united to produce a transverse perforation. Martes foina foina (Erxleben). Type l o c a l i t y : German; Range: Europe, except southern Spain, probably eastwards into Russia (Ellerman, Morrison-Scott 1951)• The type form, with longer hairs of back tipped with sepia, the general hue of upper parts drab (Miller 1912a). Condylobasal length: males 8.19 (8.90 - 8.46) n = 5; females 7.77 (7.40 - 8.00) n = 5. Martes foina bosniaca Brass 1911 (in Aus der Reich der Pelze) Type l o c a l i t y : Bosnia, Yugoslavia. Range and characteristics uncertain. Martes foina bunites (Bate 1906). Type local i t y : Kontopalo, Kania, Crete. Range: Crete. Near to M. f. intermedia but with shorter fur 51 of a duller, more uniform colour, without any gloss, t a i l less bushy and shorter (Bate 1906). Condylobasal length: males 7.72 (7.64 - 7.84) n = 2; females 7.00 n = 1. Martes foina intermedia (Severtzov 1873) Type l o c a l i t y : Basin of the Chu<, - Tallas and Naryn, eastern Turkestan. Range: Russian and Chinese Turkestan, Tianshan, Afghanistan, Baluchistan, western Persia and Kashmir (Ellerman, Morrison-Scott 1951). Bannikov (1953) says i t occurs i n Mongolia. The most distinctive of a l l subspecies of M. foina; compared to M. f. foina are small in size, light in colour and with comparatively l i t t l e fur (Bobrinskoy et al.. 1944). Condylobasal length: males 9.04 (7.60 - 8.46) n = 15} females 7.48 (7.37 - 7.80) n = 9. Martes foina kozlovi Ognev 1931 Type l o c a l i t y : Kam (valley of River Mekong, eastern Tibet. Range: the Tibetan Plateau and Himalayas. Characteristics un-certain. Condylobasal length: males 8.10 n = l'j females 7»52 n = 1. Martes foina mediterranea (Barrett-Hamilton I898) Type l o c a l i t y : Sierra de Jerez, Cadiz, Spain. Range: southern Spain. Compared to M. f. foina with the longer hairs of back tipped with light yellowish brown, the general hue of the upper parts lighter, yellower, and less drab (Miller 1912a). 52 Martes f o i n a m l l l e r i F e s t a 1914. Type l o c a l i t y : Aghios I s i d i r o s , I s l a n d of Rhodes, eastern Medi-terranean. Range: I s l a n d of Rhodes. C h a r a c t e r i s t i c s u n c e r t a i n . Martes f o i n a n e h r i n g i (Satunin 1906), i n M i t t . Kauk. Mus. T i f l i s 2 : 120,292. Type l o c a l i t y : T I f l i s , Transcaucasia. Range and c h a r a c t e r i s t i c s u n c e r t a i n . Condylobasal l e n g t h : males 8.30 (8.10 - 8.51) n = 4; females 7.58 (7.21 - 7.90) n = 4. Martes f o i n a rosanowi Martino 1917 • Type l o c a l i t y : north-western slope of the Chatyrdag Mountains, Crimea. Range: Crimean P e n i n s u l a . C h a r a c t e r i s t i c s u n c e r t a i n . Condylobasal l e n g t h : male 7.87 n = 1; females 7.57 (7.42 - 7.68) n * 3. Martes f o i n a s y r i a c a (Nehring 1902) Type l o c a l i t y : Wadi Syr (which runs i n t o Wadi K e f r e n , a t r i b u t a r y of the lower Jordan), S y r i a . Range and c h a r a c t e r i s t i c s u n c e r t a i n . Martes f o i n a toufoeus (Hodgson 1842). I n the Journal of the A s i a t -i c S o c i e t y of Bengal, I I , p. 281). Type l o c a l i t y : L l a s a , T i b e t . Range: T i b e t , Himalayas and Afgh a n i s t a n (Jerdon 1874). Colour of head and body above i s l i g h t y e l l o w i s h gray; the e x t r e m i t i e s are b l a c k i s h , the c h i n , t h r o a t and breast pure white ( H o r s f i e l d 53 1851). Possibly quite similar to M. f. intermedia and M. f. Kozlovi. Ellerman, Morrison-Scott remark: "Despite Pocock's contention that this i s allied to M. melampus, i t looks much more like M. foina. Its range i s adjacent to that of foina, very far from melampus. From notes l e f t by him, Chaworth-Musters evidently intended to treat i t as foina. See also Pocock (1941, p. 322 footnote). We cannot trace that the form 'kansuensis' noted by him on this page was ever described." Martes zib e l l i n a (Linnaeus 1758) Type l o c a l i t y : "Asia septentrionali"; according to Ognev (1925) this must be redesignated, after Gmelin, as "Surroundings of Tobolsk, Tomsk Government, Siberia." Range: Siberia, from the middle of Pechora to the lower Leni, and from the A l t a i to the Ussuri region, also Kamchatka, Sakhaline, northern Japan (Hokkaido and the Kuriles), Korea, Manchuria, northern Mongolia and Tannu-tuba. Now highly frag-mented (Bannikov 1953, Bobrinskoy et a l . 1944) and with range much reduced (Harper 1945). According to Jerdon (1874), Blyth took a specimen i n Tibet he was inclined to identify as M. zib-e l l i n a (Bobrinskoy et. a l . 1944, Sowerby 1922-30, Ellerman, Morrison-Scott 1953, Kozhantschikov 1930, Ognev 1931). Zhitkov (1937) t i t l e s a paper "On the former distribution of the sable in Europe"; whether i t occured here formerly I do not know, not having examined the paper. Harper (1945) says that i n the past i t has extended as far west as the Kola Peninsula or even Lapland. Specimens examined: 77 from Asia. 54 Skull diagrams and photographs: de Bla i n v i l l e 1841, p i . G7 and G13; Ognev 1925, p i . 26; Ognev 1931, vol. 2, pp. 563, 587, 591, 593 and 594; Bobrinskoy et a l . 1944, f i g . 77; Bannikov 1953? p. 40. Other sources are named by Ognev 1931, vol. 2. This animal i s a forest dweller, usually l i v i n g i n h i l l y country interspersed with moss covered rock slopes. A l l of i t s range i s characterized by the presence of spruce and pine (i.e. typical taiga forest) and where these are absent, so i s the sable. It is nocturnal, and spends more time on the ground than do Martes foina and M. martes. Its food i s chiefly small nuts, berries and a l l small animals to the size of a rabbit or grouse. The den i s made i n the crevices of roots and rocks. In February they undergo a "False drive" which lasts for three or four weeks. In July i s the real "drive" which i s followed by mating. The gestation period i s nine to nine and a half months (270 - 285 days). The young number two to four, rarely five or six and are born i n Ap r i l , or early May. They are born blind. On the 31st to 34th day, their eyes open. They suckle for about two months and i n the sixth or eighth week f i r s t leave the nest. By August they are f u l l y grown. Maturity i s reached in the second year, according to one author, the third year according to another and the f i f t h according to a third. "False drive" occurs i n the f a l l , reaching i t s extremity i n October (Bobrinskoy et a l . 1944, Kozhantschikov 1930, Schmidt 1943, Ponomarev 1938). Variation i n the sable i s rather great. Ponomarev (1938) says that within a given region albinism, semi-;albinism, white spotting, and a general mottling occur. On the Island 55 of Fekliston white headedness occurs, which he attributed to the activity of a single recessive gene. Normal sable colour, he believed is a typical quantitative character controlled by sev-eral genes. The colour and shape of the throat patch he attributed to a single gene. Geographic variation i s likewise marked. According to the same author the center of darkest colouration is found on the northeast coast of Lake Baikal. Kozhantshikov (1930) reported that the characters assigned to the continental subspecies of sable are variable and insignificant but that insular forms and those from Kamtschadka were quite distinctive but probably derived from the continental form. He concluded that there were essentially only two types (subspecies ?) of sable - the continental and those remaining. Ognev (1925) came to somewhat different conclusions. In colour, at least, he concluded, decided geographic variation occurs. According to him, the animals of the mountains are small, with very dark pelage, the underfur being characteristically slaty. The sables of the forest and low country are larger, with paler fur, less slaty underfur, and generally less dense, less soft pelage. The sable of the very northern tundra-taiga i s the larg-est sable of a l l but i t s fur i s coarse and light, the underfur yellowish. Bobrinskoy et a l . (1944) state that the chief character-i s t i c s marked by geographic v a r i a b i l i t y are size, colour of fur, and quality of fur. The largest sables, he says, are found i n two l o c a l i t i e s , one being the western parts of i t s range (Altai 56 and the Ob lowlands), the other i n Kamtschadka. Medium and small sized sables occur i n central S i b e r i a and i n the Amuro-Ussuriski d i s t r i c t s . Colour and s i l k i n e s s of fur change correspondingly and are evidently conditioned by the continent-a l i t y of the climate. The darkest and s i l k i e s t l i v e i n the Transbaikal and southern Yakutia regions, whereas to the east and west of these places, the fur gradually becomes l i g h t e r and coarser5 however, i n the mountainous regions the colour again turns somewhat darker. I t i s i n t e r e s t i n g to note that i n neither the sable nor the European martens does Bergmann's Rule appear to s t r i c t l y apply (above named authors and Schmidt 194-3). Ch a r a c t e r i s t i c s : An active form; weight uncertain, but prob-ably more than i n M. martes and M. f o i n a ; length of head and body less than 57 cm.; fur usually f i n e , s o f t , brown both above and below, often grading to blackish, underfur brown to yellow-i s h , to slate and s t e e l grey; top of head l i g h t e r i n colour than spine; throat patch often absent, or at least very small, rarely reaching to fore legs; dusky salmon i n colour; upper l i p with d i s t i n c t v e r t i c a l groove; t a i l short, with hair end-ings less than one half length of head and trunk and barely reaching to end of outstretched legs; t a i l about f i v e to seven inches long; t a i l vertebrae variously numbered as 15 to 16, 13, 18, 16, 12-15 (Bobrinskoy et a l . 1944, Ognev 1931j Gray 1869, de B l a i n v i l l e 1841, Flower 1885, Schmidt 1943, Owen 1853); hair endings of t a i l l ess than quarter t a i l length; s k u l l l i g h t , never massive, elongate as i n M. americana americana, not short as i n M. f o i n a , narrow and deeper with nose somewhat produced; 57 condylobasal length 70 - 90 mm., rarely 96 mm.; zygomatic width about 48 mm., about one half condylobasal length; zygo-matic arches highest at rear, whence they slope suddenly down-ward and forward, tend to be narrow, long; sagittal crest heavy, only rarely absent; rostrum sides convergent, but very narrow and long; nasal bones with only slight constriction; frontal profile not steeply sloping, concavity of profile not pronounced; supraorbital constriction moderate, pointed, rather deep, always present; postorbital process midway between point of greatest constriction and anterior root of zygoma, i s usually weak, often lacking; postorbital region short and convex; palate relatively narrow and long, often with a distinct azygos notch on i t s pos-terior edge; anteorbital foramen small; braincase high, narrow and extended; bullae long and placed close together so that twice the distance between carotid canals equals the greatest length of the bulla; the distance between the outer edges of the jugular foramen i s less than the greatest length of the bulla; distance between carotid canals i s less than one half the distance between the foramen lacerum posteriore and the eustachian foramen; mastoid process f a i l to extend beyond edge of meatus; general dentition weak; upper premolars not crowded; PM^  rather large, 4 strongly convex on inner, strongly concave on outer side; PM with inner cusp placed at right angles to trenchant part, large and thick, but with width less than total length, although nearly 4 equal to width of trenchant part: PM longer than width of upper molar; M^  wide, inner moiety large (as i n M. martes) twice as large as outer part and one third longer, waist strongly 58 constricted; with well developed inner tubercle (hypoconid) at base of main cusp; unlike M. martes and M. foina the last several caudal vertebrae a l l possess transverse process (accord-ing to Schmidt 194-3) baculum 38 mm. long, of the Martes type, bending gradually upwards throughout i t s di s t a l third, ending often in a biforked ti p , according to Ognev (1931) hut equally often the ends of bifurcations uniting to produce a transverse perforation. Martes zib e l l i n a zibellina (Linnaeus 1758) Type l o c a l i t y : "Surroundings of Tobolsk, Tomsk Government, Siberia" (Ognev 1925). Range: Pechora basin, northern Urals, and Ob Plain (Ognev 1925) and 1931» Bobrinskoy et a l . 1944, Ellerman, Morrison-Scott 1951). With long nose, short braincase. Colour dull and pale even in dark individuals, varying from cinnamon drab to dark brown, the underfur retaining i t s lightness of t i n t . In the Urals tends to be smaller, with darker denser fur, with a large bright throat patch (Ognev 1925). Condylobasal length; males 8.59 (8 .39 - 8.92) n = 6; females 7.83 (7.73 - 7.93) n = 4. Martes zibellina arsen.levi Kuznetzov 1941 (according to Bobrin-skoy et a l . 1944). Type l o c a l i t y : Ussuri Basin, eastern Siberia. Range and characteristics uncertain. Martes zib e l l i n a averini Bashanov 1943. Type l o c a l i t y : Katon-Karagai region, southern A l t a i . 59 Range and characteristics uncertain (Ellerman, Morrison-Scott 1953), but occurs i n Mongolia, according to Bannikov (1953). Martes zibellina brachyura (Temminck 1844) Type l o c a l i t y : Yeso, Japan. Range: Hokkaido Island, Japan, and the Kurile Islands. Inferior to the continental forms i n fineness and length of fur; back and t a i l dark brown, sides and limbs lighter; hair of feet long, concealing the claws; t a i l short, only, 3»5 inches in length (Temminck). Martes zibellina hamgyenensis Kishida 1927 (In Dobuts Zasshi Tokyo 39 : 509). Type l o c a l i t y : Korea. Range: so far as known, the range of the species i s Korea, possibly north to near the type l o c a l i t y of M. z. arsen.ievi. Characteristics uncertain. Martes zib e l l i n a kamtshadalica (Birula 1918) i n Comptes Rendus Mus. Zool. Acad. Sci. Petrograd, vol. 821) Type l o c a l i t y : Kamtchadka. Range: the Kamtchadkan Peninsula. A strongly characterized sable, being much larger and with more robust dentition than any other subspecies. Colour ranges from warm sepia to mars brown, throat patch variable (Ognev 1925)• Condylobasal length; males 8.95 (8.50 - 9*50) n = 22; females 8.08 (7.86 - 8.41) n = 15. Martes zibellina princeps (Birula 1922) In the Ann. Mus. Zool. Acad. Sci., St. Petersberg 22:) 6 0 Type l o c a l i t y : Baragusin Mountains, Transbaikalia. Range: the mountain forests of Transbaikalia, i.e. i n the Baragusin H i l l s and the branches of the Stanavoy Mountains, and through parts of Mongolia. Formerly between the Olekma and Witim Rivers. A medium sized sable, near that found i n the Sajan Mountains, but with much longer bullae and braincase. Fur long, dense and soft, of a dark black-brown, with underfur bluish grey and with brown at base of t i p . Throat patch reduced, often invisible (Ognev 1 9 2 5 , Bannikov 1 9 5 3 ) . Condylobasal length: males 8 . 1 0 ( 7 * 8 2 - 8 .35) n = 9 } females 7 . 4 8 ( 7 . 4 3 - 7 . 5 3 ) n = 5. Martes zibellina sahalinensis Ognev 1 9 2 5 Type l o c a l i t y : Wedernikovo, Sakhalin Island. Range: Sakhalin Island. Skull distinctive, fine and weak, with strong contraction i n the interorbital region and with short bullae. Dentition weak. Colour, as i n M. z. kamtshadallca, but commonly more cinnamon (Ognev 1 9 2 5 ) . Condylobasal length: males 7 . 9 7 ( 7 . 6 3 - 8 . 2 0 ) n = 1 3 ; females 7.41 ( 7 . 2 6 - 7 . 5 9 ) n = 6 . Martes zibellina sa.lanensis Ognev 1 9 2 5 . Type l o c a l i t y : Orsyba River, northern Sayan Mountains, middle Siberia. Range: the mountain forests of northwestern Mongolia including the Usa and Kasyr Rivers. Skull short, both i n nasal and cran-i a l regions. Colour dark brown, the underfur pale yellowish. Throat patch varies from dusky brown to b r i l l i a n t salmon (Ognev 1 9 2 5 , Allen 1 9 3 8-40, Bannikov 1953). 61 Condylobasal length: males 8.24 (8.16 - 8.40) n = 6; females 7.36 (7.32 - 7-40) n = 2. Martes zibellina schantarica Kuznetzov 1941 (according to Bobrin-skoy et a l . 1944). Type l o c a l i t y : Shantar Islands, lower Amur, eastern Siberia. Range: the Shantar Islands. According to Bobrinskoy i s very like M. z. sahalinensls. Martes zibellina tungusensis Kuznetzov 1941 (according to Bobrinskoy et a l . 1944). Type l o c a l i t y : Basins of the Nizhaya and Podkamennaya Tungush, middle Siberia. Range: presumably as the type l o c a l i t y . Characteristics un-certain. Martes zib e l l i n a yeniseensis Ognev 1925. Type l o c a l i t y : Forest on plain along Yenesei River, Krasnoiarsk d i s t r i c t , eastern Siberia. Range: Taiga between the Angerra and Sayan foothills (Bobrinskoy et a l . 1944, Ellerman, Morrison-Scott 1951, Ognev 1925). Near to M. z. zibellina but with shorter nasal region, longer cranial region, wider zygomatic arches and postorbital processes. The colour of the fur i s a more dusky warm brown. Condylobasal length: males 8.27 (8.00 - 8.70) n = 13; females 7.43 (7.20 - 7.70) n = 7. Ognev (1925) suggested that another distinct, but un-named, subspecies was to be found in the Tung-uska River region, near Turuhansk. Bobribskoy 1944 suggests dis-tinctive forms may occur i n the A l t a i and in southeastern Yakutia. 62 Martes melampus (Wagner 1841) Type l o c a l i t y : Southern Hondo, Japan (Thomas 1905a) Range: Hondo, Shikoku, Kiushii and Tsushima Islands, Japan and western Korea. Specimens examined: eight from Asia. Skull diagrams and photographs: Temminck 1844, figs . 3 and 4. This appears to be a poorly understood animal; at least I have been able to locate almost nothing on i t s structure, l i f e history, ecology or true relationships to the other forms. Some data may possibly be obtained from Okada 1938, or Kuroda 1940 which papers I have not seen. Ellerman, Morrison-Scott have this to say: "In the London material, this has the t a i l an average about 4 4-47 per cent of the head and body length (resembling zibellina therefore i n rather short t a i l ) ; a white throat patch seems f a i r l y constant, and, at least in winter, the head tends to be paler than the back, a l l characters remin-iscent of zibellina except the throat patch. But the bullae seem to be definitely of the martes—foina type, and do not seem to resemble those of zibellina. The forelimbs are clearly con-trasted blackish, more so than i n our zibellina skins, therefore the conclusion has been reached that melampus is an isolated and valid species, partly combining the characters of the other two groups. So far as colour i s concerned, i t i n no way resembles the subgenus Charronia...." Described by Gray 1865 and 1869 as: fur soft, brown or yellow brown, underside scarcely paler, shoulder and outside of thigh blacker; feet blackish; head, chin and upper part of throat dark red brown; blackish streak from orbit to nose. Throat and sides of neck yellow or white, 63 crown paler. Characteristics. Length of head and body less than 57 cm.; fur a rich golden brown to a deep brown; top of head lighter i n colour than the spine; throat patch usually present, white in colour; upper l i p presumably with a distinct v e r t i c a l groove; t a i l short, with hair endings about 44 to 47 per cent length of head and body; skull very massive, foina shaped, or perhaps intermediate between foina and martes, not quite so broad, low and short as in M. foina. Condylobasal length 84 to 85 mm.; zygomatic width greater than half condylobasal length, averaging about 48 mm.; zygomatic arches highest at rear, whence they slope suddenly downward and forward; relatively broad and short; sagittal crest strong, rarely absent; rostrum sides convergent, the rostrum being wide and short; nasal bone's with no constriction; nasal profile rather sloping, with pronounced concavity; supraorbital con-striction slight; postorbital process near middle or slightly nearer point of greatest constriction than anterior root of zygoma, is strong, coming to a distinct point. Postorbital ' region long and straight; palate relatively broad (.09 mm.) and short (86 mm.) with distinct azygos notch on rear surface; anteorbital foramen small; braincase broad, not relatively extended; bullae short, as i n M. foina, and rather less i n -flated; widely spaced, so that the distance between the outer edges of the jugular foramen i s greater than the greatest length of the bullae; so that the distance between the carotid canals equals about five-sevenths of the distance between foramen lacerum posteriore and the eustachian tube foramen; mastoid 64 process extends somewhat beyond edge of meatus; general denti-tion stronger; upper pre-molars sometimes rather crowded; 3 4 PM convex on inner side, concave on outer; PM with inner cusp heavy, i t s width equalling i t s length and equalling three-fourths of the width of the trenchant part placed diagonally to 4 1 axis of main cusp; length of PM greater than width of M ; 1 M large, massive, inner moiety large, intermediate between that of M. martes and M. foina, one and a half to two times as large as outer moiety, and twice as long; Mi with inner cusp well developed; M2 the same; baculum presumably of the Martes type, though as far as I know, never positively determined as such. Martes melampus melampus (Wagner 1841) Type l o c a l i t y : Japan (Wagner 1841), Southern Hondo, Japan (Thomas 1905a, b, and c). Range: Hondo, Shikoku and Kiushiu Islands, Japan. With a rich golden brown pelage, the back usually much browner and with a bright orange throat patch (Thomas 1905a, b, and c). Condylo-basal length: males 8.52 (8.50 - 8.55) n = 2; females 7.50 n = 1. Martes melampus coreensis Kuroda and Mori 1923. Type l o c a l i t y : Tenan, southern Chusei d i s t r i c t , Korea. Range: Korea. Near to M. m. melampus but general colour of upper parts yellow washed with fox-red instead of rich orange yellow and the ashy white of head extending farther to the nape instead of confined to the forehead. Separated from 65 M. m. tsuensis by much paler colouration throughout. Martes melampus tsuensis (Thomas 1897) Type l o c a l i t y : Kamoze, Tsuehima Islands, Japan. Range: Tsuehima Islands, Japan. Si m i l a r to the t y p i c a l form i n a l l characters but colour of f u r . General colour above i n winter d i r t y yellowish brown instead of r i c h golden brown; muzzle and l i p s black instead of brown; throat patch white instead of yellow. In summer pelage much darker than the t y p i c a l form, a l l brown parts being black as are the whole of the limbs; throat patch yellow (Thomas 1897). Condylobasal length: females 7.80 n - 1. Martes americana ('.."Kerr.: 1792) including "M. caurina" (Merriam 1890) Type l o c a l i t y : eastern North America. Range: H o l a r c t i c North America. Because t h i s species w i l l be treated i n d e t a i l else-where, only those aspects needed for comparison to Eurasian spe-cies and the f i s h e r w i l l be u t i l i z e d here. Because the two sub-species to be considered d i f f e r morphologically, comparison w i l l be made f o r both of these. Ch a r a c t e r i s t i c s : Active forms, body long and slim with short limbs. Length of head and body less than 57 cm.; weighs about 700 - 1200 gms. i n males, 600 - 800 gms. i n females; outer fur ., dark brown, under fur longer i n americana than i n caurina, l i g h t yellow i n both; top of head l i g h t e r i n colour than i s spine; throat patch i n americana usually present, varying from white to 66 orange, but usually yellow, i n caurina, always present, usually large (extending to base of forelegs), yellow or orange; upper l i p with distinct vertical groove; t a i l i n americana short with hair endings about one half a head and body length; i n caurina, t a i l somewhat shorter equalling less than one half the head and body length; t a i l vertebrae average 17 (Cooper, 1953) i n ameri-cana, 14 in caurina. According to Hall (1926a) i t possesses an abdominal skin gland. Skull i n americana light, narrow, elon-gate^ nose rather produced; braincase not massive, especially in posterior part; in caurina more massive, broader, f l a t t e r , nose less produced; braincase more massive, especially i n pos-terior part; condylobasal length 67 mm. to 88 mm.; zygomatic width about 48 mm. i n both, i n americana being about one half condylobasal length, i n caurina slightly more; arches i n americana highest at rear, whence they slope suddenly downward and forward, being narrow and long; in caurina regularly curved, broadest and high-est near their middle, broader and shorter; sagittal crest in americana only rarely absent, i n caurina often weak or absent; rostrum, i n americana narrow and long, rostral sides parallel; i n caurina short and wide, sides convergent; i n americana anter-ior nares oval, nasal bones with moderate constriction; frontal profile rather sloping, concavity pronounced; in caurina anterior nares more heartshaped, nasal bones with no constriction, frontal profile less sloping, concavity less pronounced; supra-orbital constriction in americana moderate, always present, i n caurina, slight, sometimes absent; postorbital process i n ameri-cana midway between point of greatest constriction and anterior 67 root of zygoma, usually weak; postorbital region short and convex; i n caurina much nearer point of greatest constriction than anterior root of zygoma and stronger, postorbital region long and straight; palate relatively long and narrow, often with azygos notch on posterior surface i n americana, i n caurina shorter and broader, usually truncate behind; anteorbital foramen relatively large, i n both; braincase markedly elongate, more so than i n any other marten, in americana; less extended, broader and lower i n caurina; bullae in americana long, narrow, highly inflated and situated near to each other, so that twice the distance between the carotid canals i s the same or slightly less than the greatest length of bulla; distance between the outer edges of the jugular foramen is less than the greatest length of bullae; distance between the carotid canals i s one half the distance between the foramen lacerum posteriore and the eustachian foramen; i n caurina the bullae are less inflated, shorter, broader and more widely separated so that of the three measurements above,the f i r s t two are greater than the greatest length of bulla, and the distance between the carotid canals equals two-thirds the distance between the other two foramina; in americana the mastoid process does not extend beyond the meatus, i n caurina i t does; general dentition i n americana weaker, upper premolars only occasionally crowded, in caurina stronger, premolars often crowded; i n both PM strongly convex 4 on inner, concave on outer side; PM i n americana with inner cusp small, width being much less than length and only half the width of the trenchant part, placed nearly at right angles to 68 1 4 main axis of tooth, is as long as M is wide; i n caurina PM has the inner cusp large and thick, i t s width nearly equal i t s length and more than one half the width of trenchant part and placed diagonally to main axis of tooth, not as long as M1 i s wide; M^" i n americana small, quadrate, always smaller than i n Old World forms, inner moiety small, very l i t t l e larger or longer than outer moiety and one third smaller than in caurina, i n the 1 latter form M i s large, massive, inner moiety one and a half to two times as large as outer moiety and twice as long; M with an inner cusp present i n both, more often absent i n ameri-cana than i n caurina; M2 with inner cusp weak or absent i n americana, present and well developed i n caurina; baculum in both of the Martes type, bending gradually upwards throughout i t s d i s t al third, ending in two prongs, or more, usually uniting to produce a transverse perforation. Two subspecies are accepted i n this paper. Martes americana americana ('.Kerr,. 1792) Type l o c a l i t y : eastern North America. fiange: North America west to the Coast Range, excepting the Rocky Mountains south of central Idaho and Montana; also New-foundland, Cape Breton Island and Prince Edward Island. Martes americana caurina (Merriam 1890) Type l o c a l i t y : Gray's Harbor, Chehalis County, Washington. Range: the Coast Mountains, Coast Ranges, Cascade Mountains, Sierra Nevada and the Rocky Mountains, south of central Idaho and Montana, also Vancouver Island, the Alexander Archipelago and the Queen Charlotte Islands. 69 Martes pennanti (Erxleben 1777) Type l o c a l i t y : Eastern Canada Range: Holarctic North America As i n the case of M. americana, only those aspects important i n comparison to other species are made here. Characteristics: Active; elongate with short limbs; length of head and body usually greater than 57 cm. (ranging from 50 to 62 cm.); weight from 1500 to 4500 grams; fur coarse, less soft than in the martens; colour dark brown to blackish; outer fur brown-black with often a white subterminal area on each hair; usually without a throat patch, when present small and white; generally darker below than above; t a i l , without hair tips about one half the length of head and body, at least one third the length; t a i l relatively long, with variously 20 to 21 vertebrae (Kneeland 1859» Gerrard 1862); according to Hall (1926a) lacking an abdominal skin gland; skull not massive, although large; skull elongate, narrow, deep; nose rather pro-duced; braincase not massive especially behind; condylobasal length usually more than 95 mm., varying from 92 to 122 mm.; zygomatic width from 55 to 60 mm., being more than one half con-dylobasal length; arches highest at rear, whence they slope suddenly downward and forward, are broader and shorter; sagittal crest strong, only rarely absent; rostrum long, narrow; rostral sides parallel; anterior nares oval; nasal bones slightly con-stricted; frontal profile rather sloping, concavity pronounced; supraorbital constriction moderate, always present; postorbital process about midway between point of greatest constriction and anterior root of zygoma, usually weak; postorbital region (in 70 front of point of greatest constriction) short and convex; palate relatively long (65 mm.) and narrow (10 mm.) usually with distinct azygos process on rear; antorbital foramen small braincase l i g h t , extended; bullae long, but rather widely spaced so that twice the distance between the carotid canals i s greater than the greatest length of bulla; the distance between the outer edges of the jugular foramen i s the same or slightly more than the greatest length of bulla; the distance between the carotid canals equals two thirds the distance between the foramen lacerum posteriore and the eustachian fora-men; the mastoid process extends beyond the edge of the meatus general dentition weak; upper premolars not crowded; PM^  con-4 vex on inner side, concave on outer; PM with inner cusp small width less than i t s length and only half that of trenchant part placed diagonally to main axis of tooth, i s longer than M1 is wide; M1 i s light, quadrate but with inner moiety small as i n M. americana americana, only slightly larger than outer lobe; waist moderately constricted; M^  has well developed inner cusp at base of main cusp; M2 likewise with strong inner cusp; baculum of the Martes type, gradually bending upward throughout i t s d i s t al third, and with either two prongs at the t i p , or these united to form a transverse perforation. Of supposedly three subspecies, here not accepted. Martes flavigula (Boddaert 1785) = Charronia flavigula Pocock 1918 = Lamprogale flavigula Ognev 1928 = Martes guatkinsi Hors f i e l d 1851. Type l o c a l i t y : Unknown, but considered by Pocock to have been 71 fixed by tradition as Nepal. Range: The Amur and Ussuri regions of eastern Siberia, south through most of China, Tibet, Burma, Assam, thence westwards to Kashmir and North-West Frontier, and south into Indo-China, Siam, Malay States, Formosa, Sumatra, Java, Borneo and Banka; also the N i l g i r i H i l l s , Coorg and Travancore, southern India (Sowerby 1916, Ellerman, Morrison-Scott 1951, Brongersma 194-1). Brongersma adds that other unverified records exist i n the literature for Palawan, Great Naturna, Balabac, and Calamianes, Cuyo, Subu, Sibutu, the Paternosters and Sebuko Islands. Jerdon (1874-) says i t occurs on Ceylon, but later authors do not authenticate this. Specimens examined: 32 from Asia. Skull diagrams and photographs: Blanford 1888-91, p. 157; Ognev 1931, vol. 2, pp. 635, 636 and 637; Pocock 1936b, f i g . 2; Brongersma 194-1, p i . 4 and f i g . 1; Colbert and Hooijer 1953, pp. 51 and 52. Other sources are given by Ognev 1931, vol. 2. According to Allen (1938-40) i t i s an animal of wooded mountainous country, f a i r l y well distributed over most of south-ern and central China south of the Gobi. The same author records their feeding on honey bees, which are apparently a favourite food. They are also reported to feed on nectar and to run down and k i l l fawns of the barking deer. See also Appelman 1940, and Hutton 194-9, hot examined during this study. Characteristics: Length of head and body usually more than 57 cm. (Varying from 50 to 80 cm.); general colour black-brown above and yellow below; underchin white, chest, neck and under 72 belly yellow, occasionally tending to white; throat patch almost always extending so as to include the belly; upper l i p never with vertical groove; t a i l long, but not bushy, i t s length (without hair tips) being two thirds the head and body length; t a i l vertebrae 24 (Blanford 1888-91, Owen 1853); soles of feet covered with hair (excepting i n insular forms); skull massive, swollen, broad and f l a t , with short nose; condylobasal length 90 to 110 mm., occasionally as small as 70 mm.; zygomatic width about 60 mm., always greater than one half condylobasal length; arches regularly curved, broadest and highest at their middle, broader and short; sagittal crest absent or very weak; rostrum wide and short with convergent sides; nasal bones strongly constricted; frontal profile not steep, concavity not pronounced; supraorbital", constriction moderate, pointed, always present; postorbital process about midway between point of greatest constriction and anterior root of zygoma, often absent, always weak; postorbital region short and convex; palate wide and short, truncated at hinder margin; anteorbital foramen small; braincase massive, especially at hinder end; bullae short and widely spaced so that twice the distance between the carotid canals or the distance between the outer edges of the jugular foramina are both greater than the greatest length of bulla; dis-tance between the carotid canals is two thirds the distance between the foramen lacerum posteriore and the eustachian foramen; bulla with only two chambers, never three or more as i n M. martes or M. foina; mastoid processes do not extend beyond the edge of the meatus; general dentition weaker; upper premolars not crowded; PMJ strongly convex on inner, concave on outer side; 73 PM with inner cusp small, width less than length and much less than width of trenchant part, placed at right angles to axis of main part, longer than width of M1; M1 small to moderate i n size, inner moiety small, of about same length and size as outer moiety, waist only slightly constricted; the details of the lower molars are unknown to me; baculum distinctive, being easily separated from a l l other species of the genus, as pointed out by Pocock, curving up abruptly i n i t s distal sixth, i t s t i p being nearly v e r t i c a l , length 65 - 80 mm., the tip i s divided into four subsymetrically arranged processes, each with a rounded condyle-like head (for details see Pocock 1918, 1936b and 1941, and Ognev 1931)• The two most recent classifications of the species (in-cluding M. guatkinsi) have been made by Pocock (1936b), Ellerman, Morrison-Scott (195D and Chasen (1940). A l l of these authors, as well as considering M. flavigula and M. guatkinsi distinct, divide M. flavigula into eight subspecies (using quite distinct terminology for two of them). The classification of the latter two authors i s given below, subjugating M. guatkinsi to subspeci-f i c status. Martes flavigula flavigula (Boddaert 1785) Type l o c a l i t y : fixed by Pocock as Nepal. Range: Kashmir to Tibet and upper Burma and southern China, north to Shensi, Kensu (Ellerman, Morrison-Scott 1951) Pocock 1936b). Characterized by the narrowness of the inner lobe of M^ , which is the same width throughout and i n no way widened or flattened on i t s innermost margin (Bonhote 1901). Muzzle 7 4 and top of head black or dark brown; shoulders and back tawny brown to nearly white. Abdomen brown to whitish; throat rich yellow to nearly white; hair on rear foot extending back to and beyond the plantar pad. In India the soles are without hair. (Jerdon I874, Bonhote 1901, Pocock 1936b). Condylobasal length, males 10.06 (9.05 - 10.60) n = 19; females 8.97 (8.20 - 9.70) n = 20. Martes flavigula aterrima (Pallas l 8 l l ) = M. f. borealis Radde (1862) of Pocock. Type l o c a l i t y : between the Uth and Amur Rivers, eastern Siberia, northern China and Korea (Ellerman, Morrison-Scott 1951? Pocock 1936b). Rather larger than typical flavigula and the colour of the back is much more yellow, the colour of the throat much more whitish (Bonhote 1901, Pocock 1936b). According to Howell (1929) individual colour variation i s exceedingly high i n the area occupied by this subspecies. Condylobasal length: males 10.72 (10.20 - 11.00) n = 6; females 9.63 (9.25 - 10.03) n = 2. Martes flavigula chrysospila Swinhoe 1866 = M. f. xanthospila Swinhoe (1870) of Pocock. Type l o c a l i t y : Mountain forests of Formosa. Range: Formosa. Described by Bonhote (1901) as intermediate between the southern forms (peninsularis and henrici) and the typical form to the north. Small in size, with only moderately long fur; differs from typical flavigula by the head being deeper brown; back of the foreleg with a whitish patch (Bonhote 75 1901, Pocock 1936b). Condylobasal length: males 9.20 n = 1; females 8.20 (7.60 -8.80) n = 2. Martes f l a v i g u l a h e n r i c i (Schinz 1845) Type l o c a l i t y : Sumatra. Range: Sumatra, Banka and part of Borneo (Chasen 1940). According to Pocock (1936b) occurs only on Sumatra. In short-ness of fur and naked soles, resembles.peninsularis; separated from i t by i t s smaller s k u l l and darker colouration, es p e c i a l l y on the abdomen (Bonhote 1901, Pocock 1936b). Brongersma (1940) says a specimen of Mustela l u t r e o l i n a from Sumatra i n the Leiden Museum was o r i g i n a l l y m i s - i d e n t i f i e d as M. f. h e n r i c i . Whether t h i s casts doubt on the v a l i d i t y of the subspecies, or of the presence of the species on the island I do not know. Condylobasal length: males 9.23 (8.75 - 9.70) n = 4; females 8.10 n = 1. Martes f l a v i g u l a indochinensis Kloss 1916. Type l o c a l i t y : Klong Menao, southeastern Slam (Ellerman, Morrison-Scott 1 9 5 D , Kohtak, southeastern Siam (Pocock 1936b). Range: Siam, northern Tenasserim and Annam (Pocock 1936b). Distinguished from t y p i c a l f l a v i g u l a and aterrima by a naked area of skin between the lobes of the plantar pad on the hind-foot, by the shorter winter coat, with a more yellowish t i n t , and by the paler patch beneath the eye (Pocock 1936b). Condylobasal length: males 10.00 (10.00 - 10.00) n = 2; females 9.03 (8.80 - 9.20) n = 3 . 7$ Martes flavigula peninsularis (Bonhote 1901). Type l o c a l i t y : Bankachon or Bankasun, Victoria Point, Tennas-serim. Range: southern Tenasserim through the Malay Peninsula (Pocock 1936b, Ellerman, Morrison-Scott 1951). Similar to indochinensis i n having the skin above the hind plantar pads naked, but differs i n having the head brown, not black, shoulders and back less yellow and the abdomen dark brown. The coat i s short and thin, varying l i t t l e seasonally (Bonhote 1901, Pocock 1936b). Condy-lobasal length: males 9.75 ((.70 - 9-80) n = 2; females 8.67 (8.40 - 9.00) n = 4. Martes flavigula robinsoni (Pocock 1936a). Type lo c a l i t y ; Tjbodas, Java. Range: Java (Pocock 1936a). Distinguished from the Malayan and other insular forms i n having the crown, nape and shoulders buffy grey and the line separating the white from the brown on the cheek poorly delineated (Pocock 1936 a and b). Condylobasal length: females 9.10 n = 1. Martes flavigula saba Chasen and Kloss (1932). Type l o c a l i t y : near Sandakan, Bri t i s h North Borneo (Pocock 1936b) Range: Parts of Borneo. Distinguished from henrici by i t s slightly smaller skull and i t s slightly smaller carnassial teeth (Pocock 1936b). Condylobasal length: males 8.82 (8.60 - 8.95) n = 4; females 8.00 (7.80 - 8.20) n = 4. 77 Martes flavigula guatkinsi Horsfield 1851 = M. guatkinsi of authors. Type l o c a l i t y : Madras, India. Range: N i l g i r i H i l l s , Coorg and Travancore, southern India. The baculum, according to Pocock (1936b) differs somewhat from that of the other subspecies of M. flavigula i n that the four distal processes are quite short, the l e f t of the anterior pair being represented merely by a small tubercle. Skull distinc-tive in being depressed, as in an otter's, much fl a t t e r and less convex on i t s dorsal surface than i n the other subspecies °f flavigula, the frontal region being less swollen, with the occiput less sloped and the zygomata less arched. PM"*" may be absent. Colour dark, as in M. f. saba uniformly dark above; unlike saba metatarsal area above the plantar pad not naked (Bonhote 1901, Pocock 1936b). Condylobasal length: males 8.90 n = 1; females 8.60 (8.20 -9.00) n = 2. Six to eight subspecies of Martes flavigula have been described from China, Mongolia and eastern Siberia. Pocock in his 1936(b) revision, reduced the number of these to two, but Howell 1929, Allen 1938-1940 and Brongersma 1941 say even these two f a l l within the range of individual variation and name a l l of them simply as M. flavigula flavigula admitting the possi-b i l i t y of the existence of the northern race, M. f. aterrima. According to Allen, Thomas has concluded that M. f.  indochinensis i s an invalid form and Brongersma (1941) concluded that three insular forms (henrici, saba and roblnsoni) a l l f e l l 78 within the range of individual variation and represented only a single subspecies M.~f. henrici (called M. f. l a s i o t i s by Brongersma). These forms are supposedly characterized by slightly smaller size, thinner pelage and naked feet. Martes flavigula guatkinsi has been to the present considered a distinct species, in part at least because i t i s isolated geographically from M. flavigula. Jerdon (1874) says, however, that while Horsfield applied the name to the south Indian race, originally i t was given to a Himalayan specimen. Tate 19.4-7 says that M. guatkinsi i s only "sometimes treated as a separate species". Pocock (1936b and 194-1) considered the species distinct and pointed out what he considered important distinctions. L i t t l e is gained, however, by spli t t i n g M. f l a v i -gula into two species and I have here considered i t a single form. To summarize the above, i t appears that while Pocock (1936b, 1941), Ellerman, Morrison-Scott (195D and Chasen (1940) consider nine forms of the species to exist, the recent tendency has been towards simplification and these forms may be reduced to five as follows: Martes flavigula flavigula Martes flavigula chrysospila Martes flavigula henrici Martes flavigula peninsularls Martes flavigula guatkinsi 79 RELATIONSHIPS OF THE MARTENS The eventual answer to the puzzling problem of the relationship of the world's marten w i l l be answered, l i k e l y through the study of genetics and paleontology. Some material concerning the latter i s given elsewhere. Concerning the genetics of marten, I know only that E r l i c h (1949)"reports that M. foina like most other carnivores, has a 2n chromosome number of 38 with heterogamic (x - y type) males. I presume that a l l species of the genus possess the same sort of chromosome struc-ture . This discussion is concerned primarily with a comparison of the morphology of the various species. It should be read in conjunction with the discussions elsewhere concerned with f o s s i l marten,zoogeography, etc. I have elsewhere discussed the "specific" status of the various species of marten i n the world. I have pointed out that Severtsov (I876) considered the Tienschan Mountains of Eurasia to be the "homeland" of a l l marten and that here M. foina had been crossed with M. martes and M» zibellina, but that most other workers have found few or no hybrids elsewhere. Schmidt (1943) concluded that the separation between M. martes and M. zibellina was considerably greater than that between M. martes and M. foina. His distinction would thus be: A. M. zibellina B. M. martes and M. foina. Bonhote (1901) related M. flavigula upon the basis of i t s skull conformation to M. foina. Pocock's work on the baculum 80 of these species led to the relegation of M. flavigula to a distinct subgenus. When considering the New World forms i n relation to the Old, we may immediately distinguish the fisher and thus arrive at: A. M. zibellina B. M. martes and M. foina C. M. flavigula (including M. guatkinsi) D. M. pennanti We have not yet considered the relationship of M. a.  americana and M. a. caurina to these, and i t i s this problem that we are primarily concerned with here. The "ecology" of the various species here suggests a clue. Marshall 194-2, correctly I believe, points out that M. americana, M. martes and M. zibellina have, according to available information, so similar a pattern of behaviour and response as to be essentially the same. M. foina, M. flavigula and M. pennanti are, on the other hand, quite significantly different i n their biology from each other and from the preceding, and hence might be grouped thus: A. M. americana, M. martes and M. zibellina B. M. foina C M. flavigula D. M. pennanti This scheme differs somewhat from that above with regard to M. americana, M. foina, M. martes and M. zibellina . It i s of interest to compare them to that given by Gray (1865 and 1869), the last worker to so classify a l l of the world's species of marten. His scheme follows: 81 A. M. martes, M. zibellina, M. americana and M. melampus B. M. foina C. M. flavigula D. M. pennanti Aside from the introduction of the species M» melampus, not yet considered, i t w i l l be seen that Gray's scheme l i e s nearest to the ecological one of Marshall and the one here accepted. The classification currently accepted by taxonomists is essentially this, so that M. foina i s grouped with M. martes  M. zibellina, M. americana and M. melampus as the subgenus Martes; M. pennanti as the subgenus Pekania, and M. flavigula (including M. guatkinsi) the subgenus or genus Charronia or Flavigula. It is thus within the subgenus Martes that we must search primarily for relationship. I think we may accept that M. flavigula (upon the basis of i t s peculiar baculum) and M. pennanti (upon the basis of i t s size and biology) are distinctive from a l l others, although the degree of difference between the subgenera Charronia and Martes i s probably greater and more fundamental than that between Pekania and Martes. There has been disagreement here in the past. Poland (1892) believed M. pennanti to be closely related to M. flavigula, so much so that he called the latter the "Afghan Fisher" and remarked on the disjunction i n the dis-tribution of the two. Schmidt (194-3) considered M. flavigula the Asiatic counterpart of M. pennanti, but these workers did not know of, or i n Schmidt's case did not take seriously, the fundamental distinction of the baculum i n M. flavigula and their suggested relationship cannot be accepted. 82 Returning to the subgenus Martes, I can offer l i t t l e concerning M. melampus, Ellerman, Morrison-Scott (195D say-that i t s colouration i s near M. zibellina, but that i t s skull i s nearer M. foina and M. martes, and they conclude that i t i s an "isolated and valid species". Other detail is given subsequently. It has already been shown that early workers i n North America often considered North American marten to be merely rep-resentatives of M. zibellina or M. martes but never of M. foina or M. melampus. Between 1771 and 1869> of those workers who named New World marten referable to Old World species, eleven referred them to M. zibellina and fourteen to M. martes (five of these referring them to both species). The honours then, have l a i n about equally between M. martes and M. zibellina and not to other species. We may devise from this again a scheme of rela-tionship: A. 1. Martes martes, M. zibellina and M. americana 2. M. foina 3. M. melampus (possibly with M. foina) B. 1. M. pennanti C. 1. M. flavigula The earlier workers on the problem, Gray (1865), Brandt (1855), Baird (1857), Allen (1869), and Coues (1077) came to the conclusions referred to above. Gray distinguished M. americana  americana from the Old World forms on the shape and size of the upper molar. Brandt and Allen employed only characteristics of pelage. Baird employed both cranial and pelage characteristics as did Coues. A l l related M. americana to M. martes or M. zib-e l l i n a and Coues, the last of the group, reviewed the findings 83 of the earlier workers and found closest relationship to M. zib-e l l i n a of the two. Coues and the earlier workers were unfortunate i n that they did not recognize that a distinctive marten occurred on the west coast of North America (M. a. caurina) and i t must be assumed, built their case around representatives of M. a. ameri-cana. In 1890 Merriam described caurina and said that i n cranial and dental characters i t departed from M. a. americana i n the direction of M. zibellina. Rhoads (1902) reviewed the problem i n some detail. M.a.  americana he considered total l y distinct from a l l other forms on the basis of the smallness of the inner lobe of M"*", as f i r s t out-lined by Gray and the smallness of the lower molar and the small size of the inner cusp of M2. In general shape of skull M. a.  caurina was nearest to M. foina but i n the large size of the inner lobe of the upper molar and the large size of the lower f i r s t molar this form was very similar to M. martes and M. zib- bellina, not at a l l to M. americana americana to the east. In shape of skull, then, Rhoads related M. zibellina to M. americana and M. martes; M. foina to M. a. caurina. On the basis of dentition, however, M. a. americana was distinct, and M. a. caur-rina, M. martes and M. zibellina closely related and M. foina i t s e l f distinctive. He concluded of M. a. caurina that " i t s homologies connect i t far more closely with the Eurasian than the American fM. a. americana] type of marten.... [It] i s an isolated member of the martes - zibellina group". 84 Lydekker (1901 - 1904) stated that he believed the marten of North America were so nearly related to M. martes and M. zi b e l l i n a that they should possibly be considered only varieties of these. He added no detail. Ognev (193D reviewing the marten of Russia said that M. zibellina lay intermediate i n cranial and dental character-i s t i c s between M. martes and "M. americana". Subsequent des-cription of "M. americana" makes i t clear that Ognev was refer-ring to M. a. americana and that he was giving no consideration to M. a. caurina. Allen in 1938-40 said only that "Probably ... the sable (martes zibellina) ... finds i t s counterpart In the Ameri-can Pine Martens (M. americana and races), but the Beech Marten (M. foina) i s not represented i n the New World". My own interpretation of the relationship of martens, based upon examination of skulls of the species involved, leads me to conclude that the above writers are nowhere really wrong, and that M. a. caurina i s most nearly related to M. martes and M. zibellina. M. a. americana is distinctive with some features allying i t to M. zibellina and that M. melampus is to some degree related to martes and zibellina. When I f i r s t began my study of skulls of a l l the species of marten, I was impressed by the fact that two basic types of skull existed. At the one extreme was the short, broad, and rounded skull, with short widely spaced bullae, and at the other, the long closely spaced bullae. The species of marten, placed in order, going from the former type to the 85 latter, would f a l l i n this sequence: M. melampus, M. pennanti, M. foina, M. flavigula, M. a. caurina, M. martes, M. z i b e l l i n a and M. a. americana (see Figure 7). The greatest division between the two groups occurs between M. melampus and M. pennanti and M. martes and M. zibellina. I f e l t that I could at that point divide martens into two basic divisions. I was puzzled, however, by the fact that both M. flavigula and M. pennanti f e l l so clearly within the f i r s t group and yet be so distinctive other-wise as to merit subgeneric status. Perhaps M. pennanti should not be so distinguished, but M. flavigula i n view of i t s distinc-tive baculum, biology, distribution and antiquity, as shows i n the palaeontological record could not be so ignored. It seemed that the two schemes of classification were mutually exclusive. In attempting to resolve the problem, help was given by Alston (1879) who remarked on the following findings of Blasius, with which Alston himself agreed: that the differences of proportion in the skulls of M. martes and M. foina are less conspicuous when a skull of an aged M. foina is compared to that of a young M.  martes, than when individuals of the same age were contrasted. The implication that was made was that at a younger stage M. martes is more similar to M. foina than i t is- later on i n i t s l i f e . I myself have been fortunate enough to examine five skulls of very young marten (perhaps two or three months old) two referable on geographic grounds to M. a. americana and three to M. a. caurina. I have been led to conclude from these that at this very young stage, the skulls of the two races are nearly indistinguishable, and what is more, both are marked to a greatly Figure 7 (To face page 86.) Figure 7« Ratio of canine width to b u l l a length i n the skulls of the world's marten. This r a t i o approximates an index combining the character-i s t i c s of s k u l l length, height and breadth. S K U L L L O N G , N A R R O W , H I G H , T O G E T H E R 86 exaggerated degree by a skull conformation of the foina - caurina type with short widely spaced bullae, broad low braincase and short rather broad rostrum, rather than the elongate skull of the zib e l l i n a americana type. I conclude that the short, broad and low skull i s char-acteristic of a l l martens i n their youth; that i n some (e.g. M. pennanti, foina, flavigula, caurina and martes) i t i s lost only to a slight degree with maturity. In the remaining species (M. zibellina and M. americana americana) while i n youth they have the same skull type, with maturity i t becomes elongate, high, narrow, and with long bullae. Thus, again referring to Figure 7, i t i s hypothesized that the pennanti - melampus group are neotenic; that i s , the development of the skull i s either primitive or retarded, while the americana - zibellina end of the scale i s a result of hypermorphosis or acceleration i n devel-opment, or possibly i t is a primitive condition (i.e. neoteny and hypermorphosis are mutually exclusive conditions; which of the two exists i s not known). While I have no means at present of testing this hypothesis for the time being I accept i t . If i t i s to be accepted, certain conclusions follow. The most important of these i s this, that while a l l species possess skull types that f a l l within two extreme kinds, forming a gradient, two of these do not belong, because of other distinctions (M. pennanti and M. flavigula). It follows that the neotenic or hypermorphotic condition has arisen several times i n quite distinctive groups; 87 thus we may make a classification of species, something like this: Skull broad Skull long and short and narrow Subgenus Martes: melampus, foina, caurina, martes, zibellina, americana Subgenus Pekania: pennanti Subgenus Flavigula: flavigula We come now to the problem: i s the sequence shown i n skull shape within the subgenus Martes a real one? To attempt this one must consider dentition. Figure 8 depicts pie dia-grams of ratios, given i n percentage, made from measurements taken from teeth of the various species. These are essentially the ratios devised by Brongersma (1941) and referred to previous-ly . These diagrams show basic similarity i n the combinations of the six dental ratios depicted for the following species: M. melampus, M. a. caurina, M. martes and M. zibellina . Of these M. a. caurina is apparently the most distinctive. M. a. americana and M. foina are totally distinctive, and strangely, are to a surprising degree similar i n dental conformation though distinctive i n skull shape. I assume that these relationships are real and i n view of the lack of other evidence, that within the subgenus Martes, M. melampus, M. a. caurina, M. martes and M. zibellina are closely related species. M. foina and M. a. americana are distinctive i n dental characteristics and within Figure 8 face page 88) Figure 8. Pie diagrams depicting certain characteristics of the skulls of the world's marten. The indices used are, with the exception of one, those devised by Brongersma (1941). The values graphed are not absolute, but rather percentages, the smallest mean equalling 0 i n the pie cut, the largest 100. The pie cuts are as follows: 1. Width of inner lobe of P 4 x 100 / width of trenchant part of P 4 2. Width of inner lobe of P 4 x 100 / length of P 4 3. Breadth of M1 x 100 / length of P 4 4. Lateral length of M1 x 100/mesial length of M 5. Mesial length of M1 x 100 / breadth of M1 6. Breadth of P 4 / length of P 4 7. Canine width at base x 100 / bulla length 88 themselves are similar i n this respect, but are so different cranially and i n their biology as to be l i k e l y distinctive within themselves. On zoogeographic grounds, and pelage structure and colour, M. a. americana shows a degree of rela-tionship with M. zibellina and M. martes, but for the time being i t i s here considered otherwise. Thus a scheme of re-lationship within the subgenus might be drawn up thus: Skull broad, short Skull long, narrow P 4 small, with small inner lobe; M1 with foina americana small inner moiety 4 P large, with large inner lobe; M^" with melampus caurina martes zibellina large inner moietsy k hierarchical classification designed to show these relationships may be prepared as follows. It must be remembered that this does not attempt to represent phylogeny, though to a degree i t might. For the time being i t attempts rather to depict morphological similarity of crania. 89 Subgenus Martes: 1. M. melampus, americana caurina, martes and zibellina 2 . M. foina 3. M. americana Subgenus Pekania: 1. M. pennanti Subgenus Charronia: 1. M. flavigula The baculum of M. pennanti is so similar to that of the species of the subgenus Martes, that i t s relationship to that subgenus must be considered much closer than is the rela-tionship between the subgenera Martes and Charronia. There i s one problem more to be considered. The evi-dence appears such that M. a. americana and M. a. caurina are as morphologically distinct between themselves as either i s compared to any of the Old World species. In fact, M. a. caurina appears much more closely related to M. melampus, M. martes or M. zibellina than M. a. americana does to M. a. caurina. And yet, as Wright has shown, M. a. americana and M. a. caurina interbreed, and must be considered to represent the same biologi-cal species. If this i s so, one wonders how i t i s that the Eurasian species, so much more similar morphologically, have maintained their specific distinction i n the eyes of systematists, It has occurred to me that M. martes and M. zibellina may possibly interbreed where their ranges meet i n north central Russia and that M. zib e l l i n a and M. melampus interbreed where their ranges meet in Korea. It is significant that the ranges of these species are nowhere sympatric. One wonders, too, i f 90 M. a. caurina and hence M. a. americana might not prove to be conspecific i n the same sense with M. zibellina. If this were shown to be so, i t would reduce the number of species of marten in the world from eight as here interpreted to a maximum of four as follows: M. martes (including M. zibellina, M. melampus and M. americana) M. foina M. flavigula (including M. guatkinsi) M. pennanti. D. THE MARTENS OF THE NEW WORLD LIFE HISTORY Brief accounts of the l i f e histories of the European forms have been given i n the pages preceding. Schmidt (194-3) assures us that the natural history, reproductory and food habits of Martes foina, martes and zibellina are in every respect almost identical to those of M. americana and M. pennanti. Thus i t would appear that what is briefly stated below may be attributed also to the Eurasian forms named above. Martes americana: Birth occurs usually i n Ap r i l , less often i n May and June (l a Beree, Brassard and Bernard 1939» Schmidt 1943, Lensink 1953, Pearson and Enders 1944 a and b). According to Pearson and Enders (1944a) this period can be shortened by three or four months by means of a r t i f i c i a l lighting, and since the long gestation period has been shown by the above authors to be caused by delayed implantation of the blastocyst, they concluded 91 that implantation is controlled by day length i n spring. Oddly enough, although the long gestation period has only recently been determined, Kerr i n 1792 wrote that the period was said to be nine months. The l i t t e r size i s one to six, averaging 2.6 to 3*0 (Brassard and Bernard 1939, Markley and Bassett 1942, Marshall 1951, Schmidt 1943, Lensink 1953). The animals are born toothless, and the permanent adult dentition is attained by the 16th to 18th week (Brassard and Bernard 1939, Schmidt 1943). Both above authors give detailed descriptions of the sequence by which adult dentition i s attained. The young leave the nest at the earliest by the 44th day, but usually by the eighth week, this being most often sometime i n mid-July (Brassard and Bernard 1939, Schmidt 1943). The f u l l adult weight is approximated by the end of the third month (Rand 1948a, GrinnelL, Dixon and Lins-dale 1937, Bailey 1936, Brassard and Bernard 1939) though the animal continues to f i l l out until i t s third year (la Beree 1941). The animals f i r s t mate i n their second year, giving birth to their f i r s t l i t t e r i n their third year ( l a Beree 1941, Markley and Bassett 1942, Marshall 1951, Schmidt 1943). According to Lensink they occasionally mate for the f i r s t time at fifteen months, usually at twenty-seven months of age. They are usually considered polygamous (Marshall 1951, Schmidt 1943). One l i t t e r is produced every year thereafter until death, the oldest known animal being seventeen years old when i t died (la Beree 1941, Seton 1925-1928, Lensink 1953). Growth curves for the early stages of marten l i f e are given by Markley and Bassett 1942, Remington 1952, and Schmidt 1943. 92 As remarked on earlier, the animals occur in greatest numbers in coniferous forests, especially those dominated by spruce (Picea). Densities of the animal per unit area vary from loc a l i t y to loc a l i t y . The area covered by individuals appears to vary with sex, season and age. Marshall (195D believed the winter range of marten to be about ten to fifteen miles. Males appear to hunt over a wider area than females (Yeager 1950). Lensink (1953), reviewing the literature of de Vos (1952) and Yeager (1950), who had both reviewed the earlier literature, and including with these his own figures from interior Alaska and figures based on skulls used in the present study, found that 159 males were trapped for every 100 females. Various authors have, however, suggested that this difference may only be a result of the wider cruising radius of the male and not a real deviation from a 1 : 1 ratio. Quick (1953) in northern B r i t i s h Columbia found the ratio i n 1947 to be 1 : 1, in 1948 1.2 : 1. Lampio (1951) likewise found a nearly 1 : 1 ratio to exist i n M. martes in Finland. Lensink (1953) concluded that nearly 50$ of the marten population was made up of immature (non-breeding) animals. Lensink (1953) reviews the literature on foods eaten by Marten (Cowan and Mackay 1950, Marshall 1946, Newby 1952, Reming-ton 1950 and others) adding his own observations to these. Microtines comprised by far the largest proportion of food taken, except in Washington where squirrels appear the most eaten. Other foods include animals to the size of rabbits, insects and berries. See also Newby and Hawley (1954) not examined during this study. 93 Martes pennanti: Birth occurs from March 15 to the end of April (Douglas 194-3, l a Beree 1941, Hodgson 1937, Seton 1925-1928, Enders and Pearson 1943a and b, Schmidt 1943, Hall 1942), as i n marten probably varying with latitude. Mating occurs from three days to eighteen days after birth, probably about a week as a rule. Dates of birth given i n the literature vary from March 15 to April 27 (Douglas 1943, l a Beree 1941, Hodgson 1937, Enders and Pearson 1943a and b, Schmidt 1943, Hall 1942). The gestation period i s stated to vary from 210 to 370 days, averaging about 350 days (Douglas 1943, l a Beree 1941, Hodgson 1937, Pearson and Enders 1944a and b, Hall 1942, Schmidt 1943, de Vos 1952). According to Enders and Pearson 1943a and b, the period of delayed blastocystic implantation i s nine or more months. The l i t t e r size i s said to be one to five, averaging two or three (Douglas 1943, Hodgson 1937, Schmidt 1943, de Vos 1952, Seton 1925-1928. The features coincident to the attain-ing of adult dentition have yet to be worked out, but i n view of the similarity of the other features of the l i f e history to that of the martens i t i s probably quite similar. The young leave the nest at nine to ten weeks (Schmidt 1943) and are f u l l sized by seven months (Hodgson 1937)• Mating f i r s t occurs during the second year, the f i r s t l i t t e r being born i n the third (Douglas 1943, Hodgson 1937, Hall 1942). The animals are said to be polygamous (Douglas 1943). Fisher tend to occur i n lower altitudes than marten (Edwards 1950, de Vos 1952). De Vos considers i n rather great detail the habitat requirements of the animal. The animals 94 appear to vary i n density geographically, staying generally within an area of two to six square miles (de Vos 1952, Quick 1953). Unlike marten, there appear to be more females than males per unit area, trapping figures giving 49 males to 51 fe-males (de Vos 1952) and one male to two females (Quick 1953). Animals captured by fisher as food are largely grouse, red squirrels, hares, shrews, porcupines and others (de Vos 1952, Quick 1953). VARIATION Non-geographic variations There are several kinds of non-geographic variation, namely that which occurs by sex, by age, by season and that which i s characteristic of individuals of the same sex and age, here termed individual variation, whether genetic or otherwise. Since any attempt to describe geographic variation must be pre-ceded by some understanding of the noh-geographic sort, so that the two may be distinguished a brief account follows. Burt (1953) records the taking of about 400 marten from Fort Nelson British Columbia, which should be of use to anyone studying non-geographic variation. Sexual variation: We may consider here only the more obvious of the secondary sexual characteristics. The most important of these is size. The males i n a l l species of martens and fishers are considerably larger than the females. For martens i n terms of weight, i t may be stated as a general rule that the female weighs approximately two-thirds that of the male. The pelage 95 of the male is thicker, with longer hairs and a greater elas-t i c i t y to the fur (see Brassard and Bernard 1939? de Vos 1952, Markley and Bassett 1942, Lensink 1953? Lampio 1951). The under fur anterior to the urethral aperture on the midline of the belly l i e s forward, rather than backward i n the male (Lampio 1951, Lensink 1953). The skulls of the males are plainly heavier and more robust than those of the females. The canines and the dentition in general is heavier and the sagittal crest more l i k e l y to be highly developed i n those forms where i t occurs. For Martes  americana americana and M. a. caurina the measurements of large samples of six skull characteristics a r b i t r a r i l y chosen were compared by sex and i t was found that a l l of these averaged 1.10 2 times larger i n the male than in the female, and the X probabil-i t y that these averages came from the same supply was far beyond the 99$ level. The characters employed were condylobasal length, rostral width, upper tooth row, bulla length, rostral width / bulla length and upper molar widthplus length. The use of the above index would allow one to predict the skull size and characteristics of both sexes of any unknown form of North American marten, i f only one sex i s known. No comparable comparisons were made for Eurasian martens or M. pennanti. Hall (1934) says the figure i s about 1.6 compared to 1.1 i n marten. Most of the specimens of both marten and fisher used i n this study were already designated to sex. Where the sex of an animal was not known, i t was identified by comparison to bar graphs of condylobasal length and other measurements drawn up 96 from larger samples of known sex, from the same general area. Age Variation: As marten advance from maturity (second year) into old age, the following changes have been noted to occur: the condylobasal length increases (Lensink 1953» Wright 1953), the zygomatic breadth increases (Grinnell, Dixon and Linsdale 1937? Wright 1953)? the braincase is said to decrease i n size (Grinnell, et a l . , 1937)» the postorbital constriction narrows (Grinnell et a l . , 1937"? Wright 1953), the rostrum narrows and lengthens (Grinnell et a l . , 1937)> the naso-maxillary sutures fuse (other sutures ankylose very early) (Grinnell et a l . , 1937> Rhoads 1902, Lensink 1953? Marshall 195D > the postorbital pro-cess becomes more prominent (Grinnell et a l . , Lensink, Marshall), the baculum increases i n weight (Marshall). During the early stages of l i f e , age may be obtained by use of growth curves (Markley and Bassett 194-2, Remington 1952, Schmidt 194-3) or to the eighteenth week by the sequence of tooth eruption (Brassard and Bernard 1939, Schmidt 1943). Working with "M. c. caurina" of central Idaho, Marshall (1951) concluded that the presence of a sagittal crest i n a female or one longer i than 2 mm. i n a male implied sexual maturity (two years or older). Likewise Marshall concluded that a baculum heavier than 100 mg. implied sexual maturity. Lensink (1953), working on the marten of interior Alaska, concluded that he was able to t e l l the age of males to four years, of females to five years by means of weight of baculum and height and length of sagittal crest. His table of measurements is reproduced below. 97 Males Age Baculum weight (mgs.) Height of sagittal crest (mm.) Length of sagittal Crest (mm.) 0 - 1 Less than 130 0 0 1 - 2 130 - 205 Less than 16 1 - 2 0 2 - 3 206 - 260 16 - 19 25 - 40 Over 4 More than 260 More than 18 41 - 50 Over 5 More than 260 More than 18 More than 45 Females Age Length of sagit-t a l crest (mm.) Height of sagittal crest (mm.) Min. separation of temporal muscle scars ( mm.) 0 - 1 0 0 More than 5 1 - 2 0 0 Less than 5 2 - 3 1 - 2 0 0 0 3 - 4 2 5 - 3 7 0 0 Over 5 More than 38 0 0 Lensink's work comes closest to providing a means to age marten skulls, and i n systematic work, to break samples into age subsamples. However, one result of the present study has been to show that marten from different l o c a l i t i e s possess bacula of significantly different weights at maturity and acquire sagit-t a l crests of different sizes at different ages. At the beginning of this study, twenty-four character-i s t i c s were selected from a population of marten of restricted geographic extent (Banff region, Alberta). These were broken into two age groups by use of the following single characteristic: 98 fusion or lack of fusion of the naso-maxillary suture. Where fusion was complete, the animals (fide Rhoads 1902) were consid-ered adult; where unfused, immature. Wright (1953) has done the same and in common with him I found that the following char-acters varied significantly between the two age groups: greatest and condylobasal length of skull; zygomatic breadth, postorbital constriction, postorbital width, vomerine width, width of incisor row at base, rostral width and length and lower jaw length. Mastoid breadth, interorbital breadth, height of cranium, size of auditory bullae, and size of teeth did not vary (Wright found no variation also i n rostral width, concerning which see above, basilar length, palatal length). Of the twelve measurements used i n the present study, those found to change with age were measured only in adults, the remainder on both adult and immatures, as follows: 1. Condylobasal length. Adults only. 2. Rostral width. Adults only. 3. Upper tooth row. Adults and immatures. 4. Canine width at base. Adults only. 5. Bulla length. Adults and immatures. 6. Upper molar width. Adults and immatures. 7. Upper molar inner moiety length. Adults and immatures. 8. Palatal length. Adults and immatures. 9. Mastoid width, adults and immatures. 10. Height of skull at bullae. Adults only. 11. Length of last upper premolar. Adults and immatures. 12. Palatal width. Adults only. 99 Details of the above measurements and those given below for fisher are given i n Appendix C. In general, where large samples were available, only the adult portion (by the naso-maxillary test) were measured. In fisher, the following characteristics are known to change as old age i s attained: lightening i n pelage colour (Rhoads 1903); an increase in condylobasal length (de Vos 1952); increase i n the weight of the skull, the male about 25$? the female about 9% (Grinnell et a l . ) ; an outward bowing of the zygomatic arches (Coues 1877? Thomas 1886, Grinnell et a l . , de Vos); the postorbital constriction narrows (Coues, Thomas, Grinnell et a l . ) ; the lambdoidal crest increases i n size (Coues); the sutures of the nasal region and basi-cranium ankyl-ose (Coues, Grinnel e_t a l . , de Vos); the canines become longer (Thomas); the interorbital width increases (de Vos); the sag-i t t a l crest increases i n height and length (Coues, Thomas, Grinnel et a l . ) ; the mandible increases with depth (Hall 194-2); and the baculum increases i n weight (de Vos). According to de Vos a male with no sagittal crest i s juvenile, with one less than 62 mm. long is adult, and with one longer is old adult. Of females, only old adults bear a crest. As i n the case of marten, a large number of character-i s t i c s were measured for a sample limited geographically (Stuart Lake, Br i t i s h Columbia), these being broken into ^subadult" and "adult" on the basis of degree of fusion of the naso-maxillaries. No significant variation was found i n the upper tooth row, 100 mastoid width, palatal length, preorbital and postorbital widths, and size of teeth. These measurements were:henceforth taken on a l l "adult" and "immature" animals. The remaining characters were found to vary with age, and were measured only on "adults". These included condylobasal length, canine width at base, rostral width, bulla length, and palate length. As i n marten, where large samples were available, only "adults" were measured. Rhoads i n 1898 expressed the belief that the white patches on the breasts of fishers were lost with age. In 1903 he suggested further, that the oldest specimens are the lightest coloured, some becoming nearly white. Seasonal variation: This kind of variation is concerned chiefly with variation i n the condition of fur. Surprisingly l i t t l e is known concerning the moulting of marten and fisher. Grinnell, Dixon and Linsdale (1937) reported one complete annual moult in the marten of the Sierra Nevada. This proceeded slowly through the autumn months. It begins i n the t a i l , about the end of August. They also report a loss of hair with no renewal i n the spring months, for summer skins are less dense i n pelage than are the winter ones. Summer and winter skins, they reported, are closely alike i n colouration* the former being only slightly the paler of the two. Markley and Bassett 1942 reported that during the summer moult the colour changed from "dark brown to light tan".. The moult began on the t a i l t i p , then moved to the legs, the mid-dorsal region and the face, i n that order. 101 According to Schmidt (194-3) European marten undergo two annual moults. The hair begins to f a l l at the end of April or beginning of May and lasts until the beginning of June. Moult-ing begins on the face and works back to the t a i l . The summer coat is darker than that of the winter. Pregnant females end their summer moult much earlier than non-breeding ones. The f a l l moult begins in August, and ends by November. Here the order of hair loss i s reversed; as in Grinnell*s animals, the f i r s t part affected being the t a i l , the last the neck and head. Grinnell et a l . concluded that California fisher show only one annual moult, this occurring slowly, i n the f a l l . With the advance of summer a slight fading of tones appears. It seems l i k e l y then, that fisher and marten undergo a f a l l moult, and a spring loss of hair, at least. My own detailed examination of skins has been largely restricted to specimens taken from the Canadian Rockies. These show a summer pelage far darker and thinner than the winter one. The f a l l moult appears to occur considerably later than i t does i n the Californian speci-mens examined, time of moulting l i k e l y being, as Bissonnett (1944) shows in weasels, a function of day length. In general, where I have attempted comparison of skins, I have divided them into three groups, summer, winter (by date of capture) and moulting (by appearance). Comparison has then been made within each of the classes "summer" and "winter". Individual variation: Individual variation has been recorded for a long time in martens and i s known to be extreme, especially 1 0 2 with regard to colour of pelage. A l l writers, however, have failed to distinguish between variation that might be due to sex, age or season, which makes the situation a d i f f i c u l t one to interpret. In 1 8 7 5 Cartwright wrote that the marten "varies the most in colour, i n the same region, of any animal; some are light yellow, some are dark brown, and others are black". Mac-farlane ( 1 9 0 5 ) wrote that he had "seen several albino examples and also a considerable number of bright yellow and dark orange coloured martens particularly while stationed i n the dis-t r i c t s of Mackenzie River and Athabasca". Hardy i n 1 9 1 0 record-ed the taking of a very dark specimen from Maine. It was "the darkest furred one I ever saw south of the St. Lawrence" he reported, adding that he had examined over 2 0 , 0 0 0 pelts from the region, but none so dark as this. Cory ( 1 9 1 2 ) remarks that i n eastern marten the brown colour i s lighter or darker i n different individuals, the throat patch varying from orange to yellowish white. Markley and Bassett 194-2 reporting on marten of :Idaho and Wyoming, say that "evident at once is a great individual variation i n extent of the throat patch. Its shape, always irregular, i n several specimens was restricted to a number of detached spots on the lower throat and breast, while on others i t became a solid patch extending from the upper throat to the abdominal region. On a prime winter pelt the t a i l and legs became nearly black on some individuals." 103 A detailed account i s given by de Vos (1952). The colour of the fur, he concludes, i s "extremely variable". He presents a series of skins, selected from a collection of 450 pelts trapped from November 1st to March 31st of two consecutive years. These varied from "a very pale yellow-orange marten of a partly cromistic, partly albinistic type, "with an entirely white head" and with white underparts, feet and t a i l the same colour of the upper parts and with no throat patch to a "very dark marten with a large number of white guard hairs the underparts ... of a similar colour ... with a large orange throat patch." He also records a dark, possibly melanistic specimen, but no albinos, though Ponomarev (1938) found them i n Russian sables. Ponomarev reported also a tendency toward geographic colour v a r i a b i l i t y and de Vos likewise found the marten of Algonquin Park, Ontario, to be, on the average, slight-l y paler. De Vos concluded that "the v a r i a b i l i t y i n a local series indicates that one should be on his guard when making comparisons between series from different l o c a l i t i e s , and that one preferably should compare large series." Wright i n 1953» working with 300 specimens designated M. a. abietinoides from Montana and Idaho, likewise found great v a r i a b i l i t y ; dorsally they varied from Isabella color or Mouse Grey to Mummy Brown, and ventrally from Isabella or Mouse Grey to nearly black. The throat patch varied from almost complete absence to five or six inches square and from Pale Orange-Yellow to Capucine Yellow or Deep Chrome (colours from Ridgway, 1912). 104 Size variation, within a limited area, appears to be rather less than one might expect on superficial examination of skins and skulls. A sample of 100 skins taken from Ontario, when divided into male and female, but not into age groups, give the following average coefficients of v a r i a b i l i t y : total length 4.2$; t a i l length, 6.1$; hind foot 6.0$; and weight 13.2$. Another sample of 21 "adult" males from near Banff, Alberta, when measured for twenty-four skull characteristics, gave an average coefficient of v a r i a b i l i t y of 3.3$, with a maximum much higher i n only one characteristic (7*5$ for postorbital width). I believe that fishers show rather less v a r i a b i l i t y than do martens. De Vos (1952) examined 200 skins from Ontario taken between late November and late March. He remarked "Colour variations between individuals are f a i r l y striking. No such variations could be detected between local populations. The overall colour of the dorsum varies from pale grey to practi-cally black .... This is true for both sexes, although the gray colour predominates i n males. A reddish brown tinge can be noticed occasionally.... On the ventral side there i s not much variation between individuals, although light backed speci-mens are slightly paler than dark ones. No large throat patches ... were observed, although a few individuals had one or a few white spots on the throat. The largest white patch i s usually found on the belly, anterior to the anus. In some animals white spots are present i n the axillas of the front legs which may occasionally be fused on the mid ventral line and extend anteriorly for a short distance." 105 Audubon and Bachman (1851-1854) say they had examined a specimen that was "nearly white with a brown head". E l l i o t (1901a) says they vary from glossy black to gray or grayish white on the upper parts. Grinnell, Dixon and Linsdale (1937) working with Californian animals, however, concluded that colour variations were slight, being expressed chiefly i n amount of white on the underparts. Size variation i s about the same as i n marten. Merriam 1882-1884 and 1886, however, records the taking of one in New York, that was estimated to weigh 40 pounds and the skin of which was that of "a good sized otter". He records another from the Adirondacks with a skull 12.5 cm. long. Merriam con-cluded that giantism i n fishers occurred only among males. A sample of sixteen male fisher, from Ontario, not divided into age groups gives the following coefficients of v a r i a b i l i t y for external measurements: total length 6.8$; t a i l length 8.0% and hind foot 3*7%' A sample of thirty-three male adult skulls from the intermontane plateau region of middle British Columbia gives an average coefficient of v a r i a b i l i t y for fourteen characters of the skull of 4.1, none being higher than Q.6% (width of postorbital constriction). I have myself carefully examined only a relatively small number of skins (about 75 marten, about 40 fisher), although a larger number were brie f l y examined and measurements recorded. The reason for the smallness of the number examined carefully was that the colour of pelage seemed to be an unreliable solution to the problem. The degree of individual colour variation was 106 so extremely high, especially i n marten, that I have concluded that very l i t t l e can be determined by ..it, and that the rather careful descriptions of pelage colour made by earlier describers in many cases provide disappointingly l i t t l e help. As w i l l be outlined later, this does not mean that pelage colour Is not of some use to the systematist, however. Martens can be broken into several distinctive geographic colour types as follows. Those from Labrador are dark, those from Ontario and southeastern British Columbia are intermediate, those from Alaska, northern Brit i s h Columbia and the northern Prairie Provinces are paler, especially about the head. Marten from the northern Coast Ranges, Cascades, Coast Mountains and the American Rockies are dark, while those from California tend to be lighter. The situation with fisher is less distinct. While they appear to vary less i n pelage individually, they also vary much less geographically. It i s unlikely that usable samples from the west can be distinguished from those further east. In view of these findings i t was very early decided that this study would need to be based almost wholly on cranial char-acteristics. While external measurements show v a r i a b i l i t y a l -most as low as do cranial ones, these are usually limited to three or four taken at time of skinning, and hence are of restricted use. The cranial measurements used, and the conclusions derived from them follow. Methods of study of geographic variation In attempting to solve the problems presented by the martens, I was faced with that of obtaining s t a t i s t i c a l estimates 107 of the parameters to be studied quickly. The subsamples to be compared to each other were ob-tained by accretion; that i s , by taking reasonably large samples from near a given type l o c a l i t y as standard for a subspecies and adding to i t similar forms, working away from the type loc a l i t y , until I reached samples whose statistics lay nearer to those of another type lo c a l i t y . This i s admittedly an unsatisfactory technique, but presumably i s the one used by other systematists and at the time was accepted by me as the one to be used. The next problem to unravel was this: how randomly distributed were the subsamples thus obtained. Dice (1952) points out that most collections i n our natural history museums were not collected by random sampling methods and hence are not really suited to the application of s t a t i s t i c a l methods. Dice is certainly correct i n this assumption. Mine being a study carried on indoors, however, no means of correcting this situa-tion was available to me, and I have been obliged to construe my samples to be of random distribution although I know they are not. The next problem was one of homogeneity. It was possi-ble that each subsample (when age. sex, etc. were taken into consideration) could consist of more than one natural population. Ideally each subsample should have been tested for skewness or kurtosis; perhaps analysis of variance would have worked well here. Harding (1949) has reported a method that might work equally well. Time restricted me to the simple expedient of 108 calculating the coefficient of v a r i a b i l i t y i n place of the other techniques. This s t a t i s t i c does not measure homogeneity i n the true sense, nonetheless I have been obliged to consider a CV o f less than about six to indicate a normally distributed sample. Where i t has been higher than six, I have attempted more refined means of determining i t s cause. The comparison of samples may be done i n complex, satis-fying fashion, or more simply and perhaps less revealingly. I would think that the several methods of multivariate analysis available (see for example Fisher 1938, Mather 1947, Stone 1947, Storer 1950) would be the most efficient, or perhaps the simple method of trivariate analysis suggested by Burma 1948. Time has again limited this study to a simple comparison of two var-iates, these being made singly through a l l of the samples. The simple " t " test i s much used here and has been simplified by Simpson and Roe (1939) and Mayr, Linsley and Usinger (1953) by using the standard error of the difference of the means. Even simpler than these are the comparative graphs f i r s t used by Dice and Leraas (1936). Different authors have suggested different levels of probabilities should be used, Dice and Leraas suggesting the use of the x - 2 or 3 SE ; Simpson and Roe (1939) x + 3 SE : Amadon (1949) x I 3.24 SD; Hubbs and Perlmutter (1942) x -ICSD; Mayr, Linsley and Usinger (1953) x + 1.5 SD; Hubbs and Hubbs (1953) x 1 2 SE^ . Actually as satisfactory as any of these, should be the use of "1 1 1 2 "* 1 have, however, restricted myself to the use of x + 2 SE^ i n comparing subsamples and using " t " where n i n either or both was less than ten. 109 Clark (1952) suggested the use of a "coefficient of divergence" for comparing the degree of dissimilarity or similar-i t y of a number of variates of a number of subsamples simultan-eously. I have not employed his technique but feel that i t might offer a more precise means of evaluating degree of similar-i t y or otherwise than any of the techniques I have employed. A l l of the methods so far discussed are analytical. I discovered too late a method of almost synthetic viewpoint that had I known of i t would have been the major technique employed i n this study. I report i t i n some detail here, wishing that I might have used i t and hoping that some other systematists may see i t s essential worth. This technique, devised by Womble i n 1951? i s essentially a method of summing and weighing morpholog-i c a l values and mapping them two-dimensionally on a map. The method i s this: the values of each t r a i t are recalculated to a scale of 0 to 1, these representing the minimum and maximum averages for each t r a i t . Each t r a i t i s then, i f so desired, weighted according to whatever degree of genetic worth i t i s valued at, the values for each rescaled and weighted t r a i t are then mapped, a contour interval selected and isopleths or "isophenes" drawn. A random selection of slopes between iso-pleths is taken, each for convenience multiplied by 10,000 and these again mapped and isopleths again drawn. At randomly selected points, the slopes for a l l the traits are summed ( i f weighted are then divided by the sum of the weighting factors) the sums are plotted on a map and isopleths again drawn. This i s the f i n a l map. Its topography shows the rate of tra i t change 110 geographically. Ridge lines mark areas of rapid change, basins areas of uniformity. Because Womble's excellent technique could not be used, I have attempted to simplify i t by mapping the distribution of a single characteristic. The most important of these are condy-lobasal length and canine width at base / bulla length, which are presented on Enclosure A (inside back cover) and Figure 11, for martens. Too few specimens of fisher were available to allow i t s use i n these animals. I consider these diagrams close to the most Important contributions this paper has to offer, and believe that they come closer to depicting a true picture of variation than any other method I have seen. The exception of course would be Womble's method, which i f further refined so as to remove errors of sampling could be a most worthwhile tool i n the hands of the systematist. MARTENS Distribution The distribution of marten in North America appears to be governed very closely by the distribution of the northern evergreen forest (Seton 1925-1928, Grinnell, Dixon and Linsdale 1937, Bailey 1936, Dalquest 194-8, Merriam 1886, Merriam 1882-84, de Vos 1952). Lensink (1953) concludes, for example, that the marten distribution i n Alaska i s coincident with the distribution of white and black spruce, "which i s apparently the c r i t i c a l element in their habitat". Edwards (1950) believed that those coniferous forests of highest market value supported the greatest I l l number of marten. Anthony (1917) believed, however, that " i t does not appear to be particularly attached to coniferous woods, living i n them simply because such forests prevail to a great extent i n the geographical areas i t inhabits". Ponomarev1s (1944) experiments give credence to this view, i n that he con-cluded that the Eurasian martens (M. martes, M. foina and M.  zibellina) had their distribution controlled chiefly by the presence or absence of low temperatures (20° C and less). None-theless, most authors name the humid upper transition, the Canadian and the Hudsonian l i f e zones as those occupied by the marten (Rhoads 1903, Cary 1917) Grinnell and Storer 1924, Skin-ner 1927? Williams 1930, Bailey 1936, Grinnell, Dixon and Lins-dale 1937> Hall 1946, Rust 1946, Dalquest 1948, Yeager, Denney and Hammit 1949, de Vos and Guenther 1952, Durrant 1952, Sumner and Dixon 1953). Marten have been also recorded outside the evergreen type forest. While de Vos (1952) and most writers consider them scarce or absent i n hardwoods, others report them common there. Emmons (1840) stated they were common in the beech woods of Massachusetts. Adams (I873) said they occur i n the hard-woods of New Brunswick. Rhoads (1903) concluded that the hard-wood timber of New Jersey and Pennsylvania was preferred to coniferous forests. Allen (1904) reported that they occurred in beech forests throughout New England. De Vos (1952) says that they may have occurred i n the northern hardwoods of Ontario before western settlement occurred. That this relationship exists not only in North America i s shown by Jurgenson (1939) 112 who states that Martes martes prefers a mixed deciduous-coniferous forest to any other, in Russia. In mountainous regions martens have been known to fre-quent talus slides (Turner 1886, Grinnell and Storer 1924, Clarke 1940, Hayward 1952 and Marshall 1951). They are also known to occur far from timbered areas on tundra and mountain meadows (Turner 1886, Cary 1911, and Clarke 1944). However, i t i s essentially true that the distribution of marten f i t s very closely the distribution of the northern coniferous forests as comparison of the distribution of martens and forest types obviously indicates. A distribution map of martens has been prepared and is included i n this paper (Enclosure B, inside end cover). This is based on the specimens examined and an extensive, but by no means complete survey of the literature. The limits of the animal's range are based chiefly on the vegetation maps of Zon and Sparhawk 1923, Munns 1938, Preble 1908, Anderson 1934a and b, Davis 1939, Pitelka 1941, Robertson 1945, Raup 1945 and 1946, Villeneuve 1946, Camsell 1947, Munro and Cowan 1947, Anon. 1948, Villeneuve 1948, Dayton 1949 (after Shantz and Zon 1924), Gutaell 1949, Hustich 1949a and b, Braun 1950, Hare 1950, Department of Mines and Technical Surveys of Canada Map, 1950, Edwards 1950, Macleod 1950, Bandy 1952, Webb 1952, Candy 1951, Hustich 1951, Macdonald 1951, Nicholson 1951, Whitaker and Ackerman 1951, Hare 1952, Halliday 1937 and 1952, Putnam 1952, Muesbeck and Krombein 1952. To these were added the material obtained from the marten distribution maps given by Seton 1909, Cary 1911, Cory 1912, 113 Grinnell and Dixon 1926, Seton 1925-1928, Anderson 1934a, Brouilette 1934, Bailey 1936, Grinnell, Dixon and Linsdale 1937> Melven 1938, Dufresne 1942, Hamilton 1943, Hall 1946, Twining and Hensley 1947, Dalquest 1948, Burt 1948, de Vos 1952, Durrant 1952, Webb 1952, and unpublished maps for Colorado and Montana kindly prepared by Mr. Richard Denney and Dr. Philip Wright respectively. The remaining sources from which these maps and the account following has been prepared are given i n Appendix D and the account following. It i s well known that the numbers of martens throughout the continent have steadily decreased in the past several hundred years (Henderson and Craig 1932, Seton 1925-1928, Allen 1942, Yeager 1950, Minville 1946, Innis 1927 > Anon. 1927-1950, Squires 1946, Butler 1950). Alaska: A l l of forested Alaska, as far west as Norton Sound and Kotzebell Sound (Nelson and True 1887) and nearly to the mouth of the Yukon River (Dice 1921) and as far north as the head of the Nunatog River (l a t . 68°) (Nelson and True 1887) and Anaktuvak Pass in the Brooks Range (Rausch 1950 and 195D 5 the Kenai Peninsula and the Panhandle; the following at least of the Alexander Archipeligo: Admiralty Island (Swarth 1911, Dufresne 1946, and specimens examined), Etolin Island (specimens examined), Kodiak Island (Nelson and True 1887)» Kuiu Island (Swarth 1911, Dufresne 1946, and specimens examined), Kupreanof Island (Dufresne 1942 and 1946 and specimens examined) and Revillegigedo Island (Dufresne 1946 and letter of Calvin Lensink). In 1934 marten were transplanted to Prince of Wales and 114 Chichagof Islands (Dufresne 1946). Marten were being trapped during the winter of 1948-49 (Yeager 1950). Localities not mapped include: Savioyok Valley, Brooks Range (Rausch 1 9 5 D , Bering Sea, Coast of Alaska ( E l l i o t 1905a) and Salcha (U.S. Fish and Wildlife Service F i l e ) . B ritish Columbia: Occur through a l l of the following biotic areas of Munro and Cowan 1947: Cariboo Parklands, Columbia Forest, Subalpine Forest, Boreal Forest, Peace River Parklands, Coast Forest, Queen Charlotte Island, Vancouver Island (many sources and specimens examined).. Are present on both Graham and Moresby Islands of the Queen Charlotte group (specimens examined). Macleod (1950) says they have been trapped on P i t t , McCauley, Hawksbury Islands and suggests that they occur on s t i l l others. Mr. Charles Guiget says they have been reported from Hunter Island. Marten were being trapped in the winter of 1948 - 49 (Yeager 1950). California: The forested parts of the Klamath Mountains and Coast Mountains as far south as Sonoma County (Grinnell 1933, Twining and Hensley 1947, Grinnell, Dixon and Linsdale 1937) and the Sierra Nevadas as far south as Tulare County (Grinnell, Dixon and Linsdale 1937). One specimen from the U. S. National Museum (no. 32033) from San Joseph Island in lower California is named a marten, but i t is almost certainly referable to Bassariscus  astutus• Marten were being trapped i n the winter of 1948-49 (Yeager 1950). Localities not mapped include Rush Creek and Bear Creek Kellogg 1916). 115 Colorado: Everywhere In the higher mountains (above 9»500 feet) (Williams 1947, Remington 1950, Yeager, Denney and Hammlt 1949) from the northern to the southern boundaries of the state and as far east as western Arapahoe County (Yeager 1950) and western Huerjane County (Denney, l e t t e r ) . Remington (1950) reports a specimen from Camp Hale which I have possibly mistak-enly located i n Yuma County. F. C. Kleinschnitz (letter) believes the marten to be found in the following counties beyond those mapped: Alamosa, Archuleta, Chaffee, Conefos, Custer, Delia, Delores, Fremont, Gilpin, Hinsdale, Huerjano, Lake, La Plata, Mesa, Moffat, Ouray, Pitkin, Rio Grande, Saguache, San Miguel, and Teller. Marten were being trapped in the winter of 1948-49 (Yeager 1950). Localities not mapped are East Spanish Peak and Bennett's Well (Cary 1911). Connecticut: Linsley (1842) l i s t s marten as occurring i n the state but Goodwin (1935) found no authentic record of i t s former presence. He believes that i t may once have occurred i n the mountains of the northwestern and possibly northeastern part of the state, however. Two specimens from the American Museum of Natural History (41340 and 41335) collected i n Greenwich are variously labelled "M. foina" and "M. americana" and to the second is appended the statement "this may be an introduced species". My examination of these leads me to conclude that while the f i r s t specimen may be representative of M. americana, the second i s not, and is referable instead to M. foina of Eurasia. 116 District of Columbia: Captain John Smith (1607-08) recorded "martens" from the d i s t r i c t (according to Mansueti 1950). McAtee (1918) said that "according to Wm. [sic] Palmer there is a f a i r l y certain record as late as about 1880". Bailey (1926) concludes that marten once occurred, as does Mansueti (1950). Idaho: The forested mountains of Idaho, as far south as Ada, Elmore, Blaine and Fremont Counties (Baird 1857 and 1859» Suckley and Gibbs i860 and specimens examined). Davis 1 (1939) vegetation map suggests that i t may occur i n Caribou, Bonneville and Bear Lake Counties. It was s t i l l being taken i n the winter of 1948-4-9 (Yeager 1950). A lo c a l i t y not mapped i s the Kanisku region (Rust 194-6). I l l i n o i s : Kennicott (1859 and 1855) records the former presence of marten in Cook County. Sanborn (1925) says they were once present about the Chicago region. Hahn (1909) and Cory (1912) say a skeleton of a marten was held by the Chicago Academy of Sciences and that i t was said to have been collected i n the northern part of the state many years before. Shorger (194-2) concludes that they once occurred i n the strip of pine timber along the shore of Lake Michigan. Mohr (194-3) says that recent reports (U.S. Forest Service 1937) have li s t e d marten as occur-ring in the state. Mohr says these reports are incorrect and that the animal has been extinct for a century or so. Indiana: Hahn (1909/)' considered the animal a "hypothetical intruder" i n the state. Lyon (1936) believed that i f i t 117 occurred i n I l l i n o i s (see above) i t "probably" occurred i n Indiana. Shorger (194-2) concluded that i t formerly penetrated into the strip of pine timber that ran along the shore of Lake Michigan. Iowa: Scott (1937) cites papers by Goding and Osborn as l i s t i n g marten for the state, but considers these doubtful. He con-cludes that that animal "may have entered northeastern Iowa as a straggler many years ago ...." Kansas: Cockrum (letter, citing Cockrum 1952) considers the former presence of marten as hypothetical. He writes that Mead (1899) wrote "martens were rare", and that on this basis Hibbard (1933) l i s t e d i t as a former member of the state's fauna. He concludes that "I think the marten has not occurred i n Kansas within historic time." Maine: Occurs throughout the mountainous northern part of the state. Seton (1925-1928) says they are "rarely seen now i n the southern part of the state." They are reported to have once occurred as far south as Muddy River (Topsham) about 1915? and from New Gloucester over 100 years ago and are known from the shell heaps on Goose Island (Norton 1930). Coues (I877), Allen (1876b) and Anderson and Sclater (1881 and 189D report them from as far south as central Oxford County. The last specimen in the state was taken i n 1936 (Yeager 1950). Manitoba: A l l of Manitoba originally, except perhaps the extreme southwest portion and the barrens of Hudson Bay. Now very rare 118 south of latitude 53N. (Melven 1938). The most southerly-records are Aweme (Criddle 1929) and Pembina Mountains (Seton 1925-1928). R. W. Sutton (letter) reports the taking of a marten i n the poplar bluffs about the Delta Research Station i n the winter of 1951-52. Marten were regularly being trapped during the winter of 1948-49 (Yeager 1950). The chief collect-ing points are Nelson House, Lac du Brochet, and Shamatawa (L. Butler l e t t e r ) . Maryland: Scharf 1882 stated that "marten occurred i n western Maryland". Mansueti (1950) says i t "has been completely extir-pated i n Maryland for at least 70 years.... Its distribution probably was not widespread...." J. W. Aldrich (letter) reports that there appear to be no definite records for the state. Massachusetts: Formerly present i n the Berkshire Mountains of Berkshire County (Emmons 1840, Allen I869, Coues 1877, Allen 1904). Hamilton (194-3) suggests i t may s t i l l occur i n the area. Michigan: Formerly a l l of the timbered areas of the state as far south as Allegan County, and on Isle Royale. None has been taken in twenty-five years and i s now rare or absent i n the state (Burt 1948). Pruitt (1951) records their presence on Sugar Island. Minnesota: Northern Wisconsin, as far south as Polk County (specimen examined) and Crow Wing County (Swanson, Surber and Roberts 1945). Originally common, rare by 1900, and probably now extinct, although possibly s t i l l existing in Superior and 119 Chippewa National Forests (Shorger 194-2, Swanson, Surber and Roberts 1945, Gunderson and Beer 1953? Stenlund 1955). The last specimen was taken in 1953 i n St. Louis County (Stenlund 1955). De Vos says they are s t i l l present in the northern part of the state. Montanas The northern regions of the Montana and Bitterroot sections of the northern Rocky Mountains, as far east as Western Cascade County (Wright 1953) and as far south as Ravalli County (specimens examined) also present i n Madison County (TJ. S. Fish and Wildlife Service F i l e ) , Sweetgrass, Stilwater, Carbon and eastern Park Counties (specimens examined). Marten are s t i l l common i n the state (Yeager 1950, Wright letter) Nebraska: Swenk (1908) cites an undated publication of Aughey who reported that they have been taken i n the northwestern part of the state, but that they were rare. Swenk considered them extinct at time of writing. Nevada: Reported to occur in the southwest portion where the Sierra Nevadas enter the state, from Mount Rose south to Monument Peak. One specimen i s known from Marlette Lake (Hall 1946). New Brunswick: Formerly common throughout a l l of Nova Scotia (Chamberlain 1884 and 1892). At present greatly reduced and restricted to the more remote areas (Allen 1942, Morris 1948). No specimens have been taken since 1945-46 (Yeager 1950). Speci-mens not mapped include Lake Edward (Chamberlain 1884 and 1892), between Lake Edward and Salmon Lake ( E l l i o t 1901a and b), and Salmon Lake (Chamberlain 1884 and 1892). 120 Newfoundland (excluding Labrador): Originally common i n a l l the wooded areas (Reeks 1870 and 1871), but now very scarce (Bangs 1913, Allen 1942). A. Cameron (letter) says the animal is "quite rare" and confined to the more inaccessible regions. D. H. Pimlott (letter) reports that marten have, for the past two years, been reported from the lower Grand Lake and Gambo Lake areas only. New Hampshire: Found originally i n the White Mountains and northward (Allen 1904). Very nearly extirpated now, i f not completely so (Preble 1942-43). Norman Preble (letter) has sent me the following information: one specimen i s known from the state, from Coos County i n the New England Museum of Natural History (Boston Museum of Science). Common 100 years ago i n the Crawford Notch region of the White Mountains, but now rare, the last specimen being taken i n 1936. ^Alfred E. Preble observed an animal on Mount Washington i n 1930 that was l i k e l y either a marten, a fisher, or a tame Bassariscus. Another marten was reported at the same place i n the spring of 1944. A l l records from the state are i n Coos County, except one from Laconia i n Belknap County taken between 1931 and 1934 (Helenette Silver, letter, after Jackson 1922). According to Monahan (1953) have been recently reintroduced into the northern part of the state. New Jersey: "Once abundant in the mountain regions. Now wholly absent from the state. Probably exterminated f i f t y years ago" (Rhoads 1903). I know of no other record. 121 New Mexico: Present, but by no means common, i n the high mountains of the northern part of the state, especially the San Juan and Sagre de Cristo Ranges (Bailey 1931). Records known are from Chama (specimen examined), Taos (Wheeler 1875)» Truchas Peaks (Bailey 1931)» Las Vegas Mountains (Durrant 1952). The last l o c a l i t y named i s the most southerly record for the animal i n North America. An unmapped loc a l i t y for which mar-ten i s recorded is Twining (Bailey 1931). Allen (1942) believed they s t i l l existed i n the state, at time of writing; whether or not they are s t i l l present I do not know. New York: Formerly common in the Adirondacks and possibly the Catskills u n t i l 1890 (De Kay 1842, Merriam 1882-84 and 1886, Mearns 1898, Miller 1899 and 1900) but trapped almost to extinc-tion since that time (Seton 1925-1928). The most southerly records are those from the Catskills (Audubon and Bachman 1851-1854, Mearns 1898). An unmapped lo c a l i t y i s Averyville (Harper 1929). The last specimen taken was i n 1938 (Yeager 1950). North Dakota: Common a century ago i n the wooded northeastern part of the state, but have been long extinct (Bailey 1926a, Allen 1942). Known records are from Pembina County (Bailey 1926a), Walsh County (Bailey 1926a, Swanson, Surber and Roberts 1945) and Grand Forks County (Bailey 1926a). An unmapped record i s the Hair H i l l s (Bailey 1926a). Northwest Territories: A l l of the Territories south of the barrens (Ross 1861a and b and 1862, MacFarlane 1905 > Mair and Macfarlane 1908, Preble 1902 and 1908). Banfield (195D 122 records one taken seventy-five miles west of treeline near Contwoyto Lake, and Dergbol and Freuchten (1935) record the taking of one just south of Chesterfield Inlet far from the forest edge. Clarke (194-4) likewise has observed that marten have been taken far from timberline near Tuktoyoktuk. The greatest numbers of martens were taken during the period 1840 to i860, according to Seton (1925-1928), but since then the numbers available have steadily decreased (many authors). The annual catch in the years 1851-60 exceeded 30,000 pelts, but by 1930 the average had dropped to about 5,000 pelts. The chief collecting places are now Forts Simpson, Good Hope, Nor-man, Aklavik, Rae, Smith, Resolution and Providence i n that order of importance (Robinson and Robinson 1946). An unmapped record i s Fort Anderson (Preble 1908). Nova Scotia: Originally probably a l l of Nova Scotia (Smith 1940) and according to Coues (I877) and Bailey (I896) once very common. Nov/- very rare, and found only in restricted areas (Smith 1940). No specimens have been legally collected for some time, but Rand (1933) had heard of one smuggled out shortly before. Are, or were, present on Cape Breton Island (Rowan I876, Rand 1944b). Ohio: Kirtland (1838) reported that "the pine weasel i s admitted to the state's fauna on the authority of Dr. Ward, who informs me that i t was taken i n the v i c i n i t y of Chillecothe". Brayton (1882) quoting a letter of one Emory Potter, l i s t e d i t as "extinct i n Ohio". Bole and Moulthrop (1942) have examined 123 two specimens believed collected from Ashtabula County. They conclude that i t has been extinct f or about one hundred years. The Chillecothe record, i f correct, marks the most southerly marten records known f o r the area east of the Great P l a i n s . Ontario: Apparently once common through a l l of the province except f o r the barrens just south of Hudson Bay. I t now occurs no further south than the southern l i m i t s of Algonquin Park (Cross and Dymond 1929» Downing 1948, de Vos 1952) and i s made up of many small disjunct populations, the largest centering around the area fif£y to seventy-five miles west of Timmins (de Vos 1952). An unmapped l o c a l i t y i s Quinte Forest D i s t r i c t (de Vos 1952). Apparently at one time i t occurred a l l through southern Ontario, at least as f a r south as Essex County and Rondeau, as the following evidence indicates. Small and Le t t , 1884, (Trans. Ottawa F i e l d Nat. Club 6.2 : 280-283, and 150-151) and Rand (1945a) say i t occurred near Ottawa about 1840 and Seton (1925-1928) says i t was found between Lake Simcoe and Ottawa i n the 1870's. I t occurred at Whitechurch, ten miles north of Toronto (Seton 1925*1928) and near Toronto i n 1830 and "much l a t e r " ( F a u l l 1913). They were considered to have once occurred near Hamilton by Warren (1950) i n the upper Thames watershed (Richardson 1952) and Saunders (1932) was t o l d i n 1899 that they had been seen at Rondeau. De Vos (1952) says that Volume 3 of the Census of Canada, 1870-71, records f u r returns from the following r i d i n g s : Essex, Kent, Brent, Simcoe, V i c t o r i a and Peterborough. Marten skeletons have been found i n the l a t e pre-European middens of 124 Prescott, Brantford and London (Wittemburg 1919? 1939 and 1948). Marten refuges have recently been formed on the east coast of James Bay and on Akimiski Island (de Vos 1952). I do not know whether marten have ever occurred naturally on the island. Oregon: Throughout the Coast Mountains and Cascade Mountains from the northern to the southern boundaries of the state, as far east as western Crook County (Bailey 1936) and DesChutes River (eastern Wasco County) (specimen examined). Also present in the Blue Mountains of the north eastern part of the state, as far west as Grant County (Bailey 1936, Anon. 1951 and 1952) and as far south as Prairie City (Grant County) (specimen exam-ined). Marten were being trapped i n 1952 (Anon. 1952). A loc a l i t y unmapped i s Olive Lake (Bailey 1936). Pennsylvania: Once abundant in the northern mountains of the state, but were exterminated by 1900 (Rhoads 1903, Richmond and Roslund 1949). Its southern limits in the state appear to have been Crawford, Forest, Elk, Cameron, Clinton, Tioga, Sullivan, Columbia, Wyoming and Wayne Counties (Rhoads 1903? Shoemaker 1919). Rhoads (1903) also says they once occurred in Lancaster County; i f this is so i t i s , as Rhoads((I896) points out, the most southerly known record for marten i n the Appalachian Moun-tains, aside from an Ohio record (see above). Quebec, including Labrador: Commonly throughout a l l of the two, Quebec and Labrador, north to the tree line (Bell 1884, Low 1895, 1896 and 1897? Bangs 1898, Anderson 1931-32, Strong 1930, and Tanner 1944). Inland i t was abundant everywhere 125 (Stearns 1883) but north of Misstassini i t occurs chiefly i n the forested river valleys only (Low 1 8 9 5 and I 8 9 6 , Strong 1 9 3 0 ) . The numbers trapped per year have dropped from 1 2 , 0 0 0 i n 1 8 7 0 (Minville 194-6) to 1 2 0 0 i n 1 9 4 9-50 (Anon 1 9 2 7 - 5 0 ) . Originally i t occurred as far north as Chimo, as far west as George and as far east as the barrens edging the Labrador Coast (specimens examined). There are no records for i t s former occurrence i n the eastern townships though Hall ( 1 8 6 1 ) recorded i t from near Montreal, Cameron and Orkin ( 1 9 5 0 ) from Laurentides Park just north of Montreal and i t i s known to be present although scarce on the Gaspe (Cameron 1 9 5 3 and specimens examined). V e r r i l l ( 1 8 6 2 ) and Rowan ( I 8 7 6 ) reported that marten occurred on Anti-costi Island. Hunter ( 1 9 0 7 ) said he had examined specimens of these, but Newsom ( 1 9 3 7 ) said he could fihd no one on the island to confirm the story, though he did not doubt their reports that at one time they had occurred there. D. H. Pimlott (letter) writes that Mr. C. McCormick, the chief Game Warden on the island reports that the last specimen taken was trapped in 1 9 2 6 , and that tracks have not been seen since 1 9 3 1 . Saskatchewan: The northern wooded portion of the province, probably as far south as the southern limit of Aspen Parkland, and at least as far south as the southern limit of the coniferous forest. Records for the province are scarce, the most south-erly being l i e a l a Crosse in the west and the Pas Mountains (Butler, letter) and Redearth in the east (specimens examined), although I have examined a specimen from Duck Mountain i n 126 Manitoba, just across the border from Saskatchewan, about 150 miles south of the preceding l o c a l i t y . Twenty-five hundred skins were taken in 1919-20, but this has been reduced to 375 i n the winter 1949-50 (Anon. 1927-50). The chief collecting points are Fond du Lac, Isle a l a Crosse, and Pelican Narrows (L. Butler, l e t t e r ) . South Dakota: "A specimen ... was taken i n the Black H i l l s near Custer in January, 1930, which i s the only authentic record for the state. However, there i s every reason to assume that before the white trapper entered the region the pine marten roamed in the wooded area up the eastern side of the state and in the Black H i l l s as a straggler". (Over and Churchill 1945). A record of a marten taken from Pringle i n the Black H i l l s i s held in the f i l e s of the U. S. Fish and Wildlife Service. Utah: Occurs through the Wasatch and Uinta Ranges of the middle Rocky Mountains as far west as Salt Lake and Utah Coun-ties and as far south as Lost Lake, Wasatch County, i f not far-ther. They occur also in the Colorado Plateau region of the southeastern part of the state, i n San Juan and Grand Counties (Durrant 1952 and specimens examined). Small numbers of marten were collected in 1949-50 (Yeager 1950). Vermont: Originally common throughout the mountainous portions of the state (Kirk 1916, Osgood 1938). Reported scarce by 1840 (Seton 1925-1928) and nearly extinct at present (Osgood 1938, Foote 1944). Osgood (1938) says the last specimen taken was i n 127 1926. Records are known from Rutland, Bennington, and Windham counties only (Kirk 1916, Osgood 1938). Virginia: Bailey (194-6) states that "the account showing the quantity of skins and furs exported annually ... from Virginia from ... 1698 to ... 1715" l i s t s marten as being taken, and that Thomas Jefferson in his "Notes on the State of Virginia" (1801) likewise included i t . Audubon and Bachman (1851-54) report that "we have sought for i t in vain on the mountains of Virginia, where notwithstanding, we think a straggler w i l l occasionally make i t s appearance." Anthony (1928), Bailey (1946), and Brown (1952) a l l l i s t i t as a l i k e l y former inhabitant. Handley and Patton (1947) review the situation and reach the following con-clusion: "although i t s range has been given by dozens of auth-ors, even to the present day as 'extending south in the mountains to Virginia', we have been unable to find record of any specific evidence of i t s occurrence south of Pennsylvania. It i s not i n -conceivable, however, that i t might have occurred in the ... spruce forests that formerly covered eastern West Virginia and extended into the Virginia Mountains...." Washington: Occurs i n the Olympic Peninsula and south i n the Coast Mountains at least to Chehalis (specimens examined), the Cascade Mountains from the international boundary to the Columbia River, west to Camp Skagit, Mount Vernon, and Hamilton (specimens examined) and east to eastern Chelan County and Chelan (Dalquest 1948 and specimens examined). Present i n the Columbia, Bitterroot 128 and Selkirk Ranges of the northeastern part of the state (Dal-quest 194-8, Anon. 1931 and specimens examined) and the Blue Mountains of the southeastern corner (Dalquest 1948). The range i s now more restricted according to Dalquest. The last specimens were collected i n 1946-47, the season since being closed (Dalquest 1948). Wisconsin: Formerly i n the wooded portions of the northern part of the state, at least as far south as St. Croix, La Crosse, Jackson, Juneau and Brown Counties (Shorger 1942). Nearly ex-tinct by 1900 (Jackson 1908, Cory 1912, Barger 1951). Shorger (194-2) says the last specimen taken was in 1925 and believed by Barger (1951) to be completely gone from the state, but reported by de Vos (1951) as s t i l l present. Stated by Shorger (194-2) as present in 1934- on Outer Island of the Apostle Island group in Lake Superior. Wyoming: Reasonably common i n the forested areas of the Wasatch and Yellowstone sections of the Middle Rocky Mountains of the western part of the state (Cary 1917? Seton 1925-1928, Bailey 1930, Calhalane 1943, Anon. 1950, Thomas 1952b). Occurs from Yellowstone Park south to La Barge Creek, Lincoln County (speci-men examined) and as far east as western Park County (U.S.Fish and Wildlife Service card f i l e ) and Dinwoody Canyon, western Fremont County (specimen examined). No specimens were trapped in the year 1948-49 (Yeager 1950). 129 Yukon Territory: Present through a l l of the forested part of the territory, at least as far north as the Macmillan River (Mason 1924, Rand 1945d). The present status of marten i s good, and promises to remain so (Rand 1944c) although the present total take is not equal to that of the catches of some single trappers thirty-five to forty years ago (Rand 1945b). Specimens were s t i l l being taken in 1949-50 (Anon. 1927-50). A l o c a l i t y record not mapped i s the lower Yukon region (Twitchell 1921). Synoptic examination Having determined that North American marten f a l l into two distinctive morphological types (americana and caurina), whether they be of "species" or subspecies grade, and knowing that a l l New World marten are referable to these two basic groups, i t becomes important to consider the many other named groups that remain and to examine their v a l i d i t y . In 1942 Allen correctly recognized that marten f e l l into the two groups named above and considered them to be distinct "species", i.e. "Martes americana" and "Martes caurina". Wright, as pointed out elsewhere, in 1953 showed that these two supposed "species" actually intergrade and hence must be considered sub-species of M. americana. The taxonomic d i f f i c u l t i e s resulting from this conclusion are discussed elsewhere, but for the time being (insofar as this stage of the paper i s concerned), we here u t i l i z e Allen's classification, believing that i t is unsatisfac-tory, but recognizing that i t is the more convenient of the two. Reference i s made throughout this discussion to Enclosure A (inside back cover) showing the geographic 130 distribution of condylobasal length i n male marten. This was made by obtaining the mean condylobasal length i n males for each local subsample, mapping them, and joining points of equality by lines (isopleths). While more than one cranial character might have been used to advantage, lack of time made such an approach impossible, and i s perhaps not as v i t a l as at f i r s t i t might appear to be, since most geographic variation within any "species" (caurina, americana) i s reflected i n size (i.e. condylobasal length). A l l other characters (including external measurements) closely follow the pattern shown by condylobasal length. Detailed synonymies are provided i n Appendix B. Only the name combinations used are given below; for author, date and citation the reader must refer to the appendix. "Martes americana" (Kerr 1792) Detailed characteristics given elsewhere, but easily distinguished by long, narrow, high skull, small upper molar and long highly inflated bullae. Width of inner lobe of P 4 x 100 / length of P 4 always less than 25.8; breadth of M1 x 100 / length of P 4 always less than 97.15; lateral length of M1 x 100 / mesial length of M always more than 77; breadth of P (across inner lobe) x 100 / length of P 4 always less than 64; canine width at base / bullae length always less than 94. "Martes americana americana" (Kerr 1792) Synonymy: Martes (or Mustela) martes, zibellina, americanus, americana, vulpina, leucopus, huro, leucotis, sinuensis, zibel- l i n a americana, zibellina var. americana, americana var. 131 leucopus, martinus, or vison. Most writers attribute the epithet americana to Turton 1802 or 1806. However, Kerr i n 1792, was the f i r s t to employ the term and authority for the name must now be attributed to him. Type specimen: None that I know of. Type l o c a l i t y : Variously designated as "the North of ... America" and "North America, especially in Canada" (Kerr 1792); "North America" (Turton 1802 and 1806); "Country ... watered by the Missouri ..." (Rafinesque 1819a and b); "In Canada" (Kuhl 1820); "the habitat of the Huron Indians, fide Milbert, v i z . , region east and south of Georgian Bay, west peninsula of Ontario" (Cuvier 1823); "Eastern North America" (Miller 1900); "Upper Missouri?" ( E l l i o t 1901a after Rafinesque 1819a and b); "The region occupied by the small pale martens of southern Canada and the northern United States"(Rhoads 1902); "Probably came from Ontario or Quebec..." (Rhoads 1902). Although Cuvier's type l o c a l i t y i s probably the most clearly defined, i t i s now l i k e l y without martens, and most auth-ors since the turn of the last century have used Miller's (1900) designation "Eastern North America". Because this could be inter-preted to include the "races" "atrata", "brumalis" and "abieticola", i t is l i k e l y best to restrict i t to Ontario or the United States east of the Mississippi. Dr. E. R. Hall has kindly written to point out that Rafinesque's (1819a and b) Mustela vulpina, since i t was taken from the "Upper Missouri River = Idaho or southwest Montana" should be considered a synonym of "M. caurina" which occurs there and 132 indeed, should actually replace the name "caurina" for the "species". Dr. Hall mistakenly stated that Rafinesque's speci-men was taken from the "Upper Missouri". Rafinesque said only that i t was taken from the "country ... watered by the Missouri ...", the change to Hall's l o c a l i t y being made by E l l i o t (1901a). While i t is possible that M. vulpina was taken from Idaho or Montana, i t could also have been taken in the Dakotas or Iowa, and since both E l l i o t and Rhoads (1902) considered i t synonymous with "M. americana" ten years after "M. caurina" had been des-cribed, I am obliged to make the same interpretation. It i s notable, however, that Gray 1865 and 1869 considered M. vulpina to be synonymous with WM. a. abietinoides". Rafinesque's specimen was a semi-albino, so that the description of pelage cannot be safely used in identifying the specimen. He does say that the total length of the specimen was twenty-seven inches, the t a i l length nine inches. M. vulpina was thus a large marten, much nearer to "M. caurina" than to "M. americana americana", although perhaps near to "M. a. abieti- cola" so that the problem can not be said to have been closed. As Dr. Hall says in his let t e r , i f i t can be shown to have been a specimen of "M. caurina", the name caurina w i l l need to be Changed to vulpina. Pelage: The coat i s usually said to be yellowish brown, legs, feet and t a i l tipped with black and with a whitish yellowish throat patch (Emmons 1840, Miller 1900, E l l i o t 1901a, Rhoads 1902 and 1903, Cory 1912). De Vos (1952) described i t as with "head usually lighter i n colour than the body and normally 133 light gray brown, the inner sides of the ears are dull white. The dorsal parts are normally dark brown, but may range from light pale yellow to nearly black. The limbs and t a i l are normally slightly darker than the rest of the body. The ven-t r a l parts and flanks are of a lighter colour than the dorsal parts. The throat and breast patches, which may be of different sizes and shapes, and the median ventral stripe are generally yellowish-orange, but may vary from pure white to brownish-yellow or may be entirely absent." Elsewhere he says that "the colour of the fur ... i s extremely variable", ranging from very pale yellow-orange, with an entirely white head and no throat patch to very dark, with a large number of white guard hairs, with a large orange throat patch. Rhoads (1902) stated that this "subspecies" was distinc-tive from others by the paleness and yellowness of the outer fur, the clear ash of the basal under fur, and the strongly contrast-ing colours of the foreparts compared to the hinder parts. Almost a l l authors have remarked on the small size of this "race" compared to a l l others. The pelage descriptions given above are in accord with my own findings. It is generally, as Rhoads pointed out, the palest, yellowest and most thinly furred of a l l the "races" of "M. americana", although as de Vos points out, variation can be extreme. It i s likewise small, being, aside from "M. a. abieti-noides" in the Canadian Rockies, the smallest marten of the "species". 134 Geographic variation: Marten from Algonquin Park, and from extreme northwestern Ontario are, according to de Vos (1952), paler than the average Ontario skin. Those from eastern Ontario are said by Bachrach (1946) to have very short guard hairs and underfur. Marten from the north shore of Lake Superior are, according to Richardson (1829) and Rhoads (1902) very near to black i n colour, and according to Allen (1876a) are slightly larger than those from New York and Maine. Adams (1873) described this "subspecies" as i t occurs i n Nova Scotia i n some detail. He said that two types occur, a light coloured one (orange and saffron) from the hardwood forests, and a dark coloured one from the pine forests. Any large series of skins, he said, showed these two types, and intergradation between the two occurred. Rowan (1876) said that marten from the north shore of the St. Lawrence were darker than those from Nova Scotia (grading towards "M. a. brumalis"?) while those south of the St. Lawrence, i n Quebec, were paler, yellower and more thinly furred. Those from the Gaspe, however, were much darker and richer than those from New Brunswick. Rhoads (1902) reported that specimens from about Lake Superior and New Brunswick tended to largeness and darkness. Poland (1892) said that the poorest and palest skins came from the northeastern part of the United States. Rhoads (1902) came to the same conclusion. The martens from the Adirondacks of New York and from Pennsylvania are especially pale according to Rhoads (1903) and Grant (1906). The marten on Anticosti Island were described by Hunter 135 i n 1907 as being peculiar and distinct i n having in almost a l l cases the forepaws and the end of the t a i l tipped with white. They were otherwise as dark and well furred as those from Labra-dor ("M. a. brumalis"). Newsom (1937) was unable to verify this report. I am unable to add much to this discussion of pelage variation, except that I have also found that specimens from the United States are on the average considerably paler i n colour than those from Ontario and that regional populations appear to vary somewhat i n overall hue, though this i s only to be deter-mined from large samples. Geographic variation in condylobasal length i n males, as mapped on Enclosure A, i s of interest. The average condylo-basal length of the total sample of this race was 7»87 cm. (n = 118), but specimens from the northern part of Ontario aver-aged 8.10 cm. or more, while those from the area about Algonquin Park averaged 8.20 cm. and those from the Adirondacks of New York and the central Quebec-Ontario boundary less than 7.60 cm. These differences are significant at or very near the 95% l e v e l , and indicate that not only does pelage vary geographically within the "race" but that skull length and presumably total length does also. I am unable to account for the peculiar "islands" or largeness and smallness within the range of this "subspecies". De Vos (1951 and 1952) has mapped the distribution of marten i n Ontario and this map shows the Algonquin Park and Ontario-Quebec boundary populations to be disjunct and quite isolated from other nearby populations. Hamilton (194-3) maps the marten of the 136 Adirondack Mountains as being equally isolated. It i s possible, I suppose, that the differences marking the marten of these "islands" is a result of the so-called Sewell-Wright effect on small isolated populations. If this is so, i t means that they have l i k e l y occurred since these populations have become dis-junct, a period of time probably not exceeding a century and in some cases probably considerably less. Cranium: Since this is the standard against which other "sub-species" of the "species" are to be compared, l i t t l e needs to be said except that i t i s marked by a l l the characteristics of the "species" and that the skull i s exceedingly small, being equalled i n this respect only by "M. a. abietinoides", compari-son with which i s made under that "subspecies". Two "subspecies" have ranges contiguous to the range of this "subspecies", namely "M. a. brumalis" to the northeast in Quebec and Labrador and "M. a. abieticola" to the west, i n Manitoba. The samples of crania of "M. a. americana" are dis-tinct from these to the 95$ level as shown by a l l of the fourteen characters studied except the following: upper molar width, upper molar inner moiety length, narrowest width of palate, canine width at base / bullae length, and upper molar width plus inner moiety length, females only. These l i k e l y f a i l to show distinction because of the small sample size (n = 3) of female "M. a. abieticola". 137 Males Females X SD cv n X SD CV n Condylobasal length 7.86 .23 2.94 118 7.08 .24 .34 81 Rostral width 1.46 .07 5.18 120 1.28 .07 5.37 82 Upper tooth row 2.87 .10 3.61 69 2.55 .13 5.08 45 Canine width at base 1.40 .06 4.62 120 1.24 .06 5.00 82 Bulla length 1.64 .06 3.71 119 1.54 .06 3.70 82 Upper molar width .73 .08 00.53 120 .65 .03 5.38 82 Upper molar inner moiety length Palate length .45 3.92 .03 .15 7 . 3 i 3.94 119 66 .38 3.47 .02 .15 6.51 4.21 82 44 Mastoid width 3.44 .13 3.92 66 3.10 .11 3.70 44 Height of skull at hullae 2.94 .11 3.81 66 2.53 .14 5.44 44 Length of last upper PM .77 .03 4.54 118 .68 .03 4.37 80 Narrowest palate width .76 .06 7.27 110 .70 .04 6.30 68 Canine width base / bullae length : .85 .03 3.76 118 .80 .03 4.35 82 Upper molar width plus inner moiety length 1.19 .06 5.36 118 1.04 .06 5.75 82 Total length 60.00 1.96 80 53.80 2.74 33 T a i l length 17.50 1.02 80 15.60 1.15 33 Hindfoot 8.90 .31 80 7.70; .66 33 Distribution: Hamilton (1943) considered a l l marten east of the Mississippi River in the United States to belong to this '•subspecies". Cory (1912) mapped i t s range and showed i t occur-ring everywhere within the "species" range south of a line 138 connecting southern James Bay, Lake Mistassini and Godhout, and east of a line in western Ontario connecting (approximately) Bemidjii (Minnesota), Sioux Lookout and central James Bay (Ontar-i o ) . Further detail concerning these marginal limits are given subsequently. This "subspecies" has been reported to occur i n the following provinces and states, not a l l , however, being based on examination of specimens: Ontario (Downing 194-8, Cross and Dymond 1929, Anderson 194-6); Southern Quebec, Gaspe and Eastern Townships (Anderson 1934-b, Cameron 1953); New Brunswick (Morris 1948); Nova Scotia (Smith 1940); North and South Dakota (Bailey 1926a, Over and Churchill 1945); Minnesota (Swanson, Surber and Roberts 1945); Iowa (Scott 1937); Wisconsin (Barger 1 9 5 D ; I l l i n o i s (Cory 1912); Michigan (Burt 1948); Ohio (Bole and Moulthrop 1942); Pennsylvania (Williams 1930, Grimm and White-bread 1952); New Hampshire (Carpenter and Siegler 1945); Vermont (Osgood 1938); Connecticut (Goodwin 1935); Maryland (Mansueti 1950); Virginia (Handley and Patton 1947). This "subspecies" i s presumed to be the form present on Prince Edward Island and Cape Breton Island. The marten of Anticosti Island are described by Hunter (1907) as possessing a rich dark pelage that suggests that i t l i e s nearer to "M. a.  atrata" or "M. a. brumalis". The vegetation of Anticosti is reported by Halliday (1937) and Hustich (1949b) to be most closely related to that of the Mingan area directly to the north, so that I make the assumption that the nearest relative of the Anticosti marten may be "M. a. brumalis", the Labrador marten. 139 My examination of specimens and of the literature leads me to believe that the marten of the following provinces and states can be properly referred to this "race": central, eastern and southern Ontario, southern Quebec, Eastern Townships and Gaspe, New Brunswick, Nova Scotia, Prince Edward and Cape Breton Islands, and doubtfully Anticosti Island, eastern and central Minnesota, Wisconsin, Iowa, I l l i n o i s , Michigan (including Isle Royale), Indiana, Ohio, Pennsylvania, New York, Vermont, New Hampshire, Maine, Massachussetts, Connecticut, and other adjacent states where they may have once occurred. Specimens from western Minnesota, western Ontario and North and South Dakota may be referable to this "race", or may be more nearly allied to "M. a. abieticola" as discussed below. The marten from the Black H i l l s of South Dakota, a specimen of which was recorded taken by Over and Churchill (194-5) and which i s liste d i n the f i l e s of the U. S. Fish and Wildlife Service is considered by the above named authors to be representative of "M. a. americana" but may really represent "abieticola" or as l i k e l y "M. caurina" which occurs only 175 miles to the southwest of the Black H i l l s . To the northeast and into northern Quebec and Labrador, "M. a. americana" grades into "M. a. brumalis", as the isopleth map clearly shows. Just where the range of "americana" ends and where "brumalis" begins is a matter of opinion. The l o c a l i -ties following have provided specimens that have been, or are, referable to "M. a. brumalis": George River (Rousseau 194-9), Chimo (Anderson 1946), Matemak (Eidmann 1935)? L'anse au Loup (Bangs 1910), Fort Gjeorge, Nichuchin, Chimo and Northwest River 140 (Low I896). On the other hand the following l o c a l i t i e s have been referred i n the literature to "M. a. americana"1 Lake Edward (Bangs 1898), Lake Mistassini (Anderson 1946), James Bay region (Anderson 1934-), Straits of Belle Isle (Bhoads 1902} Ander-son 1946), James Bay region (Anderson 1934-), Straits of Belle Isle (Rhoads 1902, Anderson 1934-b, 1938 and 1940). Low (1894 and 1896) was the f i r s t to suggest that the two "races11 underwent gradual intergradation into each other. Bangs (1898) did not beli eve that such intergradation occurred, but Anderson (1934-b, 1938 and 1940) expressed belief that while the line of inter-gradation could not be accurately drawn, i t did exist. Cory (1912) mapped the southern limit of "brumalis" as extending from southern James Bay, through northern Lake Mistassini, to about Clarke City, and showed an area about f i f t y miles wide to the south of this in which i t intergraded with "americana". Anderson names the area of intergradation to be in the region north of Lake Mistassini. I have examined most of the specimens seen by Bangs and Anderson, and many of those examined by Low, and conclude with Anderson that a cline exists between the small pale "race" to the south ("americana") and the larger, darker "race" to the north (•'brumalis"). Examination of the isopleth map for condylobasal length indicates this gradation well. The area of most rapid rate of change, as exemplified by this map occurs from southern James Bay to the north shore of the Saguenay River (where the average condylobasal length increases by 2 mm. in a belt not much more than 50 or 75 miles wide). This area might be said to mark 141 a good point of division between the two "subspecies". However, Anderson considers the skull and pelage of the Mistassini speci-mens to l i e closer to "M. a. americana" than to "M. a. brumalis". Further, since the mean condylobasal length of the sample of "M. a. americana" males i s 7*86 cm. (n - 118, SD = . 2 3 ) , while that of "M. a. brumalis" i s 8.56 cm. (n = 55 j SD = . 18) , we might set the point of division midway between the two means, i.e. set the point of division at that area characterized by martens with an average condylobasal length of 8.21 cm. Referring again to the isopleth map, this isopleth is found to run from the region about the base of Hudson Bay, through the northern end of Lake Mistassini, to about Betsiamites on the St. Lawrence River. This line perhaps f i t s Anderson's interpretation better than does the previously suggested one. Actually of course, as discussed elsewhere, the problem cannot be solved. What is significant i s that "Martes a. americana" of Ontario, as one moves north into Quebec becomes gradually darker and larger until i t s maximum of largeness and darkness i s attained along the Labrador coast. Any boundary line that one may set up to mark the geographic limits of the two "races" w i l l have to be wholly arbitrary, and not particularly significant. The western boundary of "M. a. americanal1, where i t intergrades with WM. a. abieticola" is likewise d i f f i c u l t to state. It has been described variously as "the western shores of Hudson Bay" (Anthony 1928), "Ontario, west of Hudson Bay" (Cross and Dymond 1929), eastern limit of the western Hudson Bay drainage system (Anderson 1937), or "extreme western Ontario or 142 eastern Manitoba" (Anderson 1946). Cory (1912) maps the eastern limit of "abieticola" as passing from about Moorhead, Minnesota, through Lake of the Woods, Sioux Lookout to the middle reaches of the Winisk River, Ontario. He shows i t intergrading with "americana" i n a belt 75 to 100 miles wide to the east of this. Of the specimens I have examined, the most westerly i n Ontario (Goldpines and Sioux Lookout) are referable to "M. a. americana", while those most easterly in Manitoba (York Factory, Oxford House, Red River) are referable to "M. a. abieticola", though averaging slightly smaller than typical. Reference to the isopleth map shows that "abieticola" begins to grade into "americana" on about a north south line bisecting Manitoba, reaching the dimensions of "M. a. americana" somewhere west of Goldpines and Sioux Lookout. One specimen from Grand Forks, Minnesota, appears to l i e intermediate between the two "races" as are skin measurements given by Gunderson and Beer (1953) lor Minnesota marten. I would suggest that the Ontario-Manitoba boundary line marks as well as any other line the approximate area of greatest change between the two "races". The midpoint between the means of condylobasal length of the males of the two "races", 8.13 cm., i s , however, as I have mapped i t , somewhat to the east of this line. Specimens examined: 413 specimens, from Ontario, Quebec, New Brunswick, Minnesota, Michigan, New York, and Connecticut. Skull diagrams and photographs: De Kay 1842, vol. 1, p i . 11 and 19; Boddaert 1785, p i . 36; Baird 1857 and 1859, p i . 36 and 37; 143 Coues 1877, p i . 5; Osgood 1900, p i . 7; E l l i o t 1901a, f i g . 66; Cory 1912, p. 382; Pratt 1923, f i g . 126. "Martes americana brumalis" (Bangs I898) Synonymy: .Martes (or Mustela) brumalis, atrata brumalis, caurina. brumalis, americana brumalis. Type specimen: Museum of Comparative Zoology, Cambridge, Massachusetts, Bang's Collection no. 7417 (Allen 1931» Anderson 1946). Type lo c a l i t y : "Okkak, Labrador" (Bangs I898); !*:0kak, Labrador" (Miller 1900); "Okak, Labrador, Canada, now Newfoundland" (Ander-son 1938). Pelage: Long recognized as distinctively large and dark coloured, when compared to "M. a. americana" to the south (Gilpin i860, Rowan 1876, Poland 1892, Low 1895 and I896, Bangs I898, Rhoads 1902, Hunter 1907, Hardy 1910, Cory 1912, Anderson 1934b, 1938, 1940 and 1946, Tanner 1944, and 1947, Rousseau 1949). The original describer, Bangs, saw no skins of this "race", the f i r s t description of the colour being made by Rhoads (1902), who described i t as "separable from any other eastern interior form on account of i t s darker shades and i t s longer guard hairs, and distinct from the equally large Alaskan marten (M. a. actuosa) by i t s much darker colour". Rhoads' interpretation of this "subspecies" i s quite correct and I conclude with him that in colour and size i t i s 144 quite distinctive from any other "subspecies" of "M. a. americana" It is always very much darker than any other "race", being a rich glossy chocolate brown, often nearly black, with guard hairs long and glossy. In darkness of pelage i t i s approached only by the Newfoundland marten ("M. a. atrata"). Cranium: The original description of thie "race" was made by Bangs from skulls, which he described as follows: "skull large, powerful, and heavy; rostrum very short and broad; frontals highly arched; auditory bullae very large and deep; dentition extremely heavy throughout, the last uppermolar i n particular being very large; the tooth row a good deal crowded ... their large size, short wide rostrums, and enormous teeth at once dis? tinguish them." Rhoads, in 1902, validated Bangs' description, adding to i t i t s distinctive pelage characteristics and pointing out the similarity of the skull of this "race" to that of "M. a. actuosa". Rhoads, unlike Bangs, recognized too that "brumalis" should properly be considered a "subspecies" of "americana" and not a distinct "species". My examination of skulls of this "race" show i t to be very much larger than those of "M. a. americana" to the south, being equal i n size to those of "kenaiensis", "abieticola" and "actuosa", the ranges of which i t does not meet. Comparison of cranial statistics with "M. a. americana" have already been made. Comparison with "M. a. atrata" i s made under that "subspecies" topic. Compared to "abieticola", "actuosa" and "kenaiensis", they show a condylobasal length 145 significantly greater than any, and hence are the largest marten in North America. S t a t i s t i c a l significance i s also shown for upper tooth row (from male "actuosa" only), canine width at base ("kenaiensis" only) upper molar width (males of "kenaiensis" and "actuosa" only), upper molar inner moiety length (from "actuosa" and "kenaiensis), palate length (from "kenaiensis", "actuosa" and "abieticola" in males, but only "abieticola" i n females), mastoid width (from "kenaiensis" and "actuosa", males only), length of last upper premolar (from "actuosa", males only), canine width at base / bullae length (from "kenaiensis") and upper molar inner moiety length plus width ("actuosa" and "kenaiensis", males only). Thus of fourteen cranial characters considered, this "race" i s significantly different from "kenaiensis" by ten in males and three in females; from "actuosa" by eight in males and one in females; and from "abieti-cola" by two in males and one in females. The implied greater similarity to "abieticola" i s due l i k e l y to the smallness of the sample of that "race" studied (n = 18) and to the large standard deviations thus obtained. Males . Females X SD CV n X SD CV n Condylobasal length 8.56 .18 2.06 55 7.66 .15 2.01 48 Rostral width 1.63 .06 3.80 58 1.44 .06 3.88 49 Upper tooth row 3.13 .07 2.67 35 2.79 .05 1.76 23 Canine width at base 1.57 .06 4.01 58 1.38 .06 4.05 49 Bullae length 1.75 .05 3.09 58 1.62 .05 3.20 49 Upper molar width .80 .03 3.97 58 .70 .03 4.55 49 146 Males Females SD cv n X SD CV n Upper molar inner moiety length .49 . 0 3 5.66 58 .42 .02 5 . 2 5 49 Palate length 4.35 .11 2.57 3 6 3 . 8 1 . 0 8 2 . 0 5 2 3 Mastoid width 3.79 . 0 8 2.13 36 3.45 . 0 8 2.29 2 3 Height of skull at bullae 3.22 .08 2.42 36 2.89 . 0 8 2 . 8 0 2 3 Length of last upper PM .84 .03 4.14 58 .75 .02 2.93 49 Narrowest palate width .84 . 0 3 4.14 35 . 7 6 . 0 3 4.19 23 Canine width base / . 9 0 bullae length .04 4 . 4 5 58 .85 .04 4.57 49 Upper molar width 1.30 plus inner moiety length . 0 5 4.22 58 1 . 1 2 . 0 5 4.54 49 Total length 68.00. - 1 58.5 - - 1 T a i l length 19 .50 - - 1 17.90 - - 1 Hind foot 11 .50 — — 1 9.1 _ — 1 Geographic variation: Anderson (1934b) says that this "race" i s "more or less variable in any d i s t r i c t " but adds no detail. Writers who have lived in the di s t r i c t (Rowan I876, Low 1 8 9 5 and I896, Tanner 1944 and 1947) say that the largest and darkest martens are to be found in the very northern part of the "sub-species" range. Low says "the largest and darkest skins are taken along the edge of i t s northern limits", and Tanner remarks that "the darker and thicker the black spruce woods are, the darker and richer are the skins of the marten living there...." 147 We have already pointed out that as our isopleth map indicates, a steady decrease in size (and presumably of darkness) occurs from the Labrador coast southwest into the region about James Bay and the Saguenay River. The very largest specimens appear to be taken from the area about Cartwright and the region about the mouth of the Eagle River on the southern Labrador coast. Specimens from Chimo, Grand Falls and Northwest River are nearly as large, and larger than those from the adjacent coast, suggesting an "island" of largeness in the region of the Torngat Mountains, Labrador. Distribution: A l l authors appear to agree that the northern limit of this "race" i s the northern limit of marten and i s co-incident with the tree-limit in Labrador and Quebec, and the southern limit of the semi-barrens (Low 1895 and I896, Anderson 1934b, 1938 and 1940, Tanner 1944 and 1947). Low adds that north of this line i t i s found "only in the wooded stretches of the river valleys ...." While a large proportion of the speci-mens I have examined are labelled as coming from the Labrador coast i t is quite l i k e l y that many were trapped inland and later brought to the coast. The southern limits of this "race^ have been discussed under the topic "M. a. americana" and the reader should refer to this. In brief, i t may be said to grade through the length of most of the Quebec-Labrador peninsula into "M. a. americana" somewhere in the region about southern James Bay and Lake St. John. It has already been pointed out that this may be the 148 "race" that once occupied Anticosti Island. Specimens examined: 128 specimens from northern Quebec and Newfoundland Labrador. Skull diagrams: Bangs I898, p. 503; Osgood 1900, p i . 7-"Martes americana atrata" (Bangs 1897) Synonymy: Martes (or Mustela) atrata, caurina atrata, americana  atrata. Type specimen: Museum of Comparative Zoology, Cambridge, Massachusetts, Bangs1 Collection no. 5752 (Allen 1931, Anderson 1946). Type lo c a l i t y : "Bay St. George, Newfoundland" (Bangs 1897). Pelage: The characteristics of the Newfoundland marten are not well known, presumably because of i t s scarcity. Almost a l l writers (Gilpin i 8 6 0 , Bangs 1897, Miller 1900, E l l i o t 1901, Cory 1912, Bangs 1913 and Anderson 1946) describe i t as dark chocolate in colour, with lustrous pelage and orange throat patch. Most distinguish i t from the much paler coloured "M. a. americana" and ally i t with the similarly coloured "M. a. brumalis" of Labrador and Quebec. Reeks, in 1870, in comparing i t to the Labrador "race" described i t as "much lighter throughout, but especially in the head and ears". Bangs, in his original des-cription described i t as about the same size, or perhaps slightly larger than "M. a. americana" (i. e . a small marten), with 149 colouration more like that of a mink than that of a marten. "Deep chocolate becoming black o n back, head, arms, legs, rump and t a i l ; a few white hairs scattered along the back; chest and under side of neck irregularly blotched with orange; a median line of orange on belly; ears black, narrowly bordered a l l around with dull white; a patch of yellowish-white hairs i n front of the opening of ear...." Bangs later (1913) redescribed i t as a "rather small species..." as did Miller i n 1900, who said i t could always be told from "brumalis" by i t s smaller size and from "americana" by i t s much darker (never pale yellowish) pelage. Cory (1912) added that i t may always be told from "M. a. americana" by i t s more orange throat patch. I have examined no skins of this "race" and hence cannot add to the descriptions above, except to point out, as i s done below in greater detail, that Bangs and later authors were mistaken in believing that this "race" i s as small as he believed. In size i t l i e s much nearer to "M. a. brumalis". The Newfoundland marten shows, i n a l l the skulls I have studied, a l l of the characters of the "species" "M. americana" and must be considered to belong to i t . Bangs when he described the race considered i t to be a distinctive species ("Mustela  atrata"). E l l i o t i n 1901a incorrectly considered i t a subspecies of "M. caurina". Rhoads (1902) although he saw no skulls of the "race" correctly recognized from Bangs'; description that i t was closely a l l i e d to "M. a. americana" , although he made no noinencla-tural change. Allen, i n 1942, f i r s t named i t a subspecies of "M. americana" and de Vos, in 1952, although unaware of Allen's 150 change, expressed a correct interpretation of the relationship of atrata as a subspecies of "americana". Miller and Kellogg (1955) however, gave i t specific status. Bangs, in his description of the "race", did not recog-nize that the specimens on which he based his account were fe-males; i t was by comparing these to males of "M. a. americana" and "M. a. brumalis" that he concluded "atrata" to be very l i t t l e larger than "M. a. americana". Of the eleven skulls of this "subspecies" that I have examined, three were male, and i t i s evident from these that i n size "M. a. atrata" l i e s very near to "M. a. brumalis" rather than to "M. a. americana" as Bangs be-lieved. The mean condylobasal length of males of this "race" (8.39 cm., n = 3 ) 5 while not so large as that of the samples of "M. a. brumalis" from the Labrador coast, equals those (as the isopleth map shows) from extreme western Newfoundland-Labrador, which i s very much larger than those of typical "americana" from Ontario. Since in size "atrata" l i e s close to the average specimen of "M. a. brumalis" and since descriptions of i t s pelage show i t to be dark as in "brumalis" (not pale as i n 11 americana") we must conclude that i t i s closely related indeed, to the Labra-dor "race" and not to "M. a. americana". I conclude, in fact, that there have been no distinctions shown to exist between L* at rat a" and "brumalis" and that both are best considered one "subspecies", "M. a. atrata" (Bangs). In addition to the similarity of "atrata" to "brumalis" and i t s dissimilarity to "americana" stated above, I add the following based on s t a t i s t i c a l comparison of the three samples studied. Of the fourteen cranial characters studied, when 151 "atrata" i s compared to "americana" a l l but the following differ significantly to the 95$ level; upper molar width, palate length and upper molar width plus inner moiety length. On the other hand, when compared to "brumalis", "atrata" shows significance in the following only: upper molar width (females only), palate length (males only), mastoid width (females only) length of last upper premolar, canine width at base, bullae length (males only) and upper molar width plus inner moiety length. Two other characters approach but do not attain significance (rostral width, males only, canine width at base, females only). By sum-ming significances we find that "M. a. atrata" shows over twice as many significant distinctions when compared to "M. a. ameri-cana" than when compared to "brumalis". The following are cranial statistics for the Newfoundland "race": Males Females X SD CV n X SD CV n Condylobasal length 8.39 .16 1.85 3 7.37 .08 1 .04 5 Rostral width I .67 .02 1.20 3 1 .42 .03 2.25 8 Upper tooth row 3.13 .09 2.71 3 2.74 .06 2.01 8 Canine width at base 1.58 .03 1.90 3 1.35 .02 1.63 8 Bullae length 1.73 .03 1.73 3 1.59 .03 2.07 8 Upper molar width .76 .04 4.59 3 .66 .03 3-78 8 Upper molar inner moiety length .46 .01 3.02 3 .40 .03 6.23 8 Palate length 4.28 .10 2.34 3 3.68 .06 1.69 8 Mastoid width 3.79 .09 2.37 3 3.36 .06 1.63 8 Height of skull at bullae Length of last upper PM Narrowest palate width Male x SD CV n 3.21 .05 1.40 3 .82 .00 .00 3 .87 .03 3.45 3 Canine width base / .91 .00 .00 3 bullae length Upper molar width 1.23 .05 4.08 3 plus inner moiety length Female x SD CV n 2.95 .11 3.60 8 .71 .02 2.39 8 .77 .03 3-91 8 .85 .02 2.59 8 1.06 .05 4.52 8 152 Total length T a i l length Hind foot 552* 185* 88* 3 3 3 •From Bangs 1897 and Miller 1900 Geographic variation: Nothing i s known regarding this, Distribution: Restricted to forested Newfoundland, where i t i s considered the only form of marten to occur. It is possible ( i f "atrata" and "brumalis" were to be considered distinct) that this "race" (and not "brumalis") i s the one occurring on Anti-costi Island. Specimens examined: 11 specimens from Newfoundland Skull diagrams and photographs: None that I know of. 153 "Martes americana abieticola" (Preble 1902) Synonymy: Martes (or Mustela) americana abieticola Type specimen: U. S. National Museum no. 19256/34962 (Poole and Schantz 1942, also seen by myself). Type l o c a l i t y : "Cumberland House, Saskatchewan"](Preble 1902). R. Macfarlane, who collected the type for Preble, wrote (1905) that i t was obtained from an Indian chief, "I think the Chief told me that he had trapped i t i n the Pas Mountain, some sixty or seventy miles to the southward of Cumberland House". Pelage: This "race", like that preceding, i s a puzzling one, in that i t has never been properly defined and limited. Preble, when f i r s t he described the "form" properly recognized i t as a "subspecies" of "M. americana", but did l i t t l e to distinguish i t from other "subspecies" of the "species", especially "M. a. actuosa". Subsequent authors have added almost nothing to his description. Preble described the skin as follows: "type in winter pelage: general color of body rich dark yellowish brown, darkest on middle of back; legs and t a i l darker, the latter almost black at t i p ; an irregular blotch and a small spot on chest, ochrac-eous; face and cheeks greyish brown; ears edged with whitish". Anderson i n 1946 described i t simply as "a large dark brown race". Preble recognized i t to be much larger than "M. a. americana" to the east, although he considered i t rather smaller than the Alaskan marten "M. a. actuosa". My examination of Preble's skins from Manitoba leads me 154 to agree that i t i s a much larger and darker "race" than i s "M. a. americana" from Ontario. The colour tends to be richer, browner, and longer furred than the smaller "race" and browner and less grey than the pale headed Alaskan "race", though of about the same size. There i s , however, sufficient variation in the colour of the fur of the Alaskan "race" (see under that topic) especially in Mackenzie, to lead one to believe that "abieticola" f a l l s within i t s range of colour. Cranium: This i s described by Preble as "much larger than i n M. americana americana more angular, the sagittal crest being more highly developed; zygomata more bowed outward poster-ior l y ; dentition much heavier except last upper molar, which is usually about the same size." Elsewhere he remarks that "this form approaches in some of i t s characters Mustela a. actuosa from Alaska, but though smaller has heavier dentition than that species, and consequently more crowded molars." The distinctions that Preble makes between "M. a. ameri-cana" and "abieticola" are, I find, perfectly real and valid, although i t i s evident that the two "races" intergrade, and are closely related. However, Preble f a i l s to show that "abieticola" i s really distinctive from "M. a. actuosa" to the northwest. His distinctions between these two "races" are essentially these: that "abieticola" i s darker in colour, slightly smaller i n size and with slightly heavier and more crowded dentition. I have already pointed out that the difference i n colour may easily f a l l within the range of "actuosa". Further, this "race", contrary to Preble, i s not smaller than the Alaskan "race", but rather about 155 equals i t . Insofar as crowding of and heaviness of dentition is concerned, while these are d i f f i c u l t to measure quantitatively, I conclude that there i s no essential difference here between the two "races 1 1. In other words I conclude that aside from conformity to precedence, there is yet no basis for the mainten-ance of this "race", and here name them one, "M. a. actuosa" (Osgood). Comparison of this "race" to "americana" has already been made. Compared to "actuosa" of the fourtee