A new whitefish from the early Quaternary of Bluefish Basin, Yukon Territory, Canada, and its paleoenvironmental implications

A new whitefish from the early Quaternary ofBluefish Basin, Yukon Territory, Canada, and itspaleoenvironmental implications

Stephen Cumbaa, Bernard Lauriol, Noel Alfonso, Martin Ross, and Robert Mott

Abstract: A nearly complete fossil of a whitefish, Coregonus beringiaensis sp. nov. (Salmoniformes: Salmonidae: Core-goninae), the oldest known record of the genus, is described from Ch’ijee’s Bluff along the Porcupine River, Bluefish Ba-sin, Yukon Territory. The new species is closest to but distinct from species within the Coregonus clupeaformis complex,especially in its low lateral line scale count, quadrate with an extensive canal and pore complex, Y-shaped lachrymal, anda unique combination of counts and measurements of fins, scales, and body proportions. The stratigraphic position of thesource concretionary layer, at the base of unit 3, a lacustrine unit, is well below a magnetically reversed interval thoughtto represent at least the upper part of the Matuyama Chron. The specimen was found above ice-wedge pseudomorphsthought to be the first sign of the onset of cold temperatures in the early Quaternary. The fossil is older than 0.79 Ma andprobably younger than 2.58 Ma. Palynomorphs in the concretion suggest an open shrub-tundra environment and a climatecolder than that of the present, similar to conditions prevailing at or above the tree line, and a good fit for preglacial con-ditions in the early Quaternary. Sedimentary and geochemical analyses suggest that fossilization occurred in an environ-ment characterized by fine sedimentation in a cold, reducing milieu with a pH of *7.5. Alternatively, these conditionsmay reflect the postdepositional or early diagenetic environment. The lake in which the whitefish lived was possibly partof a hydrographic basin draining toward the Arctic Ocean.

Resume : Un fossile presque complet d’une nouvelle espece de poisson blanc, Coregonus beringiaensis sp. nov. (Salmo-niformes : Salmonidae : Coregoninae), le plus vieux specimen connu du genre, est decrit pour Ch’ijee’s Bluff, une localitesituee le long de la riviere Porcupine, dans le bassin de Bluefish, au Yukon. La nouvelle espece est apparentee aux especesappartenant au complexe Coregonus clupeaformis, mais elle se distingue de celles-ci, tout particulierement, de par sonfaible nombre d’ecailles le long de la ligne laterale, de l’os carre etant dote d’un complexe de canaux et de pores tres de-veloppe, un os lacrymal en forme de « Y » et d’une combinaison unique du nombre d’ecailles et de rayons des nageoires,de meme que des proportions des nageoires et certaines parties du corps. La situation stratigraphique de l’horizon conte-nant la concretion, a la base de l’unite 3, une unite lacustre, est nettement en dessous d’une inversion paleo-magnetiqueconsideree comme representant au moins la portion superieure du chron de Matuyama. Le specimen fut trouve au dessousde pseudomorphes de coin de glace, qui representent la premiere indication de l’etablissement de temperatures froides du-rant le Quaternaire inferieur. Le fossile est age d’au moins 0,79 Ma, mais moins de 2,58 Ma. Des palynomorphes dans laconcretion indiquent un environnement de toundra arbustive ouverte, et un climat plus froid que celui d’aujourd’hui, sem-blable aux conditions prevalentes a ou au-dela de la limite des arbres, ainsi qu’une bonne estimation des conditions pre-glaciales au debut du Quaternaire. Les analyses sedimentaires et geochimiques suggerent que la fossilisation est survenuedans un environnement caracterise par une sedimentation fine, dans un milieu froid et reducteur, avec un pH voisin de 7,5.Toutefois, ces conditions pourraient etre representatives de l’environnement post-sedimentaire ou de diagenese precoce. Lelac dans lequel le poisson blanc vivait appartenait probablement a un bassin hydrographique qui se drainait vers l’oceanArctique.

IntroductionWhitefishes are taxonomically contained within the fish

subfamily Coregoninae, one of three in the family Salmoni-dae. The other two subfamilies include salmon, trout, char,and grayling. Whitefishes are nearly ubiquitous in circumpo-lar cold waters, including northern North America, and arewell known from Canada even as fossils (Cumbaa et al.1981). However, most fossil whitefish specimens discoveredto date are scattered individual elements, and relatively com-plete specimens are rare. Fossils of the earliest whitefishknown from North America, Prosopium prolixus, have beenfound in Pliocene strata in the upper part of the GlennsFerry Formation, associated with deposits from Lake Idahoin southwestern Idaho, USA (Smith 1975, 1981). World-wide, the earliest known whitefish fossil appears to be Sten-

Received 27 September 2009. Accepted 10 February 2010.Published on the NRC Research Press Web site at cjes.nrc.ca on31 March 2010.

Paper handled by Associate Editor H.-D. Sues.

S. Cumbaa1 and N. Alfonso. Canadian Museum of Nature, P.O.Box 3443, Station D, Ottawa, ON K1P 6P4, Canada.B. Lauriol. Department of Geography, University of Ottawa,Ottawa, ON K1N 6N5, Canada.M. Ross. Department of Earth and Environmental Sciences,University of Waterloo, Waterloo, ON N2L 3G1 Canada.R. Mott. Geological Survey of Canada, Ottawa, ON K1A 0E8,Canada.

1Corresponding author (e-mail: [email protected]).

221

Can. J. Earth Sci. 47: 221–235 (2010) doi:10.1139/E10-015 Published by NRC Research Press

odus from the Miocene of Siberia (Reshetnikov 1995),which is also known as a fossil from Yukon Territory(Cumbaa et al. 1981).

The whitefish Coregonus cf. C. lavaretus is known fromPleistocene lacustrine deposits of the Ferdinandowian Inter-glacial at the Wola Grzymalina locality near Belchatow inPoland (Jerzmanska and Raczynski 1991). The material,several incomplete specimens identified as Coregonus cf. C.lavaretus, consists of scales, both isolated and in patches,skeletal fragments, a caudal skeleton, and the anterior halfof a skeleton covered with large, cycloid scales. The specificdetermination was made on the basis of the size, shape, andstructure of the scales and by comparing them to extant C.albula and C. lavaretus specimens. The bones of the headand pectoral girdle of these specimens were too poorly pre-served to reconstruct the shape of individual bones (Jerz-manska and Raczynski 1991).

Rzechowski (1996) and Mojski (1995) placed the strati-graphic position of the Ferdinandowian Interglacial beneaththe Wilga Glaciation (equivalent to Sanian 2?) and abovethe Sanian 1 Glaciation in the Polish system. Althoughmuch correlation remains to be done, the Ferdinandowianappears to be roughly equivalent to the late Yarmouthian In-terglacial of North America. Thermoluminescence measure-ments from three sections of the Ferdinandowian suggestages of ca. 520–550 ka, which may be better viewed as arelative rather than an absolute age (Rzechowski 1996).These Coregonus cf. C. lavaretus specimens from Polandare the earliest known Coregonus fossils from Europe and

have been thought to represent the oldest known occurrenceof the genus.

In 1997, BL recovered a roughly ovoid concretion,*70 cm � 40 cm, derived from a concretionary layer<1 m above the base of unit 3 at Ch’ijee’s Bluff on the Por-cupine River in northern Yukon Territory, Canada. The con-cretion was split and found to contain a relatively intactfossil fish. Both sides of the concretion were brought backto Whitehorse, Yukon, for further preparation and depositionin the paleontological collections of the Yukon Palaeontol-ogy Program, Yukon Heritage. This fossil specimen hassince been determined to be a whitefish and is, along withits paleoenvironment, the focus of this study. We presentdata here that indicate this fossil whitefish is the earliestknown Coregonus specimen and one of the earliest knownwhitefishes.

Stratigraphy and sedimentologyCh’ijee’s Bluff, 55 m high and 4 km long, is arguably the

most important single Cenozoic exposure in the northernYukon (Matthews et al. 1990). The basis for its significanceis the bluff’s exposure of Pliocene through late Pleistocenedeposits and the study of these deposits by a number of re-searchers from a variety of disciplines (McConnell 1891;Delorme 1968; Hughes 1969, 1972; Lichti-Federovich1974; Matthews 1975; Harington 1977, 1978; Pearce et al.1982; Westgate et al. 1985; Ross 1997). This exposure, alsoknown as Twelve Mile Bluff, Big Bluff, HH62-228, Porcu-pine River 1, Porcupine Locality 100 (or CRH 100), and

Fig. 1. Map of northern Yukon with Old Crow and Bluefish basins shown in dark grey; Ch’ijee’s Bluff fossil locality shown by whiteasterisk along Porcupine River in Bluefish Basin.

222 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

MiVl-1 by the Borden System archaeological site designa-tion (Hughes 1969, 1972; Lichti-Federovich 1974; Harington1977; Morlan 1980), is located in the middle of the BluefishBasin at 67829’30@N latitude and 139853’25@W longitude, onthe left (south) bank of the Porcupine River, 9.7 km south-southwest of the village of Old Crow, or roughly 15 kmdownstream (National Topographic System (NTS) 116 O,Fig. 1).

Hughes (1972) was the first to describe the section, andhis division of the bluff into six principal stratigraphic unitsis the basis for subsequent geological, geomorphological,and paleontological interpretations. Lichti-Federovich(1974), in consultation with Hughes, presented pollen data

from the locality (her Porcupine River 1) in a slightly re-vised form. Figure 2 shows the different lithostratigraphicunits.

Unit 1, the lowermost stratigraphic unit at Ch’ijee’s Bluff,exposed only during periods of low river level, is estimatedat 5–2.3 Ma (Pliocene – early Pleistocene) in age based oncomparative pollen studies and the discovery of cones ofthe extinct pine, Pinus matthewsii (McKown et al. 2002).Harington (1977) also reported an extensive, highly com-pressed beaver dam in the unit at the upstream end of thesection. The sedimentological analyses of unit 1 and thebase of unit 2 show a fluvial environment, characteristic ofa braided river system with tabular sheets of coarse sand andgravel with broad scour and fill structures. The occurrenceof buried in situ tree trunks and an abundance of smallbranches and other wood debris aligned with the paleoflowdirection suggest lateral migration of channels and rapidsedimentation (Ross 1997). These characteristics were seenat the downstream end of the section. The upstream part de-scribed by Matthews et al. (1990) seems to contain a higherproportion of sandy facies. The fining-upward trend in unit2 indicates a decrease in energy. The upper part of unit 2also contains lower flow-regime structures such as asymmet-ric ripples. Collectively, these characteristics and facies tran-sitions may reflect a change in slope gradient over time andthe complex interaction of small alluvial fans with alluvialplain environments. Plant macrofossils place unit 2 in the‘‘Tertiary’’ (Matthews and Ovenden 1990); the correlationchart (fig. 12 in Matthews and Ovenden 1990) places unit 2in the Late Pliocene to early Pleistocene using revised datesfor the Pliocene–Pleistocene boundary (International Com-mission on Stratigraphy 2009). The presence of ice-wedgepseudomorphs near the top of unit 2, the lowest in theCh’ijee’s Bluff section, has been interpreted as the first evi-dence of local climatic cooling in the early Quaternary.Widespread evidence for temperature decline in the regionindicates the onset of major Northern Hemisphere glaciationnear the Matuyama–Gauss boundary at 2.58 Ma (Cande andKent 1995).

Unit 3, 20 m thick, begins 17 m above the mean level ofthe Porcupine River (Fig. 3). It forms a nearly vertical faceclearly visible along the entire length of the bluff. The unitconsists of beds of dark grey to dark brownish grey siltyclay to clayey silt, indicative of a lake deposit, with occa-sional sand lenses that are more abundant near the base ofthe unit. Unit 3 also contains numerous large (1–1.5 cm)nodules of vivianite, as well as segregated ice interspersedfrom base to top. A layer of ovoid concretions, the sourceof the concretion containing the fossil whitefish, is presentat the bottom of unit 3; the largest observed was 160 cm indiameter (Fig. 4). Paleomagnetic analysis of samples takenat Ch’ijee’s Bluff was presented by Pearce et al. (1982).Their results show that the major part of unit 3 is a magneti-cally reversed interval, attributable to the Matuyama Chron,and that permafrost conditions prevailed for most of this in-terval (Pearce et al. 1982). Attribution of this interval to theMatuyama Chron indicates an age older than 0.79 Ma(Johnson 1982) for both the fossil and the lake in which itlived (Cumbaa et al. 2002). The lower part of unit 3, whichcontains the fossil-bearing concretionary layer, is normallymagnetized. Pearce et al. (1982) argued that the paleobotan-

Fig. 2. Stratigraphy of Ch’ijee’s Bluff (modified from Lichti-Federovich (1974)). m a.s.l., metres above sea level.

Cumbaa et al. 223

Published by NRC Research Press

ical data from unit 3 support an age close to that of units 1and 2; that is similar to the Pliocene – early Pleistocene agefor the lower units estimated by McKown et al. (2002). Fur-ther evidence of an early Quaternary age for at least thelower part of unit 3 is the recovery of a fossil of the extinctbeetle Micropeplus hopkinsi (Matthews et al. 1990), whichis otherwise restricted to Pliocene and older deposits (Mat-thews et al. 2003).

The sediment at the bottom of unit 3, where the concre-tion with the fossil was found, has a pH value *7.5, a light

gray olive (5Y 6/2) colour, and contains 7% organic matter.It is mainly formed by grains of quartz and feldspar, with amean grain size of 12 mm, similar to the angular to suban-gular grains of the Old Crow batholith, located northwest ofthe Bluefish Basin (Bjornson and Lauriol 2001).

An X-ray diffraction analysis by Ross (1997) of a sample ofsediment collected in unit 3 shows the occurrence of kaolinite,albite, muscovite, and vivianite. Vivianite (Fe3(PO4)2�8H2O)is an iron phosphate well-known in organic deposits and lakesediments (Rosenqvist 1970). It forms under both biotic and

Fig. 3. Magnetic stratigraphy of Ch’ijee’s Bluff. O, horizon where the whitefish concretion originated. Unit 1 is not visible in the photo-graph; it is above river level only under low-water conditions. N indicates normal magnetization or normal polarity.

Fig. 4. A large ovoid concretion at Ch’ijee’s Bluff. The 1.2 m diameter concretion is from the same layer at the base of unit 3 as the con-cretion containing Coregonus beringiaensis sp. nov. (reproduced with permission).

224 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

abiotic conditions in clay sediments (Zelibor et al. 1988). Viv-ianite nodules are a diagenetic product linked to anoxic condi-tions, commonly found in redox (reducing) environments witha pH *7.5 and an Eh (oxidation potential) *0 (Nriagu andDell 1974). Mineralogical analysis by X-ray diffraction onbulk powder samples (n = 2) of the whitefish concretionshows that iron (Fe) and silica (Si) are the main minerals, fol-lowed by calcium (Ca). The peak to the right of Si in Fig. 5 isinterpreted as phosphorus (P), since its presence is generallyassociated with fossil bone. An abundance of vivianite inPleistocene lake sediments in northeastern Russia has beenpostulated as indicating a relatively high paleoproductivity offish (Minyuk et al. 2007). Optical microscope observations(n = 3) reveal that the concretion itself contains only grains<2 mm with a brown matrix, which is probably a gel of ironand organic matter.

Materials and methods

Eleven measurements (Table 1) were made in 10ths of amillimetre on the fossil and on comparative specimens fromextant species; as well, three counts were made: lateral linescales, number of dorsal rays, and scales above lateral linefollowing Koelz (1929) and Hubbs and Lagler (1964). Rawdata were examined for normality and homoscedasticity (ho-mogeneity of variance).

Morphometric studies should compare fish populationsusing shape variates independent of the effects of size varia-tion (Reist 1985, 1986); therefore, size effects in the datawere minimized by calculating residuals from the pooledgroups regression line of logged (base 10) character andlogged standard length. As residuals are orthogonal to theregression line, they reflect the shape of each body part in-dependent of size. Discriminant function analysis (DFA)

was used on size-free factors produced by a principal com-ponents analysis (PCA) that summarized variation in the da-taset to examine group membership of the fossil specimen.Comparative specimens examined are listed in Appendix A.

Palynological analysis was undertaken by RM at the labo-ratories of the Geological Survey of Canada, Ottawa, On-tario, on matrix taken from the concretion bearing the fossilwhitefish. One sample from the concretion of *5 cm3 wastreated with dilute (10%) HCl, followed by standard palyno-logical procedures using KOH and acetolysis. The residuewas mixed with silicon oil for counting.

Changes to the definition of the base of the QuaternarySystem/Period (and the top of the Neogene System/Period)and redefinition of the base of the Pleistocene Series/Epoch(and top of the Pliocene Series/Epoch) have recently beenratified by the International Union of Geological Sciences(International Commission on Stratigraphy 2009). Accord-ingly, the base of the Pleistocene has been lowered suchthat the Pleistocene now includes the Gelasian stage/age.This boundary, formally defined by the Monte San Nicolaglobal boundary stratotype section and point (GSSP), isnow set at 2.588 Ma. We follow this convention.

Systematic paleontologyClass Actinopterygii Klein, 1885Division Teleostei Muller, 1846Order Salmoniformes sensu Nelson, 2006Family Salmonidae sensu Nelson, 2006Subfamily Coregoninae sensu Nelson, 2006Genus Coregonus Linnaeus, 1758

Coregonus beringiaensis sp. nov.

(Figs. 6, 7)

DIAGNOSIS: A species of Coregonus with the following char-acters: 12 dorsal fin rays, 11 scales above the lateral line,and *69 lateral line scales, head depth into standard length10.8. Preopercle with five, prominent, raised preopercularcanal branches; quadrate with large, radiating, fanlike poreand canal complex on the lateral surface of the anteriorlimb; and a Y-shaped lachrymal.

HOLOTYPE: Yukon Palaeontology Program Collections cata-logue number MiVl-1: 44.2, a complete specimen in partand counterpart.

LOCALITY: Canada, Yukon Territory, Ch’ijee’s Bluff on Por-cupine River, 9.7 km south-southwest of the village of OldCrow.

HORIZON: Probably early Pleistocene (Gelasian); concretion-ary zone in clay and silt lacustrine deposits, 18 m abovemean river level (1 m above base of unit 3, stratigraphic ter-minology of Hughes 1972).

ETYMOLOGY: The generic name, Coregonus, comes from theGreek, kore (pupils of the eye), + Greek, gonia (angle). Thespecific epithet, beringiaensis, refers to Beringia, the some-times ice-free land mass that connected Asia and NorthAmerica, most recently during periods of Pleistocene glacia-tion, and which formed an important pathway for faunal andfloral interchange. The fossil locality at Ch’ijee’s Bluff isnear the eastern limit of Beringia (Harington 2005).

Fig. 5. Mineralogical analysis by X-ray diffraction (XRD) of thewhitefish concretion. ? is interpreted as phosphorus.

Cumbaa et al. 225

Published by NRC Research Press

Description of specimen

Figure 6 presents the impression of the right side of thefish in lateral view after preparation. Overall body morphol-ogy, including the shape of the head, the shape and place-ment of fins, the arguable presence of an adipose fin,upturned three terminal vertebrae, and scale size and form,indicates a fish of the Order Salmoniformes, which has asingle family, Salmonidae (Nelson 2006). There are threesubfamilies: Salmoninae (salmon, trout, and char), Coregoni-nae (whitefishes), and Thymallinae (grayling). The fossilmatches the whitefish subfamily Coregoninae; on the basis

of having fewer than 16 dorsal fin rays, large scales, withfewer than 110 scales along the lateral line, and the presenceof a largely toothless mouth (Nelson 2006). Genera in Core-goninae include Prosopium Jordan, 1878; Stenodus Richard-son, 1836; and Coregonus Linnaeus, 1758. Members of thegenus Prosopium have thin, pinched snouts; and Stenodusleucichthys, the only representative of the genus Stenodus,has a distinctively elongate body and a large maxilla.Figure 7 shows details of the head (Fig. 7a), scales (Fig. 7b),and fins (Fig. 7c).

The deep body, subterminal mouth, maxilla not extendingbeyond the anterior margin of the eye, vertebral modal count

Table 1. List of morphometric characters.

BD Greatest depth of bodyCP HEIGHT Height of caudal peduncleDOR BASE Distance from origin of first dorsal ray to the insertion of last dorsal rayDOR HT Length of longest ray in dorsal fin

HL Length of headHD Depth of head behind orbitPRE DOR Distance from anteriormost part of snout to the origin of the first dorsal rayPRE PELV Distance from anteriormost part of snout to the origin of the first pelvic rayPEC VENTRAL Distance from origin of first pectoral ray to the origin of first pelvic raySL Length of axial skeletonSNOUTL Length of snout

Fig. 6. Impression of the right side of Coregonus beringiaensis sp. nov.

226 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

of *59 (Smith and Todd 1992), and orbital ring open dorsally(Norden 1961) indicate that the specimen is a member of thegenus Coregonus and possibly close to the Coregonus clupea-formis complex (C. clupeaformis, C. pidschian, and C. nel-

soni), in which the best character to distinguish specieswithin the complex is gill-raker count (McPhail and Lindsey1970). Gill rakers are not visible on the fossil specimen.

The specimen is undistorted, and a number of counts and

Fig. 7. Details of the right-side impression of Coregonus beringiaensis sp. nov.: (a) head, with crystal growth obscuring some details; AO,antorbital; ART, articular; BR, branchiostegals; CL, cleithrum; COR, coracoid; DN, dentary; FR, frontal; IO2–IO5, infraorbital series; IOP,interopercle; LA, lachrymal; MX, maxillary; OP, opercle; PF, pectoral fin; POP, preopercle; PTM, posttemporal; QU, quadrate; SCL, su-pracleithrum; SM, symplectic; and SOP, subopercle; (b) scale pattern from an 8 cm midbody section beginning under dorsal fin; anterior toleft. Note lateral line scales (marked by impression of lateral line canal) running roughly parallel to and *1 cm below the impression of thevertebrae; (c) caudal fin, posssible adipose fin, and posterior portion of dorsal fin. Note impression of upturned three terminal vertebrae. SeeFig. 1 for scale bar reference.

Cumbaa et al. 227

Published by NRC Research Press

measurements commonly used to describe extant whitefisheswere made (Koelz 1929; Hubbs and Lagler 1964) on the as-sumption that these measurements and meristics would haveclosely characterized the living specimen.

MeasurementsTotal length (TL), 418 mm; fork length, 395 mm; stand-

ard length (SL), 345 mm; head length, 76 mm (18% TL,22% SL); head depth, 30 mm; snout length, 22.5 mm; bodydepth, 89 mm (21% TL; 26% SL); length, dorsal finheight, >71 mm; dorsal fin base, *46 mm; anterior tip ofthe snout to origin, dorsal fin, 160 mm; pectoral finlength, >35.5 mm; height, caudal peduncle, *36 mm.

MeristicsLateral line scale count, 69; scale rows above lateral line,

11; dorsal fin rays, 12; pectoral fin rays, 11+ (some ob-scured); pelvic fin rays, 8+ (some obscured); branchioste-gals, at least 6 per side; vertebrae estimated at 58–59, basedon impressions plus a count of the neural arches.

Most bones of the neurocranium, skull roof, jaws, suspen-sorium, and opercular region are only partly visible, as theyare still in articulating position and thus obscured by otherbones; some remain imbedded in matrix and (or) are ob-scured by crystal growth. Fin rays are visible to a large ex-tent, but fin supports are generally obscured. Details of thevertebrae are entirely obscured by scale cover, but the up-turned direction and number of vertebrae in the extremecaudal region are sufficient to establish the specimen as asalmoniform. The matrix is too dense to obtain useful osteo-logical information from X-ray imaging.

Osteology

Neurocranium and skull roofThe partial right frontal is the only bone in this series that

can be described with any useful characteristics. It is bestviewed using the latex peel of the fossil. The partially ex-posed dorsal surface is subtriangular and anteriorly tapered.The medial edge marking the suture between the right andleft frontals is obscured by matrix; and the parietal border isbroken, as is much of the lateral edge of the frontal. Thesupraorbital canal runs in a prominently raised, bony tubefrom near the lateral corner about halfway to the midline,then bends and runs anteriorly, more or less parallel to themidline. The anterior end is obscured. There appears to beone short, laterally directed posterior branch; and there areat least two short, side-by-side, antero-medially directedbranches with open pores at the anterior end. These branchoff at roughly the point where the supraorbital canal makesits anterior bend. The morphology matches closely withCoregonus, particularly C. muksun and C. lavaretus as pic-tured in Shaposhnikova (1970).

Circumorbital series (Fig. 7a)The circumorbitals were determined with the aid of a la-

tex peel of the right-side impression of the fossil. Six bonesof the circumorbital series are present; the antorbital and in-fraorbitals 1 (lachrymal), 2, 3, 4, and 5. The area on the fos-sil along the anterior dorsal border of the orbit is broken andpartially obscured by crystal growth, and the supraorbital is

not readily identifiable. Enough structure exists along thefrontal anterior to the position of the missing dermosphe-notic to identify a gap in the circumorbital ring, proposedby Norden (1961) as diagnostic for Coregonus. The infraor-bitals are characterized by the infraorbital sensory canal sys-tem that curves around the ventral margin of the orbit; itclosely resembles the same structure in the extant taxa Cor-egonus nasus, C. lavaretus, and C. clupeaformis. The antor-bital is a narrow, rounded element closely fitted to thedorsal border of the lachrymal. It does not bear a sensorycanal. The first infraorbital (lachrymal) is the best-preserved.Overall, it is Y-shaped, expanding anteriorly to a broaddorso-anterior margin, with the short dorsally directed wingand the longer anteriorly directed wing forming an angle of*1008. The posterior end of the element narrows where itreceives the sensory canal from the second infraorbital. Theraised sensory canal appears to have three short, ventrallydirected branches in addition to openings at the anterior andposterior ends of the element. Infraorbitals 2 and 3 are rec-tangular in shape and not as complex as the lachrymal, witha somewhat sinuous, raised sensory canal running the lengthof each element close to the dorsal margin. Infraorbital 3 isconsiderably broadened ventrally relative to infraorbital 2.Infraorbitals 4 and 5 are broader yet and expand posteriorlybehind the eye. There is a damaged area of crystal growthbetween infraorbital 5 and the frontal; the dermosphenoticis missing. In number, position, and general characteristics,the infraorbital series matches well with the description byCavender (1970) of these elements in coregonines.

Opercular series (Fig. 7a)About 60% of the length of the preopercle is the vertical

ramus, which tapers gradually dorsally. The horizontal ra-mus extends anteriorly, forming an angle of *1058 withthe vertical. The preopercle has five strong, raised preoper-cular sensory canal branches, four on the horizontal ramusand one on the vertical, just dorsal to the change in angle.Overall, the canal descends the anterior margin of the verti-cal ramus and follows the bend of the preopercle to its ante-rior tip. Of the coregonines examined, the preopercle mostclosely resembles the same element in the extant taxa Core-gonus pidschian and C. lavaretus.

The opercle approaches a semicircle in overall shape butis distally broadened. The anterior margin is relativelystraight but angles *158 anteriorly above the socket of ar-ticulation with the hyomandibular. It is closest to C. nasusin overall morphology of the Coregonus species examined.Three prominent growth rings are easily observable towardthe posterior margin of the opercle and the subopercle. Thesubopercle is a long, narrow, subrectangular bone extendingalong the ventral border of the opercle from its most poste-rior point to the anteroventral margin of the opercle. The in-teropercle is present but largely obscured.

Jaws and suspensorium (Fig. 7a)This area is poorly preserved in the fossil. Only a portion

of the quadrate and symplectic is visible. The symplectic isrodlike, inserting into the notch between the limbs of thequadrate and widening at its distal end. The fan-shaped ante-rior limb of the quadrate is largely preserved, with a distinc-tive, radiating, fanlike canal and pore complex covering

228 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

most of the ventral portion of the lateral surface. This fea-ture was not seen on the extant Coregonus specimens exam-ined, with the exception of C. clupeaformis, which has avery small, similar patch just posterior to the knobby con-dyles, which articulate with the angular.

The mouth of the specimen is relatively small and doesnot extend as far as the eye. Examination of the latex peeltaken of the fossil shows that the point of the ‘‘v’’ for theinsertion of the articular on the posterior margin of the den-tary is anterior to the orbit. No teeth are visible; however,only a small portion of the dentary and maxilla can be seen;the posterior portion of the maxilla is broken off and miss-ing. The premaxilla appears to be missing. It is difficult todetermine mouth position. The mouth appears terminal (an-terior ends of the maxilla and dentary articulating andequally far forward), but comparisons of the specimen withnaturally articulated skeletons of several Coregonus taxashow that the position of the maxilla and dentary on the fos-sil specimen is consistent with a subterminal mouth position(an overbite; with mouth closed the dentary tucks in behindthe premaxilla). This is consistent with extant Coregonusspecimens. Details of the angular and articular complex can-not be discerned without additional preparation.

Comparison with extant speciesEleven measurements and three counts were made on

both the fossil and 115 specimens representing six speciesof extant Beringian coregonines; Coregonus autumnalis, C.clupeaformis, C. nasus, C. pidschian, C. lavaretus, and C.tugun (see Appendix A, Comparative materials). Measure-ments and counts were limited to those that could unequivo-cally be taken from the fossil specimen.

Results of comparison with extant Coregonus speciesAll raw data, except dorsal height and caudal peduncle

height (p = 0.104 and 0.255, respectively), were normallydistributed. Log transformation resulted in all morphometricvariables having a normal distribution and being homosce-dastic between species. Dorsal ray counts aligned the fossilspecimen with all other whitefish species, except C. tugunand C. nasus (Fig. 8). Lateral line scale counts, however,grouped the fossil specimen with C. tugun (Fig. 9). Countsfor scales above the lateral line distinguished C. nasus andC. tugun from all other species (Fig. 10).

Four factors produced by the PCA accounted for 76.3% ofthe variation in the dataset. The classification rate of theDFA was 81%, with a jacknifed validation rate of 81%.The multivariate group means were highly significantly dif-ferent using Wilks’ l, a general test statistic used in multi-variate tests of mean differences among more than twogroups (Wilks’ l p = 0.0000). The closest centroid was thatof C. pidschian in multivariate morphospace (Fig. 11), butthose of C. lavaretus and C. clupeaformis were almostequally close. Note that C. tugun was not plotted in Fig. 11,as head depth was not recorded and thus the four points rep-resenting C. tugun were omitted in the DFA.

DiscussionAnalysis of univariate data shows that the fossil specimen

is distinguishable from extant species of Coregonus by the

following combination of characters: more than 11 dorsalfin rays, fewer than 12 scales above the lateral line, andfewer than 70 lateral line scales (see Table 2 for ranges ofcounts). Head depth in standard length is >10.5. Dorsal raycounts separate the fossil specimen from C. tugun and C.nasus, whereas C. tugun and the fossil specimen form a dis-tinct group for lateral line scale counts. C. tugun has a nota-bly higher count for scales above the lateral line.

Fig. 8. Character distribution for dorsal rays. Error bars representone standard error.

Fig. 9. Character distribution for lateral line (LL) scales. Error barsrepresent one standard error.

Cumbaa et al. 229

Published by NRC Research Press

Multivariate analysis and classification confirmed that thefossil specimen was close to, but not a part of, the C. clu-peaformis complex. Results of the DFA show that the fossilspecimen (Fig. 11) was closest to the centroid of C. pid-schian but is almost equally close to the centroids of C. lav-aretus and C. clupeaformis. However, numerous individualcharacters as listed previously demonstrate that it is not con-specific with currently known members of that complex.

The fossil represents a new coregonine species as demon-strated by a combination of multi- and uni-variate analyses.It is distinct from C. tugun in scales above lateral line(Fig. 10) and dorsal rays (Fig. 8). The specimen is clearlydistinguishable from both C. clupeaformis and C. pidschian,in both lateral line scales and head depth as a ratio of SL(Table 2; Fig. 12). Head depth appears to be a valid charac-ter, as there is no distortion apparent on examination of thefossil, or as seen in a plot of head depth against standardlength in coregonine fishes (Fig. 13). In fact, the fossilspecimen is just out of the extreme range of lateral linescales for the C. clupeaformis complex (Fig. 9; Table 2),which, as currently delineated, has 70–97 lateral line scales(Scott and Crossman 1973). The fossil specimen is distinctfrom extant coregonines and constitutes a new species usingthe phylogenetic species concept (Mayr and Ashlock 1991).This phenetic study has demonstrated the unique charactercombinations of the fossil specimen.

Osteological characters unique to this specimen include adistinctive, Y-branched lachrymal; a quadrate with a radiat-ing, fanlike canal and pore complex covering most of theventral portion of the lateral surface; and a preopercle withfive, prominent, raised preopercular canal branches.

PaleoenvironmentA palynological examination of the concretion in which

the fossil specimen is enclosed was felt to be the most direct

way of assessing the paleoenvironment of the whitefish, orat least its burial environment. In general, the pollen contentof the concretion is very low, and preservation of individualgrains poor. Nevertheless, useful conclusions may be drawnfrom the results (Table 3).

Betula (birch) pollen is relatively abundant in the assem-blage. Most grains appear to be of the small, shrub-birchtype, although a few are of the larger tree-birch type. Myr-ica, possibly M. gale (sweet gale), is a minor constituent.Other shrubs represented are Alnus (alder) and Salix

Fig. 10. Character distribution for scales above the lateral line (LL).Error bars represent one standard error.

Fig. 11. Discriminant score 1 plotted against discriminant score 2for five extant and one fossil coregonine species. Specimens of C.autumnalis are represented by an open circle, those of C. clupea-formis by an ‘‘�’’, C. lavaretus by a plus sign, C. nasus by a trian-gle, C. pidschian by an inverted triangle, and the fossil specimen bya solid left-pointing arrow head (close to the centre of the upperright quadrant).

Fig. 12. Head depth as a ratio (HDR) of standard length (SL) forthe fossil specimen and five coregonine species plotted against SL.Specimens of C. autumnalis are represented by an open circle,those of C. clupeaformis by a diamond, C. lavaretus by a plus sign,C. nasus by an ‘‘�’’, C. pidschian by a triangle, and the fossil spe-cimen by a filled circle.

230 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

(willow). Among the herbs represented, the most abundantare Artemisia (sage), Rosaceae (rose family), Gramineae(grasses), and Cyperaceae (sedges). Tree taxa are poorlyrepresented, with only minor amounts of Picea (spruce) andPinus (pine). The poor condition of the tree pollen and theirform suggest that some reworking of older palynomorphs isinvolved, particularly some of the abundant undifferentiatedtrilete forms. Fern spores representing several taxa are abun-dant. Mosses, if the form identified as Bryophytes is correct,are very abundant in the assemblage.

If we assume that most of the abovementioned assem-blage is representative of the environment of the area at thetime of deposition, some inferences can be made. Openshrub tundra probably characterized the landscape, withbirch, alder, and willows the most abundant shrubs. Treeswere absent or sparse. Sage and several other herbs werepresent along with some grasses and sedges. Ferns appearto have been abundant at least locally where suitable condi-tions prevailed. Areas of bog or fen with mosses, includingSphagnum, were probably widespread. If trees were absentor sparse and shrub tundra characterized the area, the cli-mate was colder than at present and was similar to condi-tions prevailing today at or above the tree line.

When compared with the pollen profile of a section (Por-cupine River 1) along Ch’ijee’s Bluff reported by Lichti-Federovich (1974), the closest correlation is with the as-semblages near the base of unit 3. However, fern spores aremuch more highly concentrated in the concretion reportedhere than anywhere in Lichti-Federovich’s profile. Her gen-eral interpretation of the environment of unit 3 (her pollenassemblage type Vb) was of an open pine–birch–spruce for-est with herb-dominated treeless communities. We suggestthis interpretation may apply to the upper part of unit 3, butnot necessarily to the base of the unit, where Picea and Pi-nus are much less abundant, and the assemblage is more in-dicative of a shrub-tundra environment.

Ritchie (1984) and Matthews et al. (1990) updated Lichti-Federovich’s stratigraphy but showed somewhat differing in-terpretations for the base of unit 3, with higher numbers forpine and spruce. Our examination of Lichti-Fedrovich’s(1974) fig. 2 shows that the lowest samples from her unit3 equivalent were taken *2 m above the base of unit 3 andits concretionary layer (her 47’ level). Those lowest unit 3samples are dominated by shrubs, with birch the most abun-

Table 2. Mean and range values for three counts on Coregonus species.

Lateral line scales Dorsal rays Scales above lateral line

C. autumnalis 96 (82–110) 11 (10–12) 10.8 (9–12)C. clupeaformis 83.5 (70–97) 12.0 (11–13) 10.5 (9–12)C. nasus 93 (84–102) 11.5 (10–13) 11.6 (10–12)C. pidschian 86 (77–95) 12 (11–13) 10.3 (9–11)C. tugun 68.3 (65–72) 9.8 (9–10) 14.3 (14–15)C. lavaretus 89.5 (79–100) 10.5 (9–12) 10.5 (10–11)Fossil 69 12 11

Note: Lateral line and dorsal ray counts obtained from published sources except C. tugun (ourcounts). Scales above lateral line data also from specimens examined by the authors, as these areonly sporadically found in the literature.

Fig. 13. Head depth plotted against standard length (SL) for thefossil specimen and five coregonine species. Specimens of C. au-tumnalis are represented by an open circle, those of C. clupeaformisby a diamond, C. lavaretus by a plus sign, C. nasus by an ‘‘�’’, C.pidschian by a triangle, and the fossil specimen by a filled circle.

Table 3. Pollen counts (%) in the whitefish concretion.

Trees and (or) shrubsPicea 2.6Pinus 4.6Indeterminate saccates 3.3Betula type 28.8Myrica type 1.3Alnus 9.1Salix 3.9

HerbsArtemisia 10.5Compositae 0.7Ericaceae 0.7Gramineae 3.3Chenopodiaceae 1.3Rosaceae 3.3Ranunculaceae 1.3Cyperaceae 4.6Unidentified non-arboreal pollen 21.6

Ferns and mossesPolypodiaceae 5.2Trilete spores 49.8Sphagnum 13.7Lycopodiaceae 0.7Bryophyte spores? +++

Cumbaa et al. 231

Published by NRC Research Press

dant (45%), followed by alder, herbs, and willow (together15%). Pine and spruce trees together make up only 15% ofher sample. Bryophytes, grasses, and sedges together (25%)make up the rest of the flora. This composition is fairly closeto that of the fish-bearing concretion, with the exception ofthe abundance of fern spores in the concretion (Table 3).The palynology at the base of unit 3 indicates a shrub-tundraenvironment, with conditions similar to those today at ornear the tree line. White et al. (1999) noted that boreal forestand tundra floristic elements were in place in the region by2.3 Ma and pine was becoming scarce. Fission-track and pa-leomagnetic data (Froese et al. 2000; Westgate et al. 2003)reinforce that pine was abundant in the region until *3 Maand seems to be largely absent by the early Pleistocene. Thisearly Pleistocene cooling trend is consistent with our obser-vations from the basal unit 3 concretion.

Depositional environment and paleodrainageThe results of sedimentological analyses of the unit 3 con-

cretion allow further characterization of the postdepositionaland early diagenetic environment in which the fish was fos-silized. The sediment was silty of lacustrine origin, rich inorganic matter preserved under anoxic conditions, and witha pH *7.5. The presence of vivianite nodules suggests thatsedimentation was largely controlled by stratification of thewater column and anoxic bottom waters.

Matthews et al. (1990) suggested that the lake associatedwith unit 3 was possibly created by movement of one of themany faults that cross the paleo-Porcupine River system(Norris 1981). Certainly, it does not appear to be of glacio-lacustrine origin. This hypothesis is supported by the factthat there are significant differences between the unit 3 lakesediments and the higher unit 5 sediments associated withGlacial Lake Old Crow (Morlan 1980), which is poor in or-ganic material (1%–3%) and especially poor in pollen(Pearce et al. 1982).

The lake recorded by unit 3 possibly extended to both theOld Crow and Bluefish basins and was part of the paleo-Porcupine drainage system. Until the end of the Late Pleis-tocene, the Porcupine apparently flowed through the Ri-chardson Mountains, via McDougall Pass, to the BeaufortSea (Hughes 1969, 1972; Morlan 1980; Duk-Rodkin andHughes 1994). A Beaufort Sea connection to the unit 3 lakeis a very interesting possibility, given that some extant spe-cies of Coregonus (e.g., C. autumnalis, C. clupeaformis, andC. nasus) are diadromous, known to move between brackishor more fully marine waters and fresh water (Scott andCrossman 1973). Nevertheless, proof to include the OldCrow and Bluefish basins in this drainage system is scanty.Tectonic movements and differential erosion have consider-ably modified the landscape of northern Yukon Territorysince unit 3 was formed.

ConclusionsThe results of the morphological and statistical analyses

of the fossil Coregonus demonstrate that it represents a newspecies, broadly similar to, but separate from, the extant lakewhitefishes of the Coregonus clupeaformis complex.

The Ch’ijee’s Bluff whitefish, Coregonus beringiaensis sp.nov., could be as much as two million years older than the

specimens from Poland on the basis of magnetostratigraphy,paleoclimatic reconstruction, and associated fauna and flora.As such, it represents the oldest known representative of thegenus Coregonus. Specifically, we suggest an early Quater-nary (probably the Gelasian of the Pleistocene) age for thisspecimen based on its position below the reversed interval ofthe Matuyama Chron and shrub-tundra-inferred paleoenvir-onmental conditions suggested by the palynological data.

The interpreted paleoenvironment plus the apparent nor-mal polarity of sediment at the base of unit 3 suggest thestrata encompassing the whitefish fossil may have been de-posited during a cold climatic stage represented by the Old-uvai subchron (ca. 1.8 Ma; Barendregt et al. 1998). Plantmacrofossils from the underlying unit 1 at Ch’ijee’s Bluffare placed by McKown et al. (2002) between 5 and 2.3 Ma(Pliocene (Zanclean) to early Pleistocene (Gelasian)). Mat-thews and Ovenden (1990) gave a date for unit 2 based onplant macrofossil correlations that would place the top ofthis unit in the early Quaternary (*2 Ma). We presumethat the base of unit 3 is close in age to the top of unit 2.

Reshetnikov (1995) argued that the center of origin ofCoregonus coincides with its center of dispersal and centerof speciation, most likely in eastern Siberia. However, thepresence of C. beringiaensis sp. nov., to date the earliestknown Coregonus, near the eastern margin of Beringia justbefore the onset of full glaciation (and the development ofthe Bering Isthmus and freshwater links to Asia), has ob-vious implications for the evolution and radiation of this im-portant salmonid subfamily. Coregonus could certainly haveevolved in North America and spread later to Eurasia via theBering Isthmus. Perhaps more likely, the earliest Coregonus,like the Miocene to recent coregonine genus Stenodus, wasdiadromous. Stenodus is known from the Miocene of Siberiaand the Pleistocene of Yukon Territory. Faunal exchange ofdiadromous fish between Siberia and North America wouldhave been relatively straightforward given access via the pa-leo-Porcupine River system and a paleodrainage reconstruc-tion linking it to the Beaufort Sea.

AcknowledgementsThis study was supported by a Natural Sciences and Engi-

neering Research Council of Canada (NSERC) grant 7995-05to BL and by a Northern Scientific Training Program (NSTP)grant to MR. Logistical support was provided by J. Cinq-Marsof the Canadian Museum of Civilization, Gatineau, Quebec,and by E. Deschamps, whose support in the field was pro-vided in part by NSTP. R. Gotthardt, Yukon archaeologist,Yukon Tourism and Culture, Whitehorse, Yukon, providedlogistical support and took the photographs of the fossil speci-men. P. Bertrand of University of Ottawa, Ottawa, Ontario,drew the map. J. Storer, Yukon paleontologist, assisted inlogistics and had peels and a cast made for us; F. Jurak pre-pared the fossil. M. Lamothe of l’Universite du Quebec aMontreal, Montreal, Quebec, and C. Schweger of Universityof Alberta, Calgary, Alberta, provided insights on the biostra-tigraphy and sedimentology of Ch’ijee’s Bluff. G. Smith ofUniversity of Michigan, Ann Arbor, Michigan, graciouslyallowed access to preserved specimens of Siberian whitefishto augment those examined from Canada and Alaska in thecollections of the Canadian Museum of Nature, Ottawa,Ontario. B. Coad and C. Renaud of the Canadian Museum of

232 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

Nature confirmed some meristics and measurements on thefossil specimen. We thank referees D.G. Froese of Universityof Alberta and G.R. Smith of University of Michigan for theirthoughtful comments and suggested improvements and C.R.Harington of Canadian Museum of Nature and G. Zazula, Yu-kon paleontologist, Yukon Tourism and Culture, for their con-structive suggestions on earlier versions of this manuscript.Canadian Journal of Earth Sciences Associate Editor Hans-Dieter Sues and Astrid Blodgett of the Editorial Office helpedat several stages in getting this paper published.

ReferencesBarendregt, R.W., Vincent, J.-S., Irving, E., and Baker, J. 1998.

Magnetostratigraphy of Quaternary and late Tertiary sedimentson Banks Island, Canadian Arctic Archipelago. Canadian Journalof Earth Sciences, 35(2): 147–161. doi:10.1139/cjes-35-2-147.

Bjornson, J., and Lauriol, B. 2001. Meteorisation des blocs rocheuxsur les pediments dans le nord Yukon. Permafrost and Perigla-cial Processes, 12: 289–298. doi:10.1002/ppp.391.

Cande, S.C., and Kent, D.V. 1995. Revised calibration of the geo-magnetic polarity timescale for the Late Cretaceous and Ceno-zoic. Journal of Geophysical Research, 100(B4): 6093–6095.doi:10.1029/94JB03098.

Cavender, T.M. 1970. A comparison of coregonines and other sal-monids with the earliest known teleostean fishes. In Biology ofcoregonid fishes. Edited by C.C. Lindsey and C.S. Woods. Uni-versity of Manitoba Press, Winnipeg, Man., pp. 1–32.

Cumbaa, S.L., McAllister, D.E., and Morlan, R.E. 1981. Late Pleis-tocene fish fossils of Coregonus, Stenodus, Thymallus, Catosto-mus, Lota and Cottus from the Old Crow basin, northern Yukon,Canada. Canadian Journal of Earth Sciences, 18(11): 1740–1754. doi:10.1139/e81-159.

Cumbaa, S.L., Lauriol, B., Alfonso, N., and Mott, R.J. 2002. Anew whitefish from the early Pleistocene of Bluefish Basin, Yu-kon Territory, Canada. Journal of Vertebrate Paleontology, 22:47A.

Delorme, L.D. 1968. Pleistocene freshwater Ostracoda from Yu-kon, Canada. Canadian Journal of Zoology, 46(5): 859–876.doi:10.1139/z68-123.

Duk-Rodkin, A., and Hughes, O.L. 1994. Tertiary–Quaternary drai-nage of the preglacial Mackenzie Basin. Quaternary Interna-tional, 22/23: 221–241. doi:10.1016/1040-6182(94)90015-9.

Froese, D.G., Barendregt, R.W., Enkin, R.J., and Baker, J. 2000.Paleomagnetic evidence for multiple late Pliocene – early Pleis-tocene glaciations in the Klondike area, Yukon Territory. Cana-dian Journal of Earth Sciences, 37(6): 863–877. doi:10.1139/cjes-37-6-863.

Harington, C.R. 1977. Pleistocene mammals of Yukon Territory.Ph.D. thesis, University of Alberta, Edmonton, Alta.

Harington, C.R. 1978. Quaternary vertebrate faunas of Canada andAlaska and their suggested chronological sequence. NationalMuseums of Canada, Ottawa, Ont. Syllogeus, No. 15, pp. 1–105.

Harington, C.R. 2005. The eastern limit of Beringia: Mammoth re-mains from Banks and Melville Islands, Northwest Territories.Arctic, 58: 361–369.

Hubbs, C.L., and Lagler, K.F. 1964. Fishes of the Great Lakes Re-gion. Ambassador Books Ltd., Toronto, Ont.

Hughes, O. 1969. Pleistocene stratigraphy, Porcupine and Old CrowRivers, Yukon Territory. Geological Survey of Canada, Paper 69-1,pp. 209–212.

Hughes, O. 1972. Surficial geology of northern Yukon and north-eastern District of Mackenzie, Northwest Territories. GeologicalSurvey of Canada, Paper 64-36, 11 p (includes map 1319A).

International Commission on Stratigraphy. 2009. International Stra-tigraphic Chart. International Union of Geological Sciences,Trondheim, Norway.

Jerzmanska, A., and Raczynski, P. 1991. Middle Pleistocene fresh-water fishes from the interglacial deposits near Bełchatow, Po-land. Folia Quaternaria, 61/62: 109–123 (includes 6 plates).

Johnson, R.G. 1982. Brunhes-Matuyama magnetic reversal dated at790 000 BP by marine-astronomical correlations. QuaternaryResearch, 17(2): 135–147. doi:10.1016/0033-5894(82)90055-2.

Jordan, D.S. 1878. Manual of the vertebrates of the northern UnitedStates, including the district east of the Mississippi River andnorth of North Carolina and Tennessee, exclusive of marine spe-cies. 2nd ed. Jansen, McClurg & company, Chicago, Ill.

Klein, E.F. 1885. Beitrage zur Bildung des Schadels der Knochen-fische, ii. Jahreshefte Vereins vaterlandische in NaturkundeWurttembergs, 61: 107–261.

Koelz, W. 1929. Coregonid fishes of the Great Lakes. UnitedStates Bureau of Fisheries, Bulletin 43, pp. 297–643.

Lichti-Federovich, S. 1974. Palynology of two sections of LateQuaternary sediments from the Porcupine River, Yukon Terri-tory. Geological Survey of Canada, Paper 74-23.

Linnaeus, C. 1758. Systema Naturae, Xth ed. (Systema naturae perregna tria naturae, secundum classes, ordines, genera, species,cum characteribus, differentiis, synonymis, locis. Tomus I. Edi-tio decima, reformata.) Holmiae. Systema Nat. 10th ed. Vol. 1,pp. i–ii and 1–824. [Nantes and Pisces in Tom. 1, pp. 230–338;a few species on later pages. Date fixed by ICZN, Code Article3.]

Matthews, J.R., Jr. 1975. Insects and plant macrofossils from twoQuaternary exposures in the Old Crow – Porcupine region, Yu-kon Territory, Canada. Arctic and Alpine Research, 7(3): 249–259. doi:10.2307/1550000.

Matthews, J.V., and Ovenden, L.E. 1990. Late Tertiary plantmacrofossils from localities in arctic/subarctic North America: areview of the data. Arctic, 43(4): 364–392.

Matthews, J.V., Jr., Schweger, C.E., and Janssens, J.A. 1990. Thelast (Koy-Yukon) interglaciation in the northern Yukon: evi-dence from Unit 4 at Ch’ijee’s Bluff, Bluefish Basin. Geogra-phie physique et quaternaire, 44: 341–362.

Matthews, J.V., Jr., Westgate, J.A., Ovenden, L., Carter, L.D., andFouch, T. 2003. Stratigraphy, fossils, and age of sediments atthe upper pit of the Lost Chicken gold mine: new informationon the late Pliocene environment of east central Alaska. Qua-ternary Research, 60(1): 9–18. doi:10.1016/S0033-5894(03)00087-5.

Mayr, E., and Ashlock, P.D. 1991. Principles of Systematic Zool-ogy. McGraw-Hill, New York.

McConnell, R.G. 1891. Report on an exploration in the Yukon andMackenzie Basins. Geological Survey of Canada, Annual Report4, pp. 5D–163D.

McKown, A.D., Stockey, R.A., and Schweger, C.E. 2002. A newspecies of Pinus subgenus Pinus subsection Contortae fromPliocene sediments of Ch’ijee’s Bluff, Yukon Territory, Canada.International Journal of Plant Sciences, 163(4): 687–697. doi:10.1086/340425.

McPhail, J.D., and Lindsey, C.C. 1970. Freshwater fishes of north-western Canada and Alaska. Fisheries Research Board of Ca-nada, Bulletin 173, 381 p.

Minyuk, P.S., Brigham-Grette, J., Melles, M., Borkhodoev, V.Ya.,and Glushkova, O.Yu. 2007. Inorganic geochemistry of El’gy-gytgyn Lake sediments (northeastern Russia) as an indicator ofpaleoclimatic change for the last 250 kyr. Journal of Paleolim-nology, 37(1): 123–133. doi:10.1007/s10933-006-9027-4.

Mojski, J.E. 1995. Pleistocene glacial events in Poland. In Glacial

Cumbaa et al. 233

Published by NRC Research Press

deposits in north-east Europe. Edited by J. Ehlers, S. Kozarski,and P. Gibbard. A.A. Balkema, Rotterdam, the Netherlands,pp. 287–292.

Morlan, R.E. 1980. Taphonomy and Archaeology in the UpperPleistocene of the northern Yukon Territory: a glimpse of thepeopling of the New World. Ottawa National Museum of ManMercury (Mercury Series). Archaeological Survey of Canada,Paper 94, 380 p.

Muller, J. 1846. Uber den Bau und die Grenzen der Ganoiden unduber das naturliche System der Fische. Abhandlungen der Koni-glichen Akademie der Wissenschaften zu Berlin, pp. 117–216.

Nelson, J.S. 2006. Fishes of the World. Hoboken. 4th ed. John Wi-ley & Sons, Inc., N.J.

Norden, C.R. 1961. Comparative osteology of representative sal-monid fishes, with particular reference to the grayling (Thy-mallus arcticus) and its phylogeny. Journal of the FisheriesResearch Board of Canada, 18: 679–791.

Norris, D.K. 1981. Geology: Old Crow, Yukon Territory. Geologi-cal Survey of Canada, Ottawa. Series A, map 1518A, scale1 : 250 000.

Nriagu, J.O., and Dell, C.I. 1974. Diagenic formation of iron phos-phates in recent lake sediments. The American Mineralogist, 59:934–946.

Pearce, G.W., Westgate, J.A., and Roberston, S. 1982. Magnetic re-versal history of Pleistocene sediments at Old Crow, northwes-tern Yukon Territory. Canadian Journal of Earth Sciences, 19:919–929. doi:10.1139/e82-077.

Reist, J.D. 1985. An empirical evaluation of several univariatemethods that adjust for size variation in morphometric data. Ca-nadian Journal of Zoology, 63(6): 1429–1439. doi:10.1139/z85-213.

Reist, J.D. 1986. An empirical evaluation of coefficients used in re-sidual and allometric adjustment of size covariation. CanadianJournal of Zoology, 64(6): 1363–1368. doi:10.1139/z86-203.

Reshetnikov, Y.S. 1995. A review of systematics and ecology ofcoregonid fishes. Journal of Ichthyology, 35: 105–135.

Richardson, J. 1836. Fishes. In Narrative of the Arctic land expedi-tion to the mouth of the Great Fish River, and along the shoresof the Arctic Ocean in the years 1833, 1834, and 1835. Editedby G. Back. London. Narrative Arctic Land, pp. i–x and 1–663,14 pls.

Ritchie, J.C. 1984. Past and Present Vegetation of the Far North-west of Canada. University of Toronto Press, Toronto, Ont.

Rosenqvist, I.Th. 1970. Formation of vivianite in Holocene clay se-diments. Lithos, 3(4): 327–334. doi:10.1016/0024-4937(70)90039-3.

Ross, M. 1997. Lithostratigraphie et geochronologie du Pleistocenede la coupe Ch’ijee’s (station 6) dans le bassin de Bluefish, Yu-kon septentrional: implications pour la paleogeographie regional.These de maıtrise es sciences en geologie. Universite du Quebeca Montreal, Montreal, Que., 125 p.

Rzechowski, J. 1996. The Ferdynandowian Interglacial and its stra-tigraphical position in the Middle Pleistocene of Europe. In Theearly Middle Pleistocene in Europe. Edited by C. Turner. A.A.Balkema, Rotterdam, the Netherlands, pp. 279–293.

Scott, W.B., and Crossman, E.J. 1973. Freshwater fishes of Ca-nada. Fisheries Research Board of Canada, Bulletin 184, 966 p.

Shaposhnikova, G.Kh. 1970. On the taxonomy of whitefishes fromthe USSR. In Biology of coregonid fishes. Edited by C.C. Lind-sey and C.S. Woods. University of Manitoba Press, Winnipeg,Man., pp. 195–207.

Smith, G.R. 1975. Fishes of the Pliocene Glenns Ferry Formation,southwest Idaho. University of Michigan Museum of Paleontol-ogy, Ann Arbor, Mich., Papers on Paleontology No. 14, pp. 1–68.

Smith, G.R. 1981. Late Cenozoic freshwater fishes of North Amer-ica. Annual Review of Ecology and Systematics, 12(1): 163–193. doi:10.1146/annurev.es.12.110181.001115.

Smith, G.R., and Todd, T.N. 1992. Morphological cladistic study ofcoregonine fishes. In Biology and management of coregonidfishes, Proceedings of the 4th International Symposium,29 April – 3 May 1991, Wuhan, China. Edited by T.N. Toddand M. Luczynski. Polskie Archiwum Hydrobiologii 39, pp.479–490.

Westgate, J.A., Walter, R.C., Pearce, G.W., and Gorton, M.P. 1985.Distribution, stratigraphy, petrochemistry, and paleomagnetismof the Late Pleistocene Old Crow tephra in Alaska and the Yu-kon. Canadian Journal of Earth Sciences, 22: 893–906. doi:10.1139/e85-093.

Westgate, J.A., Sandhu, A.S., Preece, S.J., and Froese, D.G. 2003.Age of the gold-bearing White Channel gravel of the Klondikegoldfields. In Yukon exploration and geology 2002. Edited byD.S. Edmond and L.L. Lewis. Exploration and Geological Ser-vices Division, Yukon Region, Indian and Northern Affairs Ca-nada, pp. 241–250.

White, J.M., Ager, T.A., Adam, D.P., Leopold, E.B., Lin, G., Jette,H., and Schweger, C.E. 1999. Neogene and Quaternary quantita-tive palynostratigraphy and paleoclimatology from sections inYukon and adjacent Northwest Territories and Alaska. Geologi-cal Survey of Canada, Bulletin 543, 30 p.

Zelibor, J.L., Jr., Senftle, F.E., and Reinhardt, J.L. 1988. A pro-posed mechanism for the formation of spherical vivianite crystalaggregates in sediments. Sedimentary Geology, 59(1–2): 125–142. doi:10.1016/0037-0738(88)90103-0.

Appendix A. Comparative materialCoregonus autumnalis

Canadian Museum of Nature (CMN) CMNFI1958-0131;Northwest Territories (NWT), Cape Bathurst southwest ofBank Island; 6:295.0–341.1 mm standard length (SL).CMNFI1964-0300; NWT, Liverpool Bay, Baillie Island;3:302.2–384.3 mm SL. CMNFI1968-0416; NWT, mouth ofMackenzie River; 3:314.0–341.9 mm SL. CMNFI1968-1234; Yukon Territory (YT), Peel River, at mouth of Cari-bou River; 4:301.2–315.8 mm SL. CMNFI1970-0239;NWT, Tuktoyaktuk Island; 6:307.4–342.6 mm SL.CMNFI1975-1531; NWT, Nauyuk Lake; 6:233.2–343.6 mmSL. CMNFI1991-0120; NWT, Langton Bay; 2:275.0–301.8 mm SL.

C. clupeaformisCMNFI1964-125; Keller Lake, south of Great Bear Lake;

3:462.6–506.6 mm SL. CMNFI1968-1168; Wholdaia Lake,NWT; 1 of 4:323.2 mm SL. CMNFI1968-1775; Steve’sLake, Temlin Lake chain, NWT; 9:318.5–359.2 mm SL.CMNFI1970-0236; Tuktoyaktuk; 2:292.3–315.5 mm SL.CMNFI1972-151; Kugmallituk Bay, Tuktoyaktuk;1:269.0 mm SL. CMNFI1980-0498; King Point Harbour,NWT; 7:281.0–311.0 mm SL. CMNFI1980-1175; SurreyLake, Victoria Island; 1:355.0 mm SL.

C. lavaretus

Royal Ontario Museum (ROM) 25694; Sweden, Lake Ids-jon, 8 of 50:107.3–126.7 mm SL.

234 Can. J. Earth Sci. Vol. 47, 2010

Published by NRC Research Press

C. nasusCMNFI1964-0003; YT, Mackenzie Bay, Shingle Point;

2:280.7–302.6 mm SL. CMNFI1968-0410; YT, Teslin Lake;4:160.8–331.4 mm SL. CMNFI1968-0921; Alaska (AK), It-killik River; 2:188.5–307.2 mm SL. CMNFI1968-1244; AK,Takotna River; 4:280.0–305.9 mm SL. CMNFI1968-1270;AK, Andreafsky River; 8:294.8–408.8 mm SL.CMNFI1968-1775; NWT, Steve’s Lake; 1:368.7 mm SL.CMNFI1970-0231; NWT, Tuktoyaktuk Harbour; 2:304.6–307.8 mm SL. CMNFI1970-0235; NWT, Tuktoyaktuk Har-bour; 1:396.2 mm SL. CMNFI1970-234; NWT, TuktoyaktukHarbour; 3:292.9–328.4 mm SL. CMNFI1983-0088; NWT,Eskimo Lakes; 1:407.3 mm SL. CMNFI1983-009; NWT,Lake Laberge; 2:301.8–317.0 mm SL.

C. pidschianCMNFI1968-1255; AK, Kuzitrin River, North of Nome;

10:247.3–340.2 mm SL. CMNFI1968-1517; AK, NaknakLake, Aleutian Range; 1:434.2 mm SL. CMNFI1968-1520;AK, Colville Lake, Bristol Bay; 2:442.7–459.7 mm SL.CMNFI1968-1740; AK, Aloknagig Lake; 2:289.8–296.0 mm SL. CMNFI1979-0544; AK, Okpiksak, 65 kmsouth of Barrow; 1:293.9 mm SL.

C. tugunUniversity of Michigan Museum of Zoology (UMMZ)

218954; Siberia, tributary of Ob River; 4:95.0–102.0 mmSL.

Cumbaa et al. 235

Published by NRC Research Press