PROCEEDINGS

OF THE

WASHINGTON ACADEMY OF SCIENCES

VOL. XIII, No. 3, PP. 51-65. AUGUST 15, 1911

REMARKS ON THE FOSSIL TURTLES ACCREDITED TO
THE JUDITH RIVER FORMATION.1

BY F. H. KNOWLTON.

In my paper entitled The Stratigraphic Relations and Paleon-
tology of the ' Hell Creek beds, ' ' Ceratops beds' and their equiva-
lents, and their reference to the Fort Union Formation,2 I made,
among others, the following statements regarding the turtles of
the Judith River formation: "It has recently been confidently
asserted on eminent authority that the turtles of the 'Ceratops
beds' are very closely related to those of the Judith River formation;
in fact that several species are identical, and others so close as to
be separated with difficulty. Fortunately the work of Hatcher,
published as late as 1905, has left us with a very complete annotated
list of the Judith River forms, which makes a valuable basis of com-
parison of the two faunas. It needs, however, but a cursory exam-
ination to show that at least half of the species listed as belonging
to the Judith River do not belong to this fauna at all, but come from
the Fort Union, Arapahoe, etc., or are so fragmentary as to be
unidentifiable. . . . Any comparison of the turtles of the Ju-
dith River formation with those of the 'Ceratops beds' as tending
to support the Cretaceous age of the latter, does not make a very
impressive case."

1 Published with the permission of the Director of the United States Geological
Survey.

2 Wash. Acad. Sci., Proc., vol. xi, 1909, pp. 235, 236.

Proc. Wash. Acad. Sci., August, 1911.

51

52 KNOWLTON

Lack of space at that time precluded the full presentation of the
data on which the above statements were based. Inasmuch as
this mild remonstrance does not appear to have attracted attention,
and since the turtles are still being used to "prove" close relation-
ship between the Judith River and Lance ("Ceratops beds") for-
mations, it seems opportune to set forth the facts as they are. In
a paper3 just published Dr. O. P. Hay writes as follows: "My
study of the fossil turtles indicates that the species of these animals
rarely pass from one epoch to another. If they have ever done so
they passed from the Judith River into the Lance Creek epoch.
There are five or six species of Judith River turtles which are rep-
resented in the Lance Creek and Hell Creek beds by turtles of iden-
tical or very closely related species."

In the following pages it is proposed first to take up the species
of turtles mentioned by Hatcher4 as belonging to the Judith River
formation, and to indicate the type locality for each species as well as
its subsequent distribution so far as this is a matter of published
record. Later in this paper the species enumerated by Doctor Hay
will be similarly treated, and finally some remarks will be presented
on the results of this showing.

TURTLES or THE JUDITH RIVER FORMATION ENUMERATED BY

HATCHER.

1. Trionyx foveatus Leidy [Aspideretes foveatus (Leidy) Hay].

Type locality: "Bad Lands of the Judith River, Nebraska
Territory." Fragments collected by Doctor Hay den and named
and described by Doctor Leidy in i&$65. Subsequently Leidy
figured two of the type specimens, together with another specimen,
identified doubtfully as the same species, from Long Lake below
old Fort Clark on the Missouri River, North Dakota, the latter
belonging to the Lance formation. On this point Hatcher says:
" Considering the difference in the age of these deposits, it is quite

1 Where do the Lance Creek (" Ceratops") Beds belong, in the Cretaceous or in the
Tertiary? Indiana Acad. Sci., Proc., 1909 (issued Oct. 1910), p. 21 (of reprint).
4 U. S. Geol. Surv., Bull. 257, 1905, pp. 72-80.
» Acad. Nat. Sci., Phila., Proc., vol. viii, 1856, p. 73.

TURTLES OF JUDITH RIVER FORMATION 53

probable that had better material been at the disposal of Doctor
Leidy he would have found them to be at least specifically distinct."
Specimens identified by Cope as Trionyx foveatus were collected
by Sternberg in the Judith River area in 1876; it does not appear to
have been found by Stanton and Hatcher when they visited the
area in 1905. This species was reported by Marsh from the " Cera-
tops beds" near Denver, Colorado, but according to Hay these
specimens " belong probably to Aspideretes beecheri" a species from
the Lance formation of Converse County, Wyoming. Trionyx
foveatus has also been identified by Lambe from the Belly River
beds of the Red Deer River region in Canada, and by Barnum Brown
in the "Hell Creek beds" (= Lance formation) of Montana.

From this review it appears possible that there are two and quite
likely three species confused under the name of Trionyx foveatus,
and hence any conclusion as to which part of the "species" belongs
to the Judith River, and which to the Lance formation, must 'be a
matter of more or less doubt.

2. Trionyx vagans Cope [Aspideretes? vagans (Cope) Hay].

Type locality: Bijou Basin (Bijou Creek), 40 miles east of Den-
ver, Colorado, in beds regarded by Cross as probably Arapahoe in
age. It was also reported by Cope from near the mouth of the
Bighorn River, Montana, and near Long Lake, North Dakota,
both localities in beds belonging in all probability to the Lance
formation. Apparently on the basis of Lambe's reporting it from
the Belly River beds of the Red Deer River in Canada, it was in-
cluded by Hatcher in his list of Judith River species, though Hatcher
himself points out the obvious distinctness between the type speci-
mens and the specimens so identified by Lambe. Hay has referred
Lambe's specimens to another species, and hence all pretense of
Trionyx vagans as a Judith River species disappears.

3. Plastomenus coalescens Cope. [Aspideretes coalescens (Cope)

Hay].

Type locality: "Bad Lands south of Woody Mountain, latitude
49°," longitude about 106°, Assiniboia, Canada.

54 KNOWLTON

Much confusion has arisen concerning this species, which can only
be eliminated by a review of all the earlier references to it in the
literature of the subject. It was first mentioned but without
description in a list published by Cope in 1875,6 and was fu%
described later in the same year in an obscure paper published
as an appendix to Dawson's Report on the Geology and Resources
of the Region in the vicinity of the Forty-ninth Parallel, etc.7
Still later in 1875 this description was copied word for word, but
without reference to either of the two preceding places of publica-
tion in Cope's Vertebrata of the Cretaceous Formations of the
West.8 In the Report where it was first described the locality was
given as The Bad Lands south of Woody Mountain, latitude 49°,
a locality confirmed by Dawson in his description of the geology
of the region.9 Dawson also mentions other vertebrate remains
with which it was found associated, as well as several species of
plants, and refers the beds to the "Lignite Tertiary," and in the
latest published geological map of Canada,10 the area is still colored
as "Laramie," which in the writer's opinion is the approximate
equivalent of the Fort Union of the United States.

The confusion dates from the moment when Cope transferred the
original description to his "Vertebrata of the Cretaceous Formations
of the West," where as regards the locality he says: "This species
is found .... near the Milk River in British America," where
it was "collected by George M. Dawson .... near Woody
Mount." As a matter of fact Woody Mountain (or Woody Mount)
is more than 150 miles east of the valley of Milk River where it crosses
the international boundary, and there is not the slightest evidence
that it came from Milk River. But because it was supposed by
subsequent writers that Woody Mountain was in the Milk River
Valley, it was assumed that the age must be Judith River! Thus
Hatcher,11 who was apparently in ignorance of the original place of
publication says: "This species is founded on fragments . . . .

8 Acad. Nat. Sci., Phila., Proc., 1875, p. 9.

7 Brit. N. A. Bound. Com., Montreal, 1875. Appendix B, p. 337.

8 Rept. U. S. Geol. Surv. Terr., vol. 2, 1875, p. 92.

9 Brit. N. A. Bound. Com., Montreal, 1875, p. 105.
"Western Sheet, 1901.

11 U. S. Geol. Surv., Bull. 257, p. 74.

TURTLES OF JUDITH RIVER FORMATION 55

collected by George M. Dawson near Milk River in Canada, from beds
referred by Cope to the ' Transition series, probably the Fort Union
or Lignite epoch/ but now known to belong to the Judith River."
This error is also perpetuated by Hay in his Fossil Turtles of North
America,12 who, although obviously familiar with the original place
of publication, still gives the locality as "in the basin of Milk River,
south of Wood Mountain, Assiniboia, British Columbia."

The above exposition would seem effectively to dispose of the
claim for Judith River age of the type material of Plastomenus
coalescens, though it might possibly be considered to figure as a
Judith River species if Doctor Hay's reference to it of a specimen
identified by Lambe as Trionyx vagans is of valid standing. That
is to say Lambe collected a large, finely preserved specimen, which
he identified as Cope's Trionyx vagans, in the Belly River deposits
in the region of the Red Deer River, below Berry Creek, in Alberta.
Concerning this Doctor Hay says :12 "The present writer, regarding
Cope's type of Trionyx vagans as too small and imperfect a fragrnent
for satisfactory comparison with materials from any region, except
the type locality in Eastern Colorado, is compelled to seek for some
more probable disposition of Mr. Lambe's fine specimen. Since
the plastron of Cope's Plastomenus coalescens indicates a large tri-
onychid which lived in approximately the same region and in the
same geological period, it appears to be best to refer the Red Deer
River carapace to the same species." It appears, then, that the
study of these specimens themselves is not sufficient to determine
whether Lambe's specimen is really referable to Cope's species, so
the burden is placed on stratigraphy!

4. Plastomenus costatus Cope.

Type locality: "Bad Lands south of Woody Mountain, latitude

49°."

This species was found in the same locality and horizon as the
last; and the same confusion and compounding of error has resulted.
Simply because Woody Mountain was supposed to be in the valley

12 1908, p. 489.
1S Op. cit., p. 489.

56 KNOWLTON

of Milk River it followed that the age must be Judith River ! Upon
this assumption alone rests the claim.

This species has been found by Barnum Brown in the Lance
formation ("Hell Creek beds") 12 miles south of the Missouri
River, on Hell Creek, Montana. This locality is about 100 miles
directly south of the type locality in Canada.

5. Plastomenus punctulatus Cope.

Type locality: Bijou Creek, 40 miles east of Denver, Colorado,
in beds that are of Arapahoe age according to Whitman Cross.
Cope also states that he had the same species from beds at Long
Lake, "Nebraska" (now North Dakota), the age of which is Lance
formation.

Hatcher's reason for including this species in the Judith River
fauna is interesting. He quotes14 Cope's statement, evidently from
his Vertebrata of the Cretaceous Formations of the West, to the
effect that it was "found in association with the preceding species,"
which, in this book happens to be Plastomenus costatus, but in the
place where P. punctulatus was originally described,15 the "preced-
ing species" happens to be Trionyx vagans, the status of which has
already been considered above. As the original description was
transcribed without change from the Annual Report into the Mono-
graph the error arose as stated above and as has been pointed out
by Doctor Hay. Plastomenus punctulatus can, therefore, lay no
claim to having been found in the Judith River.

6. Plastomenus insignis Cope.

Type locality: Bijou Creek, 40 miles east of Denver, Colorado.

This species was introduced into the Judith River fauna by
Hatcher16 who, consulting Cope's Vertebrata only, and ignoring the
original place of publication, concludes that by "inference" it came
from south of Woody Mountain, and if from this locality it. was
assumed, as in the cases of the several species above considered, that

14 U. S. Geol. Surv., Bull. 257, p. 74.

15 U. S. Geol. and Geogr. Surv. Terr., Ann. Rept., 1873 (1874), p. 453.

16 U. S. Geol. Surv., Bull. 257, p. 75.

TURTLES OF JUDITH RIVER FORMATION 57

it must be of Judith River age. It is clear enough when the original
description is consulted that this species came only from Bijou
Creek, Colorado, and this is confirmed by Doctor Hay who has
examined the type now in the American Museum of Natural His-
tory and finds it labeled in Cope's handwriting as "Plastomenus
insignis. 10-9-1873, Colorado." It is therefore to be excluded from
the Judith River fauna, but not for the reasons given by Hatcher.

7. Adocus lineolatus Cope.

Type locality: Bijou Creek, 40 miles east of Denver, Colorado,
in beds believed to be Arapahoe in age. Also reported by Cope
from the mouth of the Bighorn River, Montana, and it appears in
his list of Judith River vertebrates, though according to Hatcher
this last reference is probably "due to an oversight." Fragments
that have been identified as this species have been found by Lambe
in Belly River beds of Red Deer River, Alberta, by Barnum Brown
in the Lance formation on Hell Creek, Montana, and by others in
the " Cera tops beds" (Lance formation) of Converse County,
Wyoming. Hatcher has expressed as his opinion17 that the refer-
ence of Lambe's specimens "may be incorrect" and Doctor Hay18
says: "It is the writer's opinion that it is unsafe to identify as
belonging to Adocus lineolatus specimens from the Judith River and
Laramie beds before far better materials of the species have been
collected from the type locality." He adds: "It is improbable
that the same species continued from the Judith River epoch to
the Arapahoe epoch," which latter, be it remembered, is the prob-
able age of the type material.

8. Basilemys ogmius (Cope).
Basilemys variolosa (Cope).

Type locality: "From six miles west of the first branch of Milk
River, near latitude 49°," in the province of Alberta, British America.
Because this locality is in the vicinity of Milk River it has been
assumed by Hatcher and others that the age must be Judith River,

17 U. S. Geol. Surv., Bull. 257, p. 76.

18 Fossil Turtles of North America, p. 248.

58 KNOWLTON

but a careful reading of Dawson's account19 of the geology of the
place where he actually obtained the specimens shows that in all
reasonable probability it should be referred to the "Ceratops beds,"
that is the Lance formation. Dawson considered the beds as "be-
longing probably to the base of the Lignite Tertiary," and compares
them to the beds south of Woody Mountain, adding: "The bones,
in the manner of their preservation, much resemble division /3 of
the Bad Lands south of Woody Mountain, which these beds may
possibly represent." The lithology is described as similar in the
two areas, and associated with the remains of the turtles Dawson
reports finding many bones of dinosaurs, and large Unio shells.
In any event it seems unsafe to refer these beds to the Judith River
merely on geographical position and without corroborative data.

The type specimens of Basilemys ogmius are said to be very poor,
and Doctor Hay states that : " It is doubtful whether new materials
could be identified by means of the type." However this may be,
this species has been combined with Basilemys variolosa (Cope),
which has as its type locality "Bad Lands of the Judith, Montana,"
and was described a year or more later. Material that has been
identified as Basilemys (or Adocus) variolosa has been mentioned by
Lambe from the Belly River beds of Red Deer River, Alberta, and by
Hatcher in the Judith River formation of Montana. It seems not
improbable that there may be two species represented, but in any
event it would appear that the undoubted Judith River material
has been that identified with "B. variolosa," whereas the B. ogmius,
being very poor, has not been certainly recognized.

The conclusion reached is that that part of the "species" named
"B. ogmius" came from beds that are not certainly of Judith River
age, and has not been subsequently recognized, while the part
known as "B. variolosa" came from beds of Judith River age, and
is the only part of the "species" that has been distinguished by later
authors.

9. Basilemys imbricarius (Cope).

Type locality: Judith River Basin, Montana, in beds of the
Judith River formation; it has not been found elsewhere. It was
founded on very unsatisfactory materials, according to Doctor Hay.

19 Brit. N. A. Boundary Com., 1875, p. 130 et seq.

TURTLES OF JUDITH RIVER FORMATION 59

10. Polythorax missouriensis Cope.

Type locality: Judith River Basin, Montana, in beds of the Ju-
dith River formation. It has not been reported from any other
locality or horizon.

11. Baena antiqua Lambe.

Type locality: Red Deer River, Alberta, British America, in
beds supposed to be of Belly River age; it has not been obtained
elsewhere.

12. Baena hatched Hay.

Type locality: South side of Lance Creek, opposite mouth of
Dogie Creek, Converse County, Wyoming, in beds of the Lance
formation ("Ceratops beds").

A specimen from the Belly River beds of the Red Deer River,
Alberta, was referred to this species by Lambe,20 and it was upon
this ground that it was included by Hatcher21 in the Judith River
fauna. Later, however, Doctor Hay22 referred Lambe's specimen
to a totally different genus and species (Boremys pukhra), and
hence Baena hatcheri is to be eliminated from the Judith River
fauna.

13. Neurankylus eximius Lambe.

Type locality: Red Deer River, Alberta, British America, in
beds regarded as of Belly River age; it is monotypic and has not been
collected elsewhere.

There is one naive bit of history connected with this genus that
should not be omitted. Doctor Hay in his Fossil Turtles of North
America (p. 93) remarks as follows: "Mr. Lambe placed the genus
provisionally among the Chelydridae, but as no genera of this family
are known from deposits older than the Tertiary and as the type has
some resemblance to the Baenidae, which flourish at that time, it
seems better to refer Neurankylus to this family until more is known
about it."

20 Cont. Canadian Pal., vol. 3, 1902, pt. ii, p. 43.

21 U. S. Geol. Surv., Bull. 257, p. 79.

22 Fossil Turtles of North America, 1908, p. 92.

60 KNOWLTON

TURTLES MENTIONED BY HAY AS TENDING TO SHOW IDENTITY

OR CLOSE SIMILARITY BETWEEN SPECIES or THE JUDITH

RIVER AND LANCE FORMATIONS.

Following is the list of six species given by Doctor Hay23 and of
which he says: "I shall take pains to give some details."

1. Compsemys obscura (Leidy).

Type locality: Near Long Lake, on the Missouri River below
Fort Clark, North Dakota, in beds regarded as belonging to the
Lance formation. It has otherwise been collected only in beds of
the Lance formation ("Hell Creek beds") on Hell Creek, Montana.

This species was included by Cope24 in his list of Judith River
vertebrates, but without giving any evidence in support of such
reference, and presumably on the authority of Cope's list it is in-
cluded by Osborn25 in his list showing the "distribution of the land
and fresh- water Cretaceous vertebrates in the west," as occurring
in Montana and hence "by inference" in the Judith River fauna.
Compsemys obscura as well as its companion species, C. victa, was
excluded from the Judith River fauna by Hatcher, who says, "In
no descriptions of either of these species can I find any suggestion
that remains of either have been described from Montana." That
C. obscura is properly excluded from the Judith River is shown by
Doctor Hay, who says,26 "This species is included by Cope in his
list of Judith River fossils, but the writer knows of no specimens
that confirm the statement." Curiously enough, however, this
species is the first one given by Doctor Hay27 as occurring in both
Judith River and Lance formations!

2. Compsemys victa Leidy.

Type locality: Long Lake, on the Missouri River^below Fort
Clark, North Dakota, in beds regarded as belonging to the Lance
formation, that is the same locality and formation as^the last.

28 Proc. Indiana Acad. Sci., 1909, p. (of reprint) 21.

24 U. S. Geol. and Geogr. Surv. Terr., Bull. vol. 3, 187, p. 573.

26 Cont. Canadian Paleont., vol. 3, 1902, pt. ii, p. 12.

26 Fossil Turtles of North America, 1908, p. 236.

17 Proc. Indiana Acad. Sci., 1909, p. (of reprint) 21.

TURTLES OF JUDITH RIVER FORMATION 6l

The history of this species is much the same as that of the last,
except that specimens identified under this name have a somewhat
wider distribution. It was included by Cope in his list of Judith
River vertebrates, but, as stated under the last species, was ex-
cluded from this fauna by Hatcher. It is accepted by Doctor Hay28
as of Judith River age on the ground that it was found in the region
about Milk River, Alberta. This occurrence in the "Milk River
region" is traceable to Dawson's Report on the Geology and Re-
sources in the Vicinity of the Forty-ninth Parallel, but, as in the
case of Plastomenus coalescens, a reading of the geological part of
the report shows that it came from south of Woody Mountain,
and 150 miles east of the Valley of Milk River, in beds of "Lignite
Tertiary" age.

This species was reported by Cope from supposed Arapahoc
deposits east of Denver, Colorado, and is said to be "very common
in the Lance Creek beds," of Converse County, Wyoming. It
was also found by Brown in the Lance formation ("Hell Creek
beds") on Hell Creek, Montana, and an allied but undescribed
species is reported by Doctor Hay from the Fort Union.

From this brief exposition it is clear that Compsemys victa must
be excluded from the Judith River fauna, and the only horizons in
which it is authenticated are Arapahoe and Lance formations. In
the treatment of this species in his Fossil Turtles of North America
(p. 234), Doctor Hay opines that there is so much difference in age
between Judith River, Arapahoe and Denver, and Lance formations
that it is "not improbable" that three species are represented in-
stead of one. Again would the responsibility be shifted from paleon-
tology to stratigraphy!

3. Aspideretes foveatus (Leidy) Hay.

The status of this species has been already considered under its
synonym, Trionyx foveatus, ante, p. 52.

4. Aspideretes beecheri Hay.

Type locality: East side of Lance Creek, Converse County,
Wyoming, in beds belonging to the Lance formation. Two cara-

28 Fossil Turtles of North America, p. 234.

62 KNOWLTON

paces from the Judith River of Fish Creek, Montana, have been
referred to this species by Doctor Hay, who makes the following
statement concerning them: "No characters are observed which
serve to distinguish these carapaces from that of the type of A.
beecheri. It is not improbable, however, that they belong to a
distinct species."

5. Adocus lineolatus Cope.

The status of this species has been considered, ante, p. 57.

6. Basilemys variolosa (Cope).

This has already been considered at length under its synonym
Basilemys ogmius, p. 57.

DISCUSSION OF EVIDENCE REGARDING "JUDITH RIVER" TURTLES
AS BROUGHT OUT IN THE FOREGOING REVIEW.

On combining the lists given by Hatcher and Hay it appears
that sixteen species of turtles have been accredited to the Judith
River formation. Of these sixteen species it has been demonstrated
that only three species have the type specimens from the original
Judith River area in Montana, while two additional species have
the types from the Belly River ( = Judith River?) of the Red Deer
River region of Alberta. The three species from Montana are the
following:

Trionyx foveatus

Basilemys imbricarius

Polythorax missouriensis
The two Canadian species are:

Baena antiqua

Neurankylus eximius

Of the remaining species four, as follows, have the types described
from Bijou Creek, Colorado, in beds supposed to be of Arapahoe
age:

Trionyx vagans

Plastomenus punctulatus

Plastomenu sinsignis

Adocus lineolatus

TURTLES OF JUDITH RIVER FORMATION 63

The types of two species came from south of Woody Mountain,
Assiniboia, in beds referred to the "Lignite Tertiary" by Dawson,
Cope, and others, but are now considered to belong to the Lance
formation. They are:

Plastomenus coalescens

Plastomenus costatus

From Long Lake, North Dakota, in beds of the Lance formation,
the following types are supplied:

Compsemys obscura

Compsemys victa

The types of two species were obtained from the Lance formation
of Converse County, Wyoming:

Baena hatcheri

Aspideretes beecheri

The remaining species is Basilemys ogmius which as already
shown is of mixed position. Strictly speaking the type came from
near Milk River, Alberta, in beds that are probably younger than
the Judith River, while B. variolosa, the "species" with which it
has been combined, came from the Judith River formation near
Judith River, Montana.

A further analysis of the evidence adduced indicates that the
following species should be excluded from the Judith River fauna
on the sufficient ground that they do not occur in beds of this age:

Trionyx vagans

Plastomenus costatus

Plastomenus punctulatus

Plastomenus insignis

Baena hatcheri

Compsemys obscura

Compsemys victa

To return again to the five species having their type locality in
Judith River and Belly River beds, it appears that the four follow-
ing have never been found outside the Judith River or Belly River
horizons.

Basilemys imbricarius

Polythorax missouriensis

Baena antiqua

Neurankylus eximius

64 KNOWLTON

The fifth species — Trionyx foveatus — is the only one that appears
to be distributed into higher beds, but it will be recalled on refer-
ring to the discussion of this species on page 53, that, while it has been
reported from Long Lake, North Dakota, Hell Creek, Montana,
and near Denver, Colorado, both Hatcher and Hay have expressed
the opinion that there may be two and possibly three species confused
under this name, each being confined to one of the three horizons
involved.

The following forms, although having their types from beds
higher than the Judith River formation (Arapahoe, Lance formation
etc.), appear on more or less questionable evidence to have ranged
downward into the Judith River:
Plastomenus coalescens (?)
Adocus lineolatus (?)
Aspideretes beecheri (?)

The reasons why these are regarded as questionably present in
the Judith River formation are set forth in the discussions of these
species in the preceding pages. Not a single one has a clear title.

CONCLUSIONS.

1. The Judith River fauna has had accredited to it by Hatcher
and Hay at least sixteen species of fossil turtles.

2. Of these sixteen species seven are to be excluded on the suffi-
cient ground that they do not occur in beds of Judith River age.

3. Of the nine remaining species that may have a more or less
valid claim to membership in the Judith River fauna, five have
their type localities in beds of Judith River and Belly River age, and
four in horizons above that of the Judith River.

4. Of the five typically Judith River species, four are absolutely
confined to these beds, while the remaining one which has an alleged
higher range, is shown to be probably a composite species possibly
including three species which correspond respectively to the three
horizons involved.

5. Of the four species, the types of which came from post- Judith
River beds, and which are claimed to range downward into the
Judith River, not a single one enjoys a clear title to be so regarded.

TURTLES OF JUDITH RIVER FORMATION 65

6. It is plain, then, that the comparisons that have been made
between "Judith River" turtles and those of higher horizons (Lance
formation, "Laramie," "Ceratops beds/' "Hell Creek beds,"
"somber beds," etc.), have not been made with species that really
belong to the Judith River fauna, but with forms that actually
belong to these higher beds. It is not to be wondered at that such
a comparison has resulted in showing "identity" and "striking
similarity," since it is in accord with the ancient axiom, that " Things
equal to the same thing are equal to each other."

OF !HE

UNIVERSITY

ovERDUE S'.oo ON THE°~^;°oRrH

DAY

OCT 2 1941 M

UNIVERSITY OF CALIFORNIA LIBRARY