FIELDIANA
Geology

Published by Field Museum of Natural History

VOLUME 26, No. 1

ARCTOID GENETIC CHARACTERS
AS RELATED TO THE
GENUS PARICTIS

JOHN CLARK

and

THOMAS E. GUENSBURG

JUNE 20, 1972

FIELDIANA
Geology

Published by Field Museum of Natural History

VOLUME 26, No. 1

ARCTOID GENETIC CHARACTERS
AS RELATED TO THE
GENUS PARICTIS

JOHN CLARK

Associate Curator, Sedimentary Petrology
Field Museum of Natural History

and

THOMAS E. GUENSBURG

University of Illinois

JUNE 20, 1972

PUBUCATION 1160

I. TAXONOMIC HISTORY

The genus Parictis has enjoyed a peculiarly little-complicated
taxonomic history, due perhaps to the rarity of specimens and the
restriction of its geographic and geologic range.

Scott (1893) first described Parictis primaevus, basing the type
species on Princeton Museum no. 10583, a partial left ramus mandi-
buli with P2 3 entire, and the alveoli of the other cheek teeth. The
locality is given as "John Day beds at Silver Wells, Oregon." The
specimen label gives "Camp Creek" as locality. In accordance with
custom of the time, Scott regarded the age as Miocene; he referred
the genus to the Mustelidae.

Hall (1931) redescribed the genus and species, figured the type,
dated it as late Oligocene, and referred it to the Canidae.

Clark (in Scott and Jepsen, 1936, p. 106; and in Clark 1937, p.
312) described a new species, P. dakotensis, from the Chadron For-
mation, Early Oligocene of South Dakota. He referred it to the
subfamily Cynodontinae of the Canidae, and mentioned it as a
probable ancestor of the Procyonidae.

Simpson (1945, p. 110) renamed the subfamily "Amphycyno-
dontinae," to include the two older subfamilies Cynodontinae and
Hemicyoninae. He assigned Parictis to the subfamily under its new
name.

Chaffee (1954, pp. 43-47) described the new genus and species
Campylocynodon personi, based upon a specimen from the lower
Oligocene near Wagonbed Springs, Beaver Divide, Fremont County,
Wyoming. He referred the genus to the Amphicynodontinae, and
noted its generally primitive basicranial characters. Chaffee felt
that presence of these "Procyonid" characters did not necessarily
indicate membership in the Procyonidae.

Clark and Beerbower (Clark et al., 1967, p. 27-28) divided the
genus Parictis into two subgenera, Parictis (Campylocynodon) and
Parictis (Parictis). They also described a new species, P. (C.) par-
vus, and reiterated the belief that Parictis was an ancestor of the
Procyonidae. (McGrew in 1938 had proposed that Pseudocynodic-
tis = Hesperocyon was ancestral to the Procyonidae).

2 FIELDIANA: GEOLOGY, VOLUME 26

Oligocene studies by the Carnegie Museum, Field Museum, the
American Museum, and others have, over a period of years, added
specimens of sufficient completeness to make revision of the genus
possible. The following sections of this paper will present, first,
amended descriptions of the genera and species involved; second, a
study of the significance and linkages of the anatomical characters
used in diagnosis, and; third, an interpretation of the relationships
of Parictis.

II. ACKNOWLEDGEMENTS

We wish first to thank Dr. R. Teclford of the American Museum
of Natural History, for his careful, critical review of this paper in
manuscript, and his generous expenditure of time in correspondence.
This paper is much improved as a result of his efforts, although we
are entirely responsible for the conclusions.

We wish to thank the curators in charge of vertebrate fossils for
loans from the following institutions:

1. Dr. R. Tedford of the American Museum of Natural History
for many specimens, plus excellent, enlarged stereo-photographs of
the basicranium of Parictis personi and casts of USNM 89633 and
89637.

2. Dr. Craig C. Black of the Carnegie Museum:
CM 9068, partial mandibular ramus

CM 9571, partial maxilla and rami

3. Dr. Peter Robinson, University of Colorado Museum:
CU 22749, partial mandibular ramus

4. Dr. Donald Baird, Princeton University Museum:
PU 10583, type of Parictis primaevus

PU 16265, type of Parictis (C.) parvus
PU 16695, type of Parictis (C.) personi

5. Dr. Morton Green, South Dakota School of Mines Museum:
SDSM 2476, type of Parictis dakotensis

SDSM 2567, partial mandibular ramus
Other numbers cited in the text are those of Field Museum.
The drawings are by Dr. Tibor Perenyi of Field Museum staff.

III. TAXONOMIC DESCRIPTION

For purposes of description, the classification of Romer (1966)
will be used here. Alternative classifications, based upon an analysis
of possible relationships will be presented later. Also "Oligocene"
should be read to mean "lower and middle Oligocene" throughout the
remainder of this paper.

Infra order Arctoidea

Family Canidae Gray, 1821

Subfamily Amphicynodontinae Simpson, 1946 (= Cynodontinae
Schlosser, 1911, p. 389).

Subfamily characters, fide Schlosser 1923: "Dentition f:^ M. P^
with large, posteriorly displaced inner tubercle, remaining P small
and simple. Upper M with moderate inner cingulum, large proto-
cone and metaconule, and two outer conules; protoconules never
present. M^ not much shorter than broad, triangular; M^ elliptical.
Lower carnassial with weak metaconid, and generally with large
indented trigonid.

Teilhard de Chardin (1915) gave what proves to be a more accu-
rate characterization of the subfamily, but unfortunately, he referred
to it as "Groupe de Cynodontoid^s" rather than as a subfamily with
proper nomenclatorial form. His diagnosis (de Chardin, 1915, p.
35) is repeated here, translated to English for convenient comparison:

"a. Premolars relatively simple and small, having ridges often
concave.

b. Lower carnassial little elevated, the protoconid being always
much higher than the two other tubercles of the trigonid and
than P4. Talonid generally large.

c. M2 ordinarily long, but with contour and pattern rounded,
very modernized; rather than remaining stationary or decreas-
ing, as in the Cynodictoides or the Stenoplesictoides, the tal-
onid tends to develop at the expense of the trigonid (except in
certain forms convergent toward the Stenoplesictoids, such

CLARK & GUENSBURG: ARCTOID CHARACTERS 5

as: Plesictis stenogalinus and Viverra simplicidens) ; paraconid
reduced or absent, and in any case completely separate from
the protoconid; the protoconid opposite to the metaconid, not
much larger than it, or even smaller, the internal border of the
tooth being often much higher than the external border of the
tooth (Footnote omitted here).

d. M3 present or absent.

e. Superior molars large, with a tendency to become squared and
quadritubercular.

f. Tympanic bullae ossified, simple."

Save for the description of M2, this applies well to Parictis.

Parictis Scott, 1893
Diagnosis. —

1. Amphicynodontine carnivores showing progi*essive increase
in massiveness of lower jaw, without proportionate increase in
height of teeth.

2. Progressive blunting and flattening of teeth, especially PgiJ

3. Dental formula jiyiT-f

4. Premolars transversely expanded, plan oval to subquad-
rangular.

5. Lower canine with procumbent, bulbous base constricting
suddenly to a tapering, almost vertical body.

6. M2 trigonid extremely compressed, paraconid progi*essively
reduced. Broad antero-labial shelf. Trigonid extremely low.

7. M3 tiny, single-rooted, almost circular in plan.

8. Upper canines set almost vertically, bulb less pronounced
than in lower.

9. In P^: a. Tooth low.

b. Parastyle small, scarcely more than an enlarged por-
tion of the cingulum.

c. Protocone small but well separated from the para-
cone; situated but little posterior to the parastyle.

d. Always a tiny hypocone, which may be merely a
slight enlargement of the cingulum.

10. M': a. Wide internal and external cingular shelves.

b. Small protoconule and metaconule, the latter always
double.

6 FIELDIANA: GEOLOGY, VOLUME 26

c. Tooth long antero-posteriorly.

11. M'^: Smaller and much narrower replica of M\ with ex-
ternal cingulum very close to base of metacone.

12. Carotid foramen slightly separated from Foramen Lacerum
Posterius, carotid canal following edge of internal wall of bulla, as
in Amphicynodon and Hesperocyon.

13. Dental enamel progressively more wrinkled.

14. Comparatively short-faced, large-brained carnivores.

Parictis (Campy locynodon) (Chaffee), 1954
Diagnosis. —

1. Parictids generally primitive, generic characters weakly or
partially developed.

2. Mandible thin, light, relatively shallow from top to bottom.

3. Dental enamel heavy, but smooth to very little wrinkled.

4. Size small for the genus.

5. Trigonid of M2 moderately compressed, comprising slightly
more than half the total width of the tooth.

Parictis (Campylocynodon) personi (Chaffee), 1954. Figures 1, 2.

Type Specimen. — PU 17795, skull and jaws, badly weathered.

Hypodigm. — Type specimen only.

Locality.— ^W Yx sec. 13, T. 31 N., R. 95 W., Fremont County,
Wyoming; southeast of Wagonbed Springs, Beaver Divide.

Horizon. — White River Formation, Chadronian, early Oligocene.

Description. — The specimen is, unfortunately, so poorly preserved
that only the general form of the teeth can be determined. How-
ever, it constitutes the only known specimen of the genus which pre-
serves the shape of the skull, some characters of the otic region, and
the angle and condyle of the jaw.

The general character of the dentition, plus those of the bulla and
the carotid canal, make assignment of the subgenus to the family a
certainty. The slender construction of the jaws is unlike that of the
other subgenera of Parictis, and might indeed be justification for
re-elevating Campylocynodon to generic rank. However, the sub-
generic ranking seems a useful device with which to express the
closer dental relationship of Campylocynodon to the other parictines
than to the European genera of the subfamily.

CLARK & GUENSBURG: ARCTOID CHARACTERS

Fig. 1. Parictis {Campylocynodon) personi, type specimen, PU 17795. Palate.
P* is relatively short overall, with a very short posterior blade; M* and M' are
correspondingly long.

Only the anterior portions of the auditory bullae are well pre-
served. Chaffee (1954) suggested that this might be due to partial
ossification, as in certain marsupials and insectivores, rather than to
postmortem breakage. Rounding of all broken edges precludes
direct determination. Scraps of bone canying the carotid canals,
plus a large scrap of the posterior rim of the left auditory meatus,
occur in place, suggesting that the bullae were completely ossified in
life. This would parallel the condition in Cynodon and Hesperocyon,
and, indeed, in almost all other known carnivores who possess bullae.
Unfortunately, a cursory study of recent viverrids, marsupials, and
insectivores demonstrates that extent of ossification of the bulla can

FIELDIANA: GEOLOGY, VOLUME 26

Fig. 2. Paridis (Campylocynodon) personi, type specimen, FU 17795. Basi-
cranium. Arrows indicate the carotid canal and the presumed edge of the bulla.

vary from genus to genus within any one family. The point there-
fore remains unsettled, but we believe that the inconclusive evidence
rather favors complete ossification and later breakage.

Parictis (Campylocynodon) parvus Clark and Beerbower, 1967.
Figure 3A, B

Type specimen. — PU 16265, mandibular fragment with RP2-M2,
and alveoli of RC, P1-2, M3.

Hypodigm. — Type specimen only.

Locality.— SW M sec. 25, T43N, R46W, west flank of Quinn
Draw, Shannon Co., South Dakota.

Horizon. — Red layer near base of Ahearn Member, Chadron For-
mation, Early Oligocene.

Description. — Measurements in Appendix A. Smaller size and a
very much shorter talonid of Mi are the only characters known at
present which distinguish this species from P. (C.) personi. It is
possible that this species is a junior synonym of the latter, but until

•o c c

c ^ t
03 jC ^

/.

•- c -n
° »; S

'C ± it

OJ 3 OJ

c ? 2

0) o) 2.
~ aj —

s|?

>• =.2

* ti -

P. £"3

. is

to c .
Cl, i; >

c i 2

W.2 C2

I-. g Oi

^■^<

».■£! ^
go

.2 -r a ,

"C'o.'B i

a, - 8 I
."2 "5 I

CO S Q.

Ct, w « —
o — -52

10 FIELDIANA: GEOLOGY, VOLUME 26

a more adequate hypodigm for the subgenus is discovered, it seems
to us practical to regard the two as separate.

P. (C.) parvus resembles Mustelavus in size, general lightness of
jaw, Pg-s simple, and small size of Mi talonid. It resembles other
Parictines and differs from Mustelavus in the following characters:

1. Trigonid of M2 compressed (although not so markedly as in
later species of Parictis).

2. M2 with a broad antero-external shelf and an accessory cus-
pule on that shelf, posterio-lateral to the protoconid. Both of these
are lacking in Mustelavus.

3. Possession of M3.

4. Heavily built Mi trigonid, with the three cusps subequal, and
the paraconid angled sharply inward.

The ratios of Height, hypoconid

Height, protoconid are:

P. (C.) parvus (PU 16265) .449
P. (S.) gilpini (FM 22405) .428
M. priscus (PU 13775) .345

The Ml trigonid of M. priscus is higher, more trenchant, and has
the metaconid proportionally reduced. Since the four characters in
which the two species resemble each other are all either generalized
or are characters which this paper will show to have been developed
independently in a number of cases, we feel that they bear less sig-
nificance than the four definite resemblances to Parictis.

The case for P. (C.) parvus being a member of the genus Parictis
is independent of its relationship to P. (C.) personi, and seems to us
adequate. Comparison of the types of P. (C.) personi and P. (C)
parvus reveals no reason for a generic separation of the two, admit-
ting that the fragmentary nature of one and the miserable preserva-
tion of the other make all comparisons somewhat unsure.

In view of the condition of the specimen, no one can be positive
that C. personi is or is not a member of the genus Parictis. The fol-
lowing points seem to us strongly to support the relationship:

1. Presence of M3.

2. General size relationships of P2-3, and probable absence of ac-
cessory cusps upon them (not exclusive, but their presence would be) .

3. Left P^ (right P^ seems broken) with a broad, heavy, little-
angulated protocone.

CLARK & GUENSBURG: ARCTOID CHARACTERS 11

4. P* short anteroposteriorly.

5. Lingual part of M' broad, not marked off from external por-
tion by a posterior indentation.

6. M' external wall forming a 60* angle with the anterior edge of
the tooth, as in Parictis. In M. priscus the angle is 50°. In Hes-
perocyon the anterior edge of the tooth is so curved as to make accu-
rate measurement impossible: the maximum angle measurable is 60°,
the minimum 50°. Since curvature is produced by a posterior angu-
lation of the lingual moiety of the tooth, 60° is probably the more
nearly correct.

Opposed to this evidence are the generally light proportions of the
jaw, and possibly the trigonid of M2 (which would exclude C personi
from Parictis if it were not compressed). We are unable to deter-
mine what the trigonid was like: the right M2 seems possibly to
resemble that of Mustelavus and Hesperocyon ; the left does not. We
cannot be sure whether we are looking at characters or at accidents
of preservation.

The characters of the lower dentition seem on the whole to resem-
ble those of C. parvus and of P. gilpini, and the characters of the
upper dentition generally resemble those of P. gilpini and P. mon-
tanus rather than M. priscus. We therefore regard Campylocynodon
as a subgenus of Parictis, comprising at present the two species, C
personi and C. parvus. However, it must be remembered that the
interpretation of M2 in Campylocynodon is not certain, and to that
extent the relationship to Parictis is uncertain also. P. (C.) parvus,
on the other hand, is assuredly parictine.

Parictis (Subparictis) new subgenus
Diagnosis. —

1. Parictids with mandibular characters well-developed, upper
dentition primitive for the group.

2. Mandible massive, moderate to very deep from top to bottom.

3. Dental enamel moderately wrinkled.

4. Size average to large for the genus.

5. P* hypocone extremely small; protocone small, and slender at
base.

6. M^ protocone at the apex of a V-shaped pair of ridges: anter-
ior ridge extends through protoconule to join the parastyle, and pos-
terior ridge passes through the doubled metaconule to join the me-
tastyle, as in Hesperocyon.

>. c

> c

o rt
t« o

o-^

O -rt

o .5

2 >«
a c
>.°

> Q>

. IS

O M

i- C

•2 c i

O)

rt

^

O)

—

1^
0)

a
t3

13
O

-t-»

lO

o

^

o

a

J

s

!c

o

.5f

a
o

.S-c a

D.

^

^

03

o

+->

a

j3

03

"o

^

+J

o

SR

C

's

03

V

o

>

T3

a>

S

>

CO

1

.'^

03

u

"G

IS

a

*(**

-U

a>

.

-C

^

Eh

■^'

eU

o;

S

C

O

3

o

P^

O

o
a

§

>>

12

1

b

>

M
U

'a
«

■a
s

o

0)

a

a

S

a

'5t

^

-7 O

52. iS
•S §
ecu

■IS

W

14 FIELDIANA: GEOLOGY, VOLUME 26

7. M'-^ pattern as in MS save that the tooth is much narrower
transversely.

Parictis (Subparictis) gilpini' new species. Figures 4, 5, 6A, B.

Type specimen. — F-PM 22405, anterior portion of skull with lower
jaws. Middle of Ahearn Member, Chadron Formation, SE }4, of NE
H of NW K, sec. 5, T. 4S., R. 12 E., Little Corral Draw, Pennington
County, South Dakota.

Hypodigm.—F-\JM 729, left mandible with Pg, Mi_2. Chadron
Formation, Warbonnet Creek, Hat Creek Basin, Sioux County,
Nebraska. AMNH 50241, partial mandibles with LPg-Mi-z- 15 ft.
below the purple-white layer, Chadron Formation, NW end of See-
man Hills, Niobrara County, Wyoming. USNM 19930, partial left
mandible with P2-M2. White River Oligocene, 1 mile NE of Walker
Ranch, Niobrara County, Wyoming. USNM 19932, partial left
mandible with P3-M1. White River Oligocene, Everson Ranch, 12
miles NE of Crawford, Dawes County, Nebraska. SDSM 2567,
(2 specimens) : partial right mandible with P4-M1 ; right mandibular
fragment with partial Mi. Chadron Formation, Big Badlands,
South Dakota. AMNH 63933, left mandibular fragment with Mi.
12 ft. below the purple-white layer, Chadron Formation. SW side of
Seeman Hills, Niobrara County, Wyoming. AMNH 76196, right
mandibular fragment with P4-M2; 5 ft. below the purple- white layer,
Chadron Formation, Jim Christian Hills, Niobrara County, Wyo-
ming.

Range. — W South Dakota, NW Nebraska, SE Wyoming.

Stratigraphic Range. — Chadron Formation, Ahearn Member to
near top.

Diagnosis. — 1. The description of upper dentition for the sub-
genus is derived from characters exhibited by F-PM 22405, the type
specimen of this species. This species should therefore be regarded
as type species of the subgenus, and all subgenus characters of the
upper dentition are applicable to the species also. Hypodigms of

' Named for Mr. Orville L. Gilpin of Field Museum who discovered the type.

Fig. 6. Parictis (S.) gilpini, type specimen, F-PM 22405, left mandible. A.
Crown view. Note the generally long trigonid and short talonid of Mi, with ex-
actly the reverse development in M2. B. Lateral view. The heavy jaw is strongly
rounded. The premolar series has definitely fragmented.

Iiv

15

16 FIELDIANA: GEOLOGY, VOLUME 26

the other species of the subgenus, P. (S.) dakotensis, major, and mon-
tanus, include only one specimen, CM 9571, with any upper teeth.

2. Size average for the subgenus.

3. Mandible massive but only moderately deep.

4. P series relatively narrow.

5. P4 short anterposteriorly.

6. P2-3 relatively high; base of the principal cusp rises from the
internal edge of the cingulum.

7. M2 long relative to its breadth.

Discussion. — P. (S.) gilpini seems to be a generalized species,
larger, more massive, and considerably more advanced dentally than
the species of P. (Campylocynodon) but highly variable and central
in characters to all the other species of Parictis.

The mandible is, relative to the teeth, somewhat thicker but on
the average a little shallower than that of Hesperocyon. Considerable
variation in depth of mandible occurs within even the limited hy-
podigm available, but the thickness shows much less variation.

The most satisfactory characters delimiting this species are the
combination of medium size, elongate M2, P series relatively narrow,
P4 short, and P2-3 high.

This last is a most useful character which, unfortunately, cannot
be directly measured. In P. (S.) dakotensis and P. (P.) primaevus,
the angles formed by the anterior and posterior cristae of the tooth
with the enamel line are not notably less than those of P. (S.) gilpini.
However, in the more specialized species the cingulum extends far
inward as a broad, flat area, from the middle of which a small but
steep-sided cusp arises. In P. (S.) gilpini the cusp arises immediately
inside the cingulum, which raises the tip of the unworn tooth con-
siderably higher than that attained in the more specialized forms.
Wear precludes direct measurement of cusp height, and absence of a
boundary between cusp and cingulum defeats any attempt to mea-
sure the width of the cingulum. However, the teeth of the various
taxa present a very different appearance, and are separable by eye.

Parictis (Subparictis) montanus new species. Figures 7, 8, 9A,
B.
Type specimen. — CM 9571, right mandible with C root, Pi_3,
partial P4, M1-2; left mandible with P4-M2; right maxillary fragment
with P'*-M2. Pipestone Springs Formation, late Chadronian; main |

in,

I-

1 cm

Fig. 7. Parictis (S.) montanus, type specimen, CM 9571. Crown view of maxil-
lary fragment. Arrow indicates the incipient hypocone on P*. The developing
antero-posterior protocone crest and breakdown of the primitive trigon V-crest on
M' are clearly shown.

H

1 cm

H

Fig. 8. Parictis (S.) montanns, type specimen, CM 9571. Lateral view of maxil-
lary fragment. Shows generally low crowns, short P* and long M'. Note the
extremely short posterior crest of P*.

17

n

CQ

X

o
O

o

C/3

18

CLARK & GUENSBURG: ARCTOID CHARACTERS 19

pocket, Pipestone Springs, sec. 29, T 2N, R5W., Jefferson County,
Montana.

Hypodigm. — CM 9068, right mandible with P2-M2, Pipestone
Springs Formation, late Chadronian, Little Pipestone Local Fauna
locality, sec. 9, TIN, R5W, Jefferson County, Montana. F-PM
3843, right mandible with Pz-M,. Unnamed strata of late Chadron-
ian age, probably equivalent to the Pipestone Springs Formation,
Douglas Creek valley, center, sec. 26, T12N, R12W, Powell County,
Montana.

Range. — Western Montana.

Stratigraphic Range. — Late Chadronian.

Diagnosis. —

1. Upper dentition is known only from RP*-M- of the type; all
comparisons will be stated relative to P. (S.) gilpini.

2. P*: (a) parastyle consisting of a ridge down the anterior edge
of the paracone, and an enlarged cingular buttress which does
not merge with the ridge; buttress much wider than in P. (S.)
gilpini.

(b) Protocone a definite, rounded eminence arising from
the cingulum, not a sharp crest as in P. (S.) gilpini; much
larger and displaced farther lingually than in the latter
species; notch between parastyle and protocone much deeper.

(c) Hypocone much more definite, arising from a gen-
erally heavier cingulum.

3. M*: (a) Tooth slightly shorter anteroposteriorly along its
outer margin, but considerably longer across the protocone;
internal cingular shelf expanded posteriorly; the total of these
features produces a much more rectangular outline than the
elongate, blunted triangle of M' in P. (S.) gilpini.

(b) Protocone closer to metacone, producing a smaller
trigon.

(c) A small ridge extending anteriorly from the protocone
connects that cusp with the cingulum.

(d) V-shaped ridge from the protocone reduced.

(e) A low ridge running down the lingual face of the para-
cone connects with the protoconule.

(f) A low ridge running down the lingual face of the meta-
cone connects with the metaconule.

20

CLARK & GUENSBURG: ARCTOID CHARACTERS 21

4. Size smaller than P. (S.) gilpini, but mandible proportionally
as robust.

5. P series narrow.

6. P2 3 high-crowned, as in P. (S.) gilpini.

7. P4 long in proportion to P series.

8. M2 slightly smaller proportionally than in P. (S.) gilpini.

Discussion. — This species differs from P. (S.) gilpini in size, in
the relatively long P4, and in the reduced M5. M2 of CM 9571 is
slightly smaller, but M2 in CM 9068 is scarcely over one-half the
size, relatively and absolutely, of their analogues in P. (S.) gilpini.
The tooth is so well formed in both cases that abnormality cannot be
logically presumed; the difference probably represents variation of
an evolving character within the species.

The three known specimens show less variation in all characters
than do the hypodigms of P. (S.) gilpini and P. (S.) dakotensis. F-
PM 3843 is one of two specimens listed by Konizeski (1961) as Hes-
perocyon paterculus; the other on his list, F-PM 3828, is certainly
Hesperocyon.

The upper dentition differs significantly from that of P. (S.) gil-
pini. The larger protocone and hypocone of P* must be regarded as
advances in a trend from P. (C.) personi through P. (S.) gilpini to-
ward P. (P.) primaevus.

Even more important are the changes in M'. Reduction of the
primitive V-shaped ridge from protocone to parastyle and metastyle
is accompanied by the development of ridges connecting paracone
to protoconule, metacone to metaconule, and a fore-and-aft ridge
through the protocone to the cingulum anteriorly and to a hypocone
posteriorly. These changes, incipient only in P. (S.) montanus, are
well developed in P. (P.) primaevus, as will be described later.

Parictis (Subparictis) dakotensis Clark. Figure llA, B.

Type.— SDS,M 2476, right mandible with P2-M2. Upper Chad-
ron, probably Peanut Peak Member, Big Corral Draw, Washington
(now Shannon) County, South Dakota.

Fig. 10. Hesperocyon gregarim. #F-UC 496. Left P«-M', (A) anterior and (B)
crown views, for comparison with Figures 7, 8, and 13. Note the total absence of
an incipient hypoconal swelling on the internal cingulum of P*. Note also the
primitive V-shaped crest attaching the M' protocone to the paracone and meta-
cone.

22 FIELDIANA: GEOLOGY, VOLUME 26

Hypodigm. — AMNH 50240, left mandibular fragment with P4-
M2. Chadron Formation, below the purple-white layer, Head of
Indian Creek, near Lusk, Niobrara County, Wyoming. AMNH
12244, left mandible with P2-M2. "Middle Chadron." Stebbins
Ranch between Battle and French Creeks, near Folsom, Custer
County, South Dakota. AMNH 12245, right mandibular fragment
with partial Mi, M2. Chadron Formation, Stebbins Ranch, between
Battle and French Creeks, Custer County, South Dakota. F-PM
22406, right mandibular fragment with P3^4, Ahearn Member, Chad-
ron Formation, SE 14 of NE M of NW }4, sec. 5, T. 4S., R. 12 E.,
Pennington County, South Dakota.

Range. — Western South Dakota and southeastern Wyoming.

Stratigraphic Range. — F-PM 22406 from middle of Ahearn Mem-
ber, Lower Chadron; all others from middle to upper Chadron strata.

Diagnosis :

1. Skull and upper dentition remain unknown.

2. Size generally large for the genus.

3. P series broad in proportion to the length of the teeth.

4. P2-3 low-crowned, with wide cingular shelves.

5. P4 enlongated.

6. M2 broad anteriorly, relative to its length.

7. Internal cingulum well developed on paraconid and across
paraconid-metaconid notch of Mj.

Discussion. — Size, broad premolars, broad M2, well-developed
internal cingulum on the antero-lingual surface of Mi, and low-
crowned P2-3 differentiate this species clearly from others of the sub-
genus except P. (S.) gilpini. Two specimens possess characters
sufficiently intermediate to make their assignment problematical:
F-UM 729 is within the size range of P. (S.) gilpini, and has a high-
crowned P3 and a general resemblance to other specimens of that
hypodigm, but P4 is relatively elongate and M^ is somewhat broader
than most, although still not up to average for P. (S.) dakotensis.
AMNH 76196 is intermediate in size between the two species, and
possesses an M2 unusually large but not unduly broad for P. (S.)
gilpini. Since P4 is quite short in this specimen, overall size is the
only character in which it is intermediate.

The series of specimens here assigned respectively to the species
gilpini and dakotensis present a baffling array of characters mutually
independent or only partially interdependent in their variances. The

(^ \

1

c
o

bi

O

y

>

\

•§

03

28

24 FIELDIANA: GEOLOGY, VOLUME 26

premolars vary in plan from simple oval through subquadrate to
bulbous. Lengths of all teeth posterior to P3 vary independently,
so that no one tooth can be used as an index against which others
can be evaluated. M2 displays extreme variations in plan, with the
width of the antero-external shelf related only slightly to the gen-
eral contour of the remainder of the tooth. A great variety of mea-
surable characters (see tables) show a broad spectrum of relation-
ships, with the type specimens quite definitely different but no clear
separation between their hypodigms. Presumably, this means a
genetically plastic population with various unit characters which
were controlled by independent genes and had no appreciable sur-
vival value. This would produce a large number of local and more
or less temporary races which might lead to a final differentiation of
populations (Ehrlich and Raven, 1969).

If one regards the type of P. (S.) gilpini as primitive and the type
of P. (P.) primaevus as advanced, then the type P. (S.) dakotensis fits
rather nicely as an anatomical and chronologic mid-stage (although
it happens to be larger than either end-point). The hypodigms,
however, do fatal damage to this straight-forward and simple out-
line.

First, F-PM 22406, a fragment with RP3-4, is only slightly smaller
than the type of P. (S.) dakotensis, shows almost as "advanced"
characters — the jaw is a little shallower and the cingular shelves a
little narrower — and comes from the middle part of the Ahearn
Member, the same horizon as the type of P. (S.) gilpini. Second, F-
UM 729, already discussed as intermediate in characters between the
two species, comes from the Chadron of Warbonnet Creek, Sioux
County, Nebraska, where only middle to late Chadronian sediments
are exposed. Third, AMNH 50241, which differs from the type of
P. (S.) gilpini only in a slightly larger M2 and a slightly more massive
jaw, is described as occurring in the Chadron at the northwest end
of the Seeman Hills, Niobrara County, Wyoming, "15 feet below the
purple white layer," which would place it in sediments of late Chad-
ronian age.

The two species, therefore, have completely parallel age distribu-
tions and geographic ranges. The clusters of characters upon which
they are differentiated are only partially exclusive (vide F-UM 729) .
It may in the future become advisable to declare these end-members
subspecies of P. (S.) dakotensis, but at present it seems to us useful
to recognize the generalized nature of P. (S.) gilpini by giving it
specific status.

CLARK & GUENSBURG: ARCTOID CHARACTERS 25

Incidentally, it should be pointed out that the type of P. (S.)
gilpini and F-PM 22406 (the specimen of P. (S.) dakotensis men-
tioned above) not only occurred in the same horizon but were found
within the same 10-acre area. Neither shows signs of considerable
transportation.

Parictis (Subparictis) major, new species. Figure 12A, B.
Type specimen. — F-PM 22404, left mandible with C-M,.
Hypodigtti. — Type specimen only.

Locality.— NE-l 4 of SE-K, section 12, T 42 N, R 45 W, Shannon
County, South Dakota.

Horizon. — Peanut Peak Member, Chadron Formation, latest part
of early Oligocene.
Diagnosis. —

1. Jaw as massive as in P. (S.) dakotensis, but overall much larger
and proportionally much deeper.

2. Lower edge of jaw slightly concave upward with a small but
definite chin.

3. Lower canine exceedingly large, more than twice the size of
that of P. (S.) gilpini, although the other teeth are only one-third
larger.

4. Pa and P3 subequal, and low-crowned due to low angles of
anterior and posterior cristae rather than to broad cingular shelves.

5. P4 relatively high and elongate.

6. Ml very low, metaconid notably posterior to protoconid and
much reduced; posterior wall of trigonid sloping posteriorly at a low
angle, considerably lengthening the tooth. Talonid relatively small,
cusps forming very low, rounded crests joined by a small hypoconu-
lid.

Discussion. — Discovery of more adequate specimens will prob-
ably show that this species does not belong within the genus Parictis.
It certainly differs from all others more than they do from each other.
In many ways it resembles Pachycytiodon. However, in the absence
of M2 there is no sharply definitive reason for separating P. {S.)
major from the remaining species of the genus, so we are tentatively
placing it here.

The only other North American genus resembling Parictis is
Drassonax Galbreath, 1953. Although the genus was originally
referred to the Mustelidae, we see no characters on the type and only
known specimen which separate it from the Amphicynodontinae.

26

S"

C

c

O.

Si

o

s -

C4 O

o

•^ >>

ft- c

c 5
a> o
e be

SJ *"
aS
« rt
a> bo
a c
>'S

iT 03

.=> *

B 03

•2 ">
> c
c >

27

28

FIELDIANA: GEOLOGY, VOLUME 26

Whether or not one evaluates dental characters highly as taxonomic
keys, they must be used when they are all one has: the presence of a
large, single alveolus for M3 is certainly not musteline.

However, Drassonax harpagops differs from P. (S.) major in
several characters which we believe to be significant. The following
table reveals these:

CI

laracter

D. harpagops

Parictis

P. (S.) major

1.

Pi

Small, diastemated

Proportionally
large, in cont-
inuous row,
C-P1-P2

Parictine

2.

P2/P3

P2 much smaller

Subequal

Subequal

3.

P series

Uniformly graded
in size and height
P1-P4

Broken into:
Pi small
P2-3 equal
P4 large

Parictine

4.

P4

No posterior acces-
sory cuspule

Definite poste-
rior accessory
cuspule

Parictine

5.

C

Small; elongate
cross-sectional axis
antero-external to
postero-internal

Same as Drasso-
nax

Proportionally much
heavier, root laterally
flattened to symphysis,
very little room for
incisors

6. Ml

Paraconid low, Paraconid low

forms a transversely to high, but al-
elongate bulb ways a bulb or

cone

7. Ml 43%-high proto-

Protoconid/ conid
Entoconid

8. Jaw

Heavy, moderately
deep

Varies

Heavy, shallow
to moderately
deep

Paraconid low, forms a
transverse crest

55%-low protoconid

Very deep, moderately
heavy

The characters significantly separating P. (S.) major from Dras-
sonax seem to us to be numbers 2, 3, 4, and 7 above. The others
might be specific rather than generic. This does not necessarily
mean that major will prove to be Parictis when better specimens are
known.

The jaw of P. (S.) major is deeper than that of Procyon lotor, but
not quite so massive. The cheek-tooth row is fully as long, but P4
and Ml are not nearly so broad transversely as in Procyon. In gen-
eral, the jaw presents an amazingly ursid aspect, although in details
it is significantly different.

CLARK & GUENSBURG: ARCTOID CHARACTERS 29

If P. (S.) gilpini is indeed the stem species of the group, P. (S.)
dakotensis, with a moderate increase in size and molar breadth,
represents a moderate increase in grinding capabiHty and decrease
in sectorial function of the teeth; P. (S.) major carries these tenden-
cies much further, but the details of premolar specialization differ
sufficiently to preclude a direct descent from some early population of
P. (S.) dakotensis.

Parictis (Parictis) Clark and Beerbower, 1967
Diagnosis. —

1. Parictids with the mandibular characters and upper dentition
advanced.

2. Mandible massive, only moderately deep.

3. Dental enamel strongly wrinkled.

4. Size smaller than average for the genus.

5. P2 slightly larger than P3.

6. P2-3 subrectangular, very low-crowned, with broad cingular
shelves.

7. Both antero-internal and postero-external cingula of Mi well
developed.

8. P"*: (a) Tooth relatively short and broad.

(b) Protocone forming a heavy, thick bulb, set off from
the remainder of the tooth by a pair of shallow
notches.

(c) Cingulum sharp-edged and heavy, and complete
except for the posterior-most tip of the tooth; arising
from the anterior and posterior edges of the proto-
cone.

(d) Hypocone part of the cingulum but relatively prom-
inent, with deep pockets anterior and posterior to it.

9. M*: (a) Tooth subquadrangular, large relative to P^, long

antero-posteriorly and short transversely.

(b) Paracone large and high, protocone larger and higher
than in P. (S.) montatius; metacone very low and
smaller.

(c) Anterior crest connects protocone — doubled proto-
conule — paracone; rather than protocone — doubled
protoconule — parastyle, as in P. (Subparictis).

30 FIELDIANA: GEOLOGY, VOLUME 26

(d) Posterior crest connects posterior protocone crest-
doubled metaconule — metacone.

(e) Both protoconules and metaconules much reduced.

(f) Posterior cingulum developed into a hypocone and a
postero-lingual cingular cusp, of equal size and almost
as large as the metacone.

(g) Well-developed antero-posterior crista connecting
anterior cingulum — protocone — hypocone.

(h) Internal cingular shelf sharp-edged but much re-
duced.

Parictis (Parictis) primaevus (Scott). Figures 13A, B, 14A, B

This is the type species of the subgenus and genus. Since the sub-
genus as presently recognized is monospecific, the characters listed
above for Parictis (Parictis) also characterize the species P. (S.)
primaevus.

Type specimen. — PU 10583, L. mandible with P2-3. John Day
upper Oligocene; "at Silver Wells, Oregon."

Hypodigm. — CU 22749, L. mandible with P2-M1 ; Lower Nodular
Zone, Scenic Member, Brule. NW M sec. 11, T 42 N, R 45 W, Shan-
non Co., South Dakota. F-P 27157, L maxillary fragment with P^-
M^ Base of Orella Member, Brule; Roundtop, 7 miles north of
Crawford, Dawes Co., Nebraska.

Range. — Central Oregon and Western South Dakota and
Nebraska.

Stratigraphic Range. — Base of Middle Oligocene into Upper Oli-
gocene.

Discussion. — The two specimens referred to this species may not
be co-specific with the type specimen. The difference in age between
Orellan and John Day, plus the fact that both represent animals
larger than the type, suggest this. Since the type consists merely of
a low but massive mandible bearing P2-3, diagnostic characters are
extremely limited. We have therefore referred the other mandible,
UG 22749, to this species. Our only justifications for referring the
maxillary fragment, F-P 27157, to this genus are that it is parictine
in character and size, and that its individuality is comprised of fur-
ther development of trends already manifest in the sequence P. (S.)
gilpini-P. (S.) montanus.

\:

Jt^

i

[I

.^

e
o

a .

>> u

3 o
c *-•

0) -

Z bfl

c

a> o

. CQ

c -5

led
a

to

2! c
- °

»> i.
o a
S

c E
R.2

1'=

C
cS

31

a

o
^

32

CLARK & GUENSBURG: ARCTOID CHARACTERS 33

We differ in this interpretation from Tedford (personal communi-
cation) who regards the maxillary fragment as referable to another
taxon. Since this difference has not been resolvable through cor-
respondence, we feel it proper to publish our interpretation with the
understanding that Tedford will at a future date present the evidence
supporting his viewpoint on this debatable point.

Assuming that the references are correct, we see in this species a
small carnivore with a shallow but massive jaw; broad, low, simple
premolars with a wide cingulum; a tiny but definite hypocone arising
as a cingular eminence on P*; and M' changed in pattern to a sub-
rectangular, quadritubercular tooth with accessory cuspules. P2
and P3 have become completely similar in every respect. Even P4,
by increase in the cingulum and in the total width plus decrease in
prominence of the accessory cuspule, is moving toward the general
profile of an old-fashioned circus-tent which its anterior fellows have
achieved. The enamel is heavily wrinkled on all teeth. Every
known character save those of P^ has progressed toward its logical
conclusion; it is difficult to imagine this species giving rise to any
further dental specialization, other than increase in number of molar
cusps, disappearance or flattening of P2-3, and further molarization
of P*.

IV. EVALUATION OF CHARACTERS PRESERVED

A. Introduction

The preceding sections of this paper have used a variety of char-
acters to identify taxa of varying rank. These same characters must
necessarily function to estabhsh the relationships of Parictis and of
the Amphicynodontinae to other arctoid groups ("arctoid" is used
here sensu Romer 1966, = "Ganoid" Simpson, 1946). It seems
prudent, therefore, to observe the occurrence of these characters in
other carnivores, especially in other arctoids. We may be this means
determine whether certain features or trends are unit characters or
are usually genetically linked with others. We may also be able to
slightly unravel the enormous complex of parallelisms which beset
arctoid taxonomy.

B. Dental Characters

Tables 1-6 list the characters considered, as developed in the
genera upon which they were observed. This is by no means a list
of all those genera whose dental specializations parallel in some
wise those of Parictis. Rather, selection was made first of common
genera representing various arctoid families; second, several forms
which have been regarded as related to or descended from Parictis
by various authors; third, a sprinkling of genera (e.g., Arctictis)
whose very lack of relationship might make comparison fruitful.

The tables demonstrate that no two of the characters are invar-
iably paired.

The two jaw characters, massiveness and depth, exemplify this
very well. Some animals possess deep, massive jaws; others have
deep, lightly built jaws; still others have massive but relatively shal-
low jaws. Even such apparently obvious association as crowding
with echeloning of premolars is not inevitable, for in Procyon and in
Aelurodon saevus the premolars are echeloned but not crowded.

A clearer picture emerges at the family and subfamily level. The
Mustelidae, for example, show a strong tendency to develop simple^

1 "Simple" as used in this text, refers to premolars which lack both anterior
and posterior accessory cusps upon their respective crests. The cingulum may be
variously absent, developed into a broad shelf, or into anterior or posterior cingu-
lar cusps.

34

CLARK & GUENSBURG: ARCTOID CHARACTERS 35

premolars, but none to equalize P2 and P3 in size. The Ursidae and
Amphicynodontinae, on the other hand, show a strong tendency to
develop premolars both simple and subequal. Vestiges of the poste-
rior accessory cuspules manifest themselves as sudden changes in
angle of slope of the posterior crest of P3 in some individuals of Paric-
tis (5.) gilpini.

The genetic field concept expounded by Butler (1939), and ap-
plied by Davis (1964) to the Ursidae, may well aid in understanding
both evolutionary trends within Parictis and its relationship to the
other arctoid taxa. Application of this concept to the data presented
in the tables generates an hypothesis of arctoid dental evolution
which seems to be internally consistent.

Let us presume that the proto-carnivore had a premolar denti-
tion consisting of four similar, unicuspate teeth, narrow and sec-
torial, under the control of a single genetic field. The first major
changes would be development of tri-cuspate crowns and an increase
in over-all size. Both of these started at the rear of the premolar
field and progressed forward toward the canine. Probably they
never visibly affected P|. This would produce a premolar series like
that of Viverravus, Hesperocyon, and some species of Miacis.

The next event seems to have been the separation of P[ from gen-
eral field control. Henceforward it seems to change size, shape,
number of roots, or even to disappear without reference to the char-
acteristics of the remaining teeth.

The next stage involves a further fragmentation of the premolar
field. P^:^ remain under its influence, while F\ tend to develop pari
passu with the molar series. This seems to be the stage achieved by
the earliest Amphicynodontines.

Starting from this, we see in Parictis a progressive decrease in
size of P§ until it equals or is even slightly less than P5. Both lose
their posterior accessory cusps (the anterior either were lost earlier
or were never developed), P| a little faster than Pg. P|, on the other
hand, remain disproportionately large and retain their multicuspate
character. The genus thus arrives at a genetic situation wherein F{
can be expected to fluctuate with the molars, while P| | will function
as a genetic unit and F[ will act independently.

Only a few genetic changes, which occur over the entire cheek-
tooth field, can thereafter be expected to influence all of the pre-
molars, either simultaneously or serially. Broadening of the series

AMPHICYNODONTINAE

AMPHICYNODON
(=CYNODON) PARICTIS HEMICYON

MASSIVE JAW
DEEP JAW

■
X

■

X

m

CROWDED P
ECHELONED P

P2.3 SIMPLE
P2.3 SUBEQUAL

m
■

■

BROAD P
BULBOUS CANINES

■

■

P"^ HYPOCONE
SUBRECTANGULAR M^

■

■

QUADRITUBERCULAR M^
CHIN

■

^2

■
■

2 VARIES SPECIFICALLY
□ ABSENT
Table 1

I PRESENT

3 PRESENT, POORLY DEVELOPED

36

MIACIDE - C AN IN AE

UINTACYON DAPHOENUSDAPHOENOCYONHtSPEROCYON UROCYON CANIS

MASSIVE JAW

D

■

n

n

D

DEEP JAW

Kl

m'

la

D

D

CROWDED P

D

D

D

D

D

ECHELONED P

D

m'

D

D

D

P^. 3 SIMPLE

^_ 1

D'

n

n^

■

D'

P;.3 SUBEQUAL

D

D

D

Kl

D

BROAD P

D

D

■

D

D

D'

BULBOUS CANINES

■

D

D

a

D

D

P* HYPOCON

D

D

D

D

D

D

SUBRECTANGULAR m'

D

D

n

D

D

D

QUADRITUBERCULAR M'

D

D

n

D

KI

EI

CHIN

■

D

s

D

D

D

» CINGULAR CUSP

2 P^ SIMPLE TO ACCESSORY CUSP
AS INDIVIDUAL VARIATION

3 VARIES INDIVIDUALLY

4 VARIES SPECIFICALLY

CH ABSENT

I PRESENT

13 PRESENT. POORLY DEVELOPED

Table 2

87

PROCYONIDAE

BASSARISCUS PROCYON NASUA POTOS PHLAOCYON AILURUS

MASSIVE JAW

D

■

■

D

■

■

■

DEEP JAW

D

KI

KI

■

Kl

■

■

CROWDED P

D

D

D

K

■

■

ECHELONED P

D

■

D

^'

■

■

P2.3 SIMPLE

m'

■

■

■

D^

D

P2.3 SUBEQUAL

Ei

D

D

■

D

D

BROAD P

■

■

■

■

■

■

BULBOUS CANINES

D

^

D

D

■

■

P* HYPOCONE

■

■

■

■

■

■

SUBRECTANGULAR m'

■

■

■

■

■

■

QUADRITUBERCULARM'

D

■

■

■

■

■

CHIN

D

K

D

■

■

■

1 UPPERS ONLY

2 Pj SIMPLE

P3 VERY COMPLEX

3 VARIES SPECIFICALLY

n ABSENT
H PRESENT
^ PRESENT. POORLY DEVELOPED

Table 3

38

MUSTELIDAE

OLIGOBUNIS AELUROCYON

LUTRA

GULO

TAXIDEA

CAIERA

' lAYRA'

PIESIOGUIO

MASSIVE iAW

■

■

■

■

■

■

DEEP JAW

^

D

D

IE

■

EI

CROWDED P

D

■

D

■

D

ECHELONED P

■

■

D

■

■

P}.] SIMPLE

■

■

■

■

m'

P; 3 SUBEQUAL

Kl

D

D

D

D

BROAD P

■

■

■

■

BULBOUS CANINES

la

■'

■'

■

p'hypocone

n

n

D

D

D

SUBRECTANGULAR

D

n

D

M?

D

m'

QUADRITUBERCULAR m'

n

n

D

D

D

CHIN

Ki

■

■

■

■

I LOWER ONLY
; TRAPEZOIDAL
3 ALSO P,

n ABSENT
H PRESENT
E) PRESENT. POORLY DEVELOPED

Table 4

BOROPHAGINAE-AMPHICYONINAE-VIVERRIDAE-HYAENIDAE

AELURODON

OSTEOBORUS AMPHICYON

ARCTICTIS

NANDINIA

HYAENA

MASSIVE JAW

■

■

■

D

D

m'

DEEP JAW

m'

■

K

■

■

m'

CROWDED P

D

■

m

D

D

■

ECHELONED P

m'

KI

■

D

D

D

P2.3 SIMPLE

D

■

■

■

m'

D

P2-3 SUBEQUAL

D

D

D

D

K

D

BROAD P

■

■

m'

■

■

■

BULBOUS CANINES

D

D

n

D

D

D

P* HYPOCONE

D

D

D

D

D

D

SUBRECTANGULAR m'

D

■

n

n

D

n

quadritubercularm'

D

D

n

D

D

n

CHIN

■

■

Ki

D

D

■

X LOWERS ONLY

2 UPPERS ONLY

3 UPPERS HAVE ACCESSORY CUSPULES

4 NOT IN PROPORTION TO SIZE OF TEETH
-ONLY TO LENGTH OF JAW

5 VARIES SPECIFICALLY

n ABSENT

I PRESENT

K PRESENT. POORLY DEVELOPED

Table 5

40

URS IDAE

URSUS StltNARCTOS HCIARCTOS MtLURSUS THAIARCTOS TRCM*RCTOS AllUROPOO*

MASSIVE JAW
DEEP JAW

■

■

■

n
■

D
■

■
■

CROWDED P
ECHELONED P

n

■

■

n

■
■

Pj J SIMPLE
P; ] SUBEQUAL

■

■

D'

■

■

D
D

BROAD P
BULBOUS CANINES

■
■

■
■

■
■

■
■

■
■

p'mypocone
subrectangular m>

■

■

n
■

a
■

■
■

QUADRITU6ERCULARM>
CHIN

■
■

■
■

■
■

■
■

■
■

.

Pj USUALLY

ABSENT

D ABSEN

r

P' ALWAYS SIMPLE. OFTEN ABSENT

PRESENT

Table 6

41

42 FIELDIANA: GEOLOGY, VOLUME 26

and development of a wide cingulum seem to be a related pair which
have done this independently within several taxa.

Comparison with other Arctoid taxa reveals at once that this
series of developments does not represent an inevitable genetic link-
age. The broad-toothed Mustelines reported in Table 4 all have
P2-3 unicuspate, but in none of them are the two teeth equal in size.
Crowding of premolars occurs in some but not in others; echeloning
occurs in all except Taxidea, in which the teeth are not crowded.
However, in none of them is P| apparently influenced in size or
presence by the size of the remainder of the series. It seems reason-
able that in the Mustelidae the premolar genetic field has separated
as in the Amphicynodontinae, but the Pfif field has not unified the
teeth under its control to the extent achieved in the latter taxon.

The Ursidae, on the other hand, offer a very close parallel to the
Oligocene Amphicynodontinae. P| are always large and functional,
as are Pf . In contrast, Pf if are either tiny (Selenartos, Helarctos) or
absent to vestigial (the larger genera). Pf are often absent while Pf
are present and vestigial, or present on only one side; this apparently
is a remanent trace of the former unified but gradational genetic
field.

Among the Miacidae, Uintacyon shows a remarkable parallel in
that P2-3 are both simple and subequal; they are also crowded and
echeloned. Pi is reduced and P4 both multicuspate and large. The
same fragmentation of the premolar genetic field has occurred here
as in the Ursidae and the Amphicynodontinae. However, the series
has not been broadened, although the teeth are both crowded and
echeloned. Very apparently, these latter three characters are neither
allelomorphs of premolar genetic field fragmentation, nor of each
other, nor are they inevitable ontogenetic sequelae of each other.

The Caninae proper seem never to undergo fragmentation of the
premolar field. The sequence is always graded in size and in com-
plexity increasing from front to rear. Frequently in Canis and in
Hesperocyon the premolar cuspidation varies individually and from
side to side of one individual. However, P2 is always part of a graded
size series, and always anticipates P3 in reduction of the accessory
cusps.

The Procyonidae show consistently broad premolars, but only
slight crowding in Potos and considerable crowding and echeloning
in Phlaocyon; Ailurus is questionably a Procyonid, and Ailuropoda
has been demonstrated to be an Ursid (Davis, 1964). They are
included on Table 3 for comparison, because they have previously

CLARK & GUENSBURG: ARCTOID CHARACTERS 48

been classified as Procyonids. Procyon, surprisingly, has the pre-
molars echeloned but not crowded, demonstrating that these two
unit characters are not genetically linked.

P2-3 are simple, save for vestigial accessory cusps in a few sub-
species of Bassariscus, but P3 is notably larger than P2 in all genera
excpet Potos and Phlaocyon.

The quadritubercular M^ of Procyon differs in essential details of
pattern from that of Parictis (Parictis). In Procyon lotor the proto-
cone remains very much larger than the hypocone, and retains its
primitive trigon crests, the anterior running labially to become the
anterior cingulum, and the posterior connecting labially through the
metaconule with the metacone. In Procyon cancrivorous, however,
the posterior crest of the protocone extends to the hypocone, produc-
ing a pattern quite like that of Parictis (Parictis).

From this point of view, the dental characters of Parictis resem-
ble those of the Procyonids only in the tendency to develop extra
cusps on P* and the molars, and in a general broadening of the pre-
molars. Fragmentation of the genetic field into a control of P^^
separate from that of Pj and the M series is an important develop-
ment by Parictis which has not, apparently, occurred in the Procy-
onidae. Dentally at least, Parictis more nearly resembles the Ursidae
in its genetic potential.

Comparison of Parictis with the other genera refen-ed by Simp-
son (1946, p. 140) to the subfamily reveals that the Old World genera
Amphicynodon, Pachycynodon, Plesiocyon, and probably Cephalogale
exhibit the same genetic trends to varying degrees. Equalization of
P2 and P3 is never developed to the degree attained by Parictis, but
some fragmentation of the premolar genetic field always occurs. The
various species of Hetnicyon, on the other hand, never display this
fragmentation at all. The premolars remain a graded series in the
lower jaw. They are simple, broad, and possessed of prominent cin-
gula; however, these latter characters are also well developed in such
unrelated genera as Procyon, Lutra, and Arctictis. These same pre-
molars (in Hemicyon) grade in size from P, through P4, and in the
upper dentition from Pi through P3. The tendency of P* to develop
as a straight, two-pronged shearing blade with a single median inter-
nal cusp (protocone?) differs radically from the trend already de-
scribed for Parictis. So also does the shearing alignment of the tri-
gonid cusps of Mi in both Hemicyon and Dinocyon. The Phosphorite
forms assigned by de Chardin (1915) to Cephalogale seem to parallel
the other Amphicynodontines rather than Hemicyon. For these

44 FIELDIANA: GEOLOGY, VOLUME 26

reasons we regard the Hemicyoninae as distinct from the Amphicy-
nodontinae sensu strictu (= Cynodontinae of Schlosser, 1923). The
reader must realize that this paragraph is based upon the literature,
because no specimens of Hemicyon or of the European Amphicy-
nodontines were available to us; our conclusions in this matter
should, therefore, be evaluated accordingly.

C. Basicranial Characters

In their classic taxonomic work on mammals. Flower and Lydek-
ker (1891) introduced the use of basicranial characters with dental
characters and various features of the soft anatomy. Most paleon-
tologists have employed dental characters chiefly or exclusively, due
in part to the preferential preservation of teeth and jaws. More re-
cently Segall (1943) and Hough (1948, 1953) have reintroduced to
American paleontologists the use of basicranial characters.

Few basicranial characters are preserved on the type of P. (C.)
personi. The relationships of Parictis to later forms, however, de-
pends so much upon the question of whether P. (C.) personi is actu-
ally Parictis, upon the relationships of the Amphicynodontidae in
general, and upon the relationships of such other forms as Hespero-
cyon and Mustelavus, that we feel justified in evaluating a few of the
characters most frequently used.

We have selected the following:

1. Alisphenoid canal

2. Carotid foramen and canal

3. Condylar foramen

4. Ossification of bulla

5. External auditory meatus

6. Partitions in bulla

7. Suprameatal vacuity

8. Mastoid and paroccipital processes.

These should be regarded as representative of a large constella-
tion of characters, some of which may be taxonomically more signifi-
cant than any of the above. However, all of the above have been
used by other authors.

We compared our observations with those of Hough (1948, 1953),
Flower and Lydekker (1891), Segall (1943), and less thoroughly with
those of other authors. Both study and comparison engendered con-
fusion. Variations occur at every level, from familial down to indi-

CLARK & GUENSBURG: ARCTOID CHARACTERS 45

vidual. Often a character which seems positively to differentiate
two subfamilies may vary individually within one species of one of
them. We found that ten individuals of one species was an absolute
minimum sample, and suspected that even then we were not seeing
the true range of variation. (Rarely, the excellent Field Museum
collection of Recent skulls could not furnish us with even this mini-
mal number of specimens per species). Of the nine characters listed,
only numbers 4 and 5 were reasonably consistent, with numbei*s 7
and 8 next in order of reliability. We shall discuss the characters
separately, stating the variations observed.

1. Alisphenoid canal. Flower and Lydekker ( 1891) recognized
this as absent in the Felidae and Hyaenidae, and present with rare
exceptions in the Viverridae; present in the Canidae and Ursidae
(except Aeluropoda), and absent in the Procyonidae (except Ailurus)
and Mustelidae. This we found to be true of Recent forms, with one
added observation: in Procyon, Nasua, Nasuella, Bassariscus, Bos-
saricyon, and Potos, a tiny nutrient foramen lies in the position of
the posterior opening of the alisphenoid canal; it varies from within
to anterior to the antrum of the Foramen ovale. It communicates
with a small sinus in the alisphenoid at the base of the pterygoid
wing. Might this perhaps be a vestige of the alisphenoid canal? We
found it also in F 68319 and F 85503, Felis pardalis, but absent in
three other specimens of F. pardalis and in several specimens of
Potos and Bassaricyon. We also found one unnumbered specimen
of Taxidea in which a foramen in this position penetrated into the
rear of the Foramen rotundum, resembling a tiny alisphenoid canal;
in one specimen of Lutra canadensis, F 53922, a canal on the left side,
in the position of the alisphenoid canal, leads not to the Foramen
rotundum but to the Foramen lacerum anteriu^. The specimen has
no canal on the right side. Among the Viverrids, one individual
Mungos, 5184, and two Paradoxuru^s, 62851 and 62861, lack the alis-
phenoid canal; all others that we inspected of those genera and of the
the family possess it.

In the American Oligocene genera the situation is very different.
We have studied specimens of Eusmilus sicarius, Hoplophoueus,
Dinictis, Daphoenus, Hesperocyon, and the type specimens of P. (C.)
personi and Mustelavus priscus. All of these diverse genera have a
definite alisphenoid canal, with the possible exception of Mustelavus,
in which damage precludes certain determination. The structure is,
therefore, simply a primitive structure in fissipeds, which has been

46 FIELDIANA: GEOLOGY, VOLUME 26

retained in some families and lost in others. Its atavistic recurrence
in one specimen of Lutra offers added evidence of this.

2. Carotid foramen and canal. Segall (1943, p. 35) gives an
excellent description of the anatomy of the canal for the internal
carotid artery of the dog. Sisson and Grossman (1938, p. 191) also
clearly describe this canal, pointing out its close relationship to the
petrobasilar canal (= inferior petrosal venous sinus of Segall, 1943).

The carotid canal in the Procyonidae usually lies entirely within
the wall of the bulla (always, fide Hough 1948, p. 81), but of 19 speci-
mens of Bassariscus which we examined, six had the basioccipital
forming the internal wall of the foramen and the first millimeter or so
of the canal. This was true also in three of six specimens of Bassari-
cyon, one of 19 of Polos, six of nine Nasua, five of six Nasuella, nine
of 20 Procyon lotor, and one of seven Procyon cancrivorus. Predict-
ably, the extent to which a very thin inner wall of bone may develop
from the bulla around the foramen is variable within the Procy-
onidae.

We belabor this point, first in order to correct the erroneous state-
ment by Hough, and, second, in order to indicate that this character
is of little importance in separating the Procyonidae from Hespero-
cyon. The latter genus has the carotid canal exactly in the position
of that in Bassariscus, except that the foramen lies a little closer to
the F. lacerum posterius. The canal lies completely within the bulla
for its anterior half, in F-UM418. The occipital bone forms part of
its dorso-mesial wall for the posterior part of its course. The same
seems to be the case in F-UM417, which we sectioned, but taphic
(burial) damage in this specimen makes exact determination impos-
sible.

The general procyonid arrangement of the carotid canal obtains
also among those mustelines and viverrids which fell within our pur-
view. The larger felids tend to have small but, in some individuals,
apparently functional carotid canals; in the smaller felids, as Davis
and Story (1943) noted, the internal carotid artery is imperforate
and the carotid canal small to absent. The six specimens of Crocuta
and Hyaena studied all had carotid apparatus, but it varied from
functional to very small; the carotid foramen varied in position and
in its relationship to the occipital bone; and variation occurred not
only between individuals but frequently between sides of one indi-
vidual.

Here again, apparently, we have as a primitive character a caro-
tid canal with its posterior foramen ventral to and close to the petro-

CLARK & GUENSBURG: ARCTOID CHARACTERS 47

basilar canal. Presumably, the carotid foramen primitively lies
between the bulla and the basi occipital. The procyonids, viverrids,
and mustelines have retained this structure with minor modifications;
the canids have moved it to a closer association or even junction with
the petrobasilar canal; and the felids and hyaenids are in the process
of losing it entirely.

3. Condylar foramen. This foramen, termed "Condyloid
foramen" by Flower, and "Foramen hypoglossi" by Sisson and
Grossman (1938, p. 193) is stated by Flower and Lydekker (1891,
p. 501) to be "concealed or wanting" in the Aeluroidea, present in
the Cynoidea (ibid., p. 38), and "distinct and exposed, and never
sunk into a common opening with the foramen lacerum posticiim"
(ibid, p. 557) in the Arctoidea. Segall (1943, pp. 48-49) is more
cautious and does not include this foramen as a useful character,
perhaps because it is not strictly auditory in affinities. Hough (1953,
p. 99) mentions that the condyloid foramen in Oligocene Felidae "is
large and well separated from the Foramen lacerum posterius. It is
interesting that in the true felid, Panthera atrox contemporary with
Smilodon, the two are well separated. Merriam and Stock report
that only 4 or 5 out of a total of 20 specimens have a condition
resembling that of Felis."

Hough did not look far enough. Of four Panthera leo in our col-
lection, three are "feloid," and one, F 14433, has a large condylar
foramen separate from the F. lacerum posterius. Of five Felis pardalis
observed, one, F68319, has the condylar foramen fully as separate
and prominent as does any canid. One Felis caracal, F43292, one
Lynx lynx, F51820, and one Felis temminckii, F72804, have a similar
"canoid" pattern.

Of the Viverridae, the following specimens show a "canoid"
separation of the condylar foramen:

Viverra tangalunga F 62864

Nandinia binotata F 25303, F 25306

Xenogale microdon F 25308
Paradoxurus

philippinensis F 62837

Mungos sp. F 51814

We did not by any means inspect all viverrid skulls in our collection.
This seemed enough to establish the point. Our five specimens of

48 FIELDIANA: GEOLOGY, VOLUME 26

Hyaena and Crocuta all have the condylar foramen large but sunk
within the raised rim which encloses also the F. lacerum posterius.

Separation of the condylar foramen from the antrum of the F.
lacerum posterius is a generalized or primitive condition which is
retained by all of the Arctoidea (sensu Romer, 1966). The Aeluro-
idea definitely tend to develop a raised antrum enclosing the two
foramina, with the F. condylare directed posteriorly rather than dor-
sally; but numerous individuals in every species except perhaps the
Hyaenids and some of the small cats retain the older arrangement.
The separation is primitive, not specifically "canoid" or "arctoid."

4. Ossification of bulla. The bulla was apparently ossified in
all American Oligocene carnivores except Daphoenus.

Hough (1948, p. 95) describes a demi-bulla in Daphoenus, but
does not identify the specimens upon which the description is based.
Scott (1937) also mentions such a specimen. Since no specimens
available to us retain this structure, we cannot determine the accu-
racy of Hough's statements. We can, however, determine that
Hough's (1948, p. 84; 1953, p. 97) description of the bulla in Nan-
dinia contains serious errors of fact. Reference is made to speci-
mens: F 54676, F 54415, F 73804, F 73803, F 73799, F 83642,
F 81601, F 83643, F 81602. All of these demonstrate the features
here noted. First the demi-bulla certainly does not "lie at a very
low angle"; also, the margins are not incurved, nor are they any-
where in contact with the promontorium. The postero-mesial mar-
gin of the tympanic is thick and deeply grooved for attachment of
the unossified entotympanic. Furthermore, two small bones lie
pressed against the sphenoid and anterior portion of the petrosal, as
continuations of the cartilaginous entotympanic. Very possibly they
represent ossifications of that element. Whether the failure to ossify
the entotympanic in this one genus represents retention of a primi-
tive character or a secondary development is a moot question.

Hough (1953, p. 97) is also, apparently, in error in the assumption
that there was in Hoplophoneus and Dinictis "the absence of any
trace of the septum bullae." Comparison of the Field Museum
specimens of Dinictis, UC 1443, and of Hoplophoneus, UM 701 and
P 12026, with Felis domestica, 0-1442, which has the bulla opened to
a degree similar to that of the fossils, reveals the true situation. The
anterior portion of the septum bullae in the Oligocene forms was fully
as well developed as in Felis; its posteriad extension cannot be deter-
mined. Since the septum bullae lies at the juncture of the ento-

CLARK & GUENSBURG: ARCTOID CHARACTERS 49

tympanic with the tympanic proper, these specimens prove that
both elements were developed and ossified in the Oligocene feloids.

5. External auditory meatus. Segall (1943) indicated that
the angle of inclination of, and elements composing, the external
auditory meatus are "sufficiently fundamental and consistent within
a given group." This we have abundantly confirmed, for recent
forms.

The fossil genera, however, present a different pattern. In all
three genera {Mustelavus, Parictis (Campylocynodon) , and Hespero-
cyon) the external meatus consists of a simple, thickened partial loop
of tympanic bone, not produced into a tube, closed dorsally by the
squamosal. Participation of the squamosal in Recent forms varies
from more extensive in the Procyonidae and Ursidae to less extensive
in the Mustelidae and excluded in the adult Canidae (Segall, 1943,
pp. 50-51).

Mustelavus has been variously referred to the Mustelidae and the
Procyonidae; Hesperocyon has been uniformly referred to the Can-
idae, with notes on its viverrid postcranial resemblances; and P.
(Campylocynodon) has been referred to the Amphicynodontinae of
the Canidae. Plainly, the meatus has not been used taxonomically,
and has become specialized at family level among Recent forms, but
was not so during the Oligocene.

6. Partitions in bulla. Segall (1943, p. 51) states that there is
"no septum in the bulla, except to a very slight extent in the Procy-
onidae with the exception of Bassariscus," and of the Mustelidae "an
obliquely horizontal septum in the anterior part of the bulla, always
accompanied by one or more approximately transverse ridges."

Hough (1948, p. 81) states flatly of the Procyonidae, "The bulla
is always globular, inflated, and without division into chambers.
There is no continuation of the tympanic cavity into any of the
elements, and there are no radiating rafters or septa in the interior
of the bulla."

Field Museum specimens confirm Segall and in several matters
refute Hough: The bulla in Potos and Bassaricyon is neither globular
nor highly inflated. We found no trace of an internal septum in 19
specimens of Bassariscus or in five specimens of Bassaricyon, al-
though in F 62079, Bassaricyon alleni, there is a transverse cartilag-
inous partition.

However, the situation in the other Procyonid genera differs
markedly from this. In Procyon itself, a small horizontal shelf lies

50 FIELDIANA: GEOLOGY, VOLUME 26

just ventral to the Eustachian opening; this usually terminates at
the inner wall of the carotid canal. In some individuals, two to four
cristae radiate ventrally from a point midway along the promontor-
ium. These occupy exactly the same position as the cristae in Taxi-
dea and other mustelines, but occur merely as low, sharp ridges
rather than full partitions. They seem to be commonest in juveniles
and in specimens from Florida. It is not clear to us what exigencies,
either of childhood or of life in Florida, require extra braces in the
bulla. More seriously, this individual, age, and geographic varia-
tion raises genetic and phylogenetic problems which should be stud-
ied. Both anterior shelf and cristae occur in Procyon cancrivorus as
well as in Procyon lotor.

Six specimens of Nasuella lack both shelf and cristae. Of nine
specimens of Nasua, three could not be studied due to elongate
meatus, one had no cristae or shelf, and five had a very definite
anterior shelf. In fact, Nasua narica F 34330 had a shelf exactly
like that of the canid Pseudalopex, F1377, except that it did not ex-
tend as far posteriorly.

Hesperocyon, alone of the Oligocene genera available, has the
bulla well enough preserved to reveal its structures. Three speci-
mens, F-UM418, F-UC496, and F-UC495, have been prepared to
reveal intrabuUar structures. In 495 there is no transverse anterior
shelf, but a low sharp, denticulate ridge starts at the anteromedian
corner of the bulla and extends back as a fringe along its inside ven-
tral surface for its entire length, vertical and roughly parallel to the
rim of the internal auditory meatus. A series of tiny ridges, narrow
but high, cover the bulla surface like reticules of thread for a short
distance on each side of the main ridge. These are developed again
in the depression immediately ventro-mesial to the carotid canal.
The remainder of the bulla is smooth.

Number 496 cannot be prepared to show much. It does reveal
that the reticulations cover the entire mesial wall of the bulla.

In 418, a small, sharp transverse shelf lies ventral to the Eusta-
chian canal; it extends to the median wall of the bulla, and meets the
denticulated anteroposterior ridge almost at a right angle. The retic-
ulated area in this specimen, as in 496, covers the entire wall of the
bulla from the carotid canal to the longitudinal ridge.

Briefly, in our three specimens the horizontal transverse shelf is
exactly like that of the Recent Procyonidae, but the longitudinal
ridge and reticulate area resemble nothing we have seen elsewhere.

CLARK & GUENSBURG: ARCTOID CHARACTERS 51

The bulla resembles that of Bassariscus in general size, shape, and
position of carotid foramen but diffei*s from it in possessing these
three internal structures and in lacking a produced external auditory
meatus. It differs from most Recent Canidae in possession of these
three internal structures. The partition in canids arises usually from
the inrolied dorsomedial rim of the bulla where it lies upon the prom-
ontorium, or from the carotid canal. Thence it extends as a thin
plate ventro-anteriorly to the front of the bulla. This does not, in
our opinion, homologize well with the ventral denticulated crest of
Hesperocyon.

However, a specimen of Vulpes fulvus, F 44743, has a horizontal
partition very like that of the Procyonidae and of Hesperocyon. This
specimen has the carotid canal completely roofed over by the bulla
from the level of the cochlea forward. A specimen of Lycaon, F
35122, has a horizontal anterior partition extending latero-anteriorly
from the carotid canal; a series of low, heavy rafters radiates down
the medial wall of the bulla from the area of the promontorium.

Apparently, neither a horizontal partition, nor the roofing of the
carotid canal, nor the presence of low rafters differentiates surely
between Recent Canidae and Procyonidae. This is not serious when
one has a large series of skulls from different genera, species, and even
geographic areas for study, but it vitiates these characters in cases
involving only one or a few specimens. Attempted application of
these characters to a few Oligocene specimens, where one is not sure
what constitutes a primitive character developed in many taxa,
could be exceedingly misleading. Incidentally, Hough (1948, p. 79)
was in error again with the statement that in the Canidae "there are
no rafters or radiating ridges."

The case becomes even more complex when one studies Otocyon.
We had available the following specimens: F 612, F 38189. F 38190,
F 38420, F 43334, F 73040, F 73041, F 73042, F 73043, F 38419, one
uncatalogued.

All except F 38419 and the uncatalogued specimen have the bulla
absolutely without internal partitions or rafters; it is as smooth and
empty as that of Bassariscus. The last two have a very low but
sharp anterior horizontal shelf just ventral to the Eustachian canal;
this exactly resembles the shelf of Procyon. Furthermore, in at least
six of these 11 the carotid canal lies completely within the bulla for
the anterior half of its length.

Fennecus zerda presents the ultimate situation we have met, in
variations of intrabullar structures. Of 12 specimens which we stud-

52 FIELDIANA: GEOLOGY, VOLUME 26

ied, only three have what could be regarded as a canoid shelf or
partition, and in all of these the partition seems to be merely an over-
developed anterior crista. All have one to four cristae radiating from
the carotid canal; in addition, a posterior crista extends as a flange
from the large tubular prominence which encircles the entry to the
Foramen lacerum posterius. Several possess a small, blade-like,
anterior horizontal shelf below the Eustachian tube, like that of
Procyon; others do not. Three have a thin, flat blade arising ver-
tically from the ventral floor of the bulla, set transversely; in two
others, the anteriormost crista twists a little posteriorward to join
this. In 106644, two of the cristae have needle-like extensions which
terminate in hollow globes of bone. In 93864 the anteriormost crista
joins the horizontal shelf. This crista, therefore, certainly is not
homologous with the internal partition of most other canids.

7. Suprameatal vacuity. Segall (1943, p. 50) lists this vacuity
as the fourth, and an extension of the tympanic cavity posterior to
the bulla as the sixth, of the characters which he found reasonably
consistent within the Arctoid families. We have also found this true,
with the exception Segall noted, that Potos resembles the Mustelidae
in the posterior extension of the tympanic cavity. We know of no
Oligocene carnivores which exhibit either of these cavities well devel-
oped; however we have no specimens of Bunaelurus available to us.

The exception should also be noted that in Hesperocyon there is a
small pocket which could become a suprameatal sinus. In all but
one of our specimens, this pocket is confined to the squamosal; in
F-UM387 this pocket pierces the thin squamosal plate and has as its
posterior wall a shallow indentation in the mastoid process.

8. Development of mastoid and paroccipital processes.

Although these processes vary individually in cross-sectional shape,
amount of contact with neighboring bones, and length, they remain
fairly constant within each family in terms of size relative to each
other and of general relationship to the bulla.

Among Oligocene carnivores, the situation is once more confusing.
The basicranium of Parictis (sensu strictu) is unknown. Hesperocyon,
Mustelavus, and Parictis (Campylocynodon) have the two low, sub-
equal with the paroccipital a little larger, and not in contact with the
bulla. This is also the condition in the Procyonidae, except that in
Procyon the paroccipital is longer, ventrally directed, and at its base
makes contact with the bulla. Here again, we seem to be dealing
with a primitive arrangement rather than a dependably Procyonid
trait.

CLARK & GUENSBURG: ARCTOID CHARACTERS 53

Summary of hasicranial characters

1. Alisphenoid canal. A generalized or primitive character
present in all Oligocene carnivores. Usable only to diflFerentiate
Recent families, and not consistent at superfamily or suborder level.

2. Carotid foramen and canal. Present in Hesperocyon and
P. (Campylocynodort). In the latter, all but the posterior millimeter
of the canal lies completely enclosed by the bulla; in Hesperocyon,
the anterior half is so enclosed. Presumably, possession of a poste-
riorly-located foramen and canal roofed by occipital bone is prim-
itive. Enclosure within the bulla is one advance, and joining with
the petrobasilar canal, a divergent one. Elimination is a still further
development.

3. Condylar foramen. Separation from the F. lacerum pos-
terius is primitive and characterizes all American Oligocene carni-
vores. It also characterizes the Recent Arctoidea, and numerous
individuals among various felid and viverrid genera and species.
This is not a reliable character for distinguishing family relationships
of fossil taxa.

4. Ossification of bulla. A demi-bulla occurs in Daphoenus,
a well-ossified but very thin bulla in Dinictis and Hoplophoneus, a
well-ossified bulla in Hesperocyon and Mustelavus, and either a
demi-bulla or a fragile bulla in P. (Campylocynodon) . The bulla is
complete and well ossified in all modern carnivores except the viver-
rid Nandinia which has a demi-bulla comprising the tympanic, and
an entotympanic unossified save for two medial splinters. As a
taxonomic character useful in classifying Oligocene arctoids, this
feature is meaningless.

5. External auditory meatus. Once more we have a charac-
ter useful in classifying Recent families (Segall, 1943). Also, once
more we have a character which does not apply to American Oligo-
cene carnivores. None of the Oligocene forms have a tubular,
extended meatus, and in none does the tympanic form a complete
meatal ring. Even in Dinictis and Hoplophoneus the squamosal
forms the postero-dorsal part of the auditory antrum.

The meatal ring faces antero-laterally at an angle of 20''-30° for-
ward of a true transverse plane, in Hesperocyon, Parictis {Campylo-
cynodon), Dinictis, Hoplophoneus, Nandinia, Ba^sariscus, and Pro-
yon. In the latter two genera, the tubular meatus outside the rings

54 FIELDIANA: GEOLOGY, VOLUME 26

obscures the position of the ring. In Mustelavus the direction cannot
be determined due to postmortem damage.

The tubular meatus is of no assistance because it does not occur
in American Oligocene forms. The direction of the meatal ring is
inconsequential because it is the same in all known forms.

6. Partitions in bulla. Certainly the Recent Arctoidea and
Aeluroidea have recognizably different internal structures in the bul-
lae. Certainly also the evidence is clear that within the Arctoidea
these structures are not reliable at family, generic, or specific level.
Only on a statistical basis is it safe to employ them, although in some
particular cases (notably the presence of large transverse cristae in
Mustelid bullae) a character may be usable. Generally these char-
acters are inclusive rather than exclusive: if a bulla has a "canoid"
partition it may be a canid; if it lacks such a partition it may less
probably be a canid, as in Fennecus zerda.

In the Oligocene forms of interest here, Hesperocyon has a pattern
different from any we have seen in recent Arctoids. However, the
pattern is no more different than are those of Fennecus from those of
Canis, and it includes a horizontal shelf like that of many Procyonids.
The internal pattern is not determinable in Mustelavus and P. (Cam-
pylocynodon) .

7. Suprameatal vacuity. This vacuity characterizes the
Recent Procyonoidae. The Mustelidae, as Segall (1943, p. 51)
states, have a posterior extension of the tympanic cavity which, in
the Mephitinae, communicates with a sinus in the mastoid. Segall
also notes that the canids, Nasua, and Bassaricyon "show a posterior
extension of the cavity but not to so marked an extent as in the
Mustelids." We found one specimen of Bassariscus in which the
suprameatal sinus definitely inflated the mastoid. Several specimens
of Procyon had the sinus terminating within the squamosal or merely
partially floored by the posterodorsal terminus of the tympanic ring.
This sinus pocket has been described above in Hesperocyon; it is
slightly smaller in the type specimens of Mustelavus priscus and
Parictis iC.) personi. Hough (1948, p. 87) describes a similar supra-
meatal fossa in the European Plesictis. We are once more dealing
with a character common to several Oligocene genera, but in this
case it does not occur in all. Daphoenus, Hoplophoneus, and Dinictis
do not possess this sinus, although the posterior buttress of the
tympanic ring does project out into the meatal tube.

CLARK & GUENSBURG: ARCTOID CHARACTERS 55

8. Mastoid and paroccipital processes. The development of
these two processes relative to each other seems to be fairly constant
within families, although individual, specific, and especially generic
variations in the size and shape of both may make determination
difficult. Generally, any one genus of the Recent Procyonidae,
Ursidae, and Mustelidae has the two processes of about equal length ;
in the Canidae the paroccipital is always much the larger. The
Viverridae also have the paroccipital much longer than the mastoid,
although they differ from the Canidae in having the paroccipital
closely appressed to the bulla (except in Nandinia, where the ento-
tympanic is unossified). Hesperocyon, Mustelavus, and P. (Cam-
pylocynodon) all have the two low and equally developed.

The reliability of the eight characters studied may be encapsu-
lated thus:

1.

Alisphenoid canal

Primitive, generalized

2.

Carotid foramen and canal

a. Separate foramen

b. Partial occipital roofing

Primitive
Primitive

3.

Separate condylar foramen

Primitive

4.

Ossification of bulla

a. Complete ossification

b. Peloid entotympanic

Usual by Middle Oligocene time
Present during Middle Oligocene time

5,

Tubular external meatus

Advanced; not present in Middle
Oligocene forms.

6.

Internal structures in bulla

Generally variable in Recent forms;
absent to limited presence in Oligocene
forms.

8. Relative development of mastoid Similar size in all Oligocene arctoids;
and paroccipital processes diff"ers at family level among Recent

ones.

Considering these eight characters in terms of their taxonomic
usefulness, numbers 1, 2, 3, and an equal development in 8 may well
be regarded as primitive. They are present in all Oligocene forms.
Their retention in some taxa, and atavistic recurrence in individuals
within other taxa, clearly signifies this. Numbers 4, 5, 6, and 7 are
more advanced characters, developed in some taxa and absent, dif-
ferently developed, or variable in others. Number 8 has as its arche-
type a low, similar length in the two projections; variations from this
may be regarded as specializations.

The generally high variability in detail and frequent occuiTence
of atavism may well indicate that these characters have little or no
survival significance. Genes for the more primitive conditions there-
fore continue to exist within the gene pools of the various taxa after

+

+

+

+

+
(

- May
entirely

lie almost
In bulla

7

+

+

+

+

+ ?

+

+

+

+

+

?

?

+ small

+ small

+

small

56 FIELDIANA: GEOLOGY, VOLUME 26

differentiation on more adaptive levels; given the proper genetic
combination by statistical accident, the primitive phenotypes natur-
ally recur.

Basicranial characters in three Oligocene Arctoids. The table below
summarizes the occurrence of these characters:

Character Hesperocyon P. (Campylocynodon) Mustelavus

Alisphenoid canal

Carotid foramen and
canal

Condylar foramen

Ossification of bulla

Primitive Auditory meatus

Partitions in bulla

Suprameatal sinus

Equal mastoid —
paroccipital processes + + +

The table demonstrates that on the basis of these characters one
could assign all three genera to a single family. Indeed, justification
must be offered for doing otherwise. Every one of the characters
used by Hough (1948, pp. 78-90) to demonstrate that Plesictis and
Mustelavus are Procyonids is either erroneous, misinterpreted, or so
primitive and general as to apply equally to Hesperocyon and to P.
(Campylocynodon) .

All of this signifies that the Procyonids might equally well be
derived from Hesperocyon as McGrew (1938) maintained, or from
Parictis as Clark (1937) and later Clark et al. (1967) have suggested.
The actual meaning of this similarity of characters is more important:
it substantiates Chaffee's (1954, p. 46) contention, that such an
auditory region is primitive in carnivores. By this interpretation,
the Procyonids have simply retained, with slight modifications, a
primitive pattern in the auditory region. The carotid anatomy can-
not, therefore, be used to indicate more than a common remote an-
cestory, leaving open the possibility of closer relationship between
the Amphicynodontinae and the Procyonidae. Furthermore, it does
not support either Parictis or Hesperocyon as a direct ancestor of
Bassariscus. It should be noted that the anatomy of the interior of
the bulla in Hesperocyon resembles in all but three features that of
Bassariscus. The suprameatal fossa (see Segall, 1943, p. 39) is

CLARK & GUENSBURG: ARCTOID CHARACTERS 57

present in Hesperocyon but as a shallow pit which could develop into
a suprameatal fossa.

D. Intracranial Characters.

In a further effort to evaluate the possible familial relationships
of the Amphicynodontinae, studies were made of the intracranial
venation of Hesperocyon, Daphoenus, Hyaenodon, Ictops, Merycoi-
dodon, Procyon, Ursus, Thalarctos, Homo, Equus, and other repre-
sentatives of living families. We studied the region of the tentorial
apparatus, with only confirmatory observations on the basicranial
venation reported upon by Guth (1964).

The tentorial structure is primarily controlled, naturally, by the
configuration of the brain, particularly by the extent of overgrowth
of the cerebrum over the cerebellum. Secondarily, the tentorium
develops a series of structures supportive to the venous circulation.
Changes in the tentorium and in this circulation necessarily react
upon each other. A causal determination of an individual structure
is, therefore, impossible except as it relates to cerebral development
as a final cause. However, developmental stages are easily recog-
nizable.

The "most primitive" is, apparently, the insectivore stage repre-
sented by Ictops. Here the tentorium consists dorsally of a very low,
blunt ridge which arises from obscure eminences enclosing a cruciate
depression at the midline, and gi'adually increases in size as it passes
lateroventrally. It comprises the posterior edge of the parietal, and
possibly also the anterior edge of the supraoccipital. In a sharp
angulation, a keeled crest consisting of the squamosal and part of
the petrosal passes antero-medially across the floor of the calvarium.

The sagittal sinus occurs as a groove on some skulls; on others it
leaves no trace. It bifurcates normally at the position of the ten-
torium, and the right and left sinuses transversales progress ventrally
to the squamosal-parietal sutures. Here they bifurcate; the courses
of the branches from this point on differ completely from the cor-
responding circulation of any carnivore known to us. This part of
the venous circulation is certainly not "primitive" to that of the
carnivores in the sense that the latter could be derived from it by
any known or easily hypothesized gradation of steps. We therefore
relegate further description to students of the Insectivora.

Hyaenodon (figs. 15, 16) presents a very different picture. The
tentorial structure is massive, as Guth (1964, pp. 35-38) has indi-
cated. Our specimen of H. criientus (FM-P12723) however, reveals

58

FIELDIANA: GEOLOGY, VOLUME 26

\

N

Fig. 15. Anterior view of tentorium of Hyaenodon cruentus, FM-P12723. The
sinus sagittalis divides into the sinuses transver sales, which pass lateroposteriorly
across the lip of the tentorium.

characters sufficiently different from those Guth described for H.
vulpinus to merit separate description.

The tentorium consists of a dorsal and two lateral parts, sharply
distinct. The dorsal portion is a roughly triangular wedge of bone
with a horizontal base projecting downward. The posterior face is
deeply excavated for the median lobe of the cerebellum. A triangular
median prominence projects its rounded prow forward from the ven-
tral rim ; the sagittal sinus divides immediately anterior to the dorsal
keel of this prominence, and the transverse sinuses proceed latero-
ventrally down each side of it.

The lip marking the transit of the transverse sinuses from the
cerebral to the cerebellar space also marks the boundary between the

CLARK & GUENSBURG: ARCTOID CHARACTERS

59

dorsal and lateral portions of the tentorium. Below this lip, the ridge
which forms on each side the outer boundary of the sagittal and
transverse sinuses curves laterally and ventrally as the blunt edge of
a heavy wedge of bone. This is the lateral tentorium. Its basal
portion is fluted to a sharp crest by passage of the sinus cavernosus.
Two distinct, almost separate venous systems invest this massive
bony structure. The sagittal sinus, as already mentioned, divides
in normal fashion into the two sijius transver sales, each of which
passes over the lip of the tentorium to enter the sigmoid sinus. This
is broad and but little impressed in the cranial wall. It drops quite
sharply into a very deep, partially roofed sinus postero-lateral to the
petrosal bone, which may be a deep portion of the sigmoid sinus.
The postero-internal end of this connects through the occipital sinus

I-

1 cm

-\

Fig. 16. Anterolateral view of the tentorium and temporal area of Hyaenodon
cruentus, FM-P12723. The separation of the intracranial sinuses from the in-
traosseous temporal sinus, which receives blood through the temporal foramen, is
shown.

60 FIELDIANA: GEOLOGY, VOLUME 26

(here roofed over to form a canal, as in most later forms) with the
occipital foramen and, presumably, the vertebral vein. Connection
is also clear to the Foramen lacerum posterius and the internal
jugular vein. Thus the sagittal venous system consists of the median
sagittal vein which bifurcates into the transverse venous sinuses.
These pass in a generally postero-ventral path around the cerebellum
to their respective double outlets via the vertebral and jugular veins.
The whole is shaped like a median tube leading downward via a
sloping, open-bottomed partial ring to two outlets on each side.

The second or temporal venous circulation is entirely different.
It starts as right and left temporal foramina which pierce the outer
skull wall very close to the sagittal crest. These connect with the
intraosseous transverse sinus, which leads from a median anastomosis
down an intra-osseous temporal canal (Guth's figures label this the
"Sinus transversus") to find its eventual outlet through the post-
glenoid foramen. It receives, apparently, only one internal tribu-
tary, a small sinus which enters the middle cerebral wall about 2 cm.
anterior to the tentorium. A very small commissural foramen lying
within the sinus transversus on the anterior lip of the tentorium, and
another equally insignificant on the posterior wall of the tentorium
within the shallow part of the sigmoid sinus, seem to furnish the only
transmission between the sagittal and the temporal systems.

Thus the H-shaped temporal system primarily serves the cranial
bones and temporal muscles, while the sagittal system drains the
brain. Commissural connection between the two systems is very
slight, not as reported by Guth (1964, p. 37) for H. vulpinus: "Mais
il porte sur son versant posterieur, au niveau de la large fossa suhar-
cuata, un orifice par lequel le sinus sigmoide ^tait en rapport avec le
sinus transversal."

The crania of Hesperocyon (ref: FM-UC 1405 and UC 495) and
Daphoenus (FM-UC 1426) display notable but simple modifications
of this pattern.

Expansion of both major brain segments, but most notably of
the cerebrum, has flattened the tentorium to thin triple plates of
bone. The parietal comprises the dorsal plate and the dorsal por-
tions of the lateral plates; the squamosal and probably also the alis-
phenoid partake in the construction of the ventral moiety.

An important novelty appears on the dorsum of the dorsal plate
in Hesperocyon (fig. 17A). The sinus sagittalis loops forward from
the intracranial surface and runs anteriorly, undivided, along the

1

Fig. 17. The tentorium in Oligocene Arctoids. A. .\ntprolateral view of the
tentorium in Hesperocyon gregariiia, FM-UC 1405. Note the media! channel for the
recurved sinus sagittalis. B. Anterolateral view of the tentorium in Dnphoenus
vetus, FM-UC 1426. The median structure here form.s a tran.sverse dorsal barrier,
although it retains a sagittal groove.

61

62 FIELDIANA: GEOLOGY, VOLUME 26

midline of the tentorium to pass over its lip. A series of paired sup-
porting structures, better figured than described, arise from the ten-
torial dorsum to guide the sagittal vein in its course. The cranial
wall posterior to this does not reveal where the sagittal sinus bifur-
cates to become the sinuses transver sales, but presumably this occurs
somewhere on the posteroventral surface of the dorsal tentorial plate.

The separation into temporal and sagittal systems remains,
apparently, intact. Fracture by crushing in critical areas has placed
in doubt the possible existence of a foramen on the ventro-posterior
face of the tentorium, communicating with the temporal circulation.
However, appearances are that such an opening did not occur.

Daphoenus (fig. 17B) , on the other hand, exhibits two important
innovations. A large foramen leads from the posterior end of the
sagittal sinus directly into the transverse temporal commissure.
There can be no doubt that most if not all of the blood from the
cerebral dorsum thus passed from the sagittal sinus into the lateral
temporal canals. Another, but smaller, foramen establishes com-
missural connection between the sigmoid sinus and the ventral por-
tion of the temporal canal, perforating the posterior wall of the ten-
torium. Thus blood from the sagittal sinus could drain, via the
temporal canal, either back into the sigmoid sinus and thence to the
vertebral vein, or ventrally via the postglenoid foramen to the inter-
nal jugular or (possibly) via the foramen lacerus posterivs to the inter-
nal jugular.

Despite this temporal short-circuiting of the sagittalis-sinus
transversus system, the dorsum of the tentorium in Daphoenus
exhibits a set of structures almost as complex as those of Hesperocyon.
Either the sagittalis — sinus transversus system maintained itself in
attenuated form, or the evidence of its previous existence persisted
as tentorial relict structures.

Guth (1964, pp. 40-41) mentions no such tentorial structures or
double venous circulation in Pachycynodon. However, he consist-
ently refers to the structure we identify as the "temporal canal," as
the "sinus transversus." This change in nomenclature accounts for
some, but by no means all, of the apparent differences in anatomy.

The importance of all this becomes apparent when one compares
the structures described with those in living arctoid families.

The pattern in the Canidae, Mustelidae, and Procyonidae differs
from that of Daphoenus in but two significant respects: (1) the tem-
poral foramen piercing the outer wall of the skull is absent; (2) the

CLARK & GUENSBURG: ARCTOID CHARACTERS 63

tentorium is reduced to a thin shell, lacking any dorsal structures.
The first of these seems to achieve merely a separation of cranial
from a relatively unimportant extracranial drainage. Tentorial
reduction is greatest in the Canidae; less in the Mustelidae; and least
in the Procyonidae. The tentorium of Bassariscus closely approx-
imates that of Hesperocyon in size and conformation, differing from
it only in the presence of a sagittal foramen and a postero-lateral
foramen, and the absence of dorsal structures.

The Ursidae, however, tend to retain the external temporal
foramen. Most Field Museum specimens of Ursus (= Euarctos)
americanus possess this aperture. One specimen of Tremarctos
ornatiis, and a scattering of individuals from other genera and species,
also exhibit it. Various ursids also retain on the dorsum of the ten-
torium vestigial structures which parallel those of Daphoenus and
Hesperocyon in position.

We regard the sequence we have described, Ictops — Hyaenodon —
Hesperocyon — Daphoenus — Ursus — Procyon, as representing stages
of development. It is not to be considered in any sense a phylogeny.
Obviously, none of the presently-known specimens of Parictis pre-
serve this structure in usable condition. If we may presume that
the family Amphicynodontidae is as close-knit in other sets of char-
acters as in dental ones, however, Guth's description of Pachycynodon
should give us a general key. Unfortunately, his preoccupation with
the basicranial circulation, precludes clarification of many of the
points observed here. Generally, the tentorial venous circulation
seems to be about at the stage described here for Daphoenus. No
mention is made of tentorial structures in Pachycynodon.

E. Summary of Evaluation of Characters.

Dental, basicranial, and certain intracranial sets have been sep-
arately evaluated. All three areas are complex: some developments
have probably progressed through similar stages several times,
independently, while others have been unique in certain lines. Den-
tal assemblages show unit characters which seem to arise independ-
ently in several lines; probably such characters also occur in the basi-
cranium and intracranium, if we could recognize them.

Unit dental characters which have arisen independently in several
lines comprise depth of jaw, massiveness of jaw, development of
premolar cingulum, broadening of premolars, crowding, and eche-
loning.

64 FIELDIANA: GEOLOGY, VOLUME 26

Enlargement and complication of the premolar field progressively
forward from P^ seems to have been a general arctoid trend. Frag-
mentation of the field, with P^ separating from the remainder, was
also general. Further fragmentation, with P|:f coming under a
single control separate from F^, has been only partially achieved
except in Uintacyon, Amphicynodontinae, and the Ursidae, In some
of these latter two, and only in these two, Pf and Pf achieve complete
equality of size and of secondary simplification.

The probability seems high that Uintacyon is not a direct ancestor
of either the Amphicynodontinae or the Ursidae. We must there-
fore be prepared to accept the conclusion that this complex sequence
of change in the premolar genetic field could and probably did occur
independently at least twice. Within each sequence, stages would
have taxonomic significance, but one must look to other characters
to determine that the two sequences are parallel rather than parts
of one genetic system.

The greater probability that the Ursidae may be derived from
some genus of Amphicynodontine has been suggested many times,
most recently by deBonis (1969). Parictis almost certainly is not
the genus in question: constriction of the trigonid in Mg militates
against it.

This approach to dental evolution definitely indicates that Clark
(1937, 1967) was in error when he regarded Parictis as the ancestral
Procyonid. Similarity of unit characters, molar characters, and
simplicity of premolars were misleading. He overlooked the fact
that fragmentation of the premolar field and equalization of Pfif are
nowhere as advanced in the Procyonidae as in Parictis.

The auditory characters are inconclusive. Hesperocyon, Mustel-
avus, and Parictis seem to have possessed the same structures. Very
similar patterns occur in the Procyonidae and in juvenile bears.
Chaffee's (1954, p. 46) and Hough's (1944, p. 478) suggestion that
this is a primitive Arctoid pattern seems most probable. In that
case, auditory and basicranial characters might have a very limited
taxonomic significance within the older carnivores. It might also
indicate a generally closer relationship of the Procyonidae and the
Ursidae to each other than to other Arctoid taxa. In the latter case,
one might be forced to stretch the definition of the Procyonidae to
include Hesperocyon.

The entire question of the relationship of the Amphicynodontinae
to the Procyonidae is oddly dependent upon the relationship of two
genera, Hesperocyon and Bassariscus, to each other. If Bassariscus

CLARK & GUENSBURG: ARCTOID CHARACTERS 65

is indeed a primitive procyonid from which the others have been
derived, and if Hesperocyon is ancestral to it, then plainly the Procy-
onidae do not derive from the Amphicynodontinae. Both of these
relationships have been frequently proposed. Clark's erroneous
phylogeny based upon misinterpreted dental resemblances has been
the only serious stumbling block in recent years.

The differences between the two genera, in the bulla and in P*.
are plain. None of the auditory characters of Hesperocyon seem to
exclude the possibility of an ancestral relationship. The longitudinal
partition is so unlike anything we have seen in the Canidae that we
cannot accept it as homologous with the various canid partitions.
Much of the resemblance of Hesperocyon either to the Canidae or to
Bassariscus seems to represent mutual retention of primitive char-
acters in each instance, rather than a sound, phylogenetic develop-
ment.

At present, on partly subjective gi'ounds, we favor the relation-
ship of Hesperocyon to BcLssariscus, which would exclude the Amphi-
cynodontinae from a close relationship to the Procyonidae. This
must be recognized as a tentative opinion rather than a firm hypothe-
sis.

The present study has certainly revealed grounds for sceptical
review of Hough's (1948, p. 87) reference of Plesictis and Mustelavus
to the Procyonidae.

All of Hough's reasons for refen-ing these genera to the Procy-
onidae depend either upon the mutual retention of primitive auditory
characters or upon apparently late alterations of the stylomastoid
area due to backward inflation. Her arguments could with equal
cogence justify referring Hesperocyon to the Procyonidae. She does
not mention that Plesictis has lost Mij, and Mustelavus M2, reduc-
tions which do not occur in any Procyonid; indeed, the entire family
rather tends to enlarge this tooth. Since the basicranial evidence for
a Procyonid assignment of Mustelavus is based upon misinterpreted
primitive characters, and the dental evidence is strongly opposed to
it, we continue to regard Plesictis and Mustelavus as Mustelidae.

The tentorial-venous progression which we have proposed would
rank Daphoenus as more advanced than Hesperocyon, although it is
certainly more primitive in dental characters (retention of M^) and
in failure to develop an osseous bulla. Very apparently, these three
sets of characters are not genetically related to each other. Progres-
sion in any one is not necessarily concomitant with progression in
the others. Phylogenetic taxonomies based upon any one must fall

66 FIELDIANA: GEOLOGY, VOLUME 26

into the twin errors of assuming parallel developments to be related,
and assuming advance in one set to mean an advanced evolutionary-
stage of the whole organism or taxon.

We believe that these three plus several other sets of characters
must be analysed. Four categories of genetically controlled charac-
ters: allometric, unit, linked, and separate but parallel field progres-
sions, must be differentiated. Only when this is done can the confu-
sion now extant in understanding of carnivore phylogenies be
resolved.

In this paper, we have attempted to analyse characters in one
poorly-represented genus clearly marked by a unit character (com-
pressed trigonid of M2) within a subfamily characterized by a series
of unit characters (broadened premolars, echeloned premolars, low-
crowned teeth, massive jaw). Within this genus we found a series
of stages in evolution of the premolar field, arriving at a situation
similar to that in the Ursidae. The family is characterized by a
carotid canal-bulla stage similar to that of Hesperocyon, only slightly
less developed than that of the Procyonidae and juvenile Ursidae.
A related genus (Pachycynodon) within the family apparently has
achieved a stage of venous-tentorial development roughly equal to
that of Daphoenus or the Ursidae.

V. POSSIBLE ALTERNATIVE TAXONOMIES

The genus Parictis itself can be divided into three subgenera.
The earliest of these, P. (Campylocynodon), includes two species, each
represented only by the type specimen. The latest subgenus, P.
(Parictis), includes one species represented by three specimens. These
two subgenera display characters developed in accordance with their
relative stratigi-aphic positions, which may be a real situation or an
artifact of too small and fragmentary a sample.

The intermediate subgenus, P. (Subparidis) , includes four spe-
cies; one of them, P. (5.) montanus, is clearly distinct from the others,
and occurs in a geographic area separate from them. Two, P. {S.)
dakotensis and P. (S.) gilpini, overlap completely in their strati-
graphic and geographic ranges. The fourth species, P. (S.) major,
differs so much from the others that it may be referable to another
genus. Since the type and only known specimen does not include
the diagnostic M2, determination of the relationships of P. (S.) major
must await future discovery.

One could, therefore, find justification for the following classifi-
cations:

A. Genus Campylocynodon
C. personi

C. parvus
Genus Parictis
P. gilpini
P. dakotensis
P. montanus
P. major
P. primaevus
or, alternatively, grouping as far as possible:

B. Genus Parictis
Subgenus Campylocynodon

P. (C.) personi; synonym:
P. (C.) parvus

67

68 FIELDIANA: GEOLOGY, VOLUME 26

Subgenus Paridis
P. (P.) dakotensis; synonym:

P. (P.) gilpini

P. (P.) montanus

P. (P.) major
P. (P.) primaevus

It is apparent that the classification we propose in this paper is
closer to the splitting than to the grouping philosophy. We do this
because we believe that in this case larger numbers of more complete
specimens will confirm the phylogeny suggested, and the classifica-
tion followed here best expresses that phylogeny. However, we cer-
tainly do not regard either the phylogeny or the classification as
definitely established. Our suggested classification is:

C. Genus Paridis

Subgenus Campylocynodon

P. (C.) personi

P. (C.) parvus
Subgenus Subparidis

P. (S.) gilpini

P. (S.) dakotensis

P. (S.) montanus

P. (S.) major
Subgenus Paridis

P. (P.) primaevus

At the next higher level, possibilities are too numerous for simple
tabulation. It might be reasonable to raise the Amphicynodontinae
(minus Hemicyon) to family rank, to include all arctoids with similar
temporal circulation, tentorial apparatus, and carotid canal-bulla
anatomy. This would necessarily, include Hesperocyon, which might
be in a subfamily separated dentally from the Amphicynodontinae
sensu stridu.

Otic anatomy plus the presence of M^ would exclude Daphoenus
and Daphoenocyon. The latter would then comprise a slender-den-
tition and broad-dentition pair, roughly parallel to Hesperocyon and
Paridis among the "Amphicynodontidae."

A slightly more vertical classification might place Hesperocyon
among the Procyonids, based upon its possible near relationship to

CLARK & GUENSBURG: ARCTOID CHARACTERS 69

Bassarisciis. It might even be defensible to erect a single family to
include the Amphicynodontinae as one primitive subfamily, a newly-
erected "Hesperocyoninae" as a second, and the "Procyoninae" as
a third and more advanced one.

All of this seems to us to indicate that much more knowledge is
needed before the phylogeny of the Arctoid carnivores, and with it
a reasonable taxonomy, can be established. Research should be con-
ducted along three lines simultaneously, with interchange of infor-
mation during progress: first, intensive studies of several represen-
tative living forms, after the pattern of Davis (1964) should be vig-
orously pursued; second, analyses of the genetic significance of var-
ious characters should be carried much further than the preliminary
and necessarily superficial efforts of this paper; third, exceedingly
thorough studies of all observable characters of the fossils available
should be made.

Until such work is conducted, a revision of the existing classifi-
cation would be premature. At present, we therefore adhere to it, as
presented in section III of this paper. However, we feel that revision
is highly desirable.

REFERENCES

DE Bonis, Louis

1969. Remarks sur la position systematique des Amphicyon. Acad. Sci., C.R.,
ser. D, 269, no. 18, pp. 1748-1750.

Butler, P. M.

1939. Studies of the mammalian dentition — differentiation of the postcanine
dentition. Proc. Zool. Soc, London, Series B, 109, no. 1, pp. 1-36.

Chaffee, R. G.

1954. Campylocynodon personi, a new carnivore from the Beaver Divide, Wyo-
ming. Jour. Paleontol., 28, no. 1, pp. 43-47.

DE Chardin, T.

1915. Les carnassiers des Phosphorites du Quercy. Ann. Paleontol., t. IX,
1915, pp. 103-190.

Clark, John

1937. The stratigraphy and paleontology of the Chadron Formation in the Big
Badlands of South Dakota. Ann. Carnegie Mus., 25, art. 21, pp. 261-350.

Clark, John, J. R. Beerbower, and K. K. Kietzke

1967. Oligocene sedimentation, stratigraphy, paleocology, and paleoclimatol-
ogy in the Big Badlands of South Dakota. Fieldiana: Geol. Mem., 5, 158 pp.

Davis, D. D.

1964. The Giant Panda. A morphological study of evolutionary mechanisms.
Fieldiana: Zool. Mem., 3, 339 pp.

Davis, D. D. and Story, H. E.

1943. The carotid circulation in the domestic cat. Field Mus. Nat. Hist., Zool.
Ser., 28, no. 1, pp. 1-47.

Ehrlich, p. R. and P. H. Raven

1969. Differentiation of populations. Science, 165, pp. 1,228-1,232.

Flower, W. H. and R. Lydekker

1891. An introduction to the study of mammals living and extinct. Adam
and Charles Black, London. 763 pp.

GuTH, Christian

1964. Sinus veineux et veines de I'arriere-crane de quelques carnivores fossiles.
Ann. Paleontol., 50, pp. 31-43,

Hall, E. R.

1931. Description of a new mustelid from the later Tertiary of Oregon, with
assignment of Parictis primaevus to the Canidae. Jour. Mammal., 12, no. 2,
pp. 156-158.

Hough, J. R.

1944. The auditory region in some Miocene carnivores. Jour. Paleontol., 18,
no. 5, pp. 470-479.

70

CLARK & GUENSBURG: ARCTOID CHARACTERS 71

1948. The auditory region in some members of the Procyonidae, Canidae, and

Ursidae. Bull. Amer. Mus. Nat. Hist., 92, pp. 67-118."
1953. Auditory region in fossil Felidae: its significance in phvlogeny. U.S.

Geol. Surv. Prof. Paper 243-G, pp. 95-115.

KONIZESKI, R. L.

1961. Paleoecology of an early Oligocene biota from Douglass Basin, Montana.
Geol. Soc. Amer. Bull., 72, pp. 1,633-1,642.

Lavocat, R.

1962. Revision de la faune des mammiferes Oligocenes d' Auvergne et du Velay.
Paris (Editions "Sciences et Avenir")- 153 pp.

McGrew, p. O.

1938. Dental morphology of the Procyonidae with a description of Cynarcloides,
gen. nov. Field Mus. Nat. Hist., Geol. Ser. 6, no. 22, pp. 323-339.

ROMER. A. S.

1966. Vertebrate Paleontology, 3rd ed. Univ. Chicago Pres.s. 468 pp.

ScHLOssER, Max.

1911. Grundzage der Paleontologie (Paleozoologie) von Karl A. zon Zittel, II

Abt. — Vertebrata. Neuarbeitet von F. Broili und M. Schlosser. R. Oldenburg,

Munich und Berlin, vii — 598 pp.
1923. Ibid, V — 706 pp. Reference p. 65, Eastman translation.

Scott, W. B.

1893. On a new musteline from the John Day Miocene. Amer. Nat., 27, pp.
658-659; errata p. 767.

Scott, W. B. and G. L. Jepsen

1936. The mammalian fauna of the White River Oligocene. I: Insectivora and
Camivora. Trans. Amer. Philos. Soc, N.S., 28.

Segall, Walter

1943. The auditory region of the arctoid carnivores. Field Mus. Nat. Hist.,
Zool. Ser., 29, no. 3, pp. 33-59.

Simpson, G. G.

1945. The principles of classification and a classification of mammals. Bull.
Amer. Mus. Nat. Hist., 85.

SissoN, S. and Grossman, J. D.

1938. The anatomy of the domestic animals, W. H. Saunders Co. 972 pp.

Appendix

Parictis(Campylocynodon) parvus 1
Paricfis (Subparicfis)major 2

Parictis (Parictis) primaevus 3

1. PU IA265 2. F-PM22404 3. PU 10583 3. CU 22749

JAW

T®M| ^

3 4
255

.133

TM, 62

TM,

52

l.Pl Mj

T^ '

47
e70

.671

)3 6

485

49
87

552

t> „

e70
255
27 2

274

l.P, M,
L,Pl.M3

Ci

SfCTION
L(POSTERIOR)
IBASAL BULB

R

12 5
B

Pz- 3 ALIGN

e

+

Pi

f "

CIMG.CUSPUU

25
38

658

P2

L
H

ANT «
POST ^
ACCCUSPULE
CINGULUM
CINC CUSPULC

36
59

41

60°

40°

+

610

¥. "'

37.
40°
40°

+

35
5 7
44
50°
40°

e

+

614

P3

* R
L

21

40

525

36

66

545

^ "'

35

5 2

673

H

32

4.6

35.

3.8

ANT ^

70°

60°

40°

50'

POST^

50°

40°

40°

40^

ACC CUSPULC

«

CINGULUM

+

+

+

CINC CUSPULI

+

e

Pa

* R
L

H

ANT^

POST-t

ACC CUSPULE

CINGULUM

CINGCUSPULE

26
4.7
36

70°
50°

+

+

.553

43
78
58

60°
60°

♦
♦

55>

39
68
5.2
50°
60

+

+

.573

Ml

* R

L

H PROTCD
H METCO

4 7
6 4
4.7
32

734

5 2
108
66
40

481

4.1
84

4.9
33

488

LITALONID)

25
6 4

391

35

10 8

324

32
84

381

LaOTAl)

An 2

*L R

L

TRIG W

2 7

3 4

'6
2.7

.794
592

TOTAL

Appendix

B

Parictis (Subparictis) gilpini

F-PM 22405

F-UM 72 9

AMNH 50241 USNM 19930 USNM 19932

SDSM 2567 AMNH 63933 AMNH 76196

JAW

T®M,
L.P,M2

4.7
3M

151

5.8
320

181

48
32.0

.150

4 9
30.0e

163

TM|4 6

"-^ R

5.5
8 8

.625

5 5
90

611

5.2
82

.634

48
87

.552

5.0e
8.9

^ -

i- 6,8

H "•

t" R

8.8
31.1

.283

9.0
32.0

.281

8.2
32.0

256

87
30.0e

290

L.P.Mj

L.Pi-Mj

32.5

Ci

SECTION

R

KPOSTERIOR

66

i BASAL BULB

B

P2.3ALIGN

®

Pi

* R
L

CING.CUSPULE

2

3 3

P2

W R

L

H

ANT^

POST^

ACCCUSPULE

CINGULUM

CINGCUSPULE

2.7

4.4
3.6

70=
40=
®

.614

27
4.9
43

70°
50°

.551

2.7
4.7
37
65°

60°

+

.574

2.7

4.4

34e

65°e

60°e

®

.613

P3

^ R

L

H

ANT^

POST ^
ACCCUSPULE
CINGULUM
CING.CUSPULE

2.7

4 8

3.8

60°

50°

®

.562

2.7

4.8

3.5

55°

50°

+

562

29
4 7
34e
65°e
55° e

+

.617

M 757
37 '^'

32

65°

60°

®

P4

* R

L

H

3.2
6.1
4.9

524

3.9
5.7
48

.684

3.3
T9

4.1e

.508

^ 533
60 ^"

46e

58

if 396

ANT^

70°

65°

65°

70°

50°

65°

POST^

50°

65°

60°

60°e

60°

50°

ACC CUSPULE

+

+

+

+

+

+

CINGULUM

+

+

+

®

CINGCUSPULE

®

+

Mi

! «

4.8
80

.600

4 5
85

.529

4.1
8.0

.512

45
8 3

542

^2 -

¥2 -0°

44^-

^ "*

HPROTCD

5.5

5.5

5.2

4.8

50e

4.8

5 5c

5 2

HMETCD

3.7

3.7

38e

3.1e

4 1

3 7

L(TALONID)

3
8.0

.375

3 4
8.5

.400

39
8.0

.487

3.0
8.3

.361

^ 379
82 ■*'*

ii 390

2^« 325
83 ^^^

H -

UTOTAL)

M2

^ «

3 3
4.5

.733

3.9
4 7

.851

3.7
5.0

.740

34
4.5

755

|1 78B

TRIG W

14
33

424

13
3 9

333

1.7
3.7

.459

1 6e
35

.457

^ -'

TOTAL W

Appendix

Parictis (Subparicfis)montanus

CM957I

CM 9068

F-PM 3843

JAW

T(1)M,

48
28 7

167

4 4

270

163

TM,40

L.P, M, ■■

HMi
D "

47e
85

553

4 9
8 7

563

^ ...

° R

85
28 7

296

8 7
270

322

L.P,M2

L.P1-M3

308

289

Ci

SECTION
LiPOSTERIORi
iBASAL BULB

B

P2.3 ALIGN

®

Pl

:fL R

L

CING.CUSPULE

15
25

+

600

P2

^ R
L ■*

H

25
4 2
35e

595

23
39
31

590

U "'

ANT ^

70°

60°

80°

POST ^

50°

50°

40°

ACCCUSPULE

CINGULUM

®

CINGCUSPULE

e

+

P3

f »

24
4 5

533

12
4 3

534

■^ 577
45 ^'^

H

33e

32

3 7

ANT^

60°

60°

70°

POST ^

50°

60°

50°

ACCCUSPULE

®

CINGULUM

®

CING.CUSPULE

e

+

R*

f »

H

ANT ^
POST^

2 9
60

483

29
58
4 5
60°
60°

500

^ 523
61 ^"

51

70°

70°

ACCCUSPULE

+

+

+

CINGULUM

®

CINGCUSPULE

e

+

Ml

£ -

39

76

543

40
76

526

i^ 570
79

HPROTCD

47e

4 9

59

HMETCD

30e

34

L'TALONIDi
LaOTALi

28
76

368

28
7 6

368

^ »'

An 2

^ •

3.3

4.0

825

2 7

3 3

818

TRIG W
TOTAL

1.1
33

.333

08
28

285

Publication 1150