Fossil vertebrae of an amphiumid salamander (Amphiuma sp.) are
reported from a late Pleistocene (Rancholabrean NALMA) site of coastal
Georgia. A suite of vertebral characters that identify the fossils to
Amphiuma is given, as well as illustrations of important vertebral
characters that distinguish the genus from sirenid salamanders. The
presence of Amphiuma in the paleofauna indicates an aquatic habitat was
present during the time of deposition.
Keywords: Amphiuma, Pleistocene, coastal Georgia.
The Clark Quarry local fauna of Glynn County, coastal Georgia, has
yielded a diverse Pleistocene (Rancholabrean North American Land Mammal
Age [hereafter NALMA]) vertebrate fauna (e.g., 1, 2, 3, 4, 5). Recent
radiocarbon ([.sup.14.C]) dating of bison (Bison latifrons) bone
collected at the fossil site yielded an age of 12,350 [+ or -] 70 YBP
for the fauna (6). Although the locality is best known for its
megafauna, important microfossils have recently been discovered by
screen washing in situ sediments from the site. These efforts have
yielded fossils of small vertebrates including amphibians, reptiles, and
rodents. Here we describe two relatively well-preserved Amphiuma
vertebrae from the site, and illustrate and discuss vertebral characters
and structures that clearly separate this taxon from those of the other
North American aquatic eel-like salamanders of the family Sirenidae. We
also discuss problems with the identification of isolated Amphiuma
vertebrae to the species level. In these discussions we follow the
vertebral terminology of Gardner (7). Institutional abbreviations
include GCVP, Georgia College and State University Vertebrate
Paleontology Collections; GC-H, Georgia College and State University
Recent Herpetological Skeletal Collections.
Class Amphibia Linnaeus, 1758
Order Caudata Scopoli, 1777
Family Amphiumidae Gray, 1825
Genus Amphiuma Garden, 1821
Amphiuma sp. indet.
Fossil History of Amphiuma -- Although the fossil record of the
paedomorphic salamander genus Amphiuma is limited, it dates back to the
late Paleocene (8), approximately 54 Ma B.P. Estes (8) described the
extinct taxon A. jepseni on the basis of three partial skeletons from
the late Paleocene (probably Tiffanian NALMA) Polecat Formation of
Wyoming. Although certainly present, amphiumids are not known from the
Eocene. Two records of amphiumid salamanders are reported from the
Miocene, but these may be in error. Holman (9) erected the species A.
antica on the basis of a single, poorly preserved (and poorly
illustrated) trunk vertebra from the middle Miocene (Barstovian NALMA)
of Texas. The taxon has since been shown to be invalid (nomen dubium) as
the holotype and only known vertebra does not differ in any significant
way from vertebrae of extant Amphiuma (see discussion in Gardner, 7).
Gardner (7) further suggested that the fossil may not represent an
amphiumid salamander, but we believe the specimen retains some
characters consistent with the genus Amphiuma (e.g., similar neural
crest and no aliform process). Albright (10) tentatively assigned a
poorly preserved salamander vertebra from the early Miocene (Arikareean
NALMA) of Texas to an amphiumid-like salamander. Unfortunately, the
fossil is not figured and based on the description of the specimen, it
is too fragmentary to determine a family allocation. We know of no
Pliocene records of amphiumids, but Pleistocene (Rancholabrean NALMA)
occurrences are well documented from Florida and Texas (e.g., 11) and
from Georgia (12).
Material -- Two trunk vertebrae, GCVP 8310 (Fig. 1A), and 8311.
Remarks -- The Clark Quarry vertebrae are assigned to the endemic
North American salamander family Amphiumidae based on the presence of
postzygapophyseal crests (Fig. 1A), an unequivocal vertebral
autapomorphy for the family (7). Two genera of Amphiumidae are currently
recognized, Proamphiuma Estes and the type genus Amphiuma. Proamphiuma
is known from the late Maastrichtian or early Paleocene, while Amphiuma
is known from the late Paleocene and Pleistocene to Recent (7, 11). The
vertebrae of Amphiuma (and the fossils) differ from those of Proamphiuma
in having relatively higher neural crests and postzygapophyseal crests,
deeper subcentral keels, and relatively larger anterior basapophyses
(7). Overall, the Clark Quarry vertebrae display a suite of features
that characterize Amphiuma as follows (modified from Holman, 11, 13, and
Gardner, 7): vertebrae elongate and amphicoelous; neural crests long and
well developed; neural spines short and divergent from either side of
the posterior extent of the neural crests; postzygapophyseal crests
present (worn but evident in the fossils); alar processes present
(damaged on the fossils); transverse processes broken but appear to be
unicapitate suggesting a post-anteriormost trunk position; centra
constricted medially with deep subcentral keels (evident but worn or
damaged in both fossils); anterior basapophyses elongate (preserved in
GCVP 8310 but damaged in GCVP 8311).
[FIGURE 1 OMITTED]
Amphiumid vertebrae resemble those of the sirenid genera Siren and
the diminutive Pseudobranchus in having elongate centra with alar
processes, well-developed neural crests, and well-developed subcentral
keels along the ventral aspect of the centra. In spite of these
similarities, Amphiuma vertebrae (and the fossils) differ from those of
Siren and Pseudobranchus in having less extensive alar processes,
lacking widely diverged aliform processes, having postzygapophyseal
crests, and in having differently positioned transverse processes. In
Amphiuma, the alar processes (especially the anterior ones) are shaped
differently and are much less expansive in their lateral and anterior
extent than those of the sirenids (see Fig. 1B, C). The alar processes
of the fossils are badly damaged, but they retain enough of the
structures to determine that they were amphiuma-like (Fig. 1A). The
short, divergent neural spines of Amphiuma (Fig. 1A, B) are in strong
contrast to what is seen in Siren: widely divergent aliform processes
with a thin shelf of bone between them that projects off the posterior
lateral margins of the neural arch (Fig. 1C). We do not know if this
"paraaliform" shelf of bone is unique to Sirenidae, but it
never occurs in Amphiuma. As previously mentioned, postzygapophyseal
crests are a derived character state unique to Amphiuma (see Fig. 1A,
B). Also, the transverse processes of Amphiuma are positioned
perpendicular (or nearly so) to the long axis of the centra (Fig. 1A,
B), whereas in sirenids they are oblique to the long axis, and are
directed posteriorly (Fig. 1, C).
Although the fossils are confidently referred to genus, the
specific identification of isolated Amphiuma vertebrae is problematic.
Four species of Amphiuma are known: the late Paleocene species A.
jepseni Estes and the Recent species A. means, A. tridactylum Cuvier and
A. pholeter Neill. Fossils of A. jepseni were not available for
comparison during this study, but considering the geographic and
temporal separation between the two localities (late Paleocene of
Wyoming vs. late Pleistocene of Georgia), it is highly unlikely that the
Clark Quarry fossils pertain to this taxon. During the course of this
study we did, however, critically examine isolated trunk vertebrae from
nine individuals of A. means, five of A. tridactylum, and one of A.
pholeter to determine if we could distinguish the species on the basis
of vertebral differences. Other than the fact that the trunk vertebrae
of large adult A. means and A. tridactylum were consistently larger than
those of the smaller species A. pholeter, we could not differentiate
isolated Amphiuma vertebrae to species. Compared to Recent specimens,
GCVP 8310 and 8311 represent a small amphiumid(s) with a total length of
approximately 340 mm. Although this is within the size range of adult A.
pholeter, it also is within the size range of small individuals of A.
means and A. tridactylum. Although the fossils cannot be assigned to A.
means on the basis of vertebral morphology, this is the species that
would be expected in the paleofauna on the basis of (1) the young age of
the fauna (cf. 12, 000 yrs BP), and (2) it being the only living
amphiumid species that occurs today in SE Georgia (14). Nonetheless, we
hesitate to make a species determination on the basis of temporal and
This is only the second report of a fossil amphiumid salamander in
Georgia. Hulbert and Pratt (12) reported that fossil Amphiuma means
(dentaries and vertebrae) were common in the coastal Rancholabrean Isle
of Hope fossil site of Chatham County, Georgia. They did not provide
osteological evidence for their allocation of the fossils to A. means;
presumably their identification was based on the present day
biogeographic distribution of the species. We recently examined the Isle
of Hope material, and although the vertebrae are likely too large to be
A. pholeter, they are within the size range of extant A. means and A.
tridactylum. Here we provide evidence that the vertebral morphology of
amphiumid salamanders is very distinctive at the generic level, but is
too homogeneous at the Recent species level to differentiate isolated
vertebrae at this taxonomic position.
Based on the known habitat requirements of Recent Amphiuma (e.g.,
14, 15, 16, 17), its presence in the Clark Quarry local fauna suggest
that an aquatic habitat such as (for example) a swamp, sluggish stream,
or a pond was present in the Clark Quarry area during the time of
deposition. This is supported by fossils of other vertebrates in the
Clark Quarry fauna such as ranid frogs, alligators, aquatic pond
turtles, and natricine snakes (including Thamnophis and Nerodia [see
We thank Linda Chandler for improving an earlier draft of this
paper. Special thanks must go to Dr. James Gardner for his helpful
discussions of amphiumid vertebral anatomy. We also thank the Clark
family of Brunswick, Georgia for granting Dr. Mead excavation privileges
at the fossil site, and we thank numerous GC & SU biology students
who participated in field work and/or matrix sorting.
1. Mead A J and Spell BL: New Pleistocene vertebrate locality from
Brunswick. Ga J Sci 60:45, 2002
2. Bahn RA and Mead AJ: Columbian mammoth excavations in Brunswick,
Georgia: The historical significance. Ga J Sci 62:28-29, 2004.
3. Bahn RA and Mead AJ: The Pleistocene vertebrate fauna of Clark
Quarry, Brunswick, Georgia. Ga Soc Amer Abstracts with Programs 37:4,
4. Clark J L, Clark KA, Mead AJ, Parmley D, and Bahn RA: A
preliminary description of the Pleistocene herpetofauna from Clark
Quarry, Brunswick, Georgia. Ga J Sci 63:31, 2005.
5. Clark KA, Chandler RM, Mead AJ, and Bahn RA: A preliminary
description of the Pleistocene avifauna from Clark Quarry, Brunswick,
Georgia. Ga J Sci 64:35, 2006
6. Bahn RA and Mead AJ: Occurrence of the giant Ice Age bison,
Bison latifrons, from Late Pleistocene coastal Georgia. Geol Soc Amer
Abstracts with Programs 38:84, 2006.
7. Gardner JD: The fossil salamander Proamphiuma cretacea Estes
(Caudata; Amphiumidae) and relationships within the Amphiumidae. J Vert
Paleont 23:769-782, 2003.
8. Estes R: The fossil record of amphiumid salamanders. Breviora,
No. 322, 11pp, 1969.
9. Holman JA: Amphibians and reptiles from the Gulf Coast Miocene
of Texas. Herpetologica 33:391-403, 1977.
10. Albright LB: Lower vertebrates from an Arikareean (earliest
Miocene) fauna near the Toledo Bend Dam, Newton County, Texas. J Paleont
11. Holman JA: Fossil Salamanders of North America. Bloomington:
Indiana Univ Press, 2006.
12. Hulbert RC and Pratt AE: New Pleistocene (Rancholabrean)
vertebrate faunas from coastal Georgia. J Vert Paleont 18:412-429, 1998.
13. Holman JA: A small Pleistocene Herpetofauna from Houston,
Texas. Tx J Sci 27:418-423, 1965.
14. Petranka J W: Salamanders of the United States and Canada.
Washington: Smithsonian Institution Press, 1998.
15. Gibbons JW and Semlitsch RD: Guide to the Reptiles and
Amphibians of the Savannah River Site. Univ GA Press, Athens, 1991.
16. Smith LL and Franz R: Use of Florida round-tailed muskrat
houses by amphibians and reptiles. Fl Field Natural 22:69-74, 1994.
17. Franz R: An introduction to the amphibians and reptiles of the
Katharine Ordway Preserve-Swisher Memorial Sanctuary, Putnam County,
Florida. Bull Fl Mus Nat Hist 38:1-10, 1995.
18. Mead AJ, Bahn RA, Chandler RM, and Parmley D: Preliminary
comments on the Pleistocene vertebrate fauna from Clark Quarry,
Brunswick, Georgia. Current Research in the Pleistocene. 23:174-176,
Dennis Parmley, Josh Clark, Alfred J. Mead
Department of Biological and Environmental Sciences
Georgia College & State University, Milledgeville, GA 31061
*Corresponding author (firstname.lastname@example.org)