-(7^

MEMOIES

CAENEGIE MUSEUM.

VOL. VII. NO. 2.

THE FOSSIL TURTLES OF THE UINTA FORMATION.

By Charles W. Gilmore.

The finest and most complete assemblage of the remains of fossil turtles as
yet secured from the Upper Eocene of the Uinta formation has been brought to-
gether in the Carnegie Museum through the activities of its various expeditions
to Utah. By the kindness of Dr. William J. Holland, the Dh-ector of the Museum,
I have been permitted to study this collection, and the present paper presents the
results of my investigations.

The collection comprises more than fifty individuals, and was made by field-
parties conducted by Messrs. Earl Douglass and O. A. Peterson and as an in-
cidental part of their search of the Uinta exposures for the remains of extinct
mammals. An important feature of this collection is the determination of the
exact geological horizons in which the specimens were found, thus establishing
a firm foundation for future correlative work.

The chelonian fauna of the Uinta formation is of peculiar interest, since it
marks the last appearance of several forms which had their beginning, so far as
our present records go, in the Upper Cretaceous and Lower Tertiary. Of the six
genera recognized in the present collection from the Uinta formation only three,
Anosteira, Amyda, and Testudo, are known to pass upward into the younger
Tertiaries. Anosteira is known from the Lower Oligocene of England, Amyda
reappears in the Miocene of the Atlantic coast, while Testudo is found in the
overlying Oligocene. It appears that the Uinta thus marks an important stage in
the history of the chelonian life of the Upper Eocene.

The Baenidse make their last appearance. The Dermatemydidse are rep-

101

102 MEMOIRS OF THE CARNEGIE MUSEUM

resented for the first time bj' the single genus and species Anosteira ornata Leidy.
The Emydidse, suggestive of swampy conditions, in the number of species are the
most abundant turtles in the present collection. Seven species have been recog-
nized and larger collections will doubtless add several more to the list. The soft-
sheUed river-turtles, Trionychidse, indicative of flowing water, are represented by
at least three species, one of which is as large as the existing Asiatic species. The
presence of true land-tortoises, Testudinidse, is represented by three species of
the genus Hadrianus, which includes tortoises some of which attain a length of more
than three feet, and the genus Testudo by a single species, the first recorded oc-
currence, in North America, of this genus below the Ohgocene. The discovery of
the fossil remains of the lizard-like reptile Glyptosaurus in the Uinta according to
Osborn' "hints as to the Floridan or south temperate conditions of climate."

There were a considerable number of specimens in the collection which were
too fragmentary for specific determination, and in two instances at least I am in-
clined to the opinion, that, had better material been available, distinctive characters
would have been found to show the presence of additional species new to the fauna.

I take this opportunity to protest most emphatically against the establishment
of new species of turtles based upon inadequate specimens, for it certainly cannot
serve any useful purpose to burden the literature with a lot of useless and meaning-
less names. The difiiculties encountered in the present study, in recognizing to
which species certain specimens belonged, when almost perfect individuals were
at hand, shows the futility of naming scraps with which subsequent!}^ discovered
material can never with absolute confidence be identified. There are perhaps
some few exceptions, for occasionally a fragmentary specimen is found which
shows a sculpture, or some character of such striking pecufiarity, as to make it
stand out distinctly from all previously described forms. The present study has
demonstrated that a considerable variation within the limits of a species is to be
expected, and, until the range of these variations is determined, it is quite useless
to describe new forms based upon some small part of the carapace or plastron,
which shows some slight difference from described forms, when the very next speci-
men discovered may have these same features and j'et have other characters to be
found in an adequate specimen, which show it to belong to a well-established
species.

I wish also to protest against the practice of naming species simply because
the specimen comes from a formation from which the genus to which it belongs
has not previously been recognized. In other words it is assumed that ' it is not

' Osbom, H. F., "Age of Mammals," 1910, p. 160.

GILMORE: the fossil turtles of the UINTA FORMATION 103

likely that a species known from a lower horizon continues over into a higher
horizon/ therefore a hunt for characters to sejDarate it from the other species of
the genus is instituted, with the result that minor differences are magnified to
represent specific differences, when, had the specimen come from a formation in
which species of the genus were already known, it would in all probability have
found a resting-place within one of the described species.

In order to facilitate comparisons of the descriptions here given with those of
other described forms I have closely followed the order of arrangement used by Hay
in his monographic study of the fossil turtles of North America.

At this point I wish to acknowledge the assistance rendered me while this
paper was in the course of preparation. First of all I express my gratitude to Dr.
W. J. Holland for his hearty cooperation at all times, for the privilege granted me
of studying this fine collection of fossil turtles, and for his editorial oversight of
the work. I am under obligations to Dr. W. D. Matthew, of the American Museum
of Natural History, New York, for the loan of type-specimens, and to Dr. 0. P.
Hay, to whom, because of his wide knowledge of the turtles, I am especially in-
debted for invaluable advice upon numerous occasions. The text-figures were
made by the well-known artist, Mr. Rudolph Weber, the photographs are bj' Mr.
Arthur Coggeshall, of the Carnegie Museum.

Geological Occurrence.

All of the specimens considered in the present paper are from the Uinta forma-
tion as exposed in the Uinta Basin at the southern base of the Uinta Mountains,
and from that part of the basin which lies within Uinta County, Utah. The
geological positions of the various specimens as here given were taken from the
original field-labels which accompanied each specimen, so that these determina-
tions are wholly the work of Messrs. Earl Douglass and 0. A. Peterson, whose long
experience in the field insures the accuracy of their observations.

In 1895 the Uinta formation was divided by Peterson- into three levels, or
horizons, designated as follows, A (Lower), B (Middle), and C (Upper) Uinta.

The remains of turtles have now been found in all three horizons, though
judging from the present collection, individuals occur most abundantl}'^ in Horizon
B, but the number of species recognized in the collection is about evenly divided
between Horizons B and C. Up to the present time only one species is known from
Horizon A. Douglass^ has pointed out that "the lower portions of these deposits
may be, and probably are, contemporaneous with portions of deposits in the

* Peterson, 0. A., Bull. Amer. Mus. Nat. History, VII, 1895, p. 74.
» Douglass, Earl, Bull. Geo!. Soc. of America, vol. 25, 1914, p. 418.

10-1 MEMOIRS OF THE CARNEGIE MUSEUM

Bridger and Washakie Basins and with other deposits elsewhere." If this be the
true condition, it may to some extent account for the presence of many species
common to the two formations.

Below is given a list of the identified species occurring in each of the three
subdivisions of the Uinta formation.

Horizon A (Lower Uinta).
Baena inflata sp. nov.

Horizon B (Middle Uinta).
Baena arenosa Leidy, E. hollandi sp. nov.,

B. emilice Hay, E. uintensis Hay,

B. inflata sp. nov., Hadrianus utahensis sp. nov.,

B. platyplastra sp. nov., Testudo uintensis sp. nov.,

B. (jigantea sp. nov., Amyda egregia Hay,

Echniatemys callopyge Hay, A. scutumantiquum (Cope).

Horizon C (Upper Uinta).
Baena emilice Hay, Hadrianus corsoni (Leidy),

Echmatemys douglassi sp. nov., H. robustus sp. nov.,

E. depressa sp. nov., Anosteira ornata Leidy,

E. obscura sp. nov., Amyda sp.,

E. pusilla? Hay, Glyptosaurus sp. indet.

Six genera and twenty species are recognized in the present collection, whereas
in 1908, at the time Doctor 0. P. Hay published his "Fossil Turtles of North
America" only four genera and five species were accredited to the Uinta formation.
These were as follows:

Baena emilice Hay, Hadrianus tumidus Hay,

Echmatemys callopyge Hay, Amyda crassa Hay.

E. uintensis Hay,

The two latter species have not been recognized in the present collection,
although each of the others is represented by from two to six individuals, so that
altogether six genera and twenty-two species of fossil turtles have now been
foxmd in the Uinta formation. The known geological range of these species is
graphically shown in the accompanying table.

GILMORE: the fossil turtles of the UINTA FORMATION 105

Geological Range of Recognized Species.

Name

Wasatch.

Bridger.

Uinta.

Oligocene.

A 1 B

c

v

A

B

c

Bacnidae:

Baena arenosa Leidv,

X I X

V

X
X
X
X
X

B. emilice Hay,

1

X

B. infiata sp. nov.,

' ' 1 '

X

B. platyplaslra sp. nov.,

1

B. gigantea sp. nov.,

1

1

Dermatemydidae:

'

X

X

X

Emydidse:

Echmatemys callopyge Hav,

X

X? ,

E. hollandi sp. nov.,

1

X

i. .

X

X

X? 1

E. obscura sp. nov.,

1

E. pusillaf Hay,

X

E. uinlensis Hay,

1

X

Testudinidse:

!

X

X

H. ulahensis sp. nov.,

X
X

H. robicstus sp. nov.,

Testudo uinlensis sp. nov.,

! i

X

X
X
X

Trionychidida:

Amyda egregia Hay,

1

X

A. crassa Hay,

i

-4. sculumantiquum (Cope),

X
X

Lacertilia:

Glyptosaurus sp. indet.,

X

X

Summary of Material forming the Collection of Turtles from the Uinta Formation in :

Museum.

_ . Catalog No.

Baemdac:

Baena arenosa Leidy, 2356

B. emilice Hay, 2159

B. " 3243

B- " 3253

B' " 3257

B- " 3443

B- " 3444

B. gigantea sp. nov., 3441

B. inflata sp. nov., 34%

^- " 3137

B- " 3442

B. platyplaslra sp. nov., 3227

B. sp. indet., 2372

B.f
B.f
B. '

(skull) .

.3255
.3247
.3271
.2956
.3447

C.

B, or C.

B lower.

B lower.

B lower.

A.

B.

B lower.

B.

B.

?
B.

7

B.
C.
B lower.

106 MEMOIRS OF THE CARNEGIE MUSEUM

Dermatemydida;:

Anosteira ornala Leidy, 2954 C.

Emydidse:

Echmakmys callopyge Hay, 2157 B.

E. " 2371 B.

E. douglassi sp. nov 3244 C?

E. depressa sp. nov., 2936 C.

E. hoUandi sp. nov., 3249 B.

E. obscura sp. nov 3252 C.

E. pusiUa? Hay, 3282 C.

E. uintensis Haj', 3270 B.

E. •• 2158 B, orC.

E. " 2397 B?

E. sp. indet., 2393 B.

E. " " 2361 B.

E. " " (skull and neck) 2387 B.

Testudinida;:

Hadrianus corsoni (Leidy), 3403, 3404 C.

H. robustus sp. nov., 3342 C.

H. utahetisis sp. nov., 2343 B, or C.

H. sp. indet., 2376 B.

H. " " 3256 B, or C.

Teshido uintensis sp. nov., 2331 B.

Trionychidoe:

Amyda egregia Hay, 3254 B.

A. '•? 3258 B.

A. scutumaniiquum (Cope), 3272 B.

A. " ' 3330 ?

A. sp. indet., 3254 B.

A. " " 3177 C.

A. " " 2981 C.

A. " " 3260 C.

A. " " 3134 B.

A. " " 3050 C.

A. " •' 3019 B.

A. " " 3245 C.

A. " " 3285 C.

IncertcE sedis:

Gen. and sp. indet., 2394 B.

" " " " 2374 B.

Fragments of Baena, Amyda, etc., 3250 C.

Gen. and sp. indet., 2982 C.

" " " " 2395 B.

" " " " 3445 B?

Anguidae:

Glyptosaurus sp. indet., 3405 C.

GILMORE: the fossil turtles of the UINTA FORMATION 107

Family BAENID.E Cope.

In the present collection from the Uinta formation of Utah nineteen specimens
were sufficiently well preserved to be identified as pertaining to the genus Baena.
These were found in all three subdivisions of the Uinta, being distributed as follows:
one specimen, Baena inflata, from Horizon A; thirteen from Horizon B; two from
horizon C; and three for which the data for the horizon were uncertain, or not given.

Five species of this genus are now recognized as occurring in the Uinta forma-
tion, three of which are here described as new. Only one of the recognized species,
Baena arenosa, is found to occur in other geological epochs, and no member of this
family is known to range above the Uinta.

Genus Baena Leidj^
1. Baena arenosa Leidy.
Plate XVni, fig. 1; text-fig. 1.
Baena arenosa Leidy, Proc. Acad. Nat. Sci. Phila., 1870, p. 123; U. S. Geol. Surv.
Wyoming, etc., 1870 (1871), p. 367; U. S. Geol. Surv. Montana, etc., 1871
(1872), p. 368; Contrib. Ext. Vert. Fauna West. Terrs., 1873, pp. 161, 343,
pi. 13, figs. 1-3; ?pl. 15, figs. 1-5; pi. 16, figs. 8, 9.— Cope, ?Append. LL of
Ann. Report Chief of Engineers, 1875, p. 96; ?Wheeler's Surv. 100th Merid.,
1877, p. 52, pi. 24, fig. 32.— Baur, Proc. Acad. Nat. Sci. Phila., 1891, p. 426.
—Hay, Bibhog. and Cat. Foss. Vert. N. A., 1902, p. 437; Foss. Turtles of
N. A., 1908, pp. 67-71, pi. 12; pi. 13, fig. 1; pi. 14, figs. 1-3, text-figs. 44-51.
Baena affinis Leidy, Ann. Report U. S. Geol. Surv. Wyoming, etc., 1870 (1871),
p. 367.

This species is represented in the collection by a single specimen. Cat. No. 2356.
Collected by Earl Douglass and party, June 18, 1908, from Horizon B, Uinta
formation, Upper Eocene, east of Dragon- Vernal road between White and Green
Rivers, Uinta Basin, Uinta County, Utah.

The specimen consists of a fairly complete carapace, lacking the posterior
borders and the peripherals of both sides, the plastron lacks portions of both
anterior and posterior lobes. It represents an individual of approximateh' the
same size as the type of the species (Cat. No. 103, U. S. National ]\Iuscum), with
which it has been carefully compared. This comparison shows several differences,
but such as exist are not considered of sufficient importance to separate the speci-
mens specifically. The sculpture of the carapace is rough and uneven, consisting

108

MEMOIRS OF THE CARNEGIE MUSEUM

of various longitudinal, transverse and oblique ridges, especially within the areas of
the vertebral scutes, this part of the carapace being fairly smooth in the type.

Fig. 1. Carapace of Baena arenosa Leidy. C. M. No. 2356. X
vertebral .scutes one and five.

ns., neural scute; v.s. 1-v.s. 5,

In the measurements of the vertebral scutes it also differs from the type, but
agrees almost exactly with the type of Baena affinis Leidy, which is now regarded
by Hay, following Leidy and Cope, as being a synonym of B. arenosa. The verte-
bral areas of this type are also in accord with the present specimen. In order
to show the close agreement of the present individual with the above-mentioned
specimens a table giving the comparative measurements of the vertebral scutes
of each is herewith appended.

Comparative Measurements of Vertebrals.

Length.

Width.

No. 23.56.

TypeofB.armom.

Type of .B.a^m*.

No. 2356.

Type of B. arenosa.

Type of S. affinia.

1

44

_

50

57

72

2

76

77

73

62

73

61

3

76

74

73

64

77

64

4

60

61

60

68

78

63

5

-

68

55

88

96

75

Since the anterior portion of the carapace is missing in the type of the species
and in all subsequently discovered specimens the complete anterior margin in the

GILMORE: the fossil turtles of the UINTA FORMATION 109

present specimen is therefore worthy of brief description. The nuchal scute is
small, ha\nng a length of 12 mm., and a transverse diameter of 17 mm. The
median part of the anterior border projects slightlj- bej'ond the general contour of
the shell. The nuchal is flanked on either side by small rectangular marginals.
The arrangement and proportions of the scutes in front of the first vertebral are
very similar to those of Baena antiqua Lambc.

This specimen shows many of the sutures between the costals, but on the
median dorsal surface their complete coalescence renders it impossible to differ-
entiate the neurals. There are no supernumerary costal scutes on either side of the
first vertebral such as found in many species of this genus and occasionally in
individuals pertaining to the present species.

The plastron agrees almost exactly with the type in size and proportions, and
especially in the sculpture of the surface and the course of the various sulci. The
accompanying table gives a comprehensive comparison of the principal measure-
ments of the plastra.

Comparative Meastjeements of Plastra.

Type of ^. arenosa. I Type of B. affinit.

Width of bridge

Length of anterior lobe .
Width of anterior lobe.

140 125

77 I 77

107 104

Width of posterior lobe I 96 I 98 ! 114

The discovery of the present specimen in the Uinta formation of Utah in-
creases somewhat the known geographical as well as the geological range of Baena
arenosa. The tj^pe of the species is from the Bridger deposits at the junction of
the Big Sandy and Green rivers in southeastern Wj'oming. Haj-^ has identified
this species from the Washakie, the uppermost division of the Bridger. The type
of B. affinis is from level B, of the Bridger, which represents the middle of that
formation. Both Cope and Hay have identified specimens from the Wasatch of
New Mexico as pertaining to this species, these being the most ancient known at
the present time. With a geological range beginning in the Wasatch and ending
in the middle of the Uinta, Baena arenosa enjoys the distinction of having the widest
geological distribution of any species of the genus.

2. Baena emiliae Hay.
Baena emilim Hay, 0. P., Fossil Turtles of North America, 1908, pp. 80-81, PI.
XX, fig. 1; text-figs. 67, 68.
Six specimens in the present collection are identified as pertaining to Baena

* Hay, 0. P., "Fossil Turtles of North America," Pub. Carnegie Institution, Washington, 190S, pp. 67-68.

110 MEMOIRS OF THE CARNEGIE MUSEUM

emilice Hay. The better preserved specimen, C. M. No. 3443, consists of a nearly
complete carapace and plastron, the former lacking small portions of the hinder
borders posterior to the inguinal notches. The missing margin of the right side
was apparently lost in life, as shown by the healed condition of the bone at this
point. The specimen was collected by Earl Douglass during the season of 1915 at
Wagon-hound Bend, on White River, Uinta County, Utah, from the lower part
of Horizon B. It will be seen from the table of comparative measurements given
below that this specimen has about the same dimensions as the type of the species.
It differs, however, in the more angularly rounded contour of the front lobe, a
feature in which it also is different from three of the other specimens here referred
to this species. Whether this difference represents a sexual character, or is only
an individual variation, I am unable to determine.

In many respects the present specimen is very close to the type of Ba'ena clara
from the Bridger formation, but the great length of the third vertebral scute, as
compared with the others of the series, is regarded by Hay as one of the chief dis-
tinguishing characters of the species, and together with the much shorter pos-
terior lobe, as compared with the longer lobe in B. clara, appears to show that its
closest affinities are with the present species.

An anterior portion of a carapace and plastron, C. M. No. 3253, on account
of its close general resemblance to the specimen discussed above, is provisionally
referred to the same species. This specimen is from Horizon C of the Uinta for-
mation and is the only individual in the collection, referred to the present species,
which is positively known to have come from that horizon, all of the others having
been found in strata belonging to Horizon B. It was collected by Earl Douglass
two or three miles west of Well No. 2, Uinta County, Utah.

A third specimen, C. M. No. 2159, consisting of a complete plastron and the
entire central part of the carapace, but lacking portions of both sides, is also
referred to this species. It was collected by Earl Douglass in 1908, in the Devil's
Playground, Uinta Basin, Uinta County, Utah, from Horizon B (near top) or C
(near base).

A fourth individual, C. M. No. 3243, has a nearly complete carapace and
plastron, the latter lacking the posterior lobe. It was collected by Messrs. Earl
Douglass and J. T. Goetschius, October 2, 1908, about one mile northeast of Well
No. 2, "near first gap," Uinta County, Utah, from Horizon B.

The fifth specimen, C. M. No. 3257, consists of a carapace and plastron, the
former lacking some of the posterior border. This turtle also was collected by
Earl Douglass and J. T. Goetschius, July 30, 1908, south of Kennedy's Hole and

GILMORE: the fossil turtles of the UINTA FORMATION 111

west of the Dragon- Vernal road, Uinta Basin, Uinta County, Utah, from Horizon
B, or C.

The sixth specimen, C. M. No. 3244, consists of the greater part of the carapace
and plastron. The carapace has portions of the rim missing on both the anterior
and posterior ends, the plastron lacks the anterior lobe. The specimen was
collected by Earl Douglass in 1915, at Wagon-hound Bend on WTiite River, Uinta
County, Utah, from the lower part of Horizon B.

The tjTDe of Baena emilicB is in the American ]\Iuseum of Natural History,
and was collected by Mr. O. A. Peterson in 1884, from the middle Uinta of Utah.
Geologically therefore all of the known specimens, including the tji^e, two other
specimens referred to the species by Hay, and the six specimens under con-
sideration, came from approximately the same horizon, and from neighboring
localities.

It may be shown hereafter, when larger collections shall have been made, that
more than one species is represented by the six specimens here referred to B. emilioe.
When compared with one another there are differences which appear to divide
them into three groups, as follows: Nos. 2159 and 3243, having relatively narrower
vertebral scutes and narrower plastral lobes and bridges than the tjT^e, or other
specimens here referred to B. emilioe; Nos. 3244 and 3257, having wider vertebral
scutes and a more depressed shell than the type; and No. 3443 with a wider and
more angularly rounded anterior lobe, larger intergulars, and narrower pectorals.
The latter specimen in all of these particulars is different not only from the type,
but from all of the other specimens discussed above, with the exception of the
fragmentary specimen No. 3253, which, in so far as the two can be compared,
appears to be very close to No. 3443. In nearly all other respects these specimens
agree closely with the type of the species. The differences enumerated above are
not considered important enough to warrant the separation of these turtles into
distinct species. When the considerable sexual and individual differences ob-
servable in a series of living turtles of one species and from one locality are con-
sidered, it appears to me that the specimens before me are well within the limits of
a given species. I am inclined to the belief that specimens Nos. 2159 and 3243
may be females of this species, but as to this I cannot be certain. The discovery of
more material may possibl}^ show that more than one species is represented in
these specimens, but at this time, especially in the light of a recent examination
of a large series of living turtles, I do not feel justified in the establishment of new
species on such slender distinguishing characters as have been observed. For the
present, at least, I refer all the six specimens to Baena emilicc Hay.

112

MEMOIRS OF THE CARNEGIE MUSEUM

In order to place on record the proportional variations within the species I
have prepared the table of comparative measvirements given below:

CosiPAR.\TivE Measurements of Vertebr.\ls.

Length.

Width.

Type.

No.
3443.

No.

3244.

No.
3257.

No.
2159.

No.
3243.

Type.

No.

3443.

No. 3244.

No. 3257.

No. 2159.

No. 3243.

1

55

53

51

47

38

68

73

85

65

63

2

80

81

—

82

71

77

63

79

71

87

64

56

3

92

96

97

89

87

88

75

82

80

90

68

63

4

70

69

74

72

69

67

78

77

80

91

69

65

5

70

62

52

—

—

62

98

101

90

96

COMPAKATIVB MeaSDKEMENTS OF CaRAPACE AND PlaSTRON.

Greatest length of carapace. . .
" width " "
" length of plastron . . .
" " anterior lobe . .

w-idth " "
" length posterior lobe
" vndth " "

Width of bridge

Type.

No. 3443.

368

364

294

292

320

307

80

76

110

105

83

84

122

114

160

150

io. 3443. No. 3244. No. 3257. No. 2159. No. 3243,

365e 375e
310 310

119
149

310e
75

105
98e

117

146

77
102

366
260

73

105
146

3. Baena inflata sp. nov.
Plate XIX; text-figs. 2 and 3.

Type : C. M. No. 3406, consisting of a carapace and plastron, the former lacking
the posterior end back of the middle of the fourth vertebral, the latter a small
portion of the anterior lobe; collected bj^ 0. A. Peterson in 1912.

Locality : INIcCook Canj'on, White River, Uinta County, Utah.

Horizon: Horizon A (near top), Uinta formation, Upper Eocene.

The type of this species is but little crushed and, except for the parts which
are missing, is in a beautiful state of preservation. The surface of the carapace
and plastron are everj^where covered with fine pustular elevations, forming a
shagreened surface. The pustules on the carapace are coarser than those on the
plastron. The surface of the carapace is also somewhat uneven, and laterad to
the second, third, and fourth vertebrals there are some heavy longitudinal wrink-
lings. These are most numerous laterally at the junction of the third and fourth
vertebrals. The pustular ornamentation of the carapace appears to be very
similar to that of Baena sima Hay, but not so coarse.

The greatest length of the shell is estimated to have been about 400 mm.; its
greatest width at the center is 310 mm. In outline the front of the shell is evenly,

GILMORE: the fossil turtles of the UINTA FORMATION 113

but broadly, rounded, resembling in its general contour B. sima Hay. The shell
is flat transversely in the region of the vertebral scutes, but from one border to the
other it is broadly convex. One of the distinctive features of this species is the
decided transverse inflation or swelling of the mid-costal region, which gives the
shell the appearance of being puffed out on the sides. This swelling lies largely
within the areas of the second costal scutes, and it is to this feature that the specific
name refers.

Over the posterior legs the margins of the shell begin to flare outward and
slightly upward, and at this point the border is heavy and rounded but becomes
thinner posteriorly. In front of the axillary notches the border has a thickness of
31 mm., but rapidly thins toward the center, where it measures only 8 mm., the
edge being obtusely rounded. The bones of the carapace are so thoroughly coos-
sified that but few of the sutures can now be made out. The sulci, however, can
in most instances be clearly traced.

Fig. 2. Carapace of Baena injlata. C. M. No. 340G, Type. C.S. 1, C.S. 4, costal scutes one and
four; V.S. 1, V.S. 4, vertebral scutes one and four. One-fourth natural size.

The vertebrals as in nearly all Eocene Baenidae are longer than wide. The
sides of the vertebrals, excepting the first, which is hexagonal with a very narrow
anterior end, are bracket-shaped. It will be observed, that, as in Baena emilice,
the third vertebral is the longest of the series. Along the center of vertebrals two,

114

MEMOIRS OF THE CARNEGIE MUSEUM

three, and four is a narrow low ridge, on eitlier side of which are parallel grooves
much as in B. emiliie. The principal dimensions of the vertebrals as well as those
of a second individual, C'. M. No. 3442, referred to this species, are given in the
accompanying table.

Dimensions op Vbrtebeals.

Length.

Width in front.

Greatest width.

Type.

No. 3442.

Type.

No. 3442.

Type. No. 3442.

1

2
3
4
5

61
90
99

53
92
98
90

28
62
62
68

75
63
64
70
53

72 75
74 77
78 81
78 1 81

The nuchal scute is rectangular, being about 17 mm. long and 30 mm. wide on
the free border. It is bordered on either side, as shown in specimen No. 3442,
by small triangular first marginals. The second marginal has its greatest width

Fig. 3. Plastron of Baiina inflala C. M. No. 3406, Type, restored after C. M. No. 3442. One-foui-th
natural size.

(36 mm.) on the free border. The first vertebral is bordered on either side by
small triangular supernumerary costal scutes, though there is no indication of these
in specimen No. 3442.

It is estimated that the plastron had a total length of about 365 mm. It is

GILMORE: the fossil turtles of the riNTA FORMATION 115

slightlj^ convex transversely throughout its length, and this convexity on the bridge
area continues evenly to the borders of the shell, so that these borders stand 36 mm.
above the level of the plastron at the center. The anterior lobe has its greatest
width (135 mm.) at the base; and at a point half-way to the anterior end it measures
99 mm. The sides of the lobe converge gradually from the base to the anterior
end, which appears to have been rounded. The width of the bridge is 175 mm.

The posterior lobe is tongue-shaped, shallowly, but broadly notched. Its
lengtli on the midline is 97 mm., with a width at the base of 139 mm. The notch
has a depth of 6 mm. at the center.

Excepting those of the anterior lobe all of the sulci and sutures on the plastron
can be clearly made out. The mesoplastrals widen rapidly on either side of the
midline. At the center the right scute measures 51 mm. The width of the right
hypoplastral at the center is 104 mm. ; of the left hypoplastral 91 mm. The xiphi-
plastrals are 62 mm. wide on the midline. The pectorals meet on the midline for
a distance of 75 mm.; the abdominals for 49 nam.; the femorals for 80 mm.; the
anals for 47 mm.

On the right side are three inframarginal scutes, the form of which is well
shown in Fig. 3.

A second specimen, C. M. No. 3442, belonging apparently to this species, was
collected by Earl Douglass in 1915 from the lower part of Horizon B of the Uinta
formation, at Wagon-hound Bend on White River, Uinta County, Utah. This
turtle consists of a carapace and plastron, both of which have small portions
missing from their posterior ends. In size, general contour, and the dimensions
of the dermal scutes, the specimen closely resembles the type. The inflation of the
sides of the carapace, which forms such a conspicuous feature in the tj'pe, is almost
entirely wanting in this individual. Its absence may be attributed in part, at least,
to crushing, for both sides in this respect have somewhat suffered. There are also
no supernumerary costal scutes at either side of the first vertebral, and in their
absence the first vertebral is tetragonal, whereas in the tj'pe it is hexagonal with
the narrow end in front. This specimen shows small triangular first marginals
on either side of the nuchal, and in the drawing of the type (Fig. 2), this region,
which is missing, has been restored after this specimen. It also gives the complete
form of the anterior lobe (See Fig. 3), which in its general contour closely resembles
Baena sima Hay.

The greater part of an anterior lobe, C. ]\I. No. 3137, which was collected by
Earl Douglass in the strata of Horizon B of the Uinta formation, near Well No. 2,
Uinta Basin, Utah, in 1908, is regarded as belonging to Baena inflata. It is from

116 MEMOIRS OF the' CARNEGIE MrSEtiM

an individual having the same proportions as the type, and shows on the dorsal
surface the triradiate shape of the entoplastron (See Fig. 4, 1). The entoplastron

Fig. 4. Anterior lobes of Baena inflala. 1, and ,?, superior and inferior views of C. M. No. 3137, enl.,
entoplastron; hum., humeral scute. S, inferior view of anterior lobe of C. M. No. 3442. All figures one-fourth
natural size.

has a length of at least 50 mm. ; a width of 44 mm. Transversely the lower surface
of the lobe is broadly convex. On the dorsal surface immediately posterior to the
anterior border the bone is scooped out by a shallow transverse depression. The
lateral borders are bevelled off almost perpendicularly, while in front the border
is rounded. The bone along the borders has a thickness of 13 mm., in front of the
center of only 9 mm., at the middle on the posterior broken border of 21 mm.

4. Baena gigantea sp. nov.
Plate XX, figs. 1 and 2; text-figs. 5, 6, and 7.

Type: C. ]M. No. 3441, consisting of nearly a complete shell. The carapace
lacks portions of the posterior margins on either side of the middle, a small section
of the right anterior border, and the peripherals of the right side above the bridge.
The plastron has the greater part of the posterior lobe missing. Collected by
Earl Douglass, in 1915.

Locality : Wagon-hound Bend, on White River, Uinta County, Utah.

Horizon : Lower part of Horizon B, Uinta formation. Upper Eocene.

The type of the present species is the largest species of the genus as yet dis-
covered. It is estimated that the carapace had an axial length of about 535 mm.
The greatest width, which is near the center, is about 420 mm. The bones of the
carapace are all thoroughly coossified and only the sutures defining the right half
of the mesoplastron can be detected and then only with difficulty. The shell is
oval in outline, in this respect resembling Baena clara Hay, though the oval is
somewhat more elongate than in that species. The carapace has been slightly
crushed on the right side, as may be seen by examining Plate XX, fig. 1.

The front of the carapace is decidedly projecting. The missing posterior

gilmore: the fossil turtles of the uixta formation

117

borders render it impossible to determine the character of the scallops on the
hinder end. The surfaces of both the carapace and plastron are roughened with
coarse pustular elevations, though these are more sparsely placed than in either
Baena sima or B inflata. With the exception of this pustular roughening the
surfaces are comparatively smooth, there being no longitudinal ridges or grooves,
such as are commonly found in many species of this genus from the Eocene. The
vertebral areas are a^so free from median ridges and channels.

The nuchal scute resembles in outline that of Baena hatcheri. It has a fore-
and-aft diameter of about 58 mm., and a transverse diameter of 88 mm. The

Fig. 5. Baena gigantea, carapace of C. jM. No 3441, Tyi)e. About one-fifth natural size.

unusual length of the nuchal appears to be one of the distinctive features of this
species. All of the sulci are distinctly impressed. At the left side of the nuchal
is a subrectangular first marginal, which has a length on the free border of 43 mm.
The total number of peripherals cannot be determined from this specimen.

There are the usual five vertebrals and these are relatively' wide, and differ
from those of all other Eocene Baenidse, except B. emilice, in having the fourth
considerably wider than long. The sides of the vertebrals posterior to the first are

118

MEMOIRS OF THE CARNEGIE MUSEUM

onlj' slightly bracket-shaped. The first is hexagonal, verj^ narrow in front, in this
respect closelj'^ resembling the first in Baena inflata. The surfaces within the first,
second, and third vertebral areas are flattened, but the fourth and fifth are trans-
versely broadly convex. The principal dimensions of the vertebrals are given
in the accompanying table.

Dimensions of Vertebrals.

Length.

1 Width in front.

Greatest width.

1

58

! 34

98

2

122

! 87

105

3

126

96

113

4

88

94

106

5

102e

88

152e

e, estimated.

As in Baena riparia Hay and B. hatcheri Hay, there are five costal scutes. A
small supernumerary scute is situated on each side of the first vertebral, showing a
difference from the former species by bordering on the nuchal, whereas in B.
riparia these scutes are not in juxtaposition.

On the plastron only the sutures defining the mesoplastron on the right side

Fig. 6. Bama giganlea, plastron. Tjiie, C. M. No. 3441. About one-fifth natural :

GILMORE: the fossil turtles of the UINTA FORMATION 119

are traceable, and these show them to be narrow at the micUine (18 mm.), but
expanding toward their outer extremities where the width is 112 mm. The an-
terior lobe is elongated antero-posteriorly, and turns upward with a well-defined
sweep toward the carapace, as shown in Fig. 7. Its greatest length is 137 mm.;
its greatest width 170 mm. at the base; at a point half-way to the tip measuring
117 mm. in width. The sides of this lobe gradually converge from the base to
near the anterior end, which rounds in with a shallow but broad median emargina-
tion on the anterior end. The posterior lobe is largely missing, though enough of
the base remains to show that it had a width of 160 mm. The width of the bridge
is 190 mm.

Fig. 7. Baena gigantea, lateral view of the carapace and plastron, C. M. No. 3441. T3'pe specimen,
one-fifth natural size.

The sulci defining the intergular scutes cannot be traced. The intergulars
meet on the midline for a distance of 29 mm.; the humerals 112 mm.; the pectorals
97 mm.; the abdominals 55 mm. The number of inframarginals on the bridge
cannot be determined in this specimen.

This species may be distinguished from all others of the genus by its larger size,
the great length of the nuchal scute, and differences in the relative dimensions of
the vertebral scutes. The contour of the anterior lobe of the plastron, its greater
relative length, and especially its decided upward curvature are all features which
serve to distinguish this species. In the presence of five costal scutes the type of
this species agrees with several species of the genus, especially Baena riparia and
B. hatcheri, but it differs from those forms by the decidedly longer nuchal scute and
in the apparent absence of marginal scutella. In the greater length of the third
vertebral this specimen is like B. emiliai from the same formation, but the greater
relative widths of all of the vertebrals and especially the shortness of the fourth,
together with other differences to be observed in the plastron, at once distinguish
it from that species.

120

MEMOIRS OF THE CARNEGIE MUSEUM

,"). Baena platyplastra sp. nov.
Plate XVIII, fig. 2; text-fig. 8.

Type: C. IM. No. 3227, consisting of a plastron lacking the anterior portion
of the anterior lobe, matrix cast of the carapace, at either end of which remain a
few fragmentary parts of the carapace. Collected by Earl Douglass and J. F.
Goetschius, August 5, 1908.

Locality: Northeast of Well No. 2, Uinta Basin, Uinta County, Utah.

Horizon : Horizon B, Uinta formation. Upper Eocene.

The type specimen represents one of the larger species of the genus. It is
distinguished from all other described species of Baena by the extremely flat and
thin plastral bones with sculptured inferior surfaces. Its large size and the absence
of a median emargination on the posterior lobe are features which also assist in
distinguishing this species.

The ornamentation of the plastron consists of low ridges and shallow furrows,
the former being short, sometimes straight, but usually bent or anastomosing.
The effect of the whole may be best expressed as resembling a coarse, shagreened

Fig. S. Plastron of Duma plalyplaslra, C. M. No. 3227. Type, one-fourth natural size.

leather. The sutures in the tjpe have all coalesced, but their courses are indi-
cated by ridges crossing them at right angles. These cross-ridges are especially
pronounced on the median suture between the axillary notches.

GILMORE: the fossil turtles of the UINTA FORMATION 121

It is estimated that the entire shell had a length of about 495 mm., and a height
at the center of about 170 mm. The plastron had a length of about 420 mm. The
anterior lobe at the base is 156 mm. wide. The bridge is 180 mm. wide.

The posterior lobe has a width at the base of 152 mm., and a length of 128 mm.
The lateral borders of the lobe are nearly straight and converge nearly the entire
length of the lobe, there being a slight constriction at the anal-femoral sulcus. The
posterior end of this lobe is broadly but evenly rounded and without median emar-
gination. At the anal-femoral sulcus the transverse measurement is 107 mm.

The mesoplastrals are solidly coossified with the contiguous bones and their
boundaries can only be determined by the ridges which cut them at right angles.
At the midline these bones have a width of 33 mm., at their outer ends thej^ expand
to 55 mm. in width.

As in Baena sima Hay, the median sulcus runs a very tortuous course, as is well
shown in Fig. 8. The pectorals meet on the median line for a distance of 77 mm. ;
the abdominals for 70 mm.; the femorals for 68 mm.; the anals for 85 mm. The
anal-femoral sulcus runs in from the border a short distance, then turns abrupth'
forward, then again turns at right angles toward the median line to meet the scute
of the opposite side. In the shape of the anal scutes it resembles Baena arenosa
and more especially B. clara.

Owing to the damaged condition of the bridges the number of inframarginal
scutes cannot be determined.

Family DERMATEMYDID.E Gray.

The family Dermatemydida3 is represented now for the first time in the Uinta
formation by the single genus and species, Anosieira ornata Leidy. This is the
latest recorded occurrence of this genus for North America, although in England
it is known to range upward into the Lower Oligocene.

Genus Anosteira Leidy.

6. Anosteira ornata Leidy.

Anosteira ornata Leidy, Proc. Acad. Nat. Sci. Phila., 1871, p. 102; Ann. Report
U. S. Geol. Surv. Montana, etc., 1871 (1872), p. 370; Contrib. Ext. Fauna
West. Terrs., 1873, pp. 174, 341, pi. XVI, figs. 1-6.— Hay, Bibliog. and Cat.
Foss. Vert. N. A., 1902, p. 447; Bull. Amer. Mus. Nat. Hist., XXII, 1906,
p. 157, figs. 2, 3; Fossil Turtles of North America, 1908, pp. 279-281, PI. XLIII,
figs. 1, 2; text-figs. 352-354.

122 MEMOIRS OF THE CARNEGIE MUSEUM

Anostira ornnta Cope, Ann. Report U. S. Geol. Surv. Wyoming, etc., 1872 (1873),

p. 621; Amer. Naturalist, vol. XVI, 1882, p. 989, fig. 7; Vert. Tert. Form.

West, 1884, p. 128.— DoLLO, Bull. Mus. Roy. Belgique, IV, 1886, p. 93, PI.

XI, figs. 7, 8.

A fragmentan- specimen, Xo. 295-4, collected by 0. A. Peterson August 24,
1912, from Horizon C, Uinta formation, on White River near Ouray, Uinta County,
Utah, is provisionally identified as pertaining to the above genus and species.
This specimen consists of the articulated nuchal, first and second neurals, with
portions of the abutting costals of both sides, parts of several disarticulated costals,
eleven peripherals, several of which are complete. The plastron is represented
by the right hj'poplastron lacking a portion of its outer extremity and many frag-
mentary parts.

The specimen has been carefully compared with the figures and descriptions
given by Leid}' and Hay, and especialh' ynth one of Leidj-'s cotj-pes Xo. 4062,
now in the U. S. National ]\Iuseum, and, with the exception of slight differences in
size, it agrees closely in nearly all respects. The present specimen is of about the
same size as one individual in the American ^lusevun of Xatural Historj^ described
and figured b}' Hay in his Turtles of Xorth America, but is considerably smaller
than the cotype of Leidy mentioned above.

All of the specimens described by Leidy are supposed to have come from
the lower portion of Horizon B in the Bridger as exposed in the neighborhood of
old Fort Bridger, Wyoming. The specimen described by Hay in the publication
cited is from the third division of Horizon C of the Bridger on Henry's Fork, Wj'o-
ming. The discovery of the specimen considered here now extends the geological
range of this species into the uppermost horizon of the Uinta formation.

The nuchal has a length of 15 mm., a width on the free border of 23 mm.
The free border is subacute and is not so deeply excavated in front as in the specimen
figured by Leidy. The thickness of the nuchal at the midline is 5 mm.

The first neural has a length of 13 mm., and a greatest width of 6 mm. The
bone is coffin-shaped with the widest end forward. The second neural is 9 mm.
long, and only 4 mm. wide.

All of the bones of the carapace are delicately sculptured, though those of the
anterior part of the shell appear less distinct than in most of the described specimens.
The few costals present show the usual low undulating ridges crossing them at
right angles to their shorter diameters. This sculpture is most distinct toward
their outer ends. The peripherals have their upper and lower surfaces ornamented
by the usual sharp ridges and pustular elevations.

GILMORE: the fossil turtles of the UINTA FORMATION 123

The few sulci discernible are narrow and delicately impressed. As in pre-
viously described specimens the intramarginal sulci on the nuchal and anterior
peripherals cannot be traced. The first vertebral has a greatest width of 26 mm.,
whereas in the specimen described by Hay it is only 18 mm. The sulcus forming
the posterior boundary of the first vertebral crosses the first neural as in other
described specimens. The costal sulcus on the second costal is near the center of
that bone, while in the specimen described by Hay it is very close to the posterior
border.

The right hypoplastron is 21 mm. long on the midline, and has a thickness of
5.5 mm. The sculpture on the lower surface of this bone is made up of fine ridges
arranged in a radiating pattern. There is no evidence of epidermal scutes on any
of the plastral bones found with this specimen.

Cope has recognized this species from the Upper Green River beds, so that the
evidence at hand shows that this species ranges from the lowest horizon in the
Bridger deposits to the highest horizon in the Uinta formation, the uppermost
Eocene.

Family EMYDID.E Gray.

Genus Echmatemys Hay.

In 1908 Hay^ recognized nineteen species as pertaining to the genus Ech-
matemys. Since that time he has described one new form,^ so that with the four
new species described in the present paper, twenty-four species have been recog-
nized from the fossiliferous deposits of North America. Seven of these have now
been found in the Uinta formation and increased collections will doubtless show the
presence of several more. The discovery in the present collection of Echmatemys
septaria (Cope) leads to the belief that still other species known in the older
Wasatch and Bridger beds, will sooner or later be found to continue into the upper-
most Eocene.

6. Echmatemys callopyge Hay.

Plate XXI; text-figs. 9 and 10.

Echmatemys callopyge Hay, Fossil Turtles of North America, 1908, 340-342, PI.

LIT, figs. 1, 2; text-figs. 447, 448.

Two specimens in the Carnegie Museum are identified as belonging to this
species. The better preserved specimen. No. 2371, was collected by Earl Douglass
in 1908, from Horizon B, "above second sandstone with small artiodactyls,"

' " Fossil Turtles of North America," 1908, p. 298.

« Proc. U. S. National Museum, XXXV, 1908, pp. 164-166.

124

MEMOIRS OF THE CARXEGIE MUSEUM

Uinta formation, Upper Eocene, east of Dragon- Vernal road lietwccn White and
Green rivers, Uinta Basin, Uinta County, Utah. The second specimen, Xo. 2157,
was also collected b\' Douglass from the same geological horizon near Well No. 2,
in the Uinta Basin. Like the tj-pe, both of these specimens have the carapace
somewhat crushed over toward the left side. The type of the species is said by
Hay to have come from the middle Uinta, and it appears probable that all of
these specimens were found at about the same geological level.

Hay considered the very narrow first vertebral as the chief distinguishing
character for separating this species from the others of the genus, but both of the

Fig. 9. Echmaletnys callopyge Hay. Carapace of C. M. Xo. 2371. One-fourth natural size. c. 8,
costal plate; cs. /, first costal scute; n. /, and n. 8, neurals one and eight; n.p., nuchal plate; s.p., suprapygal.

specimens before me have this scute relatively wider than in the type, although in
nearly all other respects, as is shown by the table of comparative measurements,
the specimens are remarkably similar. So far as the width of the first vertebral is
concerned these specimens are intermediate between the tj'pe of the present speci'is
and the figured specimen of Echmatemys septaria (Cope), as illustrated by Hay,
Fossil Turtles of North America, Fig. 415, p. 320. The t\-pe of the latter species
is in the U. S. National Museum (No. 4088), and consists of a fairly complete

GILMORE: the fossil turtles of the UINTA FORMATION 125

plastron, a portion of the central part of the carapace, including the third, fourth,
and fifth neurals with portions of the second, third, and fourth costals: and a small
piece of the fifth, sixth, and seventh costals with abutting peripherals. It was
collected in the badlands of South Bitter Creek, Wyoming, from the beds of the
Washakie Basin.

I have carefully compared the specimens before me with the above mentioned
type and except for differences in size, find them, so far as they can be contrasted,
remarkably similar. The broad, hatchet-shaped anterior lobe so characteristic of
Echmatemys septaria is duplicated in these specimens.

In the type of Echmatemys callopyge the front two-thirds of the first vertebral
lies wholly within the lateral borders of the nuchal plate, and, although relatively
wider, this is also true of specimen C. M. No. 2157, but specimen C. M. No. 2371
has the antero-lateral angles of the first vertebral extending across the lateral
sutures of the nuchal. In a specimen identified by Hay as pertaining to Ech-
matemys septaria (See Fossil Turtles of North America, Fig. 415, p. 320) the first
vertebral extends entirely over the lateral boundaries of this plate. From the
intermediate condition observed in the present specimens, the first vertebral
would appear to be subject to considerable variation and therefore its narrowness
cannot be relied upon as a constant specific difference. Specimen C. JNI. No. 2371

Fig. 10. Echrnatemys cdlopyye Haj'. Plastron of C. M. No. 2371. One-fourth natural size.

126

MEMOIRS OF THE CARNEGIE MUSEUM

has the surface of the carapace smooth, with the exception that, as in the type of
E. septaria, it is relieved by faint striations and growth lines, these being especially
apparent within the areas of the vertebral scutes.

Measukements of Neurals.

No.

Length.

Width.

Type.

No. 2157.

No. 2371.

Type.

No. 2157.

No. 2371.

1

52

50

50

33

35

36

2

43

40

39

41

3

58

—

43

43

—

39

4

41

41

38

37

38

34

5

41

26

35

44

41

6

30

29

40

—

39

7

29

25

25

45

—

40

^8

28

—

23

34

—

35

Measurements of Veutebrals.

No.

Length.

Width.

Type.

No. 2157.

No. 2371.

Type.

No. 2157.

No. 2371.

1

75

81

77

52

71

77

2

90±

85

85

87

86

3

88

96

87

82

95

76

4

96

86

87

100

84

5

74

—

—

100 ±

—

Principal Measurements Carapace and Plastron.

No. 2157. I No. 2371.

Greatest length of carapace

Greatest width of carapace

Greatest height of carapace

Nuchal, greatest length

Nuchal, greatest width

First marginal, greatest length . . .
First marginal, greatest height . . .

Plastron, greatest length

Anterior lobe, greatest length. . . .

Anterior lobe, width at base

Posterior lobe, greatest length. . .
Posterior lobe, width at base ....

Bridge, width

Lip, width

Entoplastron, width

Gulars, meet on the midhne

numerals, meet on the midline . .
Pectorals, meet on the midline . . ,
Abdominals, meet on the midline
Femorals, meet on the midline . .
Anals, meet on the midhne

438

4204-

415

270

276

279

153

160

133

69

66

90

86

80

40,

43

40e

32

34

28

410

422

370

116

125

107

174

180

176

135

140

—

198

205

195

155

165

162

38

44

45

62

66

75

72

67

58

35

34

36

74

76

73

110

114

102

38

41

33

66

' 62

—

Had E. callopyge not been established, I should have unhesitatingly referred
both of the specimens discussed above to Echmatemys septaria (Cope). For the

GILMORE: the fossil turtles of the UINTA FORMATION 127

present, however, it will serve all purposes to assign them to the established Uinta
species, until the discovery of additional Bridger material shall definitely determine
whether two distinct species are represented by this material, or whether E. cal-
lopyge Hay shall become a sjmonym of the earlier described E. septaria (Cope).

In order to place on record the variation within the species, some of the prin-
cipal measurements of the two specimens here considered as compared with those
of the type of the species are given in the preceding table.

7. Echmatemys uintensis Hay.

Echmatemys uintensis Hay, Fossil Turtles of North America, 1908, pp. 342, 343.

PI. LIII, figs. 1, 2.

The above species is represented in the Carnegie ^Museum collections by three
specimens. The better preserved specimen, C. IVI. No. 3270, consists of a carapace
and plastron, the former lacking a portion of the posterior end and a considerable
part of the costals and peripherals of the right side. The plastron is complete.
This specimen was collected by Earl Douglass, May 25, 1908, from Red Bluff Wash,
on the road from Bonanza to Kennedy's Hole, Uinta Basin, Utah, from Horizon B,
"transition beds. First sandstone above red layer," Uinta formation. Upper
Eocene.

The second specimen, No. 2158, consists of a carapace lacking most of the
costals and peripherals of the left side, was collected by Earl Douglass, August 22,
1908, two or three miles below Well No. 2, from Horizon B, Uinta formation, as
exposed in the Uinta Basin, Utah.

The third specimen. No. 2397, consists of considerable portions of the carapace
and plastron of a large individual, both of which are rather fragmentarj'. This
specimen was collected by Messrs. Earl Douglass and J. F. Goetschius, August 17,
1908, from Horizon B, "grey beds below red and grey beds," Badlands south of
Kennedy's Hole, Uinta County, Utah.

This species is based upon a beautifully preserved specimen, No. 11,198, in
the paleontological collection of Princeton University. It was collected in 1891 from
the middle Uinta, on White River, Utah, and until the discovery of the present
specimens was the only known representative of the species.

The specimens before me add but little to our knowledge of the species, but
I believe it important to give at this time their principal dimensions as compared
with the type in order to show the variations within the species.

128

MEMOIRS OF THE CARNEGIE MUSEUM

COMPAEATIVE MeASTTHEMENTS OF NeuRALS.

No.

Greatest length.

Greatest width.

Type.

No. 3270. 1 No. 2158.

No. 2397.

Type.

No. 3270.

No. 2158.

No. 2397.

1

60

59 57

45

43

47

_

2

43

55 45

42

48

56

3

50

55 61

69

42

44

54

55

4

50

47 48

59

43

44

45

52

5

34

- 45

53

47

41

48

60

6

30

.

38

42

45

59

7

20

j • . .

28

50

64

8

25

— 1 —

37

30

—

—

Comparative Measurements of Vertebrals.

No.

Greatest length.

1

Greatest width.

Type.

No. 3270.

No. 2158.

No. 2397.

Type.

No. 3270.

No. 2158.

No. 2397.

1

70

1 88

81

1 112

82

100

2

115

t 111

99

120

102

77

88

114

3

91

102 100

117

81

80

80

94

4

78

. i

110

94

.

98

120

5

86

'

—

132

—

Comparative Measurements of the Plastron.

Type.

Greatest length, plastron

Greatest width, pIa.stron

Length, anterior lobe

Width, anterior lobe at base ....

Width of lip

Length of entoplastron

Greatest width of entoplastron . .

Length of posterior lobe

Width of posterior lobe at base . .
Depth of posterior median notch .

Width of notch

Epiplastrals meet on midline . . . .
Hyoplastrals meet on midline . . .
Hypoplastrals meet on midline . .
Xipliiplastrals meet on midline . .

Gulars meet on midline

Humerals meet on midhne

Pectorals meet on midhne

Abdominals meet on midline. . . .

Femorals meet on midline

Anals meet on midline

295
120
200
65
70
95
150
200
16
55
44
125
120

73
115
70
52

390e
135
240
90

154
230
19
55
43
108
136
95
68
47
90
110
82
65

185
240

114
160e
100

8. Echmatemys douglassi sp. nov.

Plate XXII; text-figs. 11 and 12.

T3rpe: C. M. No. 3244, consisting of a somewhat damaged carapace with a

complete plastron. The carapace lacks portions of the peripheral borders of the

front and sides in addition to several small areas out of the costal and neural

regions. Collected by Earl Douglass, May 25, 1908.

GILMORE: the fossil turtles of the UINTA FORMATION

129

Locality: South Branch of Red Bluff Wash, above the well on the road between
Bonanza and Kennedy's Hole, Uinta Basin, Uinta County, Utah.

Horizon: Lower portion Horizon B, "Transition Beds" (Peterson), "in
sandstone same as No. 28,"' Uinta formation, Upper Eocene.

The carapace, although crushed over toward the left side, shows the shell to be
elongated with the median portion high and vaulted. The surface of the shell is
smooth. The peripherals behind the inguinal notches are moderately thin with
acute edges and with a tendencj- to flare upward. The sulci are narrow, but

Fig. U. Echmatemt/s dotiglassi. Carapace of the tj-pe, C. M. No. 3244. One-fourth natural size.
c.s. J, c.s. 8, costab one and eight; n. 2, n. S, neurals two and eight; sp., suprapygal; sp. 2, second suprapygal.

deeply impressed. The total length of the carapace in a straight line is about 470
mm. Its width is 300 mm., its height at the center is about 186 mm.

The nuchal scute is wedge-shaped with the narrow end forward. Some of
the anterior margin of this bone is missing so its length cannot be given. The
posterior end has a greatest width of 22 mm.

' No. 28 is C. M. catalog No. 3270 and is identified as Echmatemys uintcnsis Hny.

130 MEMOIRS OF THE CARNEGIE MUSEUM

The first neural is represented bj^ the posterior end onlj^, the second to the
sixth inclusive are complete, the seventh and eight are only partially preserved.
In general the neurals are hexagonal with their broadest ends forward. The an-
terior ends of the second to the fifth are concave. There is no indication of a
carina on any of the neurals. Their principal dimensions are given in the table
below:

No. Length. Width.
2 40 45

3 45 42

4 48 35

6 47 41

6 32 44

7 24 44e

29 —

Fig. 12. Echmatemijs dov^lassi. Plastron of tj^pe, C. M. No. 3244. One-fourth natural size

There are as usual eight costals. These vary but little in the width of their
proximal and distal ends. The fifth on the right side has a length of 144 mm.;
the first at the suture with the second a length of 123 mm.

GILMORE: the fossil turtles of the UINTA FORMATION

131

The peripherals are high. On the right side immediately posterior to the in-
guinal notch they extend upward 74 mm. above the margin of the shell; the most
posterior one 45 mm. ; the most anterior one 56 mm. above the margin.

The suprapygal is 41 mm. long. The second suprapygal is 52 mm. long, and
93 mm. across the middle, and 42 mm. where it joins the pygal. The pygal has a
length of 32 mm., a width of 55 mm. The posterior boundary of the fifth vertebral
crosses the second suprapygal 20 mm. anterior to its posterior margin. Both the
pygal and second suprapygal are strongly arched above the tail.

The vertebral scutes are long. The sides of the median three are slightly
bracket-shaped, the fourth, as in Echmatemys uintensis Hay, is strongly urn-shaped.
The dimensions of the vertebrals are given in the accompanying table.

Dimensions of Vertebr.\ls.

No.

Length.

Width in Front.

Greatest Width.

1

80e

97

97

2

93

77e

3

98

53

69

4

90

66

92

5

63

117

e, estimated.

There appear to be twelve marginals in the complete series of one side, and
some rise to the proximal ends of the peripherals. On the right side the eighth
peripheral is crossed by the marginal sulcus 21 mm. below the costo-peripheral
suture. The most posterior marginal scute is 53 mm. high from the margin of the
sheU.

The plastron is perfectly preserved and characters observed in the plastron
show the distinctness of Echmatemys douglassi from all other described forms. The
plastron has a maximum length of 407 mm. At the center it measures 374 mm. in
length. The front of the anterior lobe is broadly and deeply emarginated, the
emargination lying between two toothed projections which extend forward at either
side, as is well shown in Fig. 12. On account of this emargination there is no well-
defined lip. The anterior lobe is broad and at its base measures 190 mm. with a
length at the center of 95 mm. The length therefore is exactly 50 per cent, of the
width. The lateral borders of the lobe in front of the axillary notches are slightly
concave, then expanding a little to the posterior ends of the epiplastra, then turning
inward to the toothed projections at either side of the lip. The latter is wide,
measuring 94 mm., which is nearly one-half of the total width of the lobe.

The posterior lobe has a greatest length at the center of 130 mm.; a width at
the base of about 199 mm. The hinder lobe is considerably constricted at the anal-

132 MEMOIRS OF THE CARNEGIE MUSEUM

femoral sulcus, and from this point posteriorly the lateral borders converge quite
rapidly to the posterior end, which is deeply notched. The great breadth of the
lobes in E. douglassi leaves but little space for the exit of the limbs, and in front the
exit is still further closed by the upward curve of the anterior lobe as in E. caUopyge
Hay. Transversely the plastron is angularlj' concave, but how much of this de-
pression may be attributed to postmortem causes it is impossible to determine. It
may be largely sexual, and in that case this specimen would represent a male.

The entoplastron is pear-shaped and extends forward to within 11 mm. of
the anterior margin of the lobe, a most unusual position in the Emydidse. The
length of the entoplastron is 79 mm.; its width 71 mm. The bridge has a width of
167 mm. In this specimen, as shown in Fig. 12, there are intergular scutes, which
overlap the entoplastron. On the left side the gular-humcral sulcus follows the
usual course, reaching the margin of the lobe immediately posterior to the toothed
projection marking the external boundary of the lip. On the right side, however,
there is no trace of this sulcus. Intergulars are not known in any other member of
the Emydidse, and it may be that the scutes here designated as intergulars are the
gulars, and that the extra scute on the left side is supernumerary. Even should
that be the case, the position of the sulcus crossing the border on the mesiad side
of the toothed projection is unusual, and probably constitutes an individual vari-
ation, the true condition of which can only be cleared up by the discovery of ad-
ditional specimens pertaining to this species.

The pectorals do not reach the entop'astron, but at the center pass 6 mm.
behind it. Thej' have their greatest width of 60 mm. at the center. The humerals
meet along the midline for a distance of 65 mm.; the abdominals for 98 mm.; the
femorals for 08 mm.; and the anals for 55 mm.

The hy; plastrals are 85 mm. wide anteio-poster!orly at the midline, the left
being slightly more. Each extends laterally about 132 mm. The hypoplastrals
meet on the midline for a distance of 122 mm.; the xiphiplastrals for a distance of
83 mm. The notch between their h'nder ends is 20 mm. deep, with a greatest
width of 50 mm.

In the general shape and contour of the shell this species closely resembles
Echmatemys stevensoniana (Leidy), from the Bridger beds of Wyoming. It differs
from that species, however, in the deep emargination of the anterior lobe and the
failure of the pectoral scutes to reach the entoplastron.

Echmatemys douglassi is distinguished from all described species of the genus
by the deep emargination of the anterior median border of the anterior lobe; by the
short and wide anterior lobe, the length of which at the center is only fifty per cent.

GILMORE: the fossil turtles of the UINTA FORMATION 133

of the width at the base, and by the close proximity of the anterior end of the cnto-
plastron to the border of the Hp. In the unusual proportions of the anterior lobe
it most nearly resembles E. arethusa Hay from the Bridger beds, but is at once dis-
tinguished from it by the concave lip, as contrasted with the projecting lip of the
former species. In having the humero-pectoral sulcus pass behind the entoplastron
this species is distinguished from all other species of the genus with the exception
of E. lativertebralis (Cope), E. megaulax (Cope), and E. rivalis Hay.

This species is dedicated to Mr. Earl Douglass, who collected the type speci-
men, as well as the greater number of specimens comprised in this collection of

turtles.

9. Echmatemys hollandi sp. nov.

Plate XXIII, fig. 1; text-fig. 13.

Type: C. M. No. 3249, consisting of a considerable portion of the carapace,

lacking the posterior and the greater part of the j^eripherals and costals of the

left side and the outer halves of most of the remaining peripheral and marginal

Fig. 13. Echmakmys hollandi. Carapace of the type, C. M. No. 3249. One-third natural size,
c. 6, sixth costal; c.s. 1, supernumerary or first costal scute; n. 1, n. 6, first and sixth neural."!.

bones. A considerable part of the plastron is present but the under surface is so
badly shattered that nearly all traces of the sutures and sulci ha^e been obliterated.
The impression remaining in the matrix, however, gives some idea of the shape and
dimensions of the anterior and posterior lobes. Collected by Earl Douglass, 1908.

134

MEMOIRS OF THE CARNEGIE MUSEUM

Locality: Skull Butte, southwest of Well No. 2, Uinta Basin, Uinta County,
Utah.

Horizon: Horizon B, Uinta formation, Upper Eocene.

The carapace is elongated oval being more broadly rounded in front than
Echmateniys caUopyge Hay. The median portion is high and vaulted. The
length of the carapace is estimated to have been about 360 mm. ; its width about
260 mm. The height at the center is 133 mm. The surface of the shell is smooth.
The sulci are narrow, but well impressed.

The nuchal plate has the anterior border missing, so that its greatest length
cannot be determined. It has a greatest width of 82 mm., and where the lateral
sutures cross the costo-marginal sulcus it is 57 mm. wide.

The neurals back of the first are all hexagonal with the anterior ends concave.
Those preserved are all longer than broad. On either side of the anterior end of
the first neural are hollowed out depressions which give this bone the appearance of
being bluntly ridged anteriorly, but otherwise there is no indication of carinae.

Dimensions of Neurals.

Dimensions of Vertebral Scutes.

No.

No.

Length. | Width.

Length.

Width in Front.

Greatest Width.

1

45

30

1

65

46

55

2

40

36

2

80

50

77

3

41

35

3

87

40

72

4

38

35

4

49

73

5

38

36

6

36

The peripherals appear to have been high, but on account of the damaged
condition of the borders it is not possible to give their extent.

Six costals are present in this specimen, the first having a greatest antero-
posterior diameter of 63 mm.; a greatest length of 96 mm. On the upper anterior
half of the first costal there is a low, rounded, obtuse elevation or horn-like pro-
jection, which at once distinguishes this species from all other described forms. The
other costals show nothing unusual. The third has a greatest length of 106 mm.
The costo-peripheral suture between the third and fourth passes about 17 mm.
mesiad of the costo-marginal sulcus.

The first vertebral scute is unusually narrow, in this respect resembling
E. callopyge Hay. The sides of those posterior to the first are strongly urn-shaped.
The anterior end of the third is especially narrow and pointed, and extends well
forward into the second. The dimensions of the vertebrals are given in the table.
On either side of the first vertebral are supcrnumerarj^ costal scutes, and it is
largely within their boundaries that the horn-like elevations, described above, arise.

GILMORE: the fossil turtles of the UINTA FORMATION 135

The area of the first costal scute is much reduced, but it is still in contact with the
first vertebral. Supernumerary costals are not unusual in the Baenidse, though
I am not aware of their having been found before in the genus Echmatemys. When
present, they are usually confined to one side, seldom are thej' sj^mmetricalh^ paired
as in the present specimen.

The plastron is exceedingly thick and heavy. At the center it measures 30
mm. in thickness. Though much of the anterior and posterior lobes are missing
the impressions remaining in the matrix show the plastron to have had a greatest
length of about 315 mm.

The entoplastron though only partiallj^ preserved, has a greatest width of 50
mm. Its length cannot be determined.

The bridge has a width of about 145 mm.

There is no suggestion of a notch in the posterior lobe, shown by the im-
pression in the matrix, but it is not possible to state positively that such did not
exist. Judging from the impression left by the anterior lobe the lip was thick and
broad, with an abrupt depression on the dorsal surface some 35 mm. posterior to
the anterior border. The anterior end of this lobe was probably within the forward
end of the carapace.

The pair of horn-like protuberances on the front of the carapace, the presence
of a pair of supernumerary costal scutes on either side of the first vertebral serve to
distinguish this specimen from all other described species of the genus, and I
therefore take great pleasure in naming it Echmatemys hollandi for Dr. W'illiam J.
Holland, Director of the Carnegie Museum, in recognition of his activities in the
field of vertebrate paleontology.

10. Echmatemys obscura sp. nov.
Plate XXIV; text-figs. 14 and 15.

Type: C. M. No. 3252, consisting of a carapace, lacking the posterior end
back of the sixth neural; and the plastron, lacking the lip and a small portion of
the extremity of the posterior lobe. Collected by Earl Douglass, August 17, 1908.

Locality: Devil's Play Ground, south of Kennedy's Hole, Uinta Basin, Uinta
County, Utah.

Horizon: Horizon C, "gray beds below red and gray beds," Uinta formation,
Upper Eocene.

Except for the parts missing from the posterior end of the carapace, the type
specimen is well preserved and all of the sutures and sulci are clearly displayed.
The carapace is broadly rounded in front with a wide, but shallow, emargination of

136 MEMOIRS OF THE CARNECilE MUSEUM

the nuchal border. The median portion is high and vaulted. The length of the
carapace is estimated to have been about 390 mm., its breadth is 286 mm., its height
at the center is 130 mm. The plastron is broad, and, like many other species of the
genus Echmatemys, there is but little space in front of the axillary notches for the
exit of the limbs. The margins of the carapace forward of the axillary notches
are thickened and rounded, but forward it thins rapidly, coming to an obtuse edge
along the median anterior border. The peripherals of this region do not flare up-
ward, though those immediately posterior to the inguinal notches show a tendency
to do so.

The carapace is smooth, except that the median costal areas are crossed antero-
posteriorly by a series of wide, parallel, wavy, flattened ridges. In a line 12 mm.
long three of these ridges may be counted. This ornamentation is inconspicuous
unless the light strikes the surfaces at the proper angle. It is the obscure nature of
this ornamentation which has suggested the specific name.

The nuchal i^late is unusually long and narrow. At the point of its greatest
transverse diameter the sides are considerably within the boundaries of the first
vertebral. The length of the nuchal is 65 mm.; its greatest transverse diameter
is 51 mm. ; the free border measures 40 mm. The nuchal surface is without median
elevation. Echmatemys obscura is the only species of the genus which has the
nuchal plate longer than wide.

All of the neurals posterior to the first are broadly hexagonal, and all posterior
to the first are broader than long. The second and third have the anterior end
concave, those posterior being straight. Their principal measurements are given
in the accompanying table.

Dimensions of Neurals.

No. Length. Width.

1 47 35

2 38 40

3 43 43e

4 39 47

5 31 41

6 — 44

e, estimated.

The costals posterior to the first alternate in having the distal ends slightly
wider and narrower than the proximal ends.

The peripherals are moderately high, the first extending inward from the
border 50 mm., the seventh 58 mm.; the eighth 55 mm. The border above the
bridge is heavy and rounded, but posterior to the inguinal and anterior to the

GILMORE: the fossil turtles of the UINTA FOR^L\TIO^'

137

axillary notches the borders thin rapidlj^ toward the center. The costoperipheral
sutures pass along the median sides of the shell on an average of about 17 mm.
mesiad of the marginal sulcus. Beyond the bridges the sutures and sulci in some

Fig. 14. Eckmatemys obscura. Carapace of tj-pe, C. M. Xo. 3252. One-fourth natural size. Ji. /,
n. 6, neurals one and six; n.p., nuchal plate; v.s. 4, fourth vertebral scute; c. 6, sixth costal.

places approach one another as close as 4 nmi., and in other places are distant as
much as 30 mm.

The nuchal scute is narrow, measuring 8 mm. on the free border. It is 15 mm.
long antero-posteriorly.

The vertebral scutes are wider than long, the second being especially wide.
The sides of those back of the first are bracket-shaped. Their dimensions are
given in the accompanying table.

Dimensions of Vertebil\ls

No.

Length.

Width in Front.

Greatest Width.

1

75

87

87

2

83

58

102

3

79

60

81

4

—

58

The costo-marginal sulci run below the costo-periphcral sutures. The sulci
on both carapace and plastron are narrow and moderately impressed.

The plastron in life had a length of about 372 mm. The posterior end is

138

MEMOIRS OF THE CARNEGIE MUSEUM

narrow, but deeply notched. The length of the anterior lobe is about 100 mm.;
its width at the base is about 162 mm. The length is therefore onh' 61 per cent,
of the width. From the axillarj^ notch the free border runs straight forward for a
short distance then curves in regularly to the epiplastral lip, which is missing in

Fig. 15. Echmatcmys obscura. Tj-pe, C. IM. No. 3252. Plastron. One-fourth natural size.

this specimen. This free border is relatively thin and acute, thickening somewhat
as it approaches the lip. The lip appears to have been about 50 mm. wide.

The entoplastron has a length of about 60 mm. and a width of 55 mm. It is
crossed by the humero-pectoral sulcus, and is also overlapped by the gular scutes.
It is pear-shaped in outline, in this respect resembling the entoplastron of Ech-
matcmys douglassi. The bridge is 150 mm. in width.

The length of the posterior lobe is 120 mm. Its width at the base is 191 mm.
The length is therefore 62 per cent, of the width. The free borders are slightly
contracted at the femoral-abdominal sulcus, and again at the femoral-anal sulcus;
from which point the border turns in rapidly toward the center. The posterior
lobe is terminated posteriorly by rather sharp projecting points on either side of
the narrow, but rather deep, median notch. It is estimated that the notch had a
depth of 30 mm. The posterior borders of the lobe are acute. At the center of
the notch the bone has a thickness of 5 mm.

The anterior lobe curves upward toward the carapace much as in Echmatemys

GILMORE: the fossil turtles of the UINTA FORMATION 139

septaria. The plastron is flat, but in an uncrushed specimen the bridges would
doubtless curve upward to the margin of the shell. In this specimen thej' are but
little above the level of the plastron.

The hyoplastrals meet on the midline for a distance of 80 mm.; the hj-poplas-
trals for 97 mm.; the xiphiplastrals for 66 mm.

The gular scutes along the midhne are about 52 mm. long; the humerals 37
mm.; the pectorals 63 mm.; the abdominals 101 mm.; the femorals 48 mm.; and
the anals 53 mm.

Echmatemys obscura is distinguished from all species of the genus, in which
the nuchal region is known, by the extreme narrowness of the nuchal plate, it
being the only species known, in which the nuchal plate is longer than wide.
From Echmatemys cibollensis, E. megaulax, and E. euthneta (the nuchal region of all
three being unknown) the present species is distinguished: from the former by
having the gulars overlapping the entoplastron ; and from the latter two by having
the gular-humeral sulcus crossing the rear portion of the entoplastron. This
species is further distinguished by the greater relative widths of the neurals and
especially the vertebrals. The obscure, but characteristic, ornamentation of the
costal region of the carapace will also aid in recognizing this species.

11. Echmatemys depressa sp. no v.
Plate XXIII, fig. 2; text-fig. 16.

Type: C. M. No 2936, consisting of the carapace, lacking much of the anterior
margin, the peripherals, and outer halves of the costals of the left side, and most
of the peripherals posterior to the inguinal notch of the right side. The plastron
is represented by a few fragments only, though the impression in the matrix gives
some idea of its proportions. Collected by O. A. Peterson, August 5, 1912.

Locality: Six miles east of Myton, Uinta County, Utah.

Horizon: Horizon C, Uinta formation. Upper Eocene.

Although the open sutures of the type specimen give evidence of the im-
maturity of the individual, it appears to represent one of the smaller species of the
genus Echmatemys. I was first inclined to regard it as referable to the genus
Palceotheca on account of its small size and the presence of a dorsal keel, but a
comparison with the types of the two species pertaining to that genus {Palwotlieca
terrestris Cope, and P. polycypha Cope) both of which are in the U. S. National
Museum, shows differences which lead me to believe that it can with greater
propriety be referred to the genus Echmatemys. The aj^parent absence of a second
suprapygal and the extremely wide vertebrals may with other characters to be
observed in a more perfect specimen show its distinctness from that genus.

140 MEMOIRS OF THE CARNEGIE MUSEUM

In a straiji;ht lino the shell has a greatest estimated length of 135 mm.; at the
center a greatest width of 115 mm. Though depressed, the upper shell is broadly
convex in all directions, dropping off rather more rapidlj' toward the back than
toward the front. The upper surface of the carapace is smooth, the sulci lightly
impressed, and nowhere are scutal growth lines to be observed.

Fig. IG. Carapace of Eclimalenujs depressa. Type. C. M. No. 2936. One-half natural size, c.p.',
c.p. S, costal plates one and eight; n2, n7, neurals two and seven; nu.p., portion of nuchal plate; pij, pygal;
spy, suprapygal; v.s. 2, second vertebral scute.

The nuchal is only partially preserved, but this portion shows that it had a
greatest width of 26 mm. and was obtusely keeled at the center. The anterior
border, as shown in Fig. 16, is missing.

There are eight neurals, all of which are hexagonal, with the exception of the
first and eighth, the latter being subrectangular in outline. All have their antero-
lateral angles truncated, which serves at once to distinguish this form from Ech-
matemys pusilla Hay, which has the postero-lateral angles of the neurals truncated.
The neurals graduallj' decrease in size from front to back, and, excepting the first
and third, all are broader than long, as shown in the accompanying table. The
second and third neurals are sharply keeled on their posterior and anterior ends
respectively, as are the fourth and fifth, while the sixth, seventh and eighth are
keeled their entire lengths. The keel on the suprapygal is ver_y low and hardly
discernible.

All of the costals of the right side are present and perfectly preserved. They
are of moderate thickness with pointed distal ends which articulated with the periph-
erals by gomphosis. Portions of the buttresses preserved in the matrix indicate
that they articulate with the costals considerably above the costo-peripheral suture.

\

GILMORE: the fossil turtles of the UINTA FORMATION 141

Dimensions of Xeurals.

No. Length. Width.

1 17 12

2 13 16

3 14.5 14

4 13.5 14

5 13.5 14

6 11 14

7 10 12

8 9 10

The peripherals of the right side above the bridge and somewhat forward of
the axillarj'^ notch are perfectly preserved, and, as in Echmatemys pusilla Hay,
have a sharp carina beginning on the third peripheral and continuing backward
across the bridge to the hinder peripherals. The fifth peripheral has a width of
19 mm.; the sixth of 18 mm.; the seventh of 20 mm.; the eighth of 18 mm. The
lateral peripherals from the edge of the carapace to their proximal extremities have
a length of 14 mm., becoming narrower tow^ard the front of the shell.

The sulci on most parts of the carapace are very obscure, being traceable
only here and there, though w'here they cross the neurals somew^hat plainer than
elsewhere. The boundaries of the second vertebral can be partiallj^ determined,
and these indicate a very wide scute having at the center angularly pointed outer
borders. The second vertebral has a greatest width at the center of 58 mm.; an
estimated length of about 32 mm. There were four costal scutes. The costal-
marginal sulcus appears to have followed closely the course of the costo-peripheral
suture. The supracaudal scute is divided. The second suprapj'gal is absent in
this specimen.

The impression in the matrix shows the hypoplastron to have a greatest width
at the midline of 41 mm. The width of the posterior lobe at the base is about
60 mm. Its greatest length was about 47 mm. It cannot be determined whether
this lobe w^as notched on the midline. At the center the plastron has a greatest
width of 84 mm. The bridge has a wudth of about 58 mm. The inguinal buttresses
rise well above the costo-pcripheral sutures and articulate w-ith both the fifth and
sixth costals.

Echmatemys depressa is distinguished from all other species of the genus by
the greater relative widths of the vertebral scutes, the absence of a second supra-
pygal, and the presence of a dorsal keel. From Echmatemys megaulax (Cope),
which also has a dorsal keel the present species is to be distinguished by having
the sulci less deeply impressed and in having the costo-marginal sulcus follow the

142 MEMOIRS OF THE CARNEGIE MUSEUM

course of the costo-peripheral suture wliereas in the former it crosses the peripherals
on their upper third.

12. Echmatemys pusilla? Haj'.
Echmatemys pusilla Hay, Fossil Turtles of North America, 1908, pp. 337-339,

text-figs. 445, 446.

A small turtle, C. M. Xo. 3282, collected by Messrs. Earl Douglass and
Clarence Wilson, November 1, 1911, southeast of Ouray, Uinta Count}', Utah,
from Horizon C of the Uinta formation. Upper Eocene, is referred with some doubt
to Echmatemys pusilla HaJ^ The very fragmentary nature of the present specimen
renders its generic and specific affinities difficult of positive determination, but after
a careful comparison of this specimen with the type of E. pusilla, kindly loaned me
by Dr. W. D. Matthew, of the American Museum of Natural History, I am con-
vinced of the very close relationships of the two specimens, even though the dis-
covery of more perfect material may eventuallj'^ demonstrate their specific distinct-
ness.

/ 9

Fig. 17. Portions of the carapace and plastron of Echmatcmijs pusilla? Hay. C. jM. No. 32S2. (1),
plastron; (2), carapace. c2, c3, costals two and four; n2, n6, neurals two and five; v2, vertebral scute two.
Natural size.

This specimen consists of the anterior lobe of the plastron lacking the lip, a
small portion of the carapace consisting of the second neural complete, and portions
of the third, fourth, and fifth neurals with the upper portions of abutting costals,
as shown in Fig. 17. The second neural is hexagonal in outline and measures 9.5
mm. in length, with a greatest width of 10 mm.; the third is 12 mm. long and 13
mm. wide; the fourth is 13 mm. long. The second vertebral has its greatest width
of about 35 mm. at the center. The scutal areas of the carapace arc plainly grooved
by the lines of gro\\i:h of the scutes. These lines of growth are also present in the
type of the species.

GILMORE: the fossil turtles of the UINTA FORMATION 143

The entoplastron, as in the type, is long, narrow, and pointed in front. Its
greatest length is 19 mm., its greatest width 16 mm. It is overlapped by both the
gulars and pectorals. The specimen before me also agrees with the type in the
great width of the pectorals behind the entoplastron. These scutes reach backward
to the hypoplastral suture, a condition not known in any other species of the genus.
At the point where the pectoro-humeral sulcus crosses the free borders the lobe
has a width of 48 mm. The free border of the lobe is thin and acute, being bevelled
off on the upper surface.

The most important dissimilarity between the two specimens here discussed
appears to be in the wide vertebrals and in the shape of the neurals, those of the
type having the posterolateral angles truncated, whereas in the specimen from
the Uinta formation the antero-lateral angles of the third, fourth, and fifth are
thus cut off. The second neural is octagonal in the type, hexagonal in No. 3282.
In the shape of the anterior lobe of the plastron, the long pointed entoplastron
overlapped by the gulars and crossed well forward by the pectoro-humeral sulcus
and the extremely wide pectorals reaching backward nearly to the hyo-hypo-
plastral suture these specimens show a remarkably close resemblance.

Family TESTUDINID.E Gray.
The family Testudinidse is represented in the collection of cheloiiian remains
from the Uinta formation in the Carnegie Museum by the two genera Hadrianus
Cope and Testudo Linnaeus, the former genus by three, the latter by but one species.
Four species of Hadrianus are now recognized as occurring in the Uinta formation.
The discovery of Testudo in the Upper Eocene is of interest as being the first time
this genus has been found below the Oligocene in North America. In the Fayum
deposits (Upper Eocene) of northern Africa, however, the genus Testudo has been
recognized by Andrews from well-preserved specimens, which in several respects
closely resemble the species here described.

Genus Hadrianus Cope. "^

13. Hadrianus corsoni (Leidy).
Plate XXV, fig. 1; text-fig. 18.
Testudo corsoni Leidy, Proc. Acad. Nat. Sci. Phila., 1871, p. 154; Contrib. Extinct
Vert. Fauna West. Terrs., 1873, pp. 132, 339, PI. XI, figs. 1, 2; PI. XV, fig. 7;
PI. XXIX, figs. 2-4; PI. XXX, figs. 1-4.
Hadrianus odonarius Cope, Pateont. Bull. No. 2, 1872; Vert. Tort. Form. West.,
1884, p. 140, PI. XX, figs. 1-4.

144 MEMOIRS OF THE CARNEGIE MUSEUM

Hadrianus corsoni Cope, U. S. Geol. Surv. Terrs., 6th Ann. Kept., 1872 (1873),

p. 631; Hay, Bibliog. and Cat. Foss. Vert. N. A., 1902, p. 450; Fossil Turtles

of North America, 1908, pp. 376-380, PI. LX, LXI; text-figs. 473-479.

A large specimen in the present collection is identified as belonging to the
above genus and species, and represents the first recorded occurrence of Hadrianus
corsoni in the Uinta formation. This speciinen consists of a complete plastron,
the peripherals of the right side above the bridge, and a few fragments of the cara-
pace. It was collected by 0. A. Peterson, August 5, 1912, six miles east of Myton,
Uinta County, Utah, from Horizon C of the Uinta formation, Upper Eocene. It
bears the C. M. Catalog No. 3403.

The length of the plastron at the center is 740 mm.; the greatest length over
all 775 mm.; the greatest breadth about 490 mm. The posterior lobe is deeply
and widely notched and the anterior lobe is terminated in front by a wide spade-
like lip which projects prominently from the general contour of the lobe.

Hadrianus corsoni was based upon the anterior portion of a plastron consisting
of the complete lip and lobe back to and including a small portion of the anterior
end of the entoplastron. A comparison of the specimen before me with the type
shows striking similarities in the contour of the lobe and lip, and especially of the
wide anterior end of the entoplastron, which appears peculiar to this species. The
lip has a transverse width at the base of 162 mm., which is greater than that of the
type, or of any subsequently discovered representative of this species. It projects
45 mm. beyond the point where the gular-humeral sulcus crosses the free border.
This measurement is slightly greater than in the type, but less than in the type of
Hadrianus octonarius Cope, now regarded by Hay as being a synonym of the
present species. The anterior border of the lip is subacute, nearly straight, but
slightly notched on the midline as in the type. The upper surface of the lip is
slightly convex along the midline, with shallow longitudinal depressions on either
side. The lower surface is flat, but I am inclined to believe it would have been
broadly convex in life.

The anterior lobe is 345 mm. wide at the base, with a length of 252 mm. Along
the free borders on the upper surface the bone is bevelled off to an acute edge that
becomes obtusely rounded in front of the axillary notches. The lip on the superior
surface extends backward 67 mm.

The entoplastron is unusually broad and angular in front. In the figures of
the type the suture limiting the anterior border of the entoplastron runs nearly
straight across the median line, then turns abruptly backward and outward. In
the specimen here considered the anterior border is inclined more posteriorly, as

GILMORE: the fossil turtles of the UINTA FORMATION

145

in the type of H. odonarius, but it has the same angular turn backward and out-
ward as in the type of the genus and species. The length of the entoplastron on the

Fig. 18. Hadrianus corsoni (Leidy). Plastron of C. ]M. Xo. 3403. One-eighth natural size.

midline is 155 mm.; its width 184 mm. It is overlapped slightly by the gular
scutes, but is not crossed by the humero-pectoral sulcus.

The bridge is 290 mm. wide.

The length of the posterior lobe is 234 mm.; its width at the inguinal notches
386 mm. The free edges of the lobe are subacute, except in front of the notches,
where it is rounded. The width of the posterior notch is about 135 mm.; its depth
about 35 mm. There is no perceptible ridge leading back from the inguinal but-
tresses until a point midway between the buttress and the posterior end of the lobe
is reached, where the surface is concave mesiad, thus causing the ridge to stand
out quite prominently. On the outside of this ridge the surface is bevelled off
steeply near the buttress, but less and less so posteriorly, until on the posterior
outer angle of the xiphiplastral projections the slope is very gentle. The surfaces
of these projections are rugosely roughened as in the type of H. odonarius (No.
2186, U. S. National Museum), which they resemble in shape and size.

The hyoplastron on the midline has a width of 195 mm. The arrangement of

146

MEMOIRS OF THE CARNEGIE MUSEUM

the plastral scutes is not greatly different from that of other Testudinidse. The
gulars have a length on the midline of 90 mm.; pectorals of 60 mm.; abdominals of
255 mm.; femorals of 95 mm.; anals of 86 mm. The principal dimensions and
proportions of specimen C. M. No. 3403, as compared with other specimens iden-
tified as pertaining to Hadrianus corsoni Leidy, are given in the table below.

Hadrianus cor-

Type of JT. octo-

C. M. No. 3403.

soni A. M.S. B..

narius U. S.

No. 6027.

N. M. No. 2186.

778

550

740

252

175

247

345

246

360

290

215

285

234

157

240

386

240

330

155

115
134

125

184

160

162

118

135

81%

87%

79%

72%

71%

68%

80%

73%

84%

Greatest length of plastron

Greatest length anterior lobe

Greatest width anterior lobe

Width of bridge

Greatest length posterior lobe

Greatest width posterior lobe

Length of entoplastron

Width of entoplastron

Width of Up

Width of bridge to width of anterior lobe . .

Width of anterior lobe to its length

Width of bridge to length of posterior lobe .

14. Hadrianus robustus sp. nov.
Plate XXV, fig. 2; text-fig. 19.

Type: C M. No. 3342, consisting of the anterior half of the plastron, collected
by Earl Douglass, July 30, 1908.

Locality: Near Kennedy's Hole, Uinta Basin, Uinta County, Utah.

Horizon: Horizon C, Uinta formation. Upper Eocene.

The specimen upon which the present species is based represents one of the
larger species of the genus. Those parts present are in a good state of preservation,
and all of the sutures and sulci can be clearly traced. It is assigned provisionally
to the genus Hadrianus, until such time as the discovery of a more perfect specimen
makes it possible to determine its true generic affinities.

The anterior lobe is 220 mm. long, and at the base 285 mm. wide. The length
thus being 77 per cent, of the width, whereas in Hadrianus corsoni it is only 71 and
in H. majusculus 68 per cent. The lip in this species is especially prominent and
exceedingly heavy, having a thickness at the center of 46 mm. At the base it
measures 123 mm. in width, and near the anterior end 117 mm. From the point
where the gular-humeral sulcus crosses the free border the lip extends forward 52
mm. The anterior border of the lip is transversely broadly convex, and bluntly
bevelled dorso-ventrally, the longer bevel being on the lower side. On the superior
surface the lip extends posteriorly 100 mm., at this point the surface descends
perpendicularly, decreasing by one-half the total thickness of the plastron. Slightly

GILMORE: the fossil turtles of the riNTA FORMATION

147

posterior to this drop on the visceral surface there is a sharp median ridge on the
midhne which measures longitudinally 65 mm. The free borders of the anterior
lobe are bevelled off from the superior surface to a sharp edge, which becomes
obtusel}^ rounded at the base of the lip.

Fig. 19. Hadrianus robmtus.
tural size.

Anterior half of the plastron. Tj^je, C. M. No. 3342. One-sLxth na-

The entoplastron is pointed in front and wide behind. It is 134 mm. long,
and 147 mm. wide. The gular scutes overlap the anterior end.

In front of the axillary notch the plastron has a thickness of 29 mm.; im-
mediately posterior to the entoplastron of 16 mm.; and at the junction of the
hyo-hypoplastron at the center, of 15 mm. This latter measurement in the type
of Hadrianus tumidus Hay is only 9 mm. The gular scutes have their greatest
extent of 114 mm. antero-posteriorty. Within the area of the gular scutes the
surface of the lip is swollen out somewhat below the level of the epiplastral areas
bordering it. The superior surface of the lip anteriorly is broadly convex, but
flattens toward the posterior end.

The humerals meet on the midline for 105 mm.; the pectorals for 40 mm.; the
abdominals for at least 205 mm. The sulci are broad and deeply impressed. The
humero-pectoral sulcus begins just in front of the axillary notch and extends inward
and backward for a short distance, then curves forward to the entoplastron, again
turning backward to the midline, skirting, but not crossing, the entoplastron.

The pectoral scutes have quite a different shape from any of the described
species of the genus. They are narrow (28 mm.) mcsiad of the axillary notch

148 MEMOIRS OF THE CARNEGIE MUSEUM

gradually widening to a jjoint a little bej'ond where the sulcus first reaches the
cntoplastron, which measures 63 mm. antero-postcriorly, then again narrowing
to the midline, where it has a fore-and-aft extent of 40 mm. The pectoro-abdominal
sulcus runs straight across the plastron as in Hadrianus corsoni.

The hyo-hypoplastral suture runs a very tortuous course across the plastron.
The hyoplastrals meet on the midline for a distance of 136 mm. The hypoplastrals
at the center have a transverse width of 400 mm.

Tlie pectoral scutes on the midline are less than one-fifth the width of the ab-
dominals, and following Hay's analysis of the various species of the genus this
proportion would be sufficient to show the specific distinctness of the present speci-
men. In the proportions of these scutes the present species is nearest Hadrianus
tumidus Hay, which is also from the Uinta formation, and in which the pectoral
scutes are less than one-third as wide as the abdominals at the center, but it is
distinguished from that species by having a thicker plastron and by the different
form of the pectoral scutes, and the greater width posteriorly of the entoplastron.
From H. majusculus and H. corsoni the difference in the proportions of the length
to the breadth of the anterior lobe will help to separate the present form; that
is to say in Hadrianus robustus the length is 77 per cent, of the width at the base,
while it is 68 and 71 per cent, respectively in the other two species mentioned above.

15. Hadrianus utahensis sp. nov.
Plate XXVI, fig. 1; text-fig. 20.

Type: C. M. No. 2343, consisting of the plastron and portions of the periph-
erals above the bridge on the right side. The anterior portion of the lip, and
parts of the margins of both anterior and posterior lobes are missing. Collected
by Earl Douglass, July 30, 1908.

Locality : South of Kennedy's Hole, Uinta Basin, Uinta County, Utah.

Horizon: Horizon B or C, Uinta formation, Upper Eocene.

The plastron of the type specimen is estimated to have had a greatest length
of about 520 mm., and a greatest width at the center of 320 mm. The anterior
lobe is about 175 mm. long and 300 mm. wide at the base. The width of the lip,
where the gular sulci cross the free border, is 110 mm. The plastron is quite
concave, indicating that the specimen was in all probability a male.

While every thing indicates that the lip extended well forward, it probably
continued the general contour of the lobe. The free borders of the lobe are sub-
acute, being bevelled off from the upper surface, so that the edge is nearly on a
level with the ventral surface of the plastron. At the center on the broken border

gilmore: the fossil turtles of the uixta formation

149

the lip is 22 mm. thick, but posteriorly it increases to 27 mm., behind which the
plastron is deeply excavated. On the upper surface of the lip there is a slight
median elevation. Transversely the whole lip is broadly convex rounding down
at the sides to a subacute edge.

Fig. 20. Plastron of Hadriamts nlahensis. Type. C. M. No. 2343. One-fourth natural size.

The entoplastron is rhombic in form, being 97 mm. long and 95 mm. wide.
The gular scutes reach, but do not overlap, this bone. In this respect this specimen
differs from all of the described species of the genus. The bridge has a greatest
width of 210 mm.

The posterior lobe has a greatest length of about 140 mm.; a greatest width
of 230 mm. at the inguinal notches. The posterior lobe is shallowly notched on
the midline. This notch has a greatest depth of 10 mm. Immediately behind the
inguinal notches the borders of the lobe have a thickness of 27 mm. and are rounded
dorso-ventrally. The border at the anal sulcus is 18 mm. thick. The free border
of the entire lobe a short distance posterior to the inguinal notches is bevelled off

150 MEMOIRS OF THE CARNEGIE MUSEUM

from the dorsal surface to a subacute edge. There is no ridge leading back from
the buttresses as in H. tumid us Hay. In front of the notch on the dorsal surface
the plastron is transversely shallowly concave. The plastral buttresses are rela-
tively heavy.

The hyoplastrals meet on the midline for a distance of 108 mm.; hypoplastrala
for 120 mm.; xiphiplastrals for 107 mm.

The sulci on the plastron are relatively narrow but well impressed. The
gular sulci run forward and outward from the center, but, as they approach the
border, suddenly turn outward and then backward continuing in this direction
over the border upon the dorsal surface and fading out at the base of the lip. The
sulcus limiting the humerals behind runs backward and slighth^ inward from the
border in front of the axillary notch for a short distance, then turns inward and
forward to the posterior boundary of the entoplastron, but does not cross it. The
pectorals occupy 41 mm. of the midline, and have a least diameter antero-poste-
riorly of 29 mm. The abdominals meet on the midline for a distance of 155 mm.;
the femorals for 75 mm.; anals for 58 mm. The plastron at the center has a
thickness of only 7 mm.

The peripherals turn abrupth' upwards at the sides of the shell. The sulci
between the plastron and peripheral bones on the bridge, although broadly inter-
digitative, run quite a straight course antero-posteriorly. The fifth and sixth
peripherals have a width of 64 mm. Their length on account of the missing upper
extremities cannot be determined.

The present specimen is assigned provisionallj' to the genus Hadrianus,
though later it may be found, when more perfect specimens are available, that it
belongs to Testudo. It is distinguished from the described species of the genus
bj' the narrow and relatively shallow notch on the posterior lobe; the rounded and
thickened ends of the xiphiplastrals, and by having the bridge longer relative to
the length of the hinder lobe. In Hadriaiius majuscidus the posterior lobe is 85
per cent, of the length of the bridge; in H. tumidus 77 per cent.; in H. corsoni
73 per cent.; and in the present specimen only 66 per cent.

Genus Testudo Linnaeus.

16. Testudo uintensis sp. nov.

Plate XXVII; text-figs. 21 and 22.

Type: C. M. No. 2331, consisting of a carapace and plastron, the latter almost
perfectly preserved, the former lacking the peripherals of the right side, and the

gilmore: the fossil turtles of the uinta formation

151

anterior portion of the nuchal plate. Collected by Earl Douglass and party,
July 30, 1908.

Locality: South of Kennedy's Hole, and about one hundred rods west of
Dragon- Vernal road, Uinta Basin, Uinta County, Utah.

Horizon: Horizon B or C, Uinta formation, Upper Eocene.

In form the carapace of this tortoise is broad, of moderate height and strongly
arched in all directions. As now preserved the shell is somewhat more flattened
than it would have been in life. The posterior end is not so broad as in Hadrianus
but is more evenly rounded as in many species of Testudo. The areas covered by
the vertebral scutes are decidedly convex. The few anterior peripherals present
suggest that the front was Uttle, if at all, emarginated on the median line. All of
the sutures remain distinct and the sulci can be clearly traced, so there can be no
question raised as to their proper interpretation in the figures. The carapace has
a length of about 360 mm. ; a breadth of 300 mm.

The neurals and costals are highly differentiated. In this species there are
only seven neurals, as in Testudo ammon Andrews, but whether this represents a
constant character in this species or only an individual variation, as in T. ammon,
must await the discovery of additional specimens. The first neural is especially
elongated and oval; the second and sixth octagonal; the third tetragonal; the fourth,
fifth, and seventh being hexagonal. All of the neurals are longer than wide, whereas
in Hadrianus, Stylemys and most of the species of Testudo the neurals are wider
than long. The dimensions of the neurals and costals are given in the accompanj'-
ing table.

No.

Dimensions of Neurals.

No.

Dimensions of Costals.

Length.

Width.

Width of Proximal End. , Width of Distal End.

1
2
3
4
5
6
7

44
39
33
31
34
39
31

23
30
25
29
28
32
24

1
2
3
4
5
6
7
8

44
28
43
30

iS

68
47
30
60
14
49
22
35

There are two suprapygals, the anterior being bifurcate and enclosing between
its right and left limbs the lozenge-shaped second suprapygal. The first has a
diameter antero-posteriorly at the center of 32 mm., a breadth of 88 mm.; the
second is 30 mm. in length and 48 mm. in width.

The pygal is wedge-shaped, the narrower truncated end being posterior, the
anterior end is notched for the second suprapygal. The under side is transversely

152

MEMOIRS OF THE CARNEGIE MUSEUM

concave, the upper slightly convex in the same direction. The free edge is acute
with a faint median projection, the whole forming a convex covering for the tail
that apparently projected but little below the level of the carapace.

Fig. 21. Tesludo uintensis. Carapace of the type, C. M. No. 2331. One-fourth natural size, n.p.,
nuchal plate; nlowdn/, first andseventhneurals;pZos., lip of plastron; sp.,suprapygal;sp^., second suprapygal.

The costal plates are alternately wide and narrow, reaching as high a degree
of differentiation in this respect as any species of the genus. The diameters of
their proximal and distal ends are given in the accompanying table.

The nuchal plate is largely missing, as well as nearly all of the peripherals of
the right side. Those of the left side have suffered some damage, so that their
exact dimensions can not always be determined. They are however of moderate
length with thin acute edges on front and back, becoming slightly obtuse along
the sides. There are eleven peripherals in the complete series. The greatest
thickness of the ninth peripheral, the thickest of any of the posterior members, is
16 mm. The second, measured at the suture with the first, is 16 mm. thick.

The plastron has a total length of 332 mm. The width of the anterior lobe
at its base is 150 mm. It has a length of 104 mm. The lip does not extend beyond
the border of the carapace, and it projects but little beyond the general contour
of the lobe. The lower surface of the lobe is flat, while the upper is bevelled off
from back toward the front, forming rather a sharp anterior border. On the upper

GILMORE: the fossil turtles of the UINTA FORMATION

153

surface of the lip on either side of the low median ridge there are shallow longitudinal
depressions, which run forward to the slight emarginations at either side of the
center. At the gular-humeral sulcus the lateral border of the lobe is shghtly
emarginate. The upper surface, 57 mm. back from the front of the lip, is almost
perpendicularlj' excavated, thus reducing the plastron from 21 mm. to 10 mm. in
thickness. The posterior lobe is 90 mm. long; and is 163 mm. wide at the base.
The posterior notch is 50 mm. wide and 23 mm. deep. At the inguinal notch the

Fig. 22. Testudo uintensis. Plastron of tj^e, C. M. No. 2331. One-fourth natural size.

edge of the lobe forms a wall 22 mm. high. This wall diminishes in height poste-
riorly, so that at the anal-femoral sulcus it measures 15 mm. in height. The outer
face of this wall slopes off rather gradually to the subacute lateral edge, while the
inner edge slopes abruptl}' to the inner level of the floor of the carapace. The ento-
plastron is rhombic in form, 68 mm. long on the midline, and 69 mm. wide. The
hj^oplastra meet along the midline for a distance of 74 mm.; the hypoplastra for
69 mm.; and the xiphiplastra for 55 mm.

The bridge is 135 mm. wide. The plastron is quite concave and at least
suggests that this individual was a male.

The gular scutes overlap the entoplastron and on the midline have a length
of 57 mm. The humerals at the middle are 53 mm. long; the pectorals 46 mm.;
abdominals 80 mm.; femorals 42 mm.; and anals 28 mm.

154 MEMOIRS OF THE CARNEGIE MUSEUM

The vertebral scutes are longer than wide, their dimensions are given in the
table.

Dimensions of Vertebr.\l Scutes.

No. Length. W'idtli.

1 64 100

2 74 62

3 66 63

4 68 57

5 66 116

The costal scutes are wide, their outer ends joining the marginals at the costo-
peripheral suture. The costo-marginal sulcus appears to follow closely the course
of the costo-peripheral suture. The humcro-pectoral sulcus touches the ento-
plastron, but does not cross it.

The supracaudal scute is divided as in Hadrianus.

A comparison of the type of the present species with the upper Eocene tortoise,
Testudo ammon Andrews, from the Fayum deposits of Egypt, shows some striking
resemblances. Both Testudo uintensis and T. ammon are distinguished from all
other species of the genus by the octagonal shape of the second and sixth neurals,
whereas the usual arrangement is for the second and fourth neurals to be octagonal.
Both types agree in having seven neurals, although other specimens referred to
T. ammon by Andrews have the normal number of eight, and it may be found that
there is a similar variation in the present species.

Hay^ comments upon Testudo ammon as follows: "Dr. A.[C]E.[W] Andrews
(Surv. Dept., Pub. Works Ministry, Geol. Survey, Egypt, 1903; Tert. Vert. Fayum,
Egypt, 1906, p. 278, pi. 24) has described a land-tortoise from the Upper Eocene of
Egypt to which he has given the name Testudo ammon. If a true Testudo, it is the
oldest known. The published figures show that the neurals are variable in form,
but the relationships to typical Testudo are so lose that it may be accepted as
belonging to this genus. In some respects it appears to be intermediate between
Testudo and Hadrianus."

The above remarks would apply equally well to the species here described.
Up to this time the oldest known Testudo found in North America is from the
Lower Oligocene. Two species having been described, Testudo hrontops Marsh,
and T. exornata Lambe.

If, as has been inferred, Testudo has derived its ancestry from Hadrianus, the
intermediate characters observed in the present specimen are fully in accord with
its geological position. The axillary and inguinal buttresses rising but little above

» " Fossil Turtles of North America," 1908, p. 368.

GILMORE: the fossil turtles of the UINTA FORMATION 155

the costo-peripheral suture, and the presence of a divided supracaudal scute, in-
dicate its relationship with the genus Hadrianus, as now understood and defined.
On the other hand, the greatly reduced heads of the ribs, and especially the high
degree of differentiation reached by the neural and costal bones show its affinities
to Testudo.

In an attempt to find characters other than those used to separate the three
important genera constituting the family Testudinidse I have determined the width
of the bridge as compared to the total length of the plastron in all available speci-
mens pertaining to the genera Hadrianus, Stylemys, and Testudo, and find the
average to be as follows: In Hadrianus the bridge is thirty-eight per cent, of the
length of the plastron; in Stylemys it is forty-nine per cent., and in Testudo forty-four
per cent. In Testudo uintensis it is forty per cent, thus again demonstrating its
intermediate stage of development between Hadrianus and Testudo.

Cope assigned as the principal character distinguishing Hadrianus from the
other genera of the Testudinidse, "a divided supracaudal scute," and this would
perhaps appear a good reason for assigning the present specimen to that genus if it
were not known that three living species of the genus Testudo as recognized by
Boulenger, also have this scute divided. In the definition of the genus Testudo
the lip is "usually projecting abruptly from the general contour of the lobe."
The present specimen, however, is one of the exceptions, resembling the Oligocene
Testudo amphithorax in this respect.

It will be seen from this brief discussion that the present means of separating
the genera Hadrianus and Testudo is very unsatisfactory, and it would perhaps be
best to combine them until clean cut characters are found, to show that there are
two distinct genera.

Family TRIONYCHID.E Bell.

The soft shelled Trionychidse are represented in the present collection by
thirteen specimens of the genus A my da, four of which have been identified specif-
ically. These pertain to two species, hitherto kno\vn only from the Bridger
formation. Two fragmentary specimens, C. M. Nos. 2396 and 3285, judging
largely from the character of the sculpturing of the parts of the carapace present,
appear to represent undescribed species, but on account of their inadequate nature
I refrain from naming them. The genus Amyda is now represented in the Uinta
formation by the following species, Amyda crassa Hay, A. egregia Hay and A.
scutmnantiquum (Cope).

156 MEMOIRS OF THE CARNEGIE MUSEUM

Genus Am yd a Oken.
17. Amyda egregia Haj-.

Amyda egregia Hay, Fossil Turtles of North America, 1908, p. 531, plate 107,

figs. 1-3, text-fig. 691.

A specimen (C. M. No. 3254), collected by Earl Douglass from Horizon B, of
the Uinta formation, three or four miles northeast of Well 2, Uinta Basin, Utah, in
1908, is identified as belonging to Amyda egregia Hay. This specimen consists of
the articulated nuchal, the greater portion of the first neural, the first, second,
third, fourth, fifth, and sixth costals, all, excepting the first, lacking portions of
their upper extremities. In the matrix on the left side are the free ends of the
second, third and fourth costal ribs. These appear to be in their proper positions
in relation to the other parts of the shell which are preserved, and serve to give
some idea of the width and general contour of the carapace. The shell is broadly
rounded in front and considerably arched transversely. The width is estimated to
have been about 465 mm. The nuchal has a transverse extent of 255 mm. and
measures 52 mm. antero-posteriorly. Hay states that the latter measurement in
the type of the species is but 22 mm., although the figure published by him shows
it to be at least 42 mm. As in the type specimen the outer end of the nuchal over-
laps the free end of the rib of the first costal. The first neural at the anterior end
is 40 mm. wide. The sculpture of the carapace is coarse. On the inner ends of the
costals the pits and ridges form a honeycomb arrangement, but on their outer
fourths the ridges and pits are arranged in rows across the costals, toward their
ends on some of the costals the honeycomb pattern again prevails, on others the
rows persist to the smooth bevelled border.

In the form of the anterior end of the carapace, with slight emargination of
the border at the sutural junction of the nuchal, and the close resemblance of the
sculpture of the carapace, this specimen is in close agreement with the type of the
species. The chief differences observed are the considerably larger size of the
present specimen and the sloping bevel of the smooth ends of the costals, as con-
trasted with the abrupt bevel of the type.

A second specimen, C. M. No. 3255, in this collection consisting of many frag-
ments of the neurals and costals is provisionally referred to this species. This
specimen is from Horizon B, Uinta formation, and from the same locality as the
individual previously discussed. It was collected by Dr. W. J. Holland, and Earl
Douglass in 1908.

The type of Amyda egregia Hay is from the lower Washakie beds south of Hay-

GILMORE: the fossil turtles of the UINTA FORMATION 157

stack Mountain, Wyoming, so that the discovery of the present specimens in the
Uinta formation of Utah, considerably extends the geological as well as the geo-
graphical range of this species.

18. Amyda scutumantiquum (Cope).
Plate XXVI, fig. 2.

Trionyx scutumantiquum Cope, 6th Ann. Report U. S. Geol. Surv. Terr., 1872
(1873), p. 617; Amer. Naturalist, XVI, 1882, p. 988, fig. 6; Vert. Tert. Form.
West., 1884, pp. 118, 121, PL XVI, figs. 1, la; Hay, Bibliog. and Cat. Foss.
Vert. N. A., 1902, p. 454.
Amyda scutumantiquum Hay, Amer. Geologist, 35, 1905, p. 336; Fossil Turtles of
North America, 1908, pp. 521, 522, plate 100, figs. 2^, plate 101, fig. 1; text-
figs. 676, 677.

A very large specimen C. M. No. 3272 (see plate XXVI, fig. 2), consisting of a
considerable part of the anterior two-thirds of the carapace, is identified as per-
taining to the above genus and species. This specimen was collected by Earl
Douglass May 25, 1908, from the lower portion of Horizon B, "transition beds
sandstone," Uinta formation, Upper Eocene, as exposed on the south branch of
Red Bluff Wash, above the well on the stage-road between Bonanza and Kennedy's
Hole, Uinta County, Utah.

This specimen lacks all of the carapace posterior to the fifth costals, the
anterior border and left end of the nuchal, and small portions here and there of the
costals forward of the sixth. The sandstone matrix containing the impressions of
the seventh and eighth costals fortunately is preserved and serves to give a fairly
accurate idea of the dimensions of the entire shell.

In form the carapace is broadly oval with the length slightly exceeding the
breadth. The greatest width, as in the type, appears to have been at the middle.
The extreme width is about 530 mm.; the length, at the very least, was 570 mm.
The shell is broadly arched from the lateral borders to bej'ond the middle of the
costal plates. Along the middle of the back there is a pronounced longitudinal
depression which is deepest at about the middle of the shell.

The nuchal extends on each side of the midline about 180 mm.; its width on
account of the missing anterior border cannot be given. At the center on the
broken border the nuchal has a thickness of 23 mm. Its outer end is bevelled off
toward the front and outer extremity. The length of the nuchal is .74 the width of
the shell, whereas the type of the species is .76 (not .80 as stated by Hay).

158 MEMOIRS OF THE CARNEGIE MUSEUM

The first neural is exceedingly large, lioing 93 nnn. long and 50 mm. wide toward
the front. The others present diminish in size posteriorly. The second, third,
and fourth are coffin-shaped, with the widest end posterior, as is usual in the species
of this genus. The principal dimensions of the neurals are given in the table.

No. Length. Width.

1 93 50

2 62 41

3 59 38

4 58 34

All of the costal plates except the first grow wider toward their outer ends.
The expansion of the outer end of the second, as in the type, is especially pronounced.
In the angulation of the free border of the second costal it resembles that of Amyda
salebrosa Hay, more nearly than the type of the present species. At their free ends
all of the costals preserved are bevelled off to a sharp edge, except at the point
where the rib projects. None of the projecting ribs are preserved, so that the
distance they extend beyond the free border cannot be determined. Near their
outer ends at the sutural borders the costals have a thickness of from 5 to 9 mm.
The width of the proximal and distal ends of the costals are given in the table.

No. Width of Proximal End. Width of Distal End.

1 91 62

2 58 103

3 63 79

4 58 79

5 59 84e

e, estimated.

The surface of the carapace is ornamented with the usual pits and ridges, there
being two pits in a fine 10 mm. long, and occasionally three. They are large and
distinct, forming a honeycomb arrangement along the middle and on the proximal
halves of the costal plates, becoming smaller and less deeply impressed toward the
front. In small areas here and there at points about the middle of the costals the
pits are arranged in rows across the short diameters of the plates. On the distal
portions of the costals the pits are smaller and less deeply impressed, thus forming
a very distinct pattern, and, as the smooth band is approached, the ridges show a
tendency to break up into tubercles. Nowhere are the ridges wider than the pits.

Except for the considerably greater size of the specimen, and slight differences
in the general distribution of the large and small pits the present individual agrees
very closely with the type. The latter is from the Bridger beds on Cottonwood

GILMORE: the fossil turtles of the UINTA FORMATION 159

Creek, in Wyoming, and the specimen here described is the first record of the oc-
currence of this species in the Uinta formation.

A second specimen, C. M. No. 3330, was collected by Earl Douglass in 1908,
near the region of Well No. 2, Uinta Basin, Uinta County, Utah. It comes from
the Uinta formation (horizon not given) and consists of fragments of costal plates
having sculptured surfaces which are identical with those of the specimen discussed
above, and is therefore regarded as pertaining to the present species.

Order SQUAMATA.

Suborder SAURIA.

Family ANGUID.E.

Genus Glyptosaurus Auct.

19. Glyptosaurus sp. indet.

A specimen, C. M. No. 3405, consisting of the greater part of the parietal with
other fragments of the skull and lower jaws, and a few shields, is identified as per-
taining to the lizard-like reptile Glyptosaurus. It was collected by 0. A. Peterson,
August 24, 1912, on White River near Ouray, Uinta Basin, Utah, from Horizon C,
Uinta formation. Upper Eocene.

The few osseous shields present are evidently from the trunk of the bod}\
These are oblong quadrate, with a smooth, thinned out anterior end, which is over-
lapped by the next plate of the series. The lateral borders are roughened for sutural
union. The external surface, excepting the smooth area mentioned above, is
ornamented with small rounded tubercles closely arranged in more or less con-
centric rows. The cranial shields on the parietal are of irregular sizes with their
surfaces ornamented much in the same manner as the trunk-shields.

This specimen may represent an undescribed species, but at the present time,
on account of the very unsatisfactory type specimens upon which the nine de-
scribed species have been based, it is impossible to make adequate comparisons,
so that the specific determination of the present specimen must await the thorough
revision of the species of the genus. It is of interest, however, as recording for the
first time the occurrence of the genus Glyptosaurus in the Uinta formation, and also
from the fact that it occurs intermediate geologically between the oldest known
specimens from the Bridger described by Marsh, and the youngest specimen dis-
covered and described bj' Douglass from the Oligocene.**

'Douglass, Earl, "Some Oligoccnc Lizards," Annals of the Carnegie Museum, IV, 1908, pp. 278-283.

160 MEMOIRS OF THE CARNEGIE MUSEUM

Nine species of the genus have been described from the deposits of North
America. Geologically these are arranged as follows:

Glyptosaurus numlanus Douglass.

Lower Wliite River, OUgocene.

Glyptosaurus sp. indel.

Uinta formation, Upper Eocene.

Glyptosaurus sylvestris Marsh.

G. nodosus Marsh.

G. osceUatus Marsh.

G. princeps Marsh.

G. brevidens Marsh.

G. rugosus Marsh.

G. sphenodon Marsh.

G. anceps Marsh.

Bridger, Eocene.

EXPLANATION OF PLATES.

Plate XVHI.

Fig. 1. Carapace of Baena arenosa Leidy. C. M. No. 2356. One-half natural size.
Fig. 2. Plastron of Baena plahjplastra. Type. C. M. No. 3227. One-third natural size'.

Pl.\te XIX.
Fig. 1. Carapace of Baena inflata. Tj-pe. C. M. No. 3406.
Fig. 2. Plastron of the same. Both figures one-third natural size.

Plate XX.
Fig. 1. Carapace of Baena gigantea. Tj-pe. C. ]\I. No. 3441.
Fig. 2. Plastron of the same. Both figures one-fourth natural size.

Plate XXI.
Fig 1. Carapace of Echmatemys callopyge Hay. C. M. No. 2371.
Fig. 2. Plastron of the same. Both figures one-third natural size.

Plate XXII.
Fig. 1. Carapace of Echmatemxjs douglassi. Type. C. M. No. 3244.
Fig. 2. Plastron of the same. Both figures one-third natural size.

Plate XXIII.
Fig. 1. Carapace of Echmatemys hoUandi. Type. C. M. No. 3249. Viewed from the
right side. One-half natural size.

Fig. 2. Carapace of Echmatemys depressa. Type. C. 'SI. No. 2936. Natural size.

Plate XXIV.
Fig. 1. Carapace of Echmatemys obscura. Type. C. M. No. 3252.
Fig. 2. Plastron of the same. Both figures one-third natural size.

Memoirs Carnegie Museum, Vol. VII

Plate XVIII.

1. Cauai'.uk of Bdcna (ircnosd ]j:ii)v. ( '. M. No. 2'.iM\. X '..

2, Pl.vsthon ok Baaia plalijpla.itrn (iiLMOui;. Type, C. M. No. 3227. X

Memoirs Carnegie Museum, Vol

Plate XIX.

1. ("ahapack of Buciia iiifhil,

2. Plasthon ok Do. X 3.

Memoirs Carnegie Museum, Vol. VII

Plate XX.

Memoirs Carnegie Museum, Vol. VII

Plate XXI.

1. Carap.\ce of Echmatemys callopyge Hay. C. j\I. No. 2371. X -3.
2. Plastron of same. X J.

Memoirs Carnegie Museum, Vol. VII.

Plate XXII.

Erlninihiiujs dduijUissi (iiL.MOHi:. Tvi-E, ('. M. No. :V2i.-i. X
2. Pl.\stron of samk. X J.

Memoirs Carnegie Museum, Vol. VII.

Plate XXIIl.

EchwnUmu^ holhnuli ( ;ii,M(.iii:. Tvi'i:. ('. M. No. 3240. X \.
Kchmnlrnui^ dvpn-ssa (iiLMouK. Tyim:, V. M. No. 2930. X i-

Memoirs Carnegie Museum, Vol. VII.

Plate XXIV.

1. Carapace of EchiiKilcnujs obscura Gilmokk. Tvpk, C. M. No. 3252. X ^■
2. Plastron of same. X J.

Memoirs Carnegie Museum, Vol. vil.

Plate XXV.

1. Pl.\.sti{o.\ ok Hddridiius corsoiti (Lkidv). ('. M. No. o4U;i. X },.

2. Plastron of Hadrianus robustus Gilmuuk. Type, C. M. No. 3342.

Memoirs Carnegie Museum, Vol. VII,

Plate XXVI.

1. I'l.AsTHoN OF Uddn'aiuts utdhnisis (liLMoui:. Tvi-i:, (". M. No. 2343. X

2. C'AUAi'AtK OK Aiinjda sciiliDiiiintiquuni (Cope). C. M. No. 3272. X {■

Memoirs Carnegie Museum, Vol. VII.

Plate XXVII.

1. ('Aii.M'ALK OF Tctiliulo }iinUusi.s CiiLMOKi:. Tvi'i;, C. M. No. SSol. X •'..
2. Pl.\stro.\ of same. X J.

GILMORE: the fossil turtles of the UINTA FORMATION 161

Plate XXV.
Fig. 1. Plastron of Hadrianus corsoni (Leidy). C. M. No. 3403. One-fifth natural size.
Fig. 2. Plastron of Hadrianus rohustus. Type. C. j\I. No. 3342. One-third natural size.

Plate XXVI.
Fig. 1. Plastron of Hadrianus utahensis. Type. C. M. No. 2343. One-fourth natural
size.

Fig. 2. Carapace of Amyda scutumantiquum (Cope). C. M. No. 3272. One-fourth
natural size.

Plate XXVII.
Fig. 1. Carapace of Testudo uintensis. Type. C. M. No. 2331.
Fig. 2. Plastron of the same. Both figures one-half natural size.