A skeleton of Astrapotherium

LI E) RAFLY

OF THE

U N I VLRSITY

or ILLINOIS

GEOLOGICAL SERIES

OF

FIELD MUSEUM OF NATURAL HISTORY

Volume VI Chicago, October 30, 1935 No. 13

A SKELETON OF ASTRAPOTHERIUM^

By Elmer S. Riggs

Associate Curator of Paleontology

Fossil mammals of the order Astrapotheria have been known
from South America since 1853, Four genera of astrapotheres proper
have been described and generally recognized. These forms, in some
instances, are based upon series of teeth and fragments of jaws,
but the genus Astrapotherium has long been known from splendid
skulls with associated mandibles.

Until the discovery of the specimen under discussion the structure
of skeleton in all members of the order remained unknown save a
single forefoot of Par astrapotherium (Princeton University Expedi-
tions to Patagonia, 6, pi. XXXVI) and the scapula and bones of
the foreleg (Annales de Paleontologie, 1, p. 19), with an isolated
astragalus. The femur figured by Gaudry {ibid., p. 4, fig. 5) is
clearly that of Homalodotherium; the tibia figured with it is so unlike
those of the specimen in hand as to render it highly improbable
that it could have belonged to a member of the genus Astrapotherium.

Field Museum was so fortunate as to secure from the Santa
Cruz beds of middle Miocene age a skeleton of Astrapotherium
magnum (P14251) sufficiently complete to make possible a study of
the entire osteology of the animal. This skeleton is almost entire
and can be mounted as an articulated specimen, as may be seen in
figure 39. However, much of the vertebral column, ribs, and pelvis
has been so dissolved by percolating waters that only shells of the
bones remained. For this reason the vertebrae and ribs could not
be removed from the matrix, nor could the specimen be set up as a
free mount. The posture of an animal lying down has been chosen
for it and the position of the head and legs has been made to
conform to this posture as nearly as was possible. The deformation
of parts, due to compression in the matrix, could not be corrected.
The preparation was carried out by the late J. B. Abbott under

iRead before the Paleontological Society, December 28, 1934. Abstract,
Proc. G. S. A., 1934, p. 379.

No. 348 167

natural Histtry Survey
jLi1»rsury

168 Field Museum of Natural History — Geology, Vol. VI

the writer's supervision. Specimens of Astrapothericuliis (?)^ and
of Parastrapotherium were used as guides in restoring missing parts.

The mounted specimen shows that Astrapotherium magnum was
a long-bodied animal, having a moderately long neck, a deep thorax,
relatively long and strong forelegs, weaker hind quarters and slender
hind legs. The skeleton as mounted is 9 feet 5 inches (2 . 88 meters)
in length from the nose to the extremity of the pelvis and would
stand about 4 feet 6 inches (1.37 meters) in height at the shoulders.

The skull is that of a large adult animal, judging from comparison
with other specimens in this Museum as well as with figured speci-
mens from other collections. The animal has a large head, short
nasal bones, and widely open nasal passages which have been inter-
preted as indicating the presence in life of a moderately developed
proboscis. There are no upper incisors; the canine teeth both above
and below are angular and grow throughout the life of the individual.
The premolars are reduced in size and in numbers; the molars are
massive and of a lophodont pattern not unlike those of the Oligocene
aquatic rhinoceros, Metamynodon. Skulls of this species have been
figured in a number of publications and are too well known to require
detailed description here.

Vertebrae. — The atlas has moderately broad and backwardly
directed transverse processes. The axis is relatively strong and
has an elongate centrum and a broad spine, which is thickened and
vertical at the posterior margin. The odontoid process is elongate
and rounded, and, like that of the amblypods, extends quite through
the arch of the atlas. The anterior articular surfaces are subtri-
angular in outline and are strongly oblique to the axial direction.
The succeeding cervicals are of moderate length and have strong
articulations, as is common to heavy-headed animals. The trans-
verse processes in the posterior five cervicals are expanded into
broad lateral plates; the neural spines, though broken in this speci-
men, were evidently short.

The dorsal vertebrae are relatively slight and are little differentiated
throughout the series. The spines in the anterior members of the
series are short and tapering; in the posterior members they change
to a short, broad type. The zygapophyses are rather widely spaced
throughout the series, but in the thirteenth and fourteenth there
is a transition from the upward-and-downward-facing type to
the lateral-facing one common to ungulates. From that point to the

1 The identification of this specimen is in doubt as there is no skull associated
with it.

V. G

169

170 Field Museum of Natural History — Geology, Vol. VI

saq'um the zygapophyses are interlocking. The lumbar vertebrae
are surprisingly weak for so large an animal. Their spines are low
and broad; the crests, as seen from above, are bifurcate posteriorly.
The transverse processes are short and laterally directed. The
sacrum consists of five vertebrae which are firmly co-ossified but
apparently have separate spines. There is no confirmation of this
observation in other related specimens. The spines are broken close
to the arch and lost; the same is true of the specimen of Astrapotheri-
culus (?). The lateral processes are rather broad and join the ilia at
their superior margins.

The convex outline of the vertebral column in the dorsal region is
largely due to the position the specimen had taken in the matrix
and does not represent the natural outline of the back of the animal
in a recumbent position. However, the back of the animal in a
standing position would probably have been convex from the
shoulders to the mid-dorsal region and straight from that point
through the lumbar region, as is indicated in the restoration (fig.
36). There are nineteen dorsal vertebrae and five lumbars, making
a total of twenty-four dorso-lumbar vertebrae in the series.

The scapulae are elongate and tapering at the upper end, similar
in a general way to that of Homalodotherium so far as may be seen
from the Field Museum specimen (P13092) though less rounded in
outline than restored by Scott (Field Mus. Nat. Hist., Geol. Mem.,
I, No. 1, pi. 7). The spine is simple, broad at the crest, and ex-
panded at the lower end into a broad acromion process. It divides
the lateral surface of the scapula into two approximately equal
areas. The deep fossae thus formed indicate a considerable strength
and great freedom of movement in the scapular muscles.

The pelvis is relatively weaker than the shoulder-girdle, much
weaker than the same element in the specimen of Astrapothericulu^ (?).
It is more elongate antero-posteriorly than that of known toxodonts
and litoptems. More than two-thirds of its entire length lies in
front of the acetabulum. The anterior extremities of the ilia are
somewhat broken but appear to have been rounded with a rather
sharp antero-inferior angle; the mesial surfaces have contact with
the sacrum throughout their entire length.

The humerus is a relatively strong bone and almost as long as
the femur. It has a prominent tuberosity, an elongate but strongly
marked deltoid area which terminates at a point below the middle
of the shaft, a wide but simple bicipital groove, and a moderately
developed supinator ridge. The proximal articular surface extends

Skeleton of Astrapotherium

171

well around to the posterior side of the shaft, indicating a considerable
angulation in the shoulder joint.

The ulna and radiiis are of similar size and strength in the shaft.
The ulna, in its total length, is somewhat shorter than the humerus.
The olecranon is in continuous line with the shaft ; the sigmoid notch
is wide and deeply cleft at its anterior margin to receive the head
of the radius as figured by Gaudry (Annales de Paleontologie, 1,

Fig. 37. Right manus of Aslrapolherium magnum. F.M. No. P14251, x 1/3. ec., scaphoid;
(., lunar; c, cuneiform; p., pisiform; m., unciform; m., magnum; td., trapezoid; tr,, trapezium; I-V,
metacarpals.

p. 5, fig. 5). The distal articulation of this bone has the form of
a facet placed obliquely to the shaft and extending well over the
distal end. It has the unusual feature of articulating with the lunar
as well as the unciform and the pisiform bones. The radius is moder-
ately curved in the shaft and but little larger at the distal, than at
the proximal end. The proximal articulation is concavo-convex in
form and elongate in the transverse direction, covering the entire
breadth of the trochlea. The anterior margin of the radial facet
is straight; the posterior margin is angular and, fitting closely the
notched articulation in the ulna, admits of little rotary movement.
The distal end of the radius is likewise of unusual structure, having
a uniformly rounded and concave facet with possibly a slight styli-
form process. The radius articulates, in a freely rotating joint,

172 Field Museum of Natural History — Geology, Vol. VI

with the convex surface presented by the scaphoid and the greater
part of the lunar.

The articulation between the proximal ends of the ulna and
radius is similar to that of the proboscidians and the amblypods.
The radius is strong as in the latter group and shares with the ulna
the support of the body. A somewhat similar condition existed in
the earlier macrauchenids though Macrauchenia itself had lost all
rotary movement of the forearm.

The manus of Astrapotherium is quite similar to that of Coryphodon
of the Amblypoda. The important differences as derived from
direct comparison are (1) that in this genus the scaphoid and the
lunar are closely apposed and present a continuous convex surface
to the radius, while in Coryphodon the scaphoid and the lunar meet
the radius in two distinct facets; (2) that the lunar in this genus
presents a strong oblique facet on its postero-lateral surface to
articulate with a second facet on the ulna; (3) that the cuneiform
has a contact with metacarpal V. The distal row of carpals in this
genus is essentially similar to that of Coryphodon except that the
trapezium has contact with metacarpal II. The metacarpals of
this genus increase in breadth from II to V; metacarpal I is equal
in breadth to II but is much reduced antero-posteriorly. The
phalanges are entirely unknown from the forefoot.

The femora are long, slender bones, laterally curved in the upper
half of their length. The head of the femur is well constricted, the
great trochanter relatively low, the lesser trochanter rises in an
elongate but prominent crest. The shaft has a slight compound
curvature in the antero-posterior direction. This feature, together
with the backward extension of the condylar facets, indicates a
great freedom of movement at the knee. The leg in the standing
posture was apparently almost straight. The facet for the patella
extends well forward on the shaft but is little elevated above its
anterior surface. The femur figured by Gaudry, as above stated,
is that of a contemporary homalodothere and is very different from
that of this genus.

The tibia in its axial measurement is approximately three-fourths
as long as the femur. It is broad at the proximal end but laterally
compressed and angular in the upper half of the shaft, more rounded
in the lower half. The cnemial crest is prominent, the facet for artic-
ulation with the fibula is well below the overhanging condyle. The
distal articulation is irregularly convex and oblique to the axial
direction of the shaft. The internal malleolus has a distinct facet

Skeleton of Astrapotherium

173

for articulation with the astragalus. The tibia of Gaudry's figure,
referred to in connection with the femur, is so different from those
of this specimen as to indicate that it was also erroneously referred
to Astrapotherium. It is even more unlike that of the Santa Cruz
Homalodotherium. For the present, this writer is unable to identify it.

The fibula is a slender bone, straight in the shaft, moderately
expanded in the extremities, and has well-defined articular facets.

Fig. 38. Left pes of Astrapotherium mtignum. F.M. No. P14251, x 1/3. a., astragalus; c,
calcaneum; n., navicular; cb., cuboid; en., 1, 2, 3, cuneiformes; I-V, digits.

The relative length and position of femur, tibia, fibula, and pes are
more like those of the North American amblypods than like those
of any group of South American mammals. Apparently this simi-
larity is only another case of parallel development.

The pes of Astrapotherium (fig. 38) is more elongate than that of
Coryphodon and much more slender. The tarsals and metatarsals
are less uniform in their size and strength. The digits are elongate
as in the earlier amblypod genus, Titanoides. The astragalus is
short and rounded in outline, its mesial and posterior margins

174 Field Museum of Natural History — Geology, Vol. VI

indented. The proximal facet covers almost the entire surface of
the bone as is true of Coryphodon. The mesial surface bears two
distinct facets for articulation with the internal malleolus of the
tibia, separated by a narrow passage for blood vessels. The facet
for the navicular covers the entire anterior surface; this facet is
plane in the vertical axis but strongly convex in the lateral one.
It meets the anterior facet for the fibula in a sharp angle; it is con-
tinuous with the facet for the cuboid. The two inferior articular
surfaces are separated by a wide median fossa; the sustentacular
facet is smaller than that for the calcaneum.

The calcaneum is relatively small and quite overshadowed by the
astragalus in its position in the foot. The tuber calcis is short,
vertically compressed and rugose throughout the plantar surface.
Possibly there has been some vertical compression in the single
specimen preserved, but little evidence of such distortion can be
seen. The facets for articulation with the astragalus are both elongate
and stand at a right angle to each other, separated by a deep pit
in the angle. The facet for the cuboid is long and narrow but not
well preserved in this specimen.

The navicular is a very thin bone, deep in its vertical dimension
but appearing at the dorsal surface of the foot in the form of a wide
crescent, very similar to this element in Coryphodon. It articulates
proximally with the entire convex, anterior surface of the astragalus
and articulates distally with the three cuneiform bones by three
distinct facets.

The cuboid, as well as the ento- and mesocuneiform bones, is
wedge-shaped and elongate in the dorso-plantar direction. The
mesocuneiform is very much reduced, the entocuneiform is larger,
the ectocuneiform is longer than the cuboid but not so broad nor
so deep. In their proportions, these bones approach more nearly to
the Paleocene Titanoides than to the later Coryphodon.

The metatarsals are more or less broken at their distal ends.
They vary in size and in strength, the second and third being reduced
while the fifth is expanded into a rugose lateral margin very much
the same as in Titanoides. The inferior surfaces of the entire series
of metatarsals are concave and apparently supported a heavy
cushion upon which the animal walked. Only two phalanges of
the pes are preserved. They are somewhat longer than wide. There
is no evidence as to the unguals in this specimen, though they were
probably very much reduced, in conformity with the size of the
phalanges.

Skeleton of Astrapotherium 175

Measurements

Meters

Skull, length, premaxillaries to condyles 650

Mandibles, crowns of incisors to condyle 665

Atlas

Breadth across transverse processes 337

Antero-posterior breadth of superior arch 062

Axis

Length of centrum and odontoid process 171

Antero-p)osterior breadth of spine 134

Vertebrae

Cervical, length of five centra 330

Dorsal, length of nineteen centra in series 990

Lumbar, length of five centra 325

Sacral, length of centra 245 '

Scapula

Greatest length 485

Greatest breadth 222

Breadth of acromion process 184

Pelvis

Total length 520

Length from center of acetabulum to iliac crest 350

Greatest breadth of ilium 178

Humerus

Axial length 502

Greatest breadth of proximal end 162

Greatest breadth of distal end 117

Ulna, greatest length 434

Radius, greatest length 332

Third metacarpal, length 081

Femur

Greatest length 558

Greatest breadth at distal end 119

Tibia, axial length 399

Fibula, greatest length 373

Astragalus

Antero-posterior diameter 069

Transverse diameter 071

Calcaneum

Length 098

Breadth 066

Fourth metatarsal, length 080

Patella

Length 080

Breadth 049

< Estimated.

DISTINGUISHING CHARACTERS

The distinguishing characters of Astrapotherium as derived from
the entire skeleton are: nasals short and nares wide; upper incisors
wanting; canines long, angular in section and hypsodont; premolars
reduced to two; molars lophodont in structure and increasing in
size from 1 to 3; vertebral formula: cervicals 7, dorsals 19, lumbars 5,
sacrals 5, caudals not known but probably not more than 15; ulna
and radius equal in strength, head of radius notched into anterior

176 Field Museum of Natural History — Geology, Vol. VI

lip of sigmoid articulation; digits V-V, forefeet digitigrade, hind
feet plantigrade and weak; ulna having contact with lunar, the
cuneiform having contact with metacarpal V.

PROBABLE HABITS

Astrapotherium appears to have been a large-headed and long-
bodied animal of the size of a bison. The dentition indicates habits
of feeding upon low, moist-land plants. The padded feet also indicate
a forest or meadowland habit. The occurrence of the fossils in a
limited zone of the Santa Cruz beds and in a sandy, mud-bar stratum
just above the marine Patagonian beds, suggests life in a lowland
area. Likewise, the occurrence of remains of the Deseado (Oligocene)
members of this family in the lagoon deposit at La Flecha and
in channel deposits of the Deseado stage at Lake Colhue Huapi and
other localities where specimens are found disarticulated, rolled, and
water-worn, but seldom articulated, also indicates river deposition
much the same as does the occurrence of Metamynodon in the channel
deposits of the White River beds of North America. All this,
together with a dentition fitted for feeding on lush vegetation,
leads to the conclusion that Astrapotherium was a frequenter of
lagoons and of river banks and that he was probably at home in
the low grounds.

No attempt at fixing general relationships is made at this time.
The astrapotheres have many characters in common with the
litopterns, also they have many in common with the amblypods.
Ameghino was inclined to class them in that order. It is hardly
probable that the two groups, so widely separated geographically
in their later history, could have relationships nearer than through
some "protungulate" ancestry.

In accordance with a plan for placing various sections of the
collections of South American fossil vertebrates in the hands of
specialists best qualified to study them, the entire astrapothere
material in Field Museum has been assigned to Professor William
B. Scott for study and publication. This plan offers the surest
guarantee for the careful study of a unique specimen and the most
effective characterization of this little-known order of mammals.

177