-'

(US ISSN 0027-410

BuLLetln OF THE

?i?E^^|ra^^^H

Museum of

.7.-.».**^g

Comparative

Zoology

Volume 149
1 980-1 982

Ub. CO MP. ZOUL
LIBRARY

MAR 1 5 1-^^

■^RVAHD

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS 02138 U.S.A.

No. 1. The Orb-Weaver Genus Mecynogea. the Subfamily
Metinae and the Genera Fachygnatha, Glenognatha
and Azilia of the Subfamily Tetragnathinae North of
Mexico (Araneae: Araneidae). By Herbert W. Levi.
June 1980 1-74

No. 2. Zoogeography of North American Unionacae (Mollusca:
Bivalvia) North of the Maximum Pleistocene Glaciation).
By Richard I. Johnson. September 1980 77-189

No. 3. A Phyletic Study of the Lake Tanganyika Cichlid Genera
Asprotilapia, Ectodus, Lestradea, Cunningtonia, Oph-
thalmochromis, and Ophthalmotilapia. By Karel F.
Liem. February 1981 191-214

No. 4. A Case History in Retrograde Evolution: The Onca
Lineage in Anoline Lizards. H. Subdigital Fine Struc-
ture. By J. A. Peterson and E. E. Wilhams. July 1981 215-268

No. 5. The American Orb-Weaver Genera Dolichognatha and
Tetragnatha North of Mexico (Araneae: Araneidae,
Tetragnathinae). By Herbert W. Levi. July 1981 271-318

No. 6. A Revision of the Nearctic Beridinae (Diptera: Stratio-

myidae). By Norman E. Woodley. August 1981 319-369

No. 7. Rodents from the Deseadan Oligocene of Bolivia and the
Relationships of the Caviomorpha. By Bryan Patterson
and Albert E. Wood. January 1982 371-543

M.

(US ISSN 0027-41 0(

seum o

Comparative
Zoology

■••K'-'^iV/';.'..', -,-•/•■

m

IV) U 3. COM p. ZOO!
I . i Fr^ R A rr* V

The Orb-Weaver Genus Mecynogea;^f)€

Subfamily Metinae and the Genera

Pachygnatha, Glenognatha and Azilia

of the Subfamily Tetragnathinae North

of Mexico (Araneae: Araneidae)

TV

HERBERT W. LEVI

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBER 1
16 JUNE 1980

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Breviora 1952-

BULLETIN 1863-

Memoirs 1864-1938

JOHNSONIA, Department of iMollusks, 1941-

OccAsiONAL Papers on Mollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. 1. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Tere-
dinidae (iMollusca: Bivalvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts, 02138, U.S.A.

V

© The President and Fellows of Harvard College 1980.

THE ORB-WEAVER GENUS MECYNOGEA, THE SUBFAMILY
METINAE AND THE GENERA PACHYGNATHA, GLENOGNATHA
AND AZILIA OF THE SUBFAMILY TETRAGNATHINAE
NORTH OF MEXICO (ARANEAE: ARANEIDAE)

HERBERT W. LEVI'

Abstract. Several genera are desribed and il-
lustrated, webs are pictured when known, and dis-
tril)utions are plotted. These genera include: Me-
cynogea with one species, Nephila with one,
Leucauge with two, MeteUina with three, Meta
with two species (one of them new, from California
caves, similar to a Mediterranean cave Meta), the
new genus Metleucauge with one new North Amer-
ican species (others in eastern Asia), Pachygnatha
with eight species, Glenognatha with two and Azil-
ia with one North American species.

On the basis of numerous characters, it appears
that Metinae is the primitive group, with the great-
est number of plesiomorph characters, Tetragnadi-
inae are specialized in one direction, Araneinae in
another. Because of the numerous intermediate
genera, it is best to combine the family Tetragnath-
idae with Araneidae, but to consider the group a
subfamily.

INTRODUCTION

Hope oi learning about the relation-
ships among orb-weavers was the reason
for combining revisions of diverse orb-
weavers in this paper. The only genera
of the Tetragnathinae not included here
are Dolichognatha and the large genus
Tetragnatha. No surprising discoveries
were made, but evidence was provided
in support of the classification of 19th
century arachnologists.

ACKNOWLEDGMENTS

I want to thank the many colleagues
whose help made my work possible. G.

' Museum of Comparati\e Zoology, Harvard Uni-
versity, Cambridge, Massachusetts 02138.

H. Locket and A. F. Millidge, and H.
Yoshida gave me specimens of species
related to ours. Mark Stowe photo-
graphed an Azilia web; Don Boe photo-
graphed MeteUina webs. The following
made specimens available: P. Arnaud
(California Academy of Sciences), D. Bix-
ler, J. A. Beatty, G. W. Byers (Snow En-
tomological Museum, Univ. of Kansas), J.
Carico, F. Coyle, C. Dondale (Biosys-
tematics Research Institute, Agriculture
Canada), H. Dybas and J. B. Kethley
(Field Museum), W. J. Gertsch (Ameri-
can Museum of Natural History), M.
Grasshoff (Senckenberg Museum, Frank-
furt), C. E. Griswold (Essig Museum of
the University of California, Berkeley),
W. Icenogle, J- Kaspar, B. J. Kaston, J-
Kochalka, K. Kinman, R. E. Leech, D.
Otte and C. E. Dunn (Academy of Natu-
ral Sciences, Philadelphia), W. Peck (Ex-
line-Peck Collection), N. Platnick (Amer-
ican Museum of Natural History), M.
Robinson, D. A. Rossman, W. T. Sedg-
wick, W. Shear, M. Vachon and M. Hu-
bert (Museum National d'Histoire Natu-
relle, Paris), G. Uetz, J. Unzicker (Illinois
Natural History Survey), F. R. Wanless
(British Museum, Natural Historv), H. K.
Wallace, H. V. Weems and G. B. Edwards
(Florida State Collection of Arthropods),
and T. Yaginuma.

S. Peck prox'ided information on cave
localities, V. Buchsbaum Pearse on the
location of California caves. F. Coyle
read the manuscript and made helpful
coinments. Cecile Villars reworded awk-

Bull. Mus. Comp. Zool., 149(1): 1-74, June, 1980

Biilli'tiit Miiscuni of Companitii c 'Aoolo^ij, \'<)l. 149, No. 1

ward sentences. Susan Hunt mapped the
distril)uti()ns and typed \ arious drafts and
the final manuscript. The research and
publication are supported and made pos-
sible by National Science Foundation
Grant DEB 76-15568.

Publication costs of this study were
covered in part by a grant from the Wet-
more Colles Fund.

RELATIONSHIPS

One of the purposes of this revision
was to learn about the relationships of
Meta and Tetra^natJia. It was for this
reason that Meta and Pachijgnatha and
their relatives were revised together. Me-
cijnogea was first thought close to Meta.
It is clear now that it belongs to the Ara-
neinae.

The characters that were chosen for
studying relationships are the following
(Table 1):

1. Eye placement. In Meta the median
eyes are within their diameter of each
other and about two diameters at most
from the laterals. The laterals touch each
other. This condition is found in Meta
(Fig. 112), Pachy^natha (Figs. 163, 165)
and also in Zygiella (revised previously)
and various small-sized Araneinae, as
well as in Linyphiidae and Theridiidae.
It is believed the primitive condition of
Araneoidea. The lateral eyes have further
separated from the medians in many spe-
cialized Araneidae, e.g. Araneus diade-
matiis and female Nephila (Fig. 34).

The lateral eyes are rarely separated.
Separated lateral eves are characteristic
oi'Azilia (Fig. 294, 300) and Tetragnatha
(and some genera not found in the north
temperate region). It is possible that this
is the ancestral condition, since uloborids
and many other families have lateral eyes
separate.

2. Eye structure. The canoe tapetmn is
characteristic of the Araneoidea. It is
found in Meta and relatives, Zygiella,
NepJjila and in the lateral eyes of PacJi-
ygnatha.

In Araneidae there are two differences.

First, in all genera close to Aranetis (that
is, in most nearctic genera previously re-
vised), the canoe tapetiun of the posterior
median eyes has shrimk, and the rhal)-
doms are arranged in looping rows to-
ward the median (of the spider) (Fig. 11).
In Arfi,i()))e this condition is also foimd in
the lateral eyes. Secondly, in the poste-
rior eyes oi PacliygnatJia the tapetum has
disappeared, and the eye is filled with
looping rows of rhabdoms (Figs. 157,
161) which may resemble a maze (Fig.
159). In Azilia and Tetragiiatha the ta-
petum has been lost in all secondary
eyes, and in Azilia the rhabdoms are ar-
ranged in neat rows which turn at their
ends (Figs. 299, 300). This is without
doubt an improvement in eye structure
(Homann, 1971), as it is accompanied by
larger optic centers in the brain. But it is
strange that a twilight and cave spider,
Azilia, has lost the tapetimi. Glenogna-
tha emertoni has eyes like those oi Pach-
ygnatha with no canoe tapetum in the
posterior median eyes; but here the cells
in the rows of rhabdoms have a tapetum
(Fig. 262).

The shrunk tapetum is found in that
group of spiders with complicated palpi
(see below) having many palpal sclerites.
The tapetum is absent in those spiders in
which the few palpal sclerites present
have become modified. Therefore, it is
believed that evolution went in two dif-
ferent directions from the more general-
ized Meta eyes.

3. Trichobothria. It is characteristic of
the superfamily Araneoidea to have few
leg trichobothria, and one diagnostic fea-
ture is the lack of trichobothria on tarsi
where their presence might interfere
with touching and measiuing silk strands.
In general there are trichobothria only on
the dorsum and sides of tibiae, and usu-
ally there is one on the proximal half of
the metatarsus (some Linyphiidae have
more elaborate patterns of trichobothria).
Specialized trichobothria are found on
the third tibia of Mangora; they are very
long, feathered and of unknown function.
Also the fourth femur of Leucauge has

Mecynogea, Metinae, Pachygnatha and AziLiA North of Mexico • Levi

very fine, long trichobothria of unknown
function (Figs. 50, 51). Leucau^e, Pach-
ijfinatha, Glenognatha and Tetra<^natha
have several trichohothria dorsally at the
proximal end of the femur, which is per-
haps a specialization. Perhaps the pres-
ence of femoral trichobothria in Ulobor-
idae indicates a hitherto unexplored
relationship with Tetrag^natha. In some
adult Pachygnatha, spiders that make no
webs, the trichobothrium has moved to
the distal end of the metatarsus, a unique
change.

4. Male palpi. The simplest palpi in
Araneoidea are found in some tlieridiids
(Levi, 1961) and in MetelliJia and Leu-
caiige. The Metellina and Leucauge pal-
pus has only a (spherical) tegulum with
subtegulum and basal hematodocha, and
bearing on the distal end an embolus
held by the conductor (E, C in Fig. 59).
The only specialized structure in the Me-
tellina palpus is the elaborate paracym-
bium (P in Fig. 124).

The palpus has become specialized in
two different ways: 1) Some species have
acquired numerous sclerites. This is iore-
shadowed in Metleucauge (Fig. 149).
However in Araneus and most relatives
the bulb twists in such a way that the
sclerites move from the tip of the bulb to
the side facing the median. Also the fold-
ed hematodocha and tegulum face later-
ally (Figs. 13, 14). These modifications
are characteristic of the relatives of Ara-
neus; I do not know of any other group
of spiders in which they are seen. Zyg-
iella is intermediate: the tegulum has
moved laterally, the embolus is still dis-
tal in both Z. x-notata and Z. thorelli. 2)
In other genera the cymbium shrinks, the
paracymbium becomes a separate scler-
ite, and embolus and conductor become
more complicated structurally. This pal-
pus is found in Pachygnatha (Fig. 177)
and Tetragnatha.

5. Female geiiitalia. In Meta and Me-
tellina the openings into the seminal re-
ceptacles are so close to the genital
groove that the duct into the seminal re-

ceptacles may also be the fertilization
duct. Wiehle (1967) described this and
called it the semientelegyne condition. I
am not certain this is correct; there might
be two openings with adjacent ducts
(Figs. 116, 120). In other Araneoidea, the
two ducts (connecting and fertilization)
always originate (or end) together on the
seminal receptacles; when they are very
short it may be difficult to tell whether
one or two ducts are present. Again, there
are two specializations: 1) The seminal
receptacles move anteriorly, the epigynal
openings may still only be slits on the
posterior face of the base of the epigy-
num (e.g. Araneus diadematus, Erioph-
ora). Characteristic of this group, related
to Araneus, is the annulate scape. The
scape is not always present but has been
lost in heavily sclerotized species (e.g.
Mastophora, Gasteracantha, Levi,
1978c); it is also absent in Mecynogea
(Fig. 1). It is of interest that some species
of Zygiella have a scape. (These species
belong to Zygiella due to the canoe-
shaped eye tapetum and to a ventral
apophysis or modifications of the tegu-
lum of the palpus, the latter a unique syn-
apomorphy found in most members of
the genus.) 2) The second direction is the
(perhaps) secondary loss of the epigy-
num. In Pachygnatha the anterior ven-
tral cover of the epigynum grows poste-
riorly over the posterior ventral cuticle
and forms two chambers: a posterior
chamber whose dorsal wall is the ventral
cuticle which retains the same setae pat-
tern inside the chamber as outside (Figs.
182, 194). And there is an anterior cham-
ber with the seminal receptacles on each
side at the posterior edge. There appears
to be a connecting duct from the posterior
chamber into the seminal receptacles and
a fertilization duct from the receptacles
to the anterior chamber. The anterior
chamber appears open in the middle to-
ward the posterior one. This looks sur-
prisingly similar to the Meta sections
made and structures shown by Wiehle
(1967, fig. 31). Of course this might be
considered primitive. Further evidence

Bulletin Musciih} of Conipartitive Zoology, Vol. 149, No. 1

>

0)

H^

S

c

S

2

a<

'S

a

(U

a

E

>

.J

*c

U

a;

PQ

O

•<

w.

g

U CL,

cr;

U

y.

cc

0-

y.

H

0

p

1

c

,-^

c

K

CTj

eu

0

p:

a.;

C

r-i

:s

*-<

2

^

<

.^

'"~~^

v:

:/:

Q

OJ

_1;

bi

>.

N

a;

o

J

-3

■*->

<

57

w

"^

u

cu

,

X

W

^

Q

J

^

>

-^-'

^

J

r-

u

^

X

n

J

y:

Q

-U'

z

I-*

_^

<

o

M

H

'i^

-

Z

x'

c/2

sj

U

K

^

+

^

<

^

U

U

z

u

u

eu

o

C/3

— , u

C/2

z

H

h

<
H

X

z

H

g

H

eu

O

^

+1

<

^■^

E

f "1

W

K

O

"— 1

s

u

ii

J

X

cc

w

<■

-J

H

i-

1-

+ 1 +I++I + I++I++II +

+ + I I I I I I I I + I ^•

I I + I I + I + +

o+

.5 I I I I

c ^

=? aj

CL,

0; ^

u c

.0^

c

I I +

I III I I + I I I III I

I I I I I + I I I I I I + I a..

I I I I I I I I 0-.

I I III I +1 + I +1 I 11+11

I I + I ^- I I + I III

+1 I I I + I + I I + I I I + I

1/ .-

^ :/- C ^

■= -t: c

^75

1-

N

|2 0^ I

.t; C' = i! o

.~ ^ ^ 'f- QJ

3 a >. i t;

rt u ii 'n oj

a bc c ij ;-^

1-1 ^ ^T^ -'^

1^ 1;

— y.

1 1

acc

— ^JH a 5^ -5 a7c/2

+ +

+ I

+

+ +1

+ I

+ I

+ I

1; CU C

+ + I I

+ + I +

+ + I I

I + +1 I

+ 111

I I +

I I + I + +

+ I + +

+ 1 + + + I + +

•^ jZ, -y.

u y. ~i ^

'~^ be 73 r^

Qj c a-—

a •;: cb ^ a

hC

^

Oj

^_

^»,

• —

■^

^

c

c

4-*

C

C -D

0;

CD

1;

a

bC

y. ?-

.*-!

*-

-2

"""*

■^

>.

—

i^

-y.

a;

5

o

M

■4-'

c

coxal hook,
modified to

c

E

y.

y

0;

y.
y

0;

1

O

^

*XD

'o

w

O

O

;—

(M

C£>

I- X

05

C — OJ

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi

a
a
t>

l-H

H
Z

o
u

a

+ I

I +

+ 1 CV. 0-.

I O I +

+ I + +

I +

+ I

1; OJ

■* Tf

is ^

I ^- +

+ -->-•

+ + I

M

— ^ —

— 0;

0;
O

1/

Table 2. Cladogram obtained by Wagner

TREE ANALYSIS OF THE DATA IN TaBLE 1 USING
THE COMPUTER PROGRAM "WaGNER 78" DEVEL-
OPED BY J. S. Karris (a later version of Farris,
1970). All chafiacters are weighed equally
(minuses avoided, no zero, pluses arbitrarily
2's, variable characters as intermediate I's).
Prepared by M. Stowe.

Leucauge

Mecynogea

Azi lia

Pachygnatha

Tetragnatha

that the loss of the epigynum is second-
ary is the presence of an epigynum in the
related Azilia (Fig. 291) and Dolichog-
natha and the fact that in copulation only
one male palpus is used at a time. In
Meta and spiders with epigynum, palpi
are alternated; in haplogyne spiders two
palpi are used simultaneously (Gerhardt,
1921).

6. Mating behavior. The diversity of re-
productive organs suggests investigating
mating behavior. Male spiders of the hap-
logyne families use the two palpi simul-
taneously: the male's left palpus enters
the right side of the genital opening, the
right palpus enters the left side when the
spiders mate facing each other. If tetra-
gnathids are really haplogyne, one would
expect a similar copulation. This how-

6 Bulletin Musi'uin of Ci>i)i))nriiliic '/.oology. Vol. 149, No. 1

ever is not the case. Mal(^ tluMidiid and
Iin\ phiid spiders mate 1)\ alternating lett
and right palpi. The left is introduced in
the left side of the epigynum, the right
into the right side (Gerhardt, 1921). Ger-
hardt describes the eopulatory behavior
of \arious araneid genera. Ar^i()f)e and
all those species related to Araneus mate
twice only. After mating once with one
palpus, they separate and court again be-
fore using the second palpus (Gerhardt,
1926). (Gerhardt foimd it remarkable that
the males of species now placed in Nuc-
tcnca and some others can mate three to
foin- times. We now know that, unlike
Araneus or Hiipsosinga species, no scler-
ite breaks off the palpus in Nuctenea to
present further copulations.) Meta is de-
scribed by Gerhardt as using only one
palpus, after which the pair separates
(Gerhardt, 1927). Nephila maculata mates
with one palpus before separating, then
uses the other after a second courting
(Robinson, personal communication).
Pachijg,natha uses one palpus for one
lioiu' at a time before separating (Ger-
hardt, 1921). The length of mating time
is unique for Pachijgnathci; it is only min-
utes in Zygiella, Meta and TetragnatJia.
In Tetragnatha each insertion is 5 to 7
minutes, and within half an hour the pal-
pi alternate (Gerhardt, 1921). Here Meta
and Araneus species share a behavior
which is perhaps primitive and closer to
that of haplogynes, while tetragnathids
eventually alternate palpi, approaching
the perhaps more specialized behavior of
theridiids and linyphiids. Unlike Ara-
neus but like haplogynes, Tetragnatha in
sperm induction use both palpi together
(Gerhardt, 1927).

7. Clypeus. The clypeus is high (more
than twice the diameter of the anterior
median eyes) in most Araneoidea. The
clypeus is low in Metellina and Leu-
cauge (Figs. 47, 72), and in most relatives
of Araneus (but not all, e.g. Kaira, Sco-
loderus). The clypeus is high in Pachy-
gnatha (Fig. 162). Is this high clypeus
secondary (as perhaps in Scoloderus) or

is it a primitix (^ condition? The high clyp-
eus of Kaira and Scoloderus certainly
does not indicate relationship with
PacJiygnatliu.

8. Chelicerae. Chelicerae of Metellina
and Leucauge (Figs. 54, 72) are like those
of other Araneoidea, but probably weaker
and more generalized than those of Ar«-
neus, and possess only a slight proximal
boss. Araneus relatives have a proximal
l^oss, as do the enlarged chelicerae of
PacJiygnatha.

The enlarged chelicerae of the adult,
especially adidt males, are characteristic
of TetragnatJia, Pachygnatha (Figs. 163,
166) and Dolichognatha, but not of Azi/-
ia. The large chelicerae couple when the
animals mate. Enlarged chelicerae are
also found in Metellina curtisi males
(Fig. 73). The enlarged male chelicerae
are a specialization. They are most mod-
ified in Pachygnatha and Tetragnatha.

9. Legs. In contrast, in Araneus and
many related genera the sexes couple
with legs. The first coxae of males have
a hook that fits into a slit on the second
femur. The legs of males are frequently
modified in other ways, but not in the
genera described in this paper. I believe
the modifications of coxae and legs to be
specializations, synapomorphies of the
genera close to Araneus. (However, Azil-
ia males have stronger macrosetae on
legs than females.)

10. Endite tooth. An unusual modifi-
cation found in males o{ Araneus and re-
lated genera (not in any of the genera
considered here) is the lateral tooth on
the endite (gnathocoxa), facing a tooth at
the proximal end of the palpal femur. I
consider the endite tooth and the one on
the palpal femur synapomorphies of the
Araneus group of genera.

11. Labium. The labium is longer than
wide in most haplogyne spiders, proba-
bly a primiti\ e condition. A long labium
is found in Nephila (Fig. 24); in all other
genera of Araneidae and Tetragnathinae
examined, the labium is wider than long.

Mecynogea, Wetinae, PAcmcsATHA AND AziuA North of Mexico • Levi

12. Enditcs. Endites (jfnathocoxae) of
liaplos^Niies are usualK' longer than wide,
perhaps a priniitiv e condition. Of course
most mygalomorph spiders do not have
pedi palpal coxae with endites. Meta
ha\ e long endites which I believe to be
the primitix e type (Fig. 121). Endite form
also appears to ha\e specialized in two
different directions. In Leucauge (Fig.
53), Metleucaiige (Fig. 144), Pachij-
gnatha (Fig. 153) and Tetragnatha the
endite is distalK' wider than proximally,
a synapomorphic state. Perhaps the long
endites e\ ol\ ed with the elongated che-
licerae. The endite o{ Araneus and rela-
tives is as wide as long, very heavy (Fig.
6; Le\ i, 1968, fig. 65), a specialization in
another direction. The Neplula endite is
hea\ y, widest near the tip of the labium
(Fig. 24). The long endites oi Metepeira
are unusual for a genus otherwise close
to Araneus.

13. Orb- webs. Studies of araneoid orb-
webs have been of surprisingly little help
in providing data for phylogenetic stud-
ies (Levi, 1978a, 1978b). Apparendy, ara-
neoid webs, the design of which is im-
portant in determining the kinds of prey
captured, ha\ e undergone frequent and
diverse evolutionary changes as a result
of strong competition for prey (Levi,
1978b). Adult PachijgnatJia species have
no webs; close relatixes in Glenognatha
do. Tetragiiat]}(i and all Meta relatixes
have a loose web with an open hub, per-
haps a synplesiomorph condition. The
web of NepJiila is believed specialized
(Plate 2) and is quite different from that
of the related NephUengijs (M. Robinson,
pers. comm.). The Mecynogea web is
specialized: a dome that lacks viscid silk
(Plate 1); and that of Zygiella lacks viscid
silk in an upper sector of an orb. The
webs of relatives oi Araneus exhibit great
diversity (Levi, 1978b).

14. Silk liandling. An interesting char-
acter was recently found by Eberhard (in
manuscript): species related to Meta and
Tetragnatha measure webs one way;

most species, all related to Araneus, mea-
sure another way; and Nephila measures
with the fointh leg.

15. Ceca. The ventral intestinal ceca
run only to the coxae of legs in Araneus,
Zygiella and Meta. These genera have
one bladder-like cecum in front of the
dorsal apodeme (Palmgren, 1978). Tetra-
gnatha and PacJiygnatha have bulkier
ventral ceca that extend into the chelic-
erae and the pedipalpal coxae, and have
a posterior unpaired dorsal cecum be-
hind the dorsal apodeme. Leucauge has
ceca like Tetragnatha (Palmgren, 1979).

16. Resting position. Living specimens
oi Leucauge (Plate 3), Metellina (Plate 6)
and e\en Pachygnatha (Plate 7) are sim-
ilar and resemble Tetragnatha. They ap-
pear very different from Larinia (Levi,
1975, plate 1), another narrow, elongate
spider that stretches out along twigs.
Striking similarities between Leucauge
and Metellina are the lack of strong setae,
the curved metatarsi, and most important,
the forward resting position of the first
and second femora. The first and second
femora are always held backwards, fem-
oral-patellar joint flexed, in the resting
position of Araneinae (Levi, 1975, plate
1). The resting position of uloborids in
the web is like that of Leucauge and
Meta.

Holding the anterior legs pidled in, in
the Araneinae, may be the primitive con-
dition, as it is the more widespread pos-
ture in the Araneoidea; projecting femora
would represent a specialized posture.
Since the pulled-in legs permit only
small clearance between the leg articles
for the lateral eyes to function, the lateral
eyes touch each other. The great distance
of the lateral eyes from medians in many
Araneinae is no doubt an adaptation to
this resting position, making it easier for
the laterals to function. The separation of
the lateral eyes in Azilia and many Tet-
ragnathinae is made possible by the rest-
ing position: the anterior legs stretch for-
ward, lea\ ing the eyes a larger clearance
in which to function.

8 BiiUctin Mtiseum of Comparative Zoology, Vol. 149, No. 1

17. Silk g,lands. Kovoor (1972) wrote
that Arancus and Ar<i,iope have two kinds
of aciniforni silk ijjhmds, Mcta and Z?/^-
iella [Zilla, sicj only one kind. Aciniforni
glands produce silk used in wrapping
prey.

In summary, it appears that those gen-
era close to Meta have most characters
considered primitixe (Table 1). Those of
Tctra^natlia and Pacliy^tiatha are spe-
cialized in one direction, while those
close to Araneus are specialized in
another. This differs from previous con-
clusions regarding Tetragnathidae as the
most primitive group.

CLASSIFICATION

One easy and fashionable solution to
these phylogenetic problems is to de-
clare that each one of the genera or
groups of genera (Table 1) differs from
the others and has to be placed in its own
family. Unfortunately, this is not a useful
method to solve the problems of relation-
ship, and I prefer instead to place all in
the family Araneidae (perhaps the other
extreme). The family Araneidae can then
be split into three subfamilies. But this
division is not so clear-cut as that of fam-
ilies should be. The oldest name of the
included family group name has to be
used for each of these three taxa: Ara-
neinae, Metinae, Tetragnathinae. Al-
though Thorell (1869) and Simon (1894)
discussed similar problems, more genera
are now known and better imderstood
than were known to them. Although clos-
er to Araneus than to Meta, the following
genera seem to be distinct: Mangora, Ar-
giope, Mecijnogea and CyrtopJiora.
Nephila, which is closer to Meta than to
Araneus, may belong to a special group
of the Metinae. This will be followed up
in future research.

Subfamily Araneinae

Diagnosis. Secondary eyes with ca-
noe tapetimi, posterior median eyes with
the canoe narrow and looping rhabdom

rows toward the median of the spider
(Fig. 11). Lateral eyes often some dis-
tance from medians. [Cyclosa is an ex-
ception to this eye morj^hology, but the
eye description (Homann, 1971) may be
in error.] Palpus with sclerites moved to
median surface of bulb, the folded he-
matodocha and tegulum facing laterally
(Figs. 13, 14). Paracymbium a small hook
at cymbium base. The epigynum often
has an annular scape. Males mate with
one palpus at a time, the pair separating
between copulations. Complete absence
of trichobothria on femora. Chelicerae
very strong with proximal boss. Endites
short and square. Males often with hook
on coxa one and with first or second legs
modified to hold female when mating. In
resting position first and second femora
are directed posteriorly, the femoral-pa-
tellar joint flexed.

Comment. There are several genera
which may be nuclei of subgroups: 1)
Gea and Argiope; 2) Mangora; and 3)
Mecynogea and Cyrtophora.

Subfamily Metinae

Diagnosis. Secondary eyes with ca-
noe tapetum (Figs. 48, 49). Lateral eyes
always close to medians (except in Neph-
ila). Palpus usually very simple with em-
bolus held by conductor, a spherical te-
gulum and subtegulum, other sclerites
usually absent (Figs. 58, 59). Paracym-
bium often elaborate but attached to cym-
bium (Fig. 124). Epigynum very simple;
fertilization ducts may be absent. Males
mate with one palpus at a time, separat-
ing from female before using second. En-
dites long. Webs with open hub (Plates
4, 5, 6). Resting position with first and
second femora extended forward (Plates
3, 6).

Comment. Kaston (1948) mentions
lack of groox es on the sclerites above the
book-lungs. I am not certain whether this
is always so.

Zygiella is best placed with the Meti-
nae, even though some species have a
scape in the epigynum. The unique teg-

Mecynogea, Metinae, Pachygnatha and Aziua North of Mexico • Levi

Ilium of the male palpus of most species
is evidence that those species with and
without scape belong together. Also in-
termediate is the lateral position of the
tegulum in the palpus. The embolus and
conductor however are distal, not mesal.
The leg resting position of Zygiella is
like that of the Araneinae.

Nc])lula may belong in a separate
subgroup or subfamily.

Subfamily Tetragnathinae

Diagnosis. Canoe tapetum lost in at
least posterior median eyes (Figs. 157,
161, 300), usually lost in all secondary
eyes (Fig. 299); rhabdoms arrange them-
selves in looping rows. Palpus similar to
that of Metinae but cymbium reduced,
and paracymbium usually a free sclerite
(Fig. 177) (not in Azilia). Epigynum very
simple or secondarily lost. After using
one palpus repeatedly, the mating male
alternates to use the other palpus. En-
dites long (Figs. 153, 257, 298). Tricho-
bothria on base of femora (not in Azilia).
Male chelicerae usually large, engaged
with those of female when mating. Webs
usually with open hub (not Azilia). The
resting spider has the femora extended
forward.

Comments. The genera Pachygnatha,
Azilia, Mimognatha, Glenognatha, Doli-
chognatha and Tetragnatha belong here,
but Azilia appears close to Leucauge.

RELATIONSHIP TO OTHER
FAMILIES

The genera of Theridiidae closest to
Araneidae are EnoplognatJia and Rob-
ertus (and perhaps Steatoda). Both are
closest to the Metinae because the bulb
has not turned, as in Araneinae, and the
cymbium is not modified, as in Tetra-
gnathinae. The paracymbium is on the
lateral edge of the cymbium, but not at
the base. Both have a colulus. In other
theridiids the paracymbium is a hook on
the upper edge of the alveolus of the
cymbium behind the bulb. Other genera

also lack a colulus. Unlike Meta, the gen-
era Enoplognatha, Roherttis and Steato-
da have diverse sclerites, radix and me-
dian apophysis in the palpus.

At the time I studied the Theridiidae,
I presented a series from Theridula to
Enoplognatha (Levi, 1961) showing the
origin of various sclerites from the simple
palpus, including the origin of the para-
cymbium from a hook behind the bulb
(Levi, 1961, fig. 6). All the spiders with
the simplest palpi lacked a colulus. I
thought that the colulus, foimd in all oth-
er Araneoidea, was perhaps not a primi-
tive structure at all, as postulated, and not
a homolog of the cribellum. Evidence
from Glatz (1973) indicates that the mus-
cles of the cribellum do in fact corre-
spond with those of the colulus even
though Uroctea, closely related to Oe-
cobius (Oecobiidae), lacks a colulus
while Oecobiiis has a cribellum. Adult
male uloborids may lack a cribellum as
well as a colulus. Could neoteny explain
the apparent paradox that the species
with the simplest palpi lack this struc-
ture?

Characteristically all theridiids, with
rare exceptions, have each chelicera
drawn out proximally to a point hidden
by the often transparent clypeus (Levi
and Levi, 1962, figs. 9-21). This point, a
place of muscle attachment, is missing in
Meta, Araneiis and Linyphia (Lexi and
Levi, 1962, figs. 23-28) and in the spe-
cialized theridiid chelicerae of Dipoena
and Euryopis used for ant predation. But
the drawn-out chelicerae are also present
in some other araneoid families such as
Mysmenidae and Nesticidae, and in the
elongated chelicerae of Mimetiis (Mi-
metidae). Only rarely do we see indica-
tions of this drawn-out chelicera in the
Araneidae (Fig. 251). Although the proxi-
mally truncate chelicerae are a good dis-
tinguishing character of the Araneidae,
they are also foimd in Linyphiidae.

While nesticid and mimetid palpi can
l)e derived from a Meta-hke palpus, so
can the very simple palpus of the large
linyphiid Pimoa. If the complicated

10 Bulletin Museum of Comparative Zoolo^ij, Vol. 149, No. 1

palpi of various araneoid families can be
derived from a Mcta and Leucau^e-XWe
palpus, might the ancestors of other ara-
neoid families perhaps have made orh-
webs, and might not the orb perhaps be
the primitive web, the linyphiid and
theridiid web specialized? The other as
yet imsoK ed cjnestion is the relationship
between the Araneoidea and the iilobo-
rids, which, like tetragnathids (but not
other Araneoidea), also have trichoboth-
ria on the fourth femur and rest with the
first two femora stretched forward.

METHODS OF STUDY

Examination Dish. To avoid unnec-
essary transfers of spider parts from a
dish containing sand or from a dish with
depressions in black paraffin (the first has
disturbing reflections, the second insuf-
ficient light to see outlines), a new dish
was made. A stender dish (inside diam-
eter 50 mm, height 17 mm) with a groimd
glass cover was partly filled on one side
with white paraffin available from gro-
cery stores and on the other side with
black paraffin, available from biological
supply houses. To start, black paraffin
was poured into the dish temporarily di-
vided by cardboard. After a few minutes
of cooling, the white was poured in on
the other side and the cardboard re-
moved. While still soft, various size
depressions were made into paraffin to
support spiders.

Measurements. Total length mea-
surements of Pachij^natha were made,
without the chelicerae, measiuing from
clypeus to posterior tip of the abdomen.
About 10 specimens of each sex from dif-
ferent localities were measured to obtain
total length. While going through the
a\ailable collections, measiuements of
noticeably larger and smaller specimens
were also recorded. Often, smaller and
larger specimens were foimd throughout
the range of the species. If there were
geographic size differences, these were
noted.

Genitalia. The female genital area of
PacJiy^natha was cut off, the tissue
around the genital structure removed
with sharpened needles, and the exposed
genitalia submerged in clove oil for clear-
ing and examination. Examination was
done optimally when only partly cleared.
After the clo\ e oil was removed by wash-
ing in alcohol, the genital area was
mounted from alcohol on a microscope
slide into Hoyer's medium and examined
in a Leitz-made Smith interference mi-
croscope.

Eyes. The eyes were examined in re-
flected light for tapetum, then cut off.
The tissue behind was carefully teased
away while placed above a light back-
ground, until the pigment outlining the
rhabdoms became visible. This was done
under a dissecting microscope. The illus-
trations are diagrammatic and composite,
showing the tapetum in reflected light,
the rows of rhabdoms in transmitted
light. The left posterior median eyes are
illustrated looking down vertically into
the eye. The right posterior median is il-
lustrated at its normal angle to the left
eye (Fig. 300). The left lateral eyes are
illustrated looking vertically into the
eyes, as if both were facing the same di-
rection.

Neotypes of Names. Article 75 of the
International Code of Zoological Nomen-
clature (1961) restricts the use of neotype
designation "only in connection with re-
visions in exceptional circumstances
when a neotype is necessary in the inter-
est of stability of nomenclature." Because
of the contusion in the application of
names in the genus PacJiy^natJia, neo-
types were designated for two species in
accordance with qualifying conditions of
Article 75.

Tracheae. The tracheae of Pachy-
^natha and Glenognatha were examined
by mounting the spiders on a microscope
slide after most internal organs were
teased away (some with venter, some
with dorsum facing up). They were not
boiled in sodium hydroxide since these
tracheae are only lightly sclerotized.

Mecynogea, Metwae, Pachygnatha and Azilia North op^ Mexico • Levi 11

Araneinae Latreille, 1806
Mecynogea Simon

Hentzia McCook, 1894, American Spiders, 3: 244.
Type species by monotypy Epeira basilica
McCook. Name preoccupied bv Hentzia Marx,
1883.

Mecynogea Simon, 1903, Ann. Soc. Entom. Bel-
gique, 47: 25. Type species designated by Pe-
tninkevitch, 1928, Trans. Connecticut Acad. Sci.,
29: 135, M. higihba Simon, 1903, from Brazil. The
name is ieminine.

Allepeira Banks, 1932, Pub). Oklahoma Biol. Surv.,
4: 23. New name to replace Hentzia McCook,
preoccupied.

Diagnosis. Mecynogea differs from
the related tropical Cyrtophora in the
closely grouped eyes (Figs. 5, 8), in the
more slender legs (Figs. 5, 7, 12) and in
the shape of the abdomen whose sides
are always more or less parallel (Figs. 5,
6). Mecynogea differs from most other
araneid genera in the close spacing of the
eyes, in the elongated abdomen, and in
the slender, almost linyphiid-like legs
with macrosetae. It differs from Leucauge
and Mangora in its lack of the long trich-
obothria.

Coloration. Carapace light in most
species, with lateral black longitudinal
bands on thora.x and a median longitudi-
nal black line (Fig. 5). Abdomen with
white or silver bands and black marks on
dorsum and sides (Fig. 5). Venter dark,
usually with a pair of longitudinal white
or silver bands which may be broken
(Fig. 6).

Structure. Carapace highest in head
region with a shallow, indistinct thoracic
depression, very narrow in front (Fig. 5).
Height of clypeus less than diameter of
anterior median eyes (Fig. 8). Eyes sub-
equal in size, anterior median eyes some-
times slightly larger than others, espe-
cially in males; anterior lateral eyes
sometimes slightly smaller. Eyes closely
grouped, anterior median eyes their di-
ameter apart, less than their diameter
from laterals (Figs. 5, 8). Posterior medi-
an eyes their diameter or slightly more
apart, the same distance from laterals.
Posterior median eyes with characteristic
narrow canoe tapetum and loops of rhab-

doms as in species related to Araneus
(Fig. 11). Chelicerae strong, endites al-
most as wide as long (Fig. 6). Legs slen-
der, with long macrosetae, but no long
trichobothria (Figs. 5, 7, 12). First leg
longest, third shortest, second and fourth
subequal in length, second leg of male
being slightly longer than fourth. Abdo-
men almost three times as long as wide,
sometimes with anteriodorsal humps,
and with rounded posterior overhanging
spinnerets (Figs. 5-7), widest anterior of
middle (Fig. 5), at a slight angle to pro-
soma (Fig. 7). Males only slightly smaller
than females. No leg articles modified
(Fig. 12).

Genitalia. Female epigynum always
has a projecting hood with a depression
posteriorly containing the openings (Figs.
1-4). There is no scape. The seminal re-
ceptacles appear to be two dumbbell-like
structures grown together (Fig. 3).

Femur of male palpus without pro.xi-
mal tooth (present in most Araneinae),
and the endite lacks the facing tooth. Pal-
pal patella with only one macroseta (Figs.
12, 14). The palpal structures (Fig. 15) are
difficult to homologize with those of oth-
er araneids; they appear close to those of
Cyrtophora and Leucauge. The homol-
ogies of Exline (1948) are almost certain-
ly in error. Mecynogea lemniscata has a
conductor (C in Fig. 15) and, unlike Cyr-
tophora species examined, there is no
median apophysis. The filamentous em-
bolus (E) is held by a structure which is
probably not a terminal apophysis as
thought by Exline (1948). This embolus
is surprisingly similar to that of Leucauge
venusta. Separated above the tegidum
(T) at the base of the part which forms
the embolus is a sclerotized loop, per-
haps the radix. The filiform emliolus of
Cyrtophora moluccensis is also support-
ed over part of its length by a basal out-
growth, and the palpus also has a scler-
otized plate which is the radix. There is
another lobe of uncertain homology
above the embolus (Fig. 15).

Species. All species examined are
very similar, but differ in structure of the

12 Bulletin Musciiin of C<ntijuinttivi' 7AH)lo<s,{f, Vol. 149, No. 1

Plate 1. Mecynogea lemnlscata (Walckenaer), above, female under web with three egg-sacs above; below, detail of
web showing very small mesh. Web with dew.

Mecynogea, Metinae, Pachygnatha and Azi LI a North ok Mexico • Levi 13

Map 1. Distribution of Mecynogea lemniscata (Walckenaer).

genitalia and in the shape of" the egg-sac.
There may be half a dozen to a dozen
species, all American, described in var-
ious genera. Most species are tropical or
south temperate.

Natural History. Mecynogea makes
webs like Cyrtophora (Plate 1) but no
Mecynogea species are known to be co-
lonial. The web is a small-mesh horizon-
tal dome, lacking sticky silk. It has a bar-
rier web above and below (Plate 1 and
Exline, 1948).

Relationship. The canoe tapetum in
the posterior median eyes and the loop-
ing rows ofrhabdoms (Fig. 11) place Me-
cynogea in the Araneinae. Only Cyrtoph-
ora has a similar web, and the genitalia
of Cyrtophora and Mecynogea can be
derived from a common ancestor. The
two genera are not close to each other
however, although there is one species
[M. guianensis (Keyserling)] {=Cyrtoph-

ora granimica Simon) which has inter-
mediate genitalia. Mecynogea is not
close to Araneus and other Araneinae.
The palpal parts can at present not be
readily homologized with those of palpi
of the Araneus group. There is a resem-
blance oi Mecynogea with Mangora: the
shape of the abdomen and its angle to the
prosoma, the many erect macrosetae on
the legs (Fig. 7) and the narrow head.
Whether this is superficial or an indica-
tion of relationship is not presently
known. We will know more about the re-
lationship after the tropical American
Cyrtophora and Mangora have been
studied.

Mecynogea lemniscata (Walckenaer)
Plate 1; Figures 1-15; Map 1

Linyphia lemniscata Walckenaer, 1841, Histoire
Naturelle des Insectes Apteres, 2: 263. Name for
illustration fig. 25, J. Abbot, 1792. Drawings of

14 Bulletin Museum of CUtinixiiddve Zoology, Vol. 149, No. 1

tlu' Insects of Georgia in America in tlic British
Mnsemn, Natural Historx. Photocopy at Museuni
of Coinparati\ c Zooiog)', examined.

Limjphia confertu Hentz, 1850, J. Boston Soc. Nat-
nr. Hist., 6: 30, ph 4, fiti;. 7. Specimens from Ahi-
hama, clestroved.

Epeira basilica McCook, 1878, Proe. Acad. Natur.
Sci., PhiUidelphia, p. 133, figs. 1-3, ?. Specimens
from Colorado Ri\ er, southwest of Austin, Te.xas
in the Academy of Natural Sciences, Philadel-
phia. One specimen from Te.xas marked E])cira
basilica found and labeled hy me as probable
type.

Hentzia basilica: — McCook, 1893, American Spi-
ders, 3: 244, pi. 14, fig. 2, pi. 23, fig. S, 9. Com-
stock, 1913, Spider Book, p. 417, fig. 430; 1940,
re\. ed., p. 431, fig. 4.30. Bonnet, 1957, Bibliogra-
phia Araneorum, 2(3): 2157.

Cyrtophora basilica: — Simon, 1895, Histoire Na-
turelle des Araignees, 1(4): 771.

Allepeira basilica: — Banks, Newport and Bird,
1932, Publ. Univ. Oklahoma, Biol. Surv., 4: 23.
Roewer, 1942, Katalog der Araneae, 2: 778.

Allepeira conferta: — Archer, 1940, Paper Alabama
Mus. Natur. Hist., 14: 24.

Allepeira lemniscata: — Chamberlin and Ivie, 1944,
Bull. Univ. Utah, Biol. Sen, 8(5): 93. Exline, 1948,
Ann. Entomol. Soc. Amer., 41: 309, figs. 1-14,
9, (5, web, egg-sac.

Allepeira affinitata Ki-aus, 19.55, Abhandl. Senck-
enberg. Naturforsch. Gesell., 493: 26, figs. 54-.56,
9, 6. Male holotype from El Salvador in the
Senckenberg Museum, Frankfurt, examined.
NEW SYNONYMY.

Note. 1 follow here Exline (1948), who is the
first rexisor, in using Walckenaer's name lemnis-
cata. The use of any other name for this species,
although perhaps applied with more certainty,
would not help tlie name stability. Abbot described
the web of the spider, on the page facing the illus-
tration, named by Walckenaer.

An old fashioned, unnecessarily detailed descrip-
tion of this species is found in Exline (1948).

Description. Living specimens. Legs,
carapace, sternum greenish. Three lon-
gitudinal black bands on carapace (Fig.
5); sternum black on each side (Fig. 6).
Femora with indistinct dark longitudinal
lines. Median dorsal abdominal band lat-
erally framed by a red line enclosing or-

ange-yellow areas. Anterior of orange
area a white patch; orange area divided
by a median bluish black line and a trans-
verse black mark; posteriorly orange area
grades into median l)luish l)lack (Fig. 5).
A white, wavy longitudinal line on each
side of band; sides of abdomen bluish
green (Fig. 7).

Female from northern Florida. Total
length, 7.2 mm. Carapace, 2.7 mm long,
2.1 mm wide. First femiu-, 3.6 mm; pa-
tella and tibia, 3.8 mm; metatarsus, 3.2
mm; tarsus, L3 mm. Second patella and
tibia, 3.0 mm; third, 2.0 mm; fourth, 3.2
mm.

Male from northern Florida. Total
length, 5.5 mm. Carapace, 2.4 mm long,
L8 mm wide. First femur, 3.7 mm; pa-
tella and tibia, 3.7 mm; metatarsus, 3.6
mm; tarsus, L3 mm. Second patella and
tibia, 3.3 mm; third, 1.9 mm; fourth, 3.0
mm.

Variation. Female from 6.3 to 8.6 mm
total length; carapace, 2.3 to 4.0 mm long,
1.6 to 2.9 mm wide; first patella and tibia,
3.3 to 5.4 mm long. Male from 4.0 to 6.6
mm total length; carapace 1.9 to 3.5 mm
long, 1.4 to 2.4 mm wide; first patella and
tibia, 3.2 to 5.5 mm long. The smallest
specimens came from the northernmost
localities, Delaware and Virginia; the
largest from the Gulf Coast, Mexico to
Panama and Cuba.

Diagnosis. Female. Mecijnogea lem-
niscata lack bands on legs and have only
some indistinct longitudinal lines on the
femora; they differ from other species in
the shape of the depression in the epig-
ynum in posterior view (Figs. 2, 3); the
male differs from other species in the
shape of the sclerite, which holds the em-
l)olus filament and is visible in mesal
view (Fig. 13).

Figures 1-15. Mecynogea lemniscata (Walckenaer). 1-11. Female. 1-4. Epigynum. 1. Ventral. 2. Posterior. 3. Posterior,
cleared. 4. Lateral. 5. Dorsal. 6. Ventral. 7. Lateral. 8. Eye area and chelicerae. 9. Spinnerets. 10. Left lateral eyes. 11.
Posterior median eyes. 12-15. Male. 12. Lateral. 13-15. Left palpus. 13. Mesal. 14. Ventral. 15. Mesal, expanded.

Scale lines. 0.1 mm; except Figures 5-8, 12, 1.0 mm.

Abbreviations. C, conductor; E, embolus; H, hematodocha; T, tegulum.

Mecynogea, Metinae, Pachygnatha and AziLiA North of Mexico • Levi 15

If-) Ihillctiii Museum of Coniixinititc Zoolofiy, Vol. 149, No. 1

Nattiral History. Tlu> (lonic-sliai^cd
web is made in slinihs, usually in decid-
uous forest. Adult males are eolleeted in
various habitats in June and July. A Flor-
ida male was eolleeted in January. An-
derson (1978) reports that the spiderlings
of (\tj;g-saes jModueed "as early as June do
not emerge until the end of Mareh of the
following year. The young are eonfined
within the egj^-sae for 290 days without
aeeess to food." There are records of spi-
ders in urban areas. Many specimens in
Texas were found as i^rey in mud-dauber
wasp nests.

Distribution. Maryland to Missouri,
Florida to Mexico and Panama, Cuba and
Curasao (Map 1). The northernmost lo-
cality is Chesapeake City, Maryland.

Metinae Simon, 1894
Misplaced species

Pseudometa hiolo^ica Cliamherlin, 1925. Bull.
Mus. Comp. Zool., 67: 217. Female holotype from
Pacifie Grove, Hopkins Laboratory, in the Mu-
seum of Comparative Zoology, examined, = Zijfi-
iclhi x-notata (Clerek). NEW SYNONYMY.

Nephila Leacfi

Ni'phila Leaeh, 1815, Zoological Miscellany; being
Descriptions of New and Interesting Animals,
London, 2: 133. Type species liy monotypy Neph-
ila macuhita (Fabricius) from China. The name
is feminine.

Diii'^nosis. Unlike most other araneid
genera the labium of the female is longer
than wide (Fig. 24). Both sexes oi Nephi-
la differ from the related Leucauge in the
lack of femoral trichobothria. Female
Ncpliila differ from Ncphihnigiis in lack-
ing strong setae on the head; also the car-
apace length is shorter than the fburtli
patella and tibia. The long filamentous
embolus wrapped in the conductor and
projecting at right angles from the palpal
axis (Figs. 21, 22, 42, 43) is distinctive for
NepJiild, not present in Nephilengys.
The first patella and tibia is about 1.4
times carapace length, in Nephilengys
only 1.2. The first tarsus is about 1.7
times patella and til)ia length, that of
Nephilengys only 1.2.

Description. Female. Carapace very
wide in front, with a transverse groove as
a thoracic depression (Figs. 33, 36),
sometimes with two horns in posterior
head region anterior to the thoracic
dei:)ression (Figs. 35, 36). An oval depres-
sion between head and thorax in cervical
groove (Figs. 31, 36). Clypeus height
about 1 to 2 diameters of anterior median
eyes (Figs. 23, 34, 35). Eyes subequal in
size, laterals slightly smaller than medi-
ans. Secondary eyes with canoe tapetum
(without loops or rows of rhabdoms), as
in Meta and most Araneoidea, other than
Araneinae and Tetragnathinae* (Figs. 37,
38). Anterior median eyes 1.5 diameters
apart, slightly more than 3 from laterals.
Posterior median eyes more than 1.5 di-
ameters apart, more than 3.5 from later-
als. Laterals on a tubercle, anterior lat-
erals separated by more than their
diameter from posterior laterals (Figs. 33,
36). Chelicerae strong with about 3 teeth
on anterior margin, 3 on posterior (Fig.
24) and a smaller tooth facing tip of fang.
Group of setae covering face of anterior
margin (Fig. 34). Endites and labium
very long (Fig. 24). Metatarsi and tarsi of
legs longer than patellae and tibiae (Fig.
31). First legs longest, second next in
length, third shortest. Metatarsi often
with setae resembling a bottle brush
(Fig. 31). Leg tips have the sustentacular
setae less visible than in other genera
(Levi, 1978b), but they are present. Ab-
domen elongate, cylindrical, longer than
wide (Figs. 19, 20, 31, 32, 41).

Male. Smaller than female, total length
less than half total length of female (Fig.
41). Carapace with some shallow depres-
sion between head and thorax and very
shallow thoracic depression (Fig. 41).
Height of clypeus is slightly less than di-

* Nephild nuiculata has secondary eyes with ca-
noe-shaped tapetum, but the reflecting pigment
crystals are reduced. The main eyes are large with
many rhabdoms, the secondary e>es small with few.
The eyes are not like those of Araneus, Init like
those of Metd (Homann, by correspondence, Oc-
tober, 1979).

Mecynogea, Metinae, Pachygnatha and Az/lm North of Mexico • Levi 17

ameter of anterior median eyes. Anterior
median eyes slightly the largest, less than
their diameter apart, the same distance
from laterals. Posterior median eyes their
diameter apart, 1.6 diameters from later-
als. Anterior and posterior lateral eyes
tonching. There are no leg modifications;
the trochanters may appear notched. In
se\eral species stndied, the males come
in varions size classes (Gerhardt, 1929).

Genitalia. The epigynum consists of
a transverse fold (Figs. 16-18, 28-30) and
a groove in posterior view with the open-
ing on each side ventral in the groove
(Figs. 18, 30). The seminal receptacles
are heavily sclerotized (Figs. 16, 28) and
there is a fertilization duct.

The duct of the palpus is coiled within
the almost spherical tegulum (Figs. 22,
25, 43). The embolus filament is held and
wrapped within tlie conductor (Figs. 25,
26).

Natural History. Nephila females
make a large web with many radii, the
hub located above center (Plate 2), and
with the temporary spiral not removed.
The web and behavior of iV. maculata of
New Guinea have been studied by Rob-
inson and Robinson (1973, 1976).

Species. Perhaps 20 species oi Nephi-
la are found in the tropics of various parts
of the world, only one in the Americas,
N. clavipes. They have been revised by
Dahl (1912), the African species by Be-
noit (1962, 1964). Unfortunately, both pa-
pers lack essential illustrations. The long
paper by Dahl is of limited use and is
out-of-date (freely translated, p. 80):

"In order to facilitate species recognition by
the amateur, I proceeded on the following prin-
ciples: 1) the species are grouped by their occur-
rence in our [Gemian] colonies; 2) I considered
only large females, since smaller ones and males
are overlooked; 3) I have only used color marks.

I do not gi\e figures since especially in spiders
they may be misleading. Unimportant color and
marks usually show up better in illustrations than
those used for recognition of the species."

Relationship. Nephila is known to be
close to Nephilengijs. Although quite
similar in structure, Nephilengys makes
a more conservative orb-web, more like

that of other Araneidae (M. Robinson,
personal communication). Leucauge,
judging by the unusual structure of the
palpus, the base of the embolus located
between tegulum and cymbium (Figs.
25, 26 and 58, 59), is also close to Nejihi-
la.

Nephila maculata Fabricius
Figures 16-22

Aranea maculata Fabricius, 1793, Entomologia
Systematica, 2: 425. [Not A. maculata, Olivier,
1789 = Steatoda alhomaculata (De Geer).]

A. pilipes Fabricius, 1793, Entomologia Systemati-
ca, 2: 425.

Nephila maculata: — Leach, 1815, Zool. Misc., 2:
134, pi. 110. Dahl, 1912, Mitt. Zool. Mus., Berlin,
6: 35, 52. Bonnet, 1958, Bibliographia Araneo-
rum, 2: 3074.

Nephila pilipes: — Roewer, 1942, Katalog der Ara-
neae, 1: 929.

Note. Despite the original homonymy of the
name, the specific name maculata has been used
in almost all literature except Roewer.

The type species of the genus is found from
China and India to the southwestern Pacific Is-
lands. It is common in New Guinea and the Phil-
ippines.

Nephila clavipes (Linnaeus)
Plate 2; Figures 23-43; Map 2

Aranea clavipes Linnaeus, 1767, Systema Naturae,
12th ed., p. 1034. Name given to specimens horn
Jamaica illustrated by Patrik Browne, 1756, The
Civil and Natural History of Jamaica, London, p.
419, pi. 44, fig. 4.

Aranea lon^iimana Fabricius, 1781, Species Insec-
torum, Hamburg et Kilonii, 1. Aran. p. 536. Spec-
imens from Cayenne.

Nephila plumipes: — C. L. Koch, 1839, Die Arach-
niden, 6: 138, pi. 529. Not N. plumipes (Latreille).

Nephila wilderi McCook, 1893, American Spiders,
3: 251, pi. 7, figs. 1, 2, pi. 23, figs. 6, 7, 9, 6.
Syntypes from southern Atlantic and Gulf coasts
to southern California and West Indies, lost. [The
California specimens probably belonged to the
Marx collection.]

Nephila wistariana McCook, 1893, American Spi-
ders, 3: 252, pi. 23, figs. 2, 3, ?, d. Syntypes from
Caribbean Sea, Louisiana and Texas, lost.

Nephila concolor McCook, 1893, American Spiders,
3: 256, pi. 23, fig. 1, 9. Female from southern
California, erroneous G. Marx locality, lost.

Nephila clavipes: — McCook, 1893, American Spi-
ders, 3: 2.55, pi. 24, fig. 1, 9. F. P.-Cambridge,
1901, Ann. Mag. Natur. Hist. (ser. 7), 7: 325, pi.
7, figs. 1-4, 9,6; 1904, Biologia Centrali-Amer-

18

Bulletin Museum of Couiparative y.oolo^ij. Vol. 149, No. 1

Plate 2. Nephila clavlpes (Linn.) web. Web dusted with cornstarch.

icana, Araneidea, 2: 448, pi. 42, fig.s. 23, 24, 9.
Dahl, 1912, Mitt. Zool. Mus. Berlin, 6(1): 46, 73,
75, 9. Comstock, 1912, The Spider Book, p. 426,
figs. 437-441, 9,6, web. Petrunkeviteh, 1930,
Trans. Connecticut Acad. Sci., 30: 247, figs. 101,
102, 9,6. Comstock, 1940, The Spider Book, p.
440, figs. 437-441, 9,6, web. Roewer, 1942, Ka-
talog der Araneae, 1: 932. Bonnet, 1958, Biblio-
graphia Araneorum, 2: 3068. Archer, 1958, Amer.
Mus. Novitates, no. 1958: 2, figs. 1, 2, 9, 6.

Wiehle, 1967, Senckenbergia Biol., 48: 194, fig.
49, 9.
Nephila thomensis Benoit, 1963, Entom. Mitt.,
Zool. Staatsinst. Zool. Mus. Hamburg, 2(41): 5,
figs. 1-3, 9. Female holotype and paratypes from
Sao Tome, Africa, 28.111.1898, in the Zoologisch-
es Staatsinstitut und Zoologisches Museum,
Hamburg, paratype examined. NEW SYNONY-
MY.

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 19

Figures 16-22. Nephila maculata (Fabricius) female from Philippines, male from Java. 16-20. Female. 16-18. Epigynum.
16. Dorsal, cleared. 17. Ventral. 18. Posterior. 19. Ventral, legs removed. 20. Dorsal, legs removed. 21, 22. Left male
palpus. 21. Mesal. 22. Lateral.

Figures 23-27. N. clavipes (Linnaeus). 23, 24. Female. 23. Head and chelicera. 24. Labium and endites. 25-27. Left
palpus expanded. 25. Submesal. 26. Dorsal. 27. Cymbium, ventral.

Scale lines. 1.0 mm; Figures 21-22, 25-27, 0.1 mm.

Abbreviations. C, conductor: E, embolus; T, tegulum; P, paracymbium; S, subtegulum; Y, cymbium.

20 Bulletin Mitsciiiti of Conijxinitivc Zoology, Vol. 149, No. 1

Map 2. American distribution of Nephila clavipes (Linnaeus)

Description. Female from Florida.
Carapace yellow-brown coxered by white
setae; three pairs of dark brown patches;
thoracic groove also dark (Fig. 31). Che-
licerae, endites, labium dark brown with
distal edge of labium and endites light;
sternum orange, darker in middle (Fig.
32). Coxae ventrally yellow-brown, dark
brown anterior and posterior. Legs yel-
low; setae black (Fig. 31). Dorsum of ab-
domen with a dark transverse patch an-
teriorly, a light transverse mark behind,
and pairs of light spots posteriorly (Fig.
31). On venter two parallel longitudinal
light lines with intervening irregular
white marks (bordered by dark) (Fig. 32).
Indistinct white marks in a marbelized
pattern on sides. Living specimens are
greenish. Legs with tufts of hair on distal

ends of femur and tibia of legs 1, 2 and
4, but not always present (Fig. 31). Ab-
domen overhangs spinnerets slightly
(Fig. 32). Total length, 23.0 mm. Carapace,
8.4 mm long, 5.4 mm wide. First femur,
14.2 mm; patella and tibia, 14.4 mm;
metatarsus, 18.0 mm; tarsus, 3.6 mm. Sec-
ond patella and tibia, 11.4 mm; third, 5.8
mm; fourth, 11.0 mm.

Male. Carapace orange-yellow. Dark
marks on each side of head in depression
(Fig. 41). Chelicerae orange-yellow. Ster-
num orange with white pigment in cen-
ter, some dark opposite each coxa. Legs
orange-yellow. Dorsum of abdomen has
a shiny, dark cardiac mark and scattered
small white pigment spots in two longi-
tudinal bands on each side (quite vari-
able). Posterior tip of abdomen dark as in

Mecynogea, Metinae, Pachygnatha andAz/ua North of Mexico • Levi 21

Figures 28-43. Nephila clavlpes (Linnaeus). 28-38. Female. 28-30. Epigynum. 28. Dorsal, cleared. 29. Ventral. 30.
Posterior. 31. Dorsal. 32. Ventral. 33-36. Carapace. 34, 35. Head and chelicerae. 33, 34. (Florida). 35, 36. (Iquitos, Peru).
37. Left lateral eyes. 38. Posterior median eyes. 39, 40. Juvenile. 39. Ventral. 40. Dorsal. 41-43. Male. 41. Dorsal. 42, 43.
Left male palpus. 42. Mesal. 43. Lateral.

Scale lines. 1.0 mm; Figures 37, 38, 42, 43, 0.1 mm.

22 Bulletin Miiscuin of Cojuixirativc Zoology, Vol. 149, No. 1

Leucaiigc. Sides with white dots on or-
ange. X'eiiter ,ti;ra\ in eenter and on eaeh
side ot longitndinal white hnes. Wliite
hnes in tiun made np of white pigment
dots on orange. Total length, 5.8 mm. Car-
apace, 2.4 mm long, 1.7 mm wide. First
femur, 5.3 mm; patella and tibia, 5.0 mm;
metatarsus, 6.8 mm; tarsus, 2.2 mm. Sec-
ond patella and tibia, 4.2 mm; third, 1.9
mm; fourth, 3.3 mm.

Variation. In South American speci-
mens the leg brushes are less distinct.
Most siuprising are the two horns (Figs.
35, 36) present on the carapace of females
from the Amazon Basin; females from Rio
de Janeiro and southwestern Ecuador,
Trinidad and Peru have small horns.

Archer called "the tufted race [with
strong leg-brushes] which occurs in Ja-
maica, other parts of the Caribbean re-
gion. North America, Middle America
and South America N. c. fasciculata (De
Geer)," restricting N. c. clavipes to Cuba
and Hispaniola. This was done despite
the fact that Linneaus named a Jamaican
spider clavipes. Specimens with horns
were placed by Archer in Nephila cor-
nuta (Pallas) and into Archer's new sub-
genus NotJwnepJiihi. Archer ascribes F.
P. -Cambridge's observation of specimens
with small horns to juveniles. However,
Archer's observations are in error and F.
P. -Cambridge and Dahl are correct in
saying there are adult intermediates with
small horns.

In posterior view the median piece of
the epigynum in most specimens other
than those from Florida is wider than the
one illustrated (Fig. 30). The dorsum of
the al)domen may be slightly sclerotized
in larger males (Fig. 41). The sizes of in-
dividuals differ enormously, a variation
which does not appear to be geographic.
Females measured 19 to 34 mm total
length; carapace, 6.8 to 12.1 mm long, 4.9
to 8.6 mm wide; first patella and tibia,
12.5 to 19.4 mm. Southeastern males
measured 4.6 to 10.5 mm total length;
carapace, 1.8 to 4.5 mm long, 1.5 to 3.1
mm wide; first patella and tibia, 3.4 to
10.3 mm. A male from British Guiana was

2.6 mm total length; carapace, 1.4 mm
long, 1.1 mm wide.

Dia^)iOsis. As far as is known this is
the only species of Nephila in the Amer-
icas. Females can readily be separated
from other American orb-weavers and
species of Nephila from other parts of the
world by the abdomen coloration, the
dorsimi having an anterior black trans-
verse bar, followed by both a white bar
which fades out on the sides posteriorly
and by the two irregular rows of spots
(Fig. 31). The venter also has a white bar
posterior to the epigynum (Fig. 32). The
epigynum has cornicular sculpturing on
the ventral median portion, while the
posterior part is smooth, with the open-
ing on each side (Figs. 29, 30). The male
NepJiila can readily be recognized by the
long conductor of the palpus which en-
closes the embolus. The shape of the teg-
idum, the convolutions of the duct inside,
and the long conductor with a twist near
the tip (Figs. 42, 43) are diagnostic for the
species.

Natural History. NepJiila clavipes oc-
curs only in the warmer parts of the
southeastern United States in mesic to
moist deciduous hardwoods and swamps.
In Simberloffs studies of recolonization
of islands whose fauna was destroyed,
Nephila clavipes was the first orb-weaver
and among the first spiders to appear
(Simberloff and Wilson, 1970; Simber-
loff, 1976). Adult males have been col-
lected from May to September in the
southeastern states; one male from Or-
ange Comity, Florida was collected in
December. But males are found in Feb-
ruary and March in the West Indies.
Adult females can be found all year. The
web has been described and illustrated
by Comstock (1912, 1940). It is found at
1 to 6 m heights. It is 60-90 cm in di-
ameter, the hub near the top, and con-
tains many radii (Plate 2). The temporary
spiral is not removed and a barrier web
may be next to it. Only parts of the viscid
spiral are replaced at a time; it is not re-
made daily like that of most orb-weavers.
The silk is yellow. The important behav-

Mecynogea, Metinae, Pachygnatha and Az/lm North of Mexico • Levi 23

ior studies are those of Robinson and
Mirick, 1971, Robinson and Rol)inson,
1974, and Christenson and Goist, 1979.

Distribution. Southeastern United
States to northern Argentina (Map 2).
There are several records from CaHfor-
nia, but all are from the Marx Collection,
whose locality data is often in error.
There are no recent finds from California.
The most northwestern record is Show-
low, Arizona, 28 August 1965, 9 (F. Mat-
zone); the most northern, Orracoke,
Blackbeards Hammock, Hyde County,
North Carolina, 25 August 1976, several
females, males (J. Coddington) and a
sight record "common near the coast in
Beaufort, North Carolina in 1966," and
"in 1976 when I visited Beaufort . . . the
same places were occupied by Argiope"
(P. Weygoldt, in letter). Another interest-
ing record is Sao Tome Island, off Gabon,
Africa.

Leucauge White 1841

Leucau^e White, 1841, Ann. Mag. Natur. Hist., 1(7):
473. Type species by monotypy Linijphia {Leu-
cauge) argyrobapta White from near Rio de Ja-
neiro. The name was first proposed as a subgenus.
The name is feminine.

Argyroepeira Emerton, 1884, Trans. Connecticut
Acad. Sci., 6: 3.31. Type species by monotypy
Epeiru liortorum Hentz (=Leucauge venustci).

Opus O. P.-Cambridge, 1889, Biologia Centrali-
Americana, Araneidea, 1: 184. Type species des-
ignated by F. P.-Cambridge, 1903, O. lugens.

Plesiometa F. P.-Cambridge, 1903, Biologia Cen-
trali-Americana, Araneidea, 2: 438. Type species
by monotypy and by original designation P. ar-
gyra-

Note. The specimens of L. argyrobapta (White)
are lost. The identity of the species is not known.

Diagnosis. Leucauge differs from
Nepliila in the presence of long, feathered
trichobothria, proximally and on the an-
terior surface of the femur of the fourth
leg (Figs. 50, 51, 67). Also the labium of
Leucauge is wider than or as wide as
long, rather than longer than wide (Fig.
53). Close to Leucauge are two genera,
Alcimosphenus (with two species in the
West Indies) and the neotropical Mecij-
nometa. Both also have the rows of trich-

obothria on the fourth femur. Alcimo-
sphenus has more colored pigment and
is darker. Mecijnometa is silver and
black, the abdomen having a tail that ex-
tends beyond the spinnerets, and the
fourth coxae spaced farther apart. It is
possible that these two genera cannot be
kept separate from Leucauge.

Description. Carapace relatively wide
in front, with a deep transverse thoracic
depression, deeper on each side (Fig.
50). A round or oval diagonal depression
between head and thorax in the cei^vical
groove on each side (Fig. 50) (also pres-
ent in Nephila). Eyes subequal in size.
Anterior median eyes about their diam-
eter apart, one to two diameters from lat-
erals. Posterior median eyes about one
diameter apart, 1.5 to 2 diameters from
laterals. Secondary eyes with canoe ta-
petum (Figs. 48, 49, 65, 66). Clypeus
height equals diameter of the anterior
median eye (Figs. 47, 54, 64). Sternum
has some setae, longest setae anteriorly.
Chelicerae strong with about three pairs
of teeth on the anterior, three or lour on
the posterior margin (Figs. 47, 53, 64).
Endites of the palpal coxae are longer
than wide and wider distally than proxi-
mally (Fig. 53). Legs are fairly long, the
first longest, second next, third shortest.
Abdomen longer than wide, sides sub-
parallel with characteristic silver color-
ation, broken in several longitudinal
lines (Fig. 50). Short trichobothria are
found on tibiae and base of femora, none
on metatarsi.

Male. Only slightly smaller than fe-
male, with longer legs. Median eyes
slightly larger than laterals, and eyes
grouped closer together. Clypeus height
equals radius of anterior median eyes.
Chelicerae weaker than female, with two
anterior teeth, two to four posterior. Ex-
cept for increased length, no other leg
modifications. Color is like that of the fe-
male.

Genitalia. Epigynum with two ven-
tral lateral openings leading into sclero-
tized and variable pockets (difficult to
discern) (Figs. 45, 46). Fertilization ducts

24 Bulletin Mttscum of Coinixinitivc y.oolo^ij. Vol. 149, No. 1

Plate 3. Leucauge venusta (Walckenaer), female. Upper left egg-sac, with inner silk orange colored; diameter 7 mm.
Upper right, specimen from Florida; below, specimen from Massachusetts.

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 25

present (Fig. 44). Seminal receptacles
unusually thin-walled (Figs. 44, 60) and
sometimes appear partly digested in al-
coholic collections. Thin-walled recep-
tacles are also found in Azilia; the sig-
nificance of the similarity is not known.

Palpus very simple with a large spher-
ical tegulum containing coiled duct
(Figs. 56-59, 70). Embolus with an ovate
base and a thread-like tip held by a con-
ductor; conductor coiled around thread
(Figs. 58, 59, 69, 71).

Species. Leucauge species are mainly
tropical, world-wide. They differ only
slightly in coloration and in proportions
of the epigynum; and in the shape of the
male bulb and conductor of the palpus.

Natural History. Leucauge makes
large orbs, most horizontal, with many
radii and large numbers of sticky threads
(Plate 4). Some species are colonial.

Relationship. Leucauge has at var-
ious times been placed with Tetragna-
tha. Palmgren (1979) suggests using some
internal characters to show relationship
with Tetragnatha. The short trichoboth-
ria at the base of the femora are also like
those of Tetragnatha. However, other
characteristics indicate that Leucauge is
more closely related to other genera. The
eye structure indicates the genus is close
to Meta; Leucauge is obviously a close
relative of Nephiht. The structure of the
palpus oi Nephihi and Leucauge is very
similar: the base of the embolus hidden
between cymbium and bulb is an unusu-
al specialization found in both genera.
The long endites are characteristic of
both genera, the pair of depressions be-
tween head and thorax characteristic of
both. Male Nephila have the abdomen
shaped like that of Leucauge and also
have a darker posterior end. A third ge-
nus belonging to this group is NepJiilen-
gys. Nephila and Nephilengys differ by
lacking Leucauge'^ rows of trichobothria
on the fourth femur and by having giant
females and dwarf males.

Key to Leucauge Species North of Mexico

1. Females and juveniles 2

Males 3

2(1). Epij^ynuni with median depression (Fij^.
45); middle ot venter of abdomen with tiny
silver specks and two triangular marks (Fig.
51); Ontario, eastern states to California
(Map 3) tenusta

- Epigynum with a median cone (Figs. BI-
GS); venter of abdomen with two longitu-
dinal silver lines, without silver specks in
between (Fig. 67); Florida (Map 3) __ argyra

3(1). Palpal tibia as long or longer than cymbiiun
(Fig. .56); cymbiiun without dorsal hook
(Fig. 57); Ontario eastern states to Califor-
nia (Map 3) lenusta

- Palpal tibia shorter than cymbium (Fig. 68);
cymbium with a dorsal hook (Fig. 69); Flor-
ida (Map 3) argyra

Leucauge venusta (Walckenaer)
Plates 3, 4; Figures 44-59; Map 3

Epeira venusta Walckenaer, 1841, Histoire Natu-
relle des Insectes Apteres, 2: 90. Name given to
Abbot's illustration of Georgian Spiders, p. 13,
fig. 113. Photocopy in the Museum of Compara-
tive Zoology, examined.

Epeira hortorum Hentz, 1847, J. Boston Soc. Natur.
Hist., 5: 477, pi. 31, fig. 19. Females from the
United States, destroyed.

Argyroepeira hortorum: — Emerton, 1884, Trans.
Connecticut Acad. Sci., 6: 332, pi. 37, figs. 29-32,
$, 6. Keyserling, 1893, Spinnen Amerikas, 4:
333, pi. 17, figs. 246. Emerton, 1902, Common
Spiders, p. 192, figs. 446, 447, 9.

Argyroepeira venusta: — McCook, 1893, American
Spiders, 3: 242, pi. 20, figs. 1-6, 9,6.

Leucauge venusta: — F. P. -Cambridge, 1903, Biolo-
gia Centrali-Americana, Araneidea, 2: 441, pi. 42,
figs. 1, 2, 9, c5. Comstock, 1912, The Spider Book,
p. 422, figs. 429, 434-436; 1940, re\ . ed., p. 436,
figs. 429, 434-436, 9,3, web. Roewer, 1942, Ka-
talog der Araneae, 1: 1012. Kaston, 1947, Bull.
Connecticut Geol. Nat. Hist. Sur\ ., 70: 265, figs.
836-837, 84.3-846, 9,6. Bonnet, 1957, Bibliogra-
phia Araneorum, 2: 2477.

Leucauge (Argyroepeira) mabelae Archer, 1951,
Amer. Mus. Novitates, no. 1487: 6, figs. 1, 2, 9,
6 . Male holotype from Sarasota, Sarasota County,
Florida in the American Museum of Natural His-
tory, examined. NEW SYNONYMY.

Description. Female from Florida.
Carapace, sternum, legs orange-brown
when in alcohol, green in life. Labium
and endites with some black pigment.
Dorsum of abdomen silver, made of tiny
coalescing spots; no spots in midline, and
in branches from midline (Fig. 50). On
sides posteriorly, living specimens have
yellow pigment which gives a golden ir-
idescent color. Some also have red pig-

26 Bulletin Mii.si'uni of Contixinitivc Zoology, \'()l. 149, No. 1

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 27

'9*-^^ Vj^ Leucauge argyra

Map 3. Distribution of Leucauge venusta (Walckenaer) and L. argyra (Walckenaer).

ment. Posterior of abdomen black. Venter
with two silver triangles containing fused
pigment spots, often red to orange in li\'-
ing specimens (Fig. 51). More scattered
silver pigment outside the triangles,
green and some black pigment in living
specimens. Total length, 5.9 mm. Cara-
pace, 2.2 mm long, 1.7 mm wide. First fe-
mur, 4.2 mm; patella and tibia, 4.8 mm;

metatarsus, 3.8 mm; tarsus, 1.3 mm. Sec-
ond patella and tibia, 3.7 mm; third, 1.5
mm; fourth, 2.9 mm.

Male from Florida. The abdomen nar-
rows posteriorly, imlike that of the ie-
male. Total length, 5.2 mm; carapace, 2.0
mm long, 1.7 mm wide. First femur, 5.8
mm; patella and tibia, 6.9 mm; metatar-
sus, 6.5 mm; tarsus, 1.6 mm. Second pa-

Plate 4. Leucauge venusta (Walckenaer); webs from Florida, diameter of viscid area about 28 cm. Upper web without
spider; lower web with spider in hub. Web dusted with cornstarch.

28

Bitlli'tin Museum of Conipdrativc 7a)oIo<i,\i, Vol. 149, No. 1

tella and tibia, 4.9 iniii; tliircl, 2.0 mm;
lourtli, 3.5 mm.

Variation. Indi\iduals have diiferent
amounts of black, red and yellow pig-
ment. The xentral silver patches are most
distinct in specimens from southern
states, where another Leucau^e, L. ar-
^yra, occurs (an example of character dis-
placement). The ventral marks are often
obliterated in more northern specimens.
There is considerable variation in size in
most collections. Total length of females,
3.7 to 8.0 mm. Carapace, 1.6 to 2.5 mm
long, 1.2 to 1.9 mm wide; first patella and
tibia, 3.4 to 5.5 mm long. Total length of
males, 3.2 to 5.1 mm. Carapace, 1.6 to 2.2
mm long, 1.3 to 1.7 mm wide; first patella
and tibia, 5.0 to 7.3 mm long. The degree
of sclerotization of the epigynum varies,
as does the relative length of the palpal
tibia.

Diag.nosis. Females can be separated
from L. argijra north of Mexico by the
subtriangular mark on the venter of the
abdomen (Fig. 51, Plate 3) and by the
shallow depression of the epigynum di-
vided into three superficial parts (Figs.
45, 46). Males, unlike L. argijra, have a
long palpal tibia, as long as or longer than
the cymbium (Fig. 56), and lack a dorsal
hook on the cymbium. Males can be sep-
arated from other tropical species by the
shape of the conductor (Figs. 56, 57). Ju-
veniles also have two silver triangidar
patches on the underside of the abdomen
(Fig. 51), while those of L. argyra have
two longitudinal lines. Leucauge venii.sta
is often confused withMangora maculata.
Mangora has the trichobothria on the
third patella and til^ia, Leucauge on the
fourth femur; the genitalia also differ.

Natural History. The webs have
been described bv Emerton (1902), Com-
stock (1913) and Kaston (1947). The webs
are horizontal to near vertical with an ir-
regular barrier web below. The orb has
an open hub, about 30 or more radii, a
wide free zone and more than 60 spirals
(Plate 4). There are no scaffolding
threads left in the web. The tip of the

abdomen is in the center below the open
hub, the first and second legs rest on the
radii of the free zone, third and fourth on
the mesh beyond the open hub. It is not
known how often the web is remade and
at what time of the day or night the spider
spins. The webs are found in low bushes
in wooded areas.

Males are found only from late May to
early July in the northern part of the
range, from April to June in Florida.

An egg-sac of Leiicauge ventista from
Virginia was made in June inside a col-
lecting vial between a curled up leaf and
the glass side, but only loosely attached
to the glass. The egg-sac was made of
loose, fluffy orange-white silk, was 8-9
mm in diameter (Plate 3) and contained
several hundred eggs loosely stuck to-
gether, all reddish orange in color, each
about 0.4 mm in diameter.

Distribution. Eastern United States
(Map 3), New Hampshire to southeastern
South Dakota to central Texas and south
to Panama. Coast of California. Northern-
most records are Hollis, New Hampshire,
August 1888, 9 (Fox) in the Museum of
Comparative Zoology; Longvale, Men-
docino County, California, June 1952, $
(W. J. Gertsch); and Supai, Havasu Can-
yon, Arizona, 2 April 1934, juv. (J. Lutz).

Leucauge argyra (Walckenaer)
Plate 5; Figures 60-71; Map 3

Tetragnatha argyra Walckenaer, 1841, Histoire Na-
turelle des Insectes Apteres, 2: 219, pi. 19, fig. 1,
9 . Walckenaer had specimens from Guadeloupe
in the Antilles which are lost.

LinypJiid atirulcuta C. L. Koch, 1845, Die Arach-
niden, 12: 127, pi. 1049. Female trom St. Thomas.
DOUBTFUL NEW SYNONYMY.

?LinypJiia ornata Taczanowski, 1874, Hor. Soc.
Ent. Ross., 10: 66. Female from Cayenne and St.
Lament de Maron in French Guyana in the Pol-
ish Academy of Sciences, Warsaw.

Meta argyra: — Kevserling, 1880, Verhandl. Zool.
Bot. Gesell. Wien, 30: 563, pi. 16, fig. 12, 9, 6.

Argyrocpeira argyra: — McCook, 1893, American
Spiders, 3: 243, pi. 24, figs. 2, 3. Keyserling, 1893,
Spinnen Amerikas, 4: 343, pi. 18, fig. 253, 9,6.
Simon, 1894, Histoire Naturelle des Araignees, 1:
730, fig. 806, 6.

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 29

Figures 44-59. Leucauge venusta (Walckenaer). 44-53. Female. 44-46. Epigynum. 44. Dorsal, cleared. 45. Ventral. 46.
Posterior. 47. Eyes and chelicerae. 48. Left lateral eyes. 49. Posterior median eyes. 50. Dorsal. 51. Ventral. 52. Head and
chelicera. 53. Labium and endites. 54-59. Male. 54. Eyes and chelicerae. 55-59. Left palpus. 55. Mesal. 56. Ventral. 57.
Lateral. 58, 59. Expanded. 58. Ventral. 59. Dorsal.

Scale lines. 0.1 mm; Figures 47, 50-52, 1.0 mm.

Abbreviations. C, conductor; E, embolus; P, paracymbium; S, subtegulum; T, tegulum.

Plesiometa argyra: — F. P. -Cambridge, 1903, Bio-
logia Centrali-Americana, Araneidea, 2: 438, pi.
41, figs. 15, 16, 9,3. Comstock, 1912, The Spider
Book, p. 424; 1942, rev. ed., p. 438. Bonnet, 1958,
Bibliographia Araneorum, 2: 3709.

Leucauge argyra: — Petrunkevitch, 1930, Trans.
Connecticut Acad. Sci., 30: 265, figs. 119, 120,
9, c5. Roewer, 1942, Katalog der Araneae, 1: 1009.

Note. Liiiyphia aurulenta had previously been
synonymized with Leucauge venusta. But Koch de-

30 Bulletin Museum of Conipdnifive Zoology, Vol. 149, No. 1

Plate 5. Leucauge argyra (Walckenaer), web from Puerto Rico, diameter about 50 cm (photo J. Coddington).

scribes the spider as having a black abdomen witli
golden marks, and the female as having cone-
shaped genitalia. This cannot be L. venusta.

Description. Female from Florida.
Carapace, sternum, legs orange-brown.
Chelicerae dark brown distally, and ster-
num dark brown posteriorly. Dorsum of
abdomen silvery, broken by median lon-
gitudinal and lateral lines and by branch-
es radiating from the median line. Venter
with two silver lines, one on each side,
separated by brown and black pigment.
Living specimens have orange-red
patches (Fig. 67). The first and second
legs have rows of denticles on the venter
of tibiae and metatarsi. Total length, 9.1
mm. Carapace, 3.2 mm long, 2.3 mm
wide. First femur, 8.0 mm long; patella
and tibia, 8.8 mm; metatarsus, 8.8 mm;

tarsus, L9 mm. Second patella and tibia,
6.4 mm; third, 2.4 mm; fourth, 4.8 mm.

Male from Florida. Coloration like fe-
male except silver patches are smaller.
Legs with denticles as in female and pal-
pal cymbium with a large, distally curved
spine. Femur of palpus has a slight hump
distally, facing widest place of endite.
Total length, 6.6 mm. Carapace, 3.0 mm
long, 2.3 mm wide. First femur, 8.4 mm;
patella and tibia, 9.2 mm; metatarsus,
10.0 mm; tarsus, 2.3 mm. Second patella
and tibia, 7.2 mm; third, 2.8 mm; fourth,
4.9 mm.

Variation. Individuals have a vari-
able amount of black pigment. There is
considerable variation in size. Females
vary total length, 4.5 to 10.0 mm; carapace
1.9 to 3.6 mm long, 1.3 to 2.7 mm wide;
first patella and tibia, 4.9 to 9.6 mm long.
Males vary total length, 4.1 to 6.3 mm; car-

Mecynogea, Metinae, Pachygnatha and Aziua North of Mexico • Levi 31

a^

Figures 60-71. Leucauge argyra (Walckenaer). 60-67. Female. 60-63. Epigynum. 60. Dorsal, cleared. 61. Ventral. 62.
Posterior. 63. Lateral. 64. Eye region and chelicerae. 65. Left lateral eyes. 66. Posterior median eyes. 67. Ventral. 68-71.
Male left palpus. 68. Ventral. 69. Lateral. 70, 71. Expanded. 70. Ventral. 71. Dorsal.

Scale lines. 0.1 mm; Figure 67, 1.0 mm.

Abbreviations. C, conductor; E, embolus; H, hematodocha; P, paracymbium; S, subtegulum; T, tegulum.

apace, 1.9 to 3.0 mm long, 1.6 to 2.4 mm cau^e by the cone-shaped epigynum

wide; first patella and tibia, 5.1 to 8.1 mm (Figs. 61-63), the male by the dorsal hook

long. on the cymbium (Fig. 69) and the large

Diagnosis. The female L. argyra is sclerotized sickle-shaped conductor (Figs,

separated from all other species of Leu- 68, 70). Juveniles, unlike those of L. ve-

32 Bullet ill Musciidi oj Conqidnit'n c 'Zoology, Vol. 149, No. 1

nustd, lack siKxM- speckling between the Description. Carapace yellow-white
two longitucliiial \entral lines on the ah- to li^ht Ijrown with gray to brown
donuMi. patches, eye area light without median

Natuidl History. Leucau^c argyra line; sometimes V-shaped marks (Fig.
makes a horizontal web in woods (Plate 75). Sternum black (Fig. 76). Coxae and
5) and mangro\es. It has been collected other leg articles yellow-white, more or
in sugarcane fields in Puerto Rico. Males less distinctly banded (Fig. 75). Dorsum
ha\e been collected in March, November of abdomen white with paired black
and December in Florida. patches, patches widest apart anteriorly,

Distribution. Central and southern more distinct but closer together poste-
Florida, Gulf Coast, West Indies, Mexico riorly (Fig. 75). Venter with a median lon-
to South America, most abundant in the gitudinal l)lack band, bordered on each
West Indies (Map 3). side by a white band (Fig. 76); spinnerets

and ring surrounding them dark with two
white spots on each side (Fig. 76).

Carapace is narrow in front, eyes are
Metellina Chamberlin and I\ ie, 1941, Bull. Univ. closely grouped (Figs. 72, 73, 75). Height

Metellina Chamberlin and Ivie

Utah, biol. ser., 6(3): 14. Type species Metellina of clypeus equals diameter of anterior

curtisi (McCook) by original designation. The ge- median eyes. Eyes subequal in size, an-

nenc name IS leminme. . . i. ,. i- i .i

tenor medians sometunes very slightly

Diafi,nosis. Unlike Meta and Metleu- larger or smaller than others (Figs. 72, 73,

cauge, the Metellina abdomen is longer 87). Anterior median eyes their diameter

than wide, dorsoventrally slightly flat- apart, 1.3 to 2 diameters from laterals,

tened, widest anteriorly sometimes with Posterior median eyes about their diam-

small humps (Fig. 75). The venter eter apart and the same distance to 1.5

has a median longitudinal black band diameters from laterals. Secondary eyes

with a white line on each side (Fig. 76). with canoe tapetum (Figs. 95, 96). Che-

Unlike tropical Chrysometa, there are no licerae fairly strong; three large teeth on

silver spots on the abdomen. The chelic- anterior margin, three on posterior, the

erae have only 3 teeth on the posterior middle one small (Figs. 74, 88). Endites

margin (Fig. 74, 88); Meta and Metleu- longer than wide, widest distally (Fig.

cauge have four. 76). First leg longest, second next, third

The male and female genitalia are shortest. Abdomen oval, dorsoventrally

weakly sclerotized compared to those of slightly flattened, widest anteriorly (Fig.

Meta and Metleucauge. The epigynum 75).

has at most a simple depression with a Short trichobothria are on palpi, leg

lightly sclerotized lip containing the tibiae and metatarsi, none on femur. The

openings on the ventral surface (Figs. 78, tibiae have 6 to 8 trichobothria widely

90, 102, 109). The paracymbium of the spaced dorsally, and one trichobothrium

male palpus is lightly sclerotized (Figs, dorsally on base of metatarsus.
84, 92, 104, 107, 111), as are the conduc- Males may be larger than females (M.

tor and base of eml)olus (C, E in Figs. 99, curtisi) or only slightly smaller than fe-

100, 104, 111). Unlike the tropical Chry- males; the chelicerae may be enlarged

someta the lateral side of the cymbium is {M. curtisi). Males have narrower abdo-

not divided, lobed or indented (Figs. 85, men. They lack the tooth on the proximal

93), except for the basal paracymbium. end of the palpal femur and its matching

Plate 6. Metellina curtisi (McCook). Upper photograph female: lower one web, about 15 cm diameter (lower photo by
D. J. Boe).

Mecynogea, Metinae, Pachygnatha and AziLiA North of Mexico • Levi 33

34 Bulletin Museum of Conijxinitive Zoology, Vol. 149, No. 1

Map 4. Distribution of Metellina curtisi (McCook), Metellina mimetoldes Chamberlin and Ivie and American distribution
of M. segmentata (Clerck). Open circles are cave records of M. mimetoldes.

tooth on the endite. None of coxae or legs
are modified.

Genitalia. The female epigynum has
a lightly sclerotized median ventral
depression (Figs. 78, 90, 102, 106, 109)
with lateral openings leading into heavi-
ly sclerotized seminal receptacles (Figs.
77, 79, 89, 101, 105, 108). Although
Wiehle (1967) could not find fertilization
ducts in the Meta group of species (ex-
amined by sectioning the spiders), I
think there are such ducts, but the lightly
sclerotized flattened ducts may be diffi-
cult to discern.

The male palpi are also lightly sclero-
tized. The conductor (C in Figs. 99, 100)
is a lobe from the tegulum supporting the
distal part of the embolus (E); the base of
the embolus is also a lightly sclerotized,
almost spherical lobe. The paracymbium
is large and has two branches (Figs. 84,
92, 104, 107, 111).

Species. The species, except Euro-
pean M. merianae, are difficult to sepa-
rate. I know of only 5: M. merianae, M.
mengei and M. segmentata in Eurasia;

M. curtisi and M. mimetoides in western
North America. The Eurasian M. seg-
mentata is introduced in British Colum-
bia.

Natural History. The species have a
near vertical web with an open hub
(Plate 6). The American species are
found in dark humid situations and M.
mimetoides is common in caves.

Key to Metellina Species North of Mexico

1. Females 2

Males 4

2(1). Epigynum with a median, bordered depres-
sion (Figs. 78, 80, 82, 90) 3

Epigynum with pair of shallow unhordered
depressions (Fig. 102); British Columbia
segmentata

3(2). Depression usually without median sep-
tum; width of depression about 1.2 times
length (Figs. 78, 80, 82) curtisi

- Depression with median septum and width
of depression about 2.5 times length (Fig.
90) mimetoides

4(1). Tibia of male palpus (when in ventral view)

hardly longer than wide (Figs. 91, 92)

mimetoides

- Tibia of male palpus distinctly longer than
wide (Figs. 83, 84, 104) 5

Mecynogea, Metinae, P AC hygnath a and Azi LI a North of Mexico • Levi 35

5(4). Conductor in ventral view in transverse po-
sition with distal edye shallow, concave

(F'ig. 104); British Columbia segmcntata

Conductor in ventral \iew in transverse and
niesal position, distal edge strongly concave
(Fig. 84) curtisi

Metellina curtisi (McCook)
Plate 6; Figures 72-86; 95-98; Map 4

Epeira peckluimii McCook, 1893, American Spi-
ders, 3: 189, pi. 18, figs. 5, 6, 9, d. Male, female
from Biscayne Bay, Florida (G. Mar.\) [locality in
error] and Wisconsin (G. Peckham). Two female
syntvpes in Academv of Natinal Sciences, Phila-
delphia, examined. NEW SYNONYMY.

Pachy^natha curtisi McCook, 1893, American Spi-
ders, 3: 271, pi. 26, fig. 5, 6 . Fragments of'3 male
syntypes from California in the Academy of Nat-
ural Sciences, Philadelphia, examined and la-
beled as types.

Meta curtisi: — Gertsch and Ivie, 1936, Amer. Mus.
Novitates, 858:20.

Metellina curtisi: — Chamberlin and Ivie, 1941,
Bull. Univ. Utah, biol. ser., 6(3): 15, fig. 20, 9.

'Note. The illustration of the male £. peckhami
is doubtless this species. The epigynum illustrated
does not fit, the body of the female illustrated does,
and so do the two surviving specimens. Since the
species has been cited in the literature as M. cur~tisi
I will keep this name, although E. peckhami has
page priority. According to ICZN Art. 24, the first
reviser can choose the name that provides most sta-
bility.

In collections, specimens had been erroneously
determined as Cyrtophora californiensis by Banks
and Chamberlin; C. californiensis Keyserling is a
Eustala.

Description. Female from Oregon.
Total length, 5.0 mm. Carapace, 1.9 mm
long, 1.5 mm wide. First femur, 2.5 mm;
patella and tibia, 3.3 mm; metatarsus, 2.4
mm; tarsus, 1.2 mm. Second patella and
tibia, 2,5 mm; third, 1.3 mm; fourth, 1.9
mm.

Male from Oregon. Chelicerae en-
larged (Figs. 73, 74). Three teeth on an-
terior margin; on posterior one large
tooth near base of fang and two additional
smaller teeth (Fig. 74). Total length, 4.3
mm. Carapace, 2.2 mm long, 1.6 mm
wide. First femur, 4.0 mm; patella and
tibia, 5.3 mm; metatarsus, 4.2 mm; tarsus,
1.7 mm. Second patella and tibia, 3.8
mm; third, 1.8 mm; fourth, 2.6 mm.

Variation. Males are generally slightly
larger than females. Total length of fe-
males, 3.4 to 5.5 mm; carapace, 1.4 to 2.3
mm long, 1.2 to 1.7 mm wide; first patella
and tibia, 2.7 to 3.7 mm long. The largest
specimen came from Alaska. Males 4.0 to
7.3 mm total length; carapace, 1.8 to 3.0
mm long, 1.5 to 2.1 mm wide; first patella
and tibia, 4.5 to 8.0 mm long.

Diagnosis. Metellina curtisi female
differs from M. mimetoides in the shape
of the epigynum, a depression hardly
wider than long (Figs. 78, 80, 82). The
male has the palpal tibia about twice as
long as wide (Figs. 83, 84), unlike M.
mimetoides, which has tlie tibia hardly
longer than wide.

Natural History. Adult females have
been foimd from January to September;
males from January to August, most com-
mon in April to July. Adults disappear in
fall and early winter in Oregon and Cal-
ifornia. Females are adult in September
in Alaska. Collecting sites are: walls of
shed and shrubs in British Columbia,
rain forest in Washington, and in red-
wood forest in Oregon and California.
Other specimens have been collected in
insect flight trap, in dense forest; small
webs near ground among shrubs and
grasses. Don J. Boe (personal communi-
cation) observed webs in relatively dark
places, out of sunlight, near water 23
miles north of Santa Barbara, and Big Sur,
California 60 to 250 cm above ground.
The webs were 10 to 18 cm diameter and
0 to 10° from vertical. One web was hor-
izontal, had 19 to 25 radii, 11 to 19 spirals
above, and 15 to 23 below the hub. The
hub is open (Plate 6).

Distribution. Pacific coast from
southern Alaska to southern California
(Map 4). The southernmost record is from
San Jacinto Mts., Riverside County, Cal-
ifornia, August 1958 (E. 1. Schlinger);
other unusual records are Minneapolis,
Minnesota, November 1930, 9 (W. J.
Certsch); Wisconsin, syntypes of M.
peckhami (G. W. Peckham); and Horn-

36 Bullet in Museum of Conifxirativi' Zoology, Vol. 149, No. 1

ings Mills, Ontario, 19 June 1927, 9 (T.
B. Kurata).

Metelllna mimetoides Chamberlin and
Ivie
Figures 87-94; Map 4

Metellina mimetoides Chamberlin and Ivie, 1941,
Bull. Univ. Utah, biol. ser., 6(3): 15, fig. 19, 9.
Female holotype from Mount Diablo, California,
in the American Museum of Natural History, lost.

Note. Some specimens of this species had been
determined as Cyrtophora californiensis by Banks.

Description. Female. Female has
more black pigment than Metellina cur-
ti.si. Total length, 5.0 mm. Carapace, 1.8
mm long, 1.5 mm wide. First iemiu', 2.4
mm; patella and tibia, 3.3 mm; metatar-
sus, 2.4 mm; tarsus, 1.0 mm. Second pa-
tella and tibia, 2.7 mm; third, 1.3 mm;
foiuth, 1.9 mm.

Male. The chelicerae of the male are
smaller than those of the male of M. ciir-
ti.si (Figs. 87, 88). Total length, 3.8 mm.
Carapace, 1.9 mm long, 1.6 mm wide.
First femur, 3.3 mm long; patella and tib-
ia, 4.1 mm; metatarsus, 3.5 mm; tarsus,
0.9 mm. Second patella and tibia, 3.2
mm; third, 1.5 mm; fourth, 2.0 mm.

Variation. Total length of females,
3.3 to 6.0 mm. Carapace, 1.6 to 2.7 mm
long, 1.2 to 2.0 mm wide; first patella and
tibia, 2.6 to 4.3 mm long. Total length of
males, 3.1 to 4.9 mm. Carapace, 1.5 to 2.4
mm long, 1.2 to 1.9 mm wide; first patella
and tibia, 3.0 to 5.2 mm long. All the larg-
est specimens came from caves of the
eastern part of the range of the species.
Unlike M. curtisi, the males are slightly
smaller than females and there is much
less variation in size.

Diagtio.sis. Metellina mimetoides fe-
male has a septum in the epigynum
depression, the sides of the depression
directed laterally. The total width of the
depression is more than twice its length
(Fig. 90). Unlike M. curtisi, the male has
the palpal tibia hardly longer than wide
(Figs. 91, 92), while that of M. curtisi is
about twice as long as wide. The male
chelicerae are not much longer than
those of the females and have three

eciually spaced teeth (Figs. 87, 88); in M.
curtisi the chelicerae are much larger
and the space between the distal teeth is
wider than the space between the proxi-
mal. Both sexes are more pigmented and
the markings more contrasting than those
of M. curtisi.

Natural History. This species has
been foiuid in webs on woodpiles and in
shacks, aboard ship in a Los Angeles
dock, imder an overhanging rock ledge,
and in oak-grassland in the Santa Catali-
na Mountains, Arizona. Many of the east-
ernmost records of this species come
from caves (open circles. Map 4). The
males are commonly collected while
sweeping vegetation. I suspect its habitat
is similar to M. curtisi, but its distribu-
tion is more southern.

Distribution. From southern Alaska,
Pacific coast to northern Baja California,
as far west as Utah and in caves from Ne-
vada to Oklahoma, Texas to Sonora (Map
4).

Metellina segmentate (Clerck)
Figures 99-104; Map 4

Araneits segmentatiis Clerck, 1757, Aranei Suecici,
p. 45, pi. 2, fig. 6, 9, 6. Female and male speci-
mens from Sweden.

Aninea reticulata Linnaeus, 1758, Systema Natur-
ae, p. 619. Specimens from Sweden.

Meta reticulata: — Wiehle, 1931 in Dahl, Tierwelt
Deutschlands, 23: 119, figs. 192-197, 9,6. Roew-
er, 1942, Katalog der Araneae, 2: 915.

Meta segmentata: — Locket and Millidge, 1953,
British Spiders, 2: 115, figs. 75, 76B, 77A, D. Bon-
net, 1957, Bibliographia Araneorum, 2(3): 2797.
Locket, Millidge and Merrett, 1974, British Spi-
ders, 3: 64, figs. 36a, 37b,d, 9, d.

Note. All older North American records are from
misidentified specimens of the two natixe species.

Diagnosis. Metellina segmentata dif-
fers from the two other North American
species by lacking a bordered depression
in the epigynum (Fig. 102) and by having
a wide palpal conductor in transverse po-
sition (Fig. 104).

This species can be separated from the
European M. mengei (Figs. 105-107) by
the structure of the seminal receptacles
and the shorter length of the palpal em-
bolus (Figs. 101, 104). The embolus seen

Mecynogea, Metinae, Pachvcvatha and Azilia North of Mexico • Levi 37

78

Figures 72-86. Metelllna curtisi (McCook). 72, 73. Eye region and chelicerae. 72. Female. 73. Male. 74. Male chelicera,
posterior view. 75, 76, Female. 75. Dorsal. 76. Ventral. 77-82. Epigynum. 77, 79. Dorsal cleared. 78, 80, 82. Ventral. 81.
Posterior. 77, 78. (Mendocino County, California). 79-81. (Yamhill County, Oregon). 82. (Friday Harbor, Washington). 83-
86. Left palpus; 83. Tibia, lateral. 84. Ventral. 85. Lateral. 86. Dorsal.

Figures 87-94. M. mimetoides Chamberlm and Ivie. 87. Male eye region and chelicerae. 88. Male chelicera, posterior
view. 89, 90. Epigynum. 89. Dorsal, cleared. 90. Ventral. 91-94. Palpus. 91. Tibia, lateral. 92, Ventral. 93. Lateral. 94.
Dorsal.

Scale lines. 0.1 mm; Figures 72-76, 87, 88, 1.0 mm.

38 BuUctiti Museum of Couiparatitc Zoology, Vol. 149, No. 1

through the semitransparent conductor fusca Wakkemxer i=Meta menardi). The name is

(tiuninjj; and shifting may he necessary) k-munne.

is h)nger in M. nwii^ei (Fig. 107) than in Diagnosis. Unlike Mctcll'nia and

M. .sc^mcntata (Fig. 104). The ventral Metleucau^e, the Mcta ahdomen is al-

lobe of the lateral prong of the paracym- most as high as long, in appearance like

biuni is more distinct in M. men^ei (Fig. that of the theridiid Achaearanea (Figs.

107) than in M. sc^mcntata (Fig. 104). 113-115).

There is little variation of appearance The epigynum has a median swelling

oi' the epi^ynum oi' M. se^mentata (Figs, with openings posterior (Figs. 117-119,

102, 103), hut considerable variation in 129), unlike that of Metellina and Met-

the European M. mengei (Figs. 105, 106). leucauge, which have the openings ven-

Usually the anterior, slightly more dorsal tral. Both the conductor and the embolus

lobe of the seminal receptacle is visible are sclerotized, unlike those oi' Metellina

as a distinct lobe through the translucent species, and the base of the embolus is

epigynum in M. .segmentata (Fig. 102). a complex dissected sclerite with lobes

This lobe is more posterior and dorsal in and apophyses (E in Figs. 124, 125).

M. mengei, and there appears to be one Description. Carapace glossy brown;

elongate seminal receptacle on each side, thorax lighter than head (Fig. 114). Che-

If there is doubt, clearing or dissection Ucerae dark brown. Legs brown, usually

may be needed to see differences of the indistinctly banded darker. Dorsum of

seminal receptacles in dorsal view (Figs, abdomen with paired dark patches and

101, 105). scattered white pigment spots on brown

Natural History. Metellina segmen- to gray background; posteriorly darker

tata is not native to North America; it is vvith transverse light bars; sides streaked

introduced in the Vancouver area. Habits (Figs. 113, 114). Venter gray to black,

are reported in Wiehle (1931) and Locket quite variable, usually with two longitu-

and Millidge (1953). dinal light bands consisting of tiny white

Records. British Columbia. Vancou- pigment spots. Carapace narrow in head

ver, 24 Sept. 1966, 1 9 ; 18 Oct. 1969, 16, region and eyes closely spaced (Fig. 112,

beaten from bushes (P.D.B., R. Leech 114). Height of clypeus equals 1.2 to 1.5

collection); Surrey, Eraser Valley, 18 diameters of anterior median eyes (Fig.

September 1976, 39, wooded area (J. 112). Eyes subequal in size, sometimes

Stafford, MCZ collection). secondary eyes slightly larger than ante-
rior medians. Anterior median eyes

Meta C. L. Koch slighdy less than their diameter apart, 1.2

^ , ^ , ,^,,„ , , 1 • D IT to 1.5 from laterals; posterior median

Meta C. L. Koch, 1836, Arachniden in Panzer, Fan- i • i . i i 4.U 4-U .;,. Ai .,-.T^f^v

nae Insectorum Germaniae initia, Heft 134, pi. eyes slightly less than then diameter

12. Type species by original designation Meta apart, about 1.5 from laterals. Secondary

Figures 95-98. Metellina curtisi (McCook). 95. Left lateral eyes. 96. Posterior median eyes. 97, 98. Left palpus, ex-
panded. 97. Ventral. 98. Subdorsal.

Figures 99-104. M. segmentata (Clerck). 99, 100. Left palpus, expanded. 99. Ventral. 100. Dorsal. 101-103. Epigynum.
101. Dorsal, cleared. 102. Ventral. 103. Posterior. 104. Left male palpus, ventral.

Figures 105-107. M. mengei (Blackwall) of Europe. 105, 106. Epigynum. 105. Dorsal, cleared. 106. Ventral. 107. Palpus,
ventral.

Figures 108-111. M. merianae (Scopoli) of Europe. 108-110. Epigynum. 108. Dorsal, cleared. 109. Ventral. 110.
Posterior. 111. Palpus, ventral.
Scale lines. 0.1 mm.
Abbreviations. C, conductor; E, embolus; T, tegulum.

Mecynogea, Metinae, Fachygnatha andAz//ja North of Mexico • Levi 39

P*'^?^

95

105

^^^^0^0^

109

r I

106

107

^^jP'*^^''^^^*

40 Bulletin MuseuDi of Coniixinitivc Zoology, Vol. 149, No. 1

Meta menardi

/

Meta dolloff \--j'

Map 5. Distribution of Meta dolloff new species and American distribution of M. menardi (Latreiile).

eyes with canoe tapetiim (Figs. 122, 123).
Cheliceiae with 3 teeth on anterior mar-
gin, 4 on posterior. Endites wider on dis-
tal end than on proximal end (Fig. 121).
First legs longest, second next, third
shortest. Abdomen about as high as long,
widest in middle (Figs. 113-115).

No trichobothria on femora, two rows
of trichobothria on dorsum of proximal
half of tibia (altogether 5 or 6). The first
three legs (of Meta menardi) have a sin-
gle dorsal trichobothrium on proximal
end of metatarsus.

Males are slightly smaller than fe-
males. The chelicerae have a larger an-
terior hump at their bases and are more
cui-ved distally. They lack the tooth on
the proximal end of the palpal femin- and
its counterpart on the endite, which is
present in species oi Araneus. No coxae
or legs are modified.

Genitalia. The epigynum has a me-
dian ventral lobe, hairy in Meta menardi
(hairs not shown in Figs. 117-119). The
openings are on the posterior face (Figs.
118-120). According to Wiehle (1967)
there are no separate fertilization ducts.

Wiehle made sections; I am uncertain if
this is right as I could not verify it.

The palpus is sclerotized. The conduc-
tor (C in Figs. 124, 125), a lobe of the
tegulum, holds the filamentous part of
the embolus (E). The embolus has a dis-
sected sclerotized base of different shape
in different species. The paracymbium
(P) is very large, lobed and sclerotized.

Species. Species of Meta are proba-
bly world-wide. There are only two in
North America: Meta menardi and Meta
dolloff.

Natural History. Meta species are
found in dark places, and are common in
caves. The orb- web is vertical with an
open hub.

Key to Species of Meta North of Mexico

1. Epig\nuni with a median notch in anterior
swelling (Fig. 133); palpal ct)nduetor v\'ith
a distal lobe (Fig. 135); paracymbium with
a lateral finger (Fig. 135); California ca\es

(Map 5) dolloff

Epigynum with a median, anterior, round-
ed, hairy swelling (Figs. 117-119); palpal
conductor with narrowed tip at an angle
(Figs. 124, 126); paracymbium in \entral

Mecwogea, M etin ae, Fachvg.vatha andAz/ua North of Mexico • Levi

41

127

Figures 112-127. Meta menardi (Latreille). 112-123. Female. 112. Eye region and chelicerae. 113. Lateral. 114. Dorsal.
115. Ventral. 116-120. Epigynum. 116. Dorsal, cleared. 117. Ventral. 118. Lateral. 119. Posterior. 120. Posterior, cleared.
121. Labium and endites. 122. Left lateral eyes. 123. Posterior median eyes. 124-127. Left male palpus. 124. Expanded,
ventral. 125. Expanded, lateral. 126. Ventral. 127. Lateral.

Scale lines. 0.1 mm; Figures 112-115, 121, 1.0 mm.

Abbreviations. C, conductor; E, embolus; P, paracymbium; T, tegulum.

42 Bulletin Museum of Comparative Zooloiiij, Vol. 149, Mo. 1

view with a keel laterally (Fi^s. 124, 126);
eastern North America, ca\ e-dwelliug and
epigean (Map 5) menardi

Meta menardi (Latreille)
Figures 112-127; Map 5

Aranea menunlii Latreille, 1804, Histoire Natiirelle
Generale des Crustaces et des Insectes, 7: 266.
Specimens described from le Mans, France, lost.

Meta menardi: — Thorell, 1870, Rem. Syst. Europ.
Spiders, p. 38. Emerton, 1884, Trans. Connecticut
Acad. Sci., 6: 328, pi. 34, fig. 18, pi. 37, fig. 33,
9,6. McCook, 1893, American Spiders, 3: 246,
pi. 22, figs. 4, 5, 9, d. Keyserling, 1893, Spinnen
Amerikas, 4: 313, pi. 16, fig. 231, 9, i. Emerton,
1902, Connnon Spiders, p. 190, figs. 443-445, 9,
d. Wiehle, 1927, Z. Morphol. Okol. Tiere, 8: 513,
web. Wiehle, 1931, in Dahl, Tierwelt Deutsch-
lands, 23: 128, figs. 4c, 5b, 205-209, 9,6. Niel-
sen, 1932, Biology of Spiders, 2: 502, figs. 344-
346, webs, egg-sacs. Comstock, 1940, Spider
Book, rev. ed., p. 433, figs. 431-433, 9,6, egg-
sac. Roewer, 1942, Katalog der Araneae, 1: 916.
Kaston, 1948, Bull. Connecticut Geol. Natur.
Hist. Surx'., 70: 223, figs. 700-703, 9,6. Locket
and Millidge, 1953, British Spiders, 2: 119, figs.
78, B, C, 9, 6. Bonnet, P., 1957, Bibliographia
Araneorum, 2: 2787.

Description. Female. Epigynum has
a median hairy bulge (hairs not shown in
Fig. 117). Total length, 12.5 mm. Cara-
pace, 5.8 mm long, 4.5 mm wide. First
femur, 8.9 mm; patella and tibia, 11.2
mm; metatarsus, 9.0 mm; tarsus, 3.0 mm.
Second patella and tibia, 9.4 mm; third,
6.1 mm; fourth, 7.8 mm.

Male. The abdomen is more elongate
than that of female. Total length, 10.0
mm. Carapace, 5.0 mm long, 3.7 mm
wide. First femur, 8.1 mm; patella and
tibia, 10.4 mm; metatarsus, 8.8 mm; tar-
sus, 2.9 mm. Second patella and tibia, 8.2
mm; third, 4.9 mm; fourth, 6.3 mm.

Variation. Females vary in size from
7.8 to 13.7 mm total length; carapace, 4.0
to 5.9 mm long, 3.3 to 4.5 mm wide, first
patella and tibia, 8.2 to 11.5 mm long.
Males vary in size from 7.0 to 10.1 mm
total length; carapace, 3.7 to 5.2 mm long,
3.0 to 4.0 mm wide; first patella and tibia,
8.3 to 11.6 mm long.

Diagnosis. Meta menardi females dif-

fer from M. dolloff and from other
species by tiie median swollen bidge of
the epigyniun (Fig. 117); the bulge is
hairy (not shown in illustrations). The
male differs from M. dolloff and other
species by the shape of the huge median
apophyses (P in Figs. 124, 126). Juveniles
can generally be identified by the ab-
dominal markings: a pair of anterior, dor-
sal dark patches (Figs. 113, 114) together
with two ventral parallel light lines (Fig.
115). The only other species of Meta in
North America, M. dolloff, is found in
California. Meta has been confused with
Zy giella. But the dorsal abdominal pat-
tern of Zygiella species is always differ-
ent, more like that of Metleucauge.

Natural History. Many specimens in
collections come from deep within caves,
around the mouth of caves, and from old
mines and wells; but the species is also
found in dark ravines, under overhanging
rocks, and sometimes in cool cellars.
Males and females appear mature at all
seasons. The web is at a slight angle from
the vertical, to horizontal. Kaston (1948)
reports 8 to 18 radii in an orb 15 to 30 cm
across. The hub is open. The spider
hangs in the hub or at the end of a trap-
line under a nearby rock. The egg-sac is
large, white, drop-shaped, fluffy and
translucent, about 20 mm long. It hangs
on a thread near the web; it is pictured
in Comstock (1940). Wiehle (1931) be-
lieves that this species takes at least two
years to mature.

Distribution. J'^idging from its habi-
tat, the species is probably native. Eur-
asia, eastern Canada and United States
(Map 5) from Newfoundland to Minne-
sota, eastern Oklahoma and south along
the Appalachian Mountains. The west-
ernmost localities are a cave near Rock
Port, Atchison Co., Missouri, 6 Novem-
ber 1904, 30 December 1904 (C. R. Cros-
l)y, Cornell Univ. Coll.), Three Forks
Cave, Cittin Down Mountain, Adair Co.,
Oklahoma, 1 August 1959, 9 (J. Black,
AMNH); southernmost is Baton Rouge,
Louisiana (Gibbeau, Cornell Univ. Coll.).

Mecwogea, Metinae, PACHVG.vArHA andAz/l/a North of Mexico • Levi 43

130

134

Figures 128-131. Meta bourneti Simon (Mediterranean). 128-130. Epigynum. 128. Dorsal, cleared. 129. Ventral.
Posterior. 131. Left male palpus.

Figures 132-135. M. dolloff new species. 132-134. Epigynum. 132. Dorsal, cleared. 133. Ventral. 134. Posterior.
Left palpus.

Scale lines. 0.1 mm.

130.
135.

Meta dolloff new species
Figures 132-135; Map 5

Pseudometa biologica: — Graham, 1967, Caves and
Karst, 9: 17. Not P. biologica Chamberlin (=Zyg-
iella x-notata).

Holotype. Male holotype from Empire Cave, San-
ta Cruz Co., California, 26 Augu.st 1963 (R. Gra-
ham), in the American Museum of Natural His-
tory. The specific name is a noun in apposition
after one of the localities.

Description. Female paratype. Cara-
pace, sternum, labium brown. Coloration
much like Meta meiiardi, leg banding
less distinct. Dorsum of abdomen ante-
riorly with a pair of dark patches; in be-
tween patches are two light triangles,
pointing anteriorly, one after the other;
posteriorly abdomen has dark transverse
bars. Sides streaked. Venter has two par-
allel light lines on dark gray. Head region
of carapace \ ery narrow; thoracic depres-
sion a deep pit. Eyes subequal in size.
Anterior median eyes slightly less than
their diameter apart, 1.3 from laterals.
Posterior medians their diameter apart,
1.4 from laterals. Chelicerae very strong,
with a basal boss. Three teeth on anterior

margin, two together near base of fang.
Posterior margin with three teeth close to
fang, one more distal. Total length, 14.0
mm. Carapace, 6.3 mm wide, 5.0 mm
long. First femur, 9.6 mm long; patella
and tibia, 12.9 mm; metatarsus, 10.3 mm;
tarsus, 3.5 mm. Second patella and tibia,
10.2 mm; third, 6.6 mm; foiu-th, 9.0 mm.

Male. Coloration, eye placement and
cheliceral teeth as in female. Total
length, 11.0 mm. Carapace, 5.9 mm long,
4.5 mm wide. First femur, 10.2 mm long;
patella and tibia, 13.3 mm; metatarsus,
11.7 mm; tarsus, 3.2 mm. Second patella
and tibia, 10.6 mm long; third, 6.2 mm;
fourth, 7.9 mm.

Diagnosis. Meta dolloff females, un-
like those of Meta menardi, ha\e a me-
dian notch at the anterior lip of the epig-
ynum (Fig. 133). The notch is narrower
than that of M. bourneti (Fig. 129). The
paracymbium of the male palpus has a
large ventral finger and a smaller lateral
one (Fig. 135), while that of M. bourneti
(Fig. 131) and M. menardi (Fig. 126) has
only a ventral one. The structure of the
distal projection of the base of the em-

44

Bulletin Mtisctim of Coiiiixirdtivc 'Aoolo'^ij, Vol. 149, No. 1

bolus, the tip of embolus, and the con-
ductor also differ from the two other
species.

Distribution. The only specimens
seen come from two caves in Santa Cruz
Coimt\', California. A. nearby cave in San-
ta Cruz Count\' is known as the Cave of
the Dancing Spiders, but no spiders are
available from it. It is of interest that this
species is very close to Meta hourneti
(Figs. 128-131) of the Mediterranean re-
gion.

Records. Paratypes. California. Santa
Cruz Co.: Empire Cave, Sect. 3, TllS,
R3W, Santa Cruz Quadr., 26 Aug. 1963,
3 9 (R. Graham); Dolloff Cave, Sect. 3,
TllS, R3W, Santa Cruz Quadr., 28 Aug.
1963, 9 (R. Graham).

Metleucauge new genus

Type species. Metleucauge eldorado. The generic
name is feminine.

Diagnosis. The abdomen, unlike that
of Metellina and Meta species, is ovoid,
dorsoventrally flattened, and widest in
the middle (Figs. 142, 143). The epigy-
num is a flat sclerotized structure, unlike
that oi Meta and Metellina, and has pock-
ets on the sides of the venter (Figs. 137,
138). Unlike those of related genera, the
articles of the palpus are elongated, and
the palpal trochanter has a distal spur
(Figs. 145, 151). The paracymbium is re-
duced (P in Figs. 146, 147, 149); the large
tegulum shows the coiled duct, resem-
bling Leiicaiige species; and the expand-
ed palpus shows the presence of conduc-
tor, embolus and an additional sclerite
(Figs. 148-150), unlike Leucauge, Meta
and Metellina.

Trichobothria are absent from the fe-
mur; there are two rows of dorsal tricho-
bothria on basal third of tibiae, and there
is one trichobothrium dorsally on proxi-
mal end of metatarsus.

Males slightly smaller than females,
having elongate palpal articles and a spur
distally on the long palpal trochanter
(Figs. 145, 151). Males do not have coxae
or legs modified.

/ / • ! T" • -r

\ ^ \ \ 1 1

1 \ \ \ 1 1

(TS T ^T"^"' '

1

w • ■• I

1

) --1 1 1 ■

1

/ -^ /%. ^T * - ■ — ">-- ■ - . 1

) •\-^- i \ \ 1 -.,

I —

i r / •• ' i ! ■

"~V i

/ •
i, j

I .'-'—•-7 _,o^--«- ov

C 1

\

/ VA 1 ^ \j ■•M

-\. 1

1.

J — .

\ '^

1

c' ^

*

1

T^'"'"-,-^.^

_ J

^ \ '^

i

AtlCS' \A

" 0

N!--

t^

\

Metleucauge eldorado

\

Map 6. Distribution of Metleucauge eldorado new
species.

Genitalia. The epigynum is a sclero-
tized plate with the openings in slits on
each side on the venter leading into light-
ly sclerotized seminal receptacles (Figs.
136-138). Again, I am uncertain if there
are separate fertilization ducts, but 1
think these are present (Fig. 136).

The palpus differs from that of Meta or
Metellina in having a much smaller para-
cymbium (P in Figs. 146, 147, 149), in
having the duct in the tegulum (T) coiled
as in Leucauge, and in having the con-
ductor or embolus base broken, forming
an additional sclerite (Figs. 148-150). The
homology of this sclerite is not certain.

Description. Carapace brown, thorax
light on each side, rim of thorax and tho-
racic depression darkest (Fig. 142). Che-
licerae brown. Sternum dark. Coxae

Mecynogea, Wetwae, Pachygnatha and Azili a North of Mexico • Levi 45

Figures 136-150. Metleucauge eldorado new species. 136-144. Female. 136-138. Epigynum. 136. Dorsal, cleared. 137.

Ventral. 138. Posterior. 139. Left lateral eyes. 140. Posterior median eyes. 141. Eye region and chelicerae. 142. Dorsal.

143. Ventral. 144. Labium and endites. 145-150. Left male palpus. 145. Ventrolateral. 146. Ventral. 147. Lateral. 148-150.

Expanded. 148. Ventrolateral. 149. Dorsal. 150. Mesal.

Figure 151. M. kompirensis (Bosenberg and Strand), new combination, left palpus, subventral (Japan).

Scale lines. 0.1 mm; Figures 141-145, 151, 1.0 mm.

Abbreviations. C, conductor (uncertain); E, embolus; H, hematodocha; P, paracymbium.

46 Bulletin Musctini of C-oinpaidtii c 'Aoolo^y. Vol. 149, No. 1

li^ht; distal articles of leij;s hrown, fairly mark. Posterior with li^ht transverse

distinctly handed. Dorsum of ahdonien marks. Sides of cardiac mark dark, these

variahle with folium (Fig. 142), venter dark areas with light horders on sides

black with 2 white curved subparallel (Fig. 142). Venter of abdomen hlack with

marks (Fig. 143). two silvery longitudinal lines. Secondary

Carapace narrow in front; eyes placed eyes slightly smaller than anterior medi-

close together. Height of clypeus ranges ans. Laterals on tubercles. Anterior me-

from slightly less than one diameter to dian eyes their diameter apart, about 1.2

1.5 diameters of anterior median eye. An- diameters from laterals. Posterior median

terior median eyes their diameter apart, eyes about 1.5 diameters apart, 1.5 di-

two from laterals. Posterior median eyes ameters from laterals. The height of the

1.3 diameters apart, 1.5 from laterals. Sec- clypeus is less than the diameter of the

ondary eyes with canoe tapetiun (Figs, anterior median eyes. The chelicerae

139, 140). Chelicerae with three large have three teeth on the anterior margin

teeth on anterior margin (Fig. 141), four (Fig. 141), four on the posterior. The ab-

on posterior. Endites much wider distally domen is oval, tapering to a point (Figs.

than proximally (Fig. 144). Abdomen 142, 143). Total length, 9.4 mm. Cara-

longer than wide, widest in middle or an- pace, 4.5 mm long, 3.7 mm wide. First

terior of middle (Figs. 142, 143). femur, 7.0 mm; patella and tibia, 8.9 mm;

Species. Only one species, M. eldora- metatarsus, 7.6 mm; tarsus, 2.4 mm. Sec-

do, is found in North America, in Cali- ond patella and tibia, 6.6 mm; third, 3.8

fornia. Several veiy similar related species mm; fourth, 5.5 mm.

are found in the Far East: M. yiinoha- iMale. Male is darker than female. Tho-

mensis (Bosenberg and Strand) and M. racic depression like female, with two

kompirensis (Bosenberg and Strand) depressions side by side. Anterior lateral

(Fig. 151) of Japan. Yaginuma (1958) al- eyes subequal to anterior medians; pos-

ready pointed out that the two Japanese terior eyes slightly smaller. Chelicerae

species are quite different from the Eu- and endites slightly enlarged. Chelicerae

ropean Meta and ought to be placed in with three teeth on anterior margin, three

a separate genus. on posterior. Trochanter of palpus with

spur (Fig. 145); palpal patella and tibia

,, ^, , . ^ with one macroseta. Both articles elon-

Metleucaugeeldorado new species ^^^^^ ^^^^^1 j^^^ j g^ ^^^^^^ Carapace,

Figures 136-150; Map 6 |^ ^^^^^^ ^^^^^ 3 5 ^^^^^ ^-^^ j,-^.^^ f^„^^r^

Holotype. Female from rock on river banlc, Sequoia 7.6 mm; patella and tibia, 9.4 mm; meta-

National Park, Tulare County, California, 20 July tarsus, 8.4 mm; tarsus, 2.7 mm. Second

1968 (P. Nelson) in the Museum of Comparative p.^tella and tibia, 7.2 mm; third, 3.6 mm;

^""^"^'y- . ^ fourth, 5.4 mm.

Description. Female from King's Variation. Female 8.8 to 11.3 mm to-

Canyon, California. Carapace yellow- tal length; carapace, 3.9 to 5.2 mm long,

brown with median dark, tapering. Ion- 3.3 to 4.2 mm wide. First patella and tib-

gitudinal band, as wide as eye region in ia, 7.6 to 9.6 mm long. Male 7.8 to 11.7

front and enclosing a light streak (Fig. mm total length; carapace, 3.9 to 6.2 mm

142). Margin of thorax with a narrow line; long, 3.1 to 4.8 mm wide; first patella and

on side of thorax a dark "W," light to- tibia, 9.2 to 11.5 mm long,

wards median, fading into sides. Dark Diagnosis. The shape of the epigy-

carapace areas with setae. Dark rings num (Fig. 137), the shape of the sclerites

around eyes. Clypeus light. Chelicerae, in the palpus (Figs. 146, 147) and the

endites, labium brown. Sternum dark, shape of the trochanter tooth (Fig. 145)

lightest in center. Legs banded. Dorsum distinguish the species from the Asian

of abdomen with branching, light cardiac Metleucauge.

Mecynogea, Metinae, Pachygsatha and Azilia North of Mexico • Levi 47

Natural History. Males are almost as
common in collections as females. All
adults have been collected in July and
August. The species makes an orb about
28 cm in diameter between rocks near
streams.

Distribution. California.

Records. Parati/pes. California. El-
dorado Co.: W. ofKvburz, 3 Udv 1978,
9 (D. Boe, MCZ). Glenn Co.: Brittan
Ranch, 18 Mav 1968, 6 (R. Hansen,
CAS). Humboldt Co.: 10 mi E Bridge-
ville, 20 Aug. 1959, 6 (V. Roth, W. T-
Gertsch, AMNH); 2 mi S Blocksburg,
stream under Alderpoint, 19 July 1968,
19 (H. B. Leech, CAS); 2 mi SE Hoopa,
23 Aug. 1969, Id (M. M. Bentzien,
EMUC). Kings Canvon Nad. Park: Cedar
Grove, 16 Julv 1952, 9 (W. ]. Gertsch,
AMNH); 5 Julv 1956, 9, d (V. Roth, W.
J. Gertsch, AMNH); Sequoia Natl. Park:
3 mi W Giant Forest, 9 July 1958, 9 (V.
Roth, W. J. Gertsch, AMNH). Shasta Co.:
Hatchet Creek, 10 mi W Burney, 1 Aug.
1953, 9 (W. J. and J. W. Gertsch,
AMNH); Hazel Creek, 4 July 1952, 9 (W.
J. Gertsch, AMNH). Sierra Co.: The
Cups, Sierra Citv, 6 Sept. 1959, 9, d (V.
Roth and W. J. Gertsch, AMNH). Siski-
you Co.: Ney Springs, 5 mi W Mt. Shasta,
2 Sept. 1959, 6 (V. Roth and W. J.
Gertsch, AMNH). Tulare Co.: Soda
Creek, W of Nelson's Camp, 11 July
1958, 9, 6 (V. Roth and W. J. Gertsch,
AMNH); Mclntire Creek near Camp Nel-
son, 11 July 1958, 9 (V. Roth and W. J.
Gertsch, AMNH).

Tetragnathinae Menge, 1866
Pactiygnattia Sundevall

Pachygnatha Sundevall, 1823, Specimen Academ-
iuni Geneva Araneiduni p. 16. Type species P.
clercki by monotypy. The name is feminine.

Diagnosis. Fachijgnatha differs from
the Meta group of genera in lacking ta-
petum in the posterior median eyes, and
in having the rhabdoms arranged in rows
which loop (Figs. 157, 159, 161). Unlike
Tetraffnatha, the lateral eyes still have
the canoe tapetum (Figs. 156, 158, 160).

Glenofinatha has similar eyes but has the
ventral median spiracle moved anterior-
ly. Padiijgnatlia differs from Tetragna-
tha in having the abdomen ovoid to
spherical (Figs. 171, 184, 244), and differs
from all similar genera in having the ster-
nimi siuroimd the coxae (Fig. 152). Un-
like most Araneinae and Metinae, the
height of the clypeus is equal to 2 diam-
eters of the anterior median eyes (Figs.
154, 155, 162, 163). Like Tetragnatha
and GlenognatJia, female Pachygnatha
lack an epigynimi (copulatory pores).

Description. Carapace with median
longitudinal dark band, and a band from
thoracic depression anteriolateral be-
tween thorax and head (Figs. 171, 184,
196, 208, 232, 244). Indistinct dark patch
on each side of thorax. Chelicerae and
sternum orange to brown; coxae and legs
much lighter, light yellow to orange, not
banded. Abdomen dorsum with a folium
bordered by dark marks with adjacent sil-
ver spots, folium usually with lobes pos-
teriorly. Cardiac mark sometimes out-
lined by dark pigment, or with a median
dark mark; often with paired white spots
(Figs. 171, 184, 196, 208, 232, 244).
Sides stieaked; venter indistinctly marked
(Fig. 172), sometimes having two fine
longitudinal lines of white pigment spots
which bend anteriorly and pass the book-
lung spiracles laterally (Fig. 185); usually
with indistinct dark lines behind gen-
ital opening and dark pigment above gen-
ital area (Figs. 172, 209, 221, 233, 245).
Carapace smooth without setae, l)ut
sometimes with indistinct punctate
sculpturing. Head region set off by a de-
pression on each side in groo\'e between
thorax and head (Figs. 171, 208, 244).
Eyes closely grouped, median eyes their
diameter apart, 1.5 to 2 diameters from
laterals. Height of clypeus 2 to 3 diame-
ters of anterior median eyes. Chelicerae
very large and hea\ y, always with 3 teeth
on anterior margin in female, usually 4
on posterior (3 in P. autumnalis) (Figs.
154, 155, 162). Those of male modified
and expanded, sometimes with a spur
outside, above base of fang (Figs. 215,

48 Bulletin Mitsi'uin of Conqxirativc Zoolo^ij, Vol. 149, No. 1

227). Soiitliern females oi P. furcillatd inu; into the very sliort connectinji; duct to

also lui\ e a spur in this location (Fig. a seminal receptacle. This in turn appar-

165). Labium large, endites very long ently leads into a very short fertilization

(Fig. 153), and sternum surrounding cox- duct to the anterior median chamber. The

ae varies (Fig. 152) slightly in different median chamber appears open toward

species. Sternum may be truncate or the posterior. Pachij^natha tristridta ap-

notched between posterior coxae (Fig. pears to lack seminal receptacles (Figs.

152). All legs rather tliin and lacking mac- 242, 243). The use of only one palpus in

rosetae, but with several trichol:)othria on mating (Gerhardt, 1921) rather than both

femur at base; dorsally, several on tibia simultaneously, as in haplogyne spiders,

and usually one on metatarsus (Fig. 173). is fiuther evidence that the epigynimi is

Abdomen subspherical, slightly flattened secondarily lost.

dorsoventrally (Figs. 171, 184). Book- Species. Pachygnatha species are

lung spiracles rebordered, posterior me- found in Eurasia, Africa and North Amer-

dian spiracle close to spinnerets (Figs. ica. None are known from the American

168, 172, 217, 221). tropics. There are eight species in North

Males are the same size as females, un- America, one of which is shared with

like most araneid spiders. The chelicerae Eurasia. Many European species have

are often enlarged and bent (Figs. 163, been illustrated by Wiehle (1963), others

203, 239). from the Iberian peninsula by Senglet

Genitalia. The palpus is very simple (1972).

but can be derived from a Meta- and Leu- Diagnostic Species Characters. The

caug.e-\ike structure. There is a spherical males of each species are best separated

tegidum (T in Fig. 177); from its tip there by the palpi. The shape of the conductor

is only a long embolus (E) supported by and embolus and the paracymbium are

a conductor (C). The cymbium (Y) is diagnostic in ventral view (Figs. 176, 237,

much modified and narrow, and the para- 249), but the paracymbium shows more

cymbium (P in Fig. 177) is a free sclerite. variation than the embolus tip. Chelicer-

The female genitalia can easily be de- al modifications are variable and, although

rived from the simple epigynum of Mef«. useful, are not the best diagnostic char-

A ventral lobe overhangs the original acters (Figs. 162-167).

openings; a new median opening is pos- Females may be difficult to separate,

terior and separate from the l^ook-lung Chelicerae show considerable variation,

spiracle. This leads into a chamber but are always more elaborate in P. /iir-

whose dorsal roof is a continuation of the cillata; southern furciUata females even

ventral body wall, with continuous seta- have a spur on the anterior face above the

tion (Figs. 169, 170, 182, 183). Anterior of base of the fang (Fig. 165). Trichobothria

this new chamber is a second (perhaps are present or sometimes absent on meta-

the original) opening on each side lead- tarsus (Figs. 173, 210, 222). Since their

Figures 152-161. Morphology of Pachygnatha. 152. Sternum of P. autumnalis. 153. Labium and endites of P. autum-
nalis. 154, 155. Head and chelicerae. 154. P. furcillata. 155. P. autumnalis. 156-161. Secondary eyes. 156, 158, 160. Left
laterals. 157, 159, 161. Posterior medians. 156, 157. P. furcillata. 158, 159. P. autumnalis. 160, 161. P. xanthostoma.

Figures 162-177. P. furcillata Keyserling. 162, 163, 165, 166. Eye region and chelicerae; 164, 167. Left chelicera, pos-
terior view. 162, 165. Female. 163, 164, 166, 167. Male. 162-164. (Central New York State). 165-167. (Florida). 168-173.
Female. 168. Position of genital opening. 169, 170. Internal genitalia. 169. Dorsal. 170. Lateral. 171. Dorsal. 172. Abdomen,
ventral. 173. Left third and fourth metatarsus, dorsal. 174-177. Left male palpus; 174. Ventral. 175. Lateral. 176. Embolus
and conductor, ventral. 177. Expanded, subventral.

Scale lines. 0.1 mm; Figures 152, 154, 155, 162-168, 171-173, 1.0 mm.

Abbreviations. C, conductor; E, embolus; H, hematodocha; P, paracymbium; Y, cymbium.

Mecynogea, Metinae, Pachycsatha and Az/l/a North of Mexico • Levi 49

175

50

Bulletin Mu'ieuin oj Comparatiic Zoolo^ij, \'ol. 149, No. 1

hairs are short, tlu*\' are not as easy to find
as in other groups; tliose ot tlie metatar-
sus are easiest to loeate. The pLicement
of the trieholiotliria is not as eonstant as
might he expected; tliat ot the third meta-
tarsus oi P. furcillata may l)e in tlie mid-
dle of the article or closer to the proximal
end. The placement was first carehdly
measured, hut was found too variahle to
use as diagnostic character.

It is of interest that the two specialized
Pacluj^natlui, P. autuninali.s and P. tris-
triata (judged by the shape of the cara-
pace and by the unusual structure of the
male palpus), also have the trichobothrium
distally on the metatarsus, a imique posi-
tion in the superfamily. The trichoboth-
rium is usually found proximally on the
metatarsus in Araneoidea.

The seminal receptacles are of a dis-
tinct shape in P. autumnalis (Fig. 207),
and the anterior soit parts appear diag-
nostic (Figs. 169, 170, 182, 183, 194, 195).
The genital opening may be in the mid-
dle of the abdomen (Fig. 193), while in
others of the same species more anterior,
depending on the fullness of the abdo-
men. The width of the opening appears
diagnostic, as well as its distance from an
imaginary line between the book-lung
spiracles or from the two small muscle
sclerites between the spiracles (Figs.
181, 193).

Natural History. All Pachyg,natJia
species are found in moist places on the
ground (Plate 7); adidts have lost the abil-
ity to make webs. Pachijgnatha adults
have reduced accessory claws needed for
handling silk. 1 have examined the leg
tips of a less than half-grown P. autum-
nalis and found the accessory claws pres-
ent but almost smooth, appearing like
other setae. Wiehle (1963) cites Apstein,
observing that Pachygnatha lacks aggre-
gate glands, which produce the viscid
droplets on silk.

Balogh (1934) observed young to make
webs. Wiehle (1963) questioned this ob-
servation and cites Homann raising
Pachygnatha to the third molt without
seeing webs. Balogh promised a further

description of the web, but this was nev-
er i:)ublished. \hutin (1978) again reports
finding the web of young Pacliygnatlia
(1.3 to 1.8 mm total length) in Germany.
The diameter is 4 to 6 cm; there are 13
to 17 radii, and an open hub. It is built
horizontally between plants, at a height
of 2 to 8 cm. The web is fine and hard to
see. Young Pachygnatha have more dis-
tinct accessory claws than adults, and
possess aggregate and flagelliform glands
like other ecribellate orl)-weavers. Mar-
tin's specimens examined by Thaler and
Homann (personal communication) were
found to lack tapetum in the posterior
median eyes, thus undoubtedly belong-
ing to PachygnatJia, probably P. degeeri.

Key to Female PAcmcsATHA

1. Posterior median eyes 2 to 3 times diameter
of posterior laterals (Figs. 155, 202); only 3

teeth on posterior margin ot chelicerae

autumnalis

- Posterior median eyes 1 to 1.5 times diam-
eter ot posterior laterals (Figs. 154, 214); 4
teeth on posterior margin of chelicerae 2

2(1). Genital area covered by a thin, triangular
sclerite longer than wide (Fig. 241, 245);
dorsum of abdomen with two curved lon-
gitudinal lines, not wavy behind (Fig. 244)
tristriata

- Genital area without such sclerite (Figs.
229, 233); dorsal abdominal lines wavy be-
hind or broken (Fig. 232) 3

3(2). Total length less than 4.2 mm

xanthostoma

- Total length more than 4.5 mm 4

4(3). Trichobothrium dorsally on fourth metatar-
sus (Fig. 234); Alaska, Canada clercki

- Fourth metatarsus never with trichoboth-
rium (Fig. 173) 5

5(4). Head with bulge on each side (Figs. 154,
162, 165); chelicerae elongated and curved
(Figs. 162, 165); trichobothrium on third
metatarsus dorsal and near middle of article

(Fig. 173) furcillata

Head without bulges (Figs. 178, 190); che-
licerae fairly straight (Figs. 178, 190); third
metatarsus trichobothrium posterodorsal,
twice as far from distal end of article as from
proximal end (Figs. 186, 198) 6

6(5). Genital opening one to 1^2 times its width
distant from two tiny sclerites between
book-lung spiracles (Fig. 181); in middle or
closer to sclerites between book-lungs than
to spinnerets (Fig. 181) brevis

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 51

Plate 7, Above, Pachygnatha autumnalis Keyserling, female; below, P. tristriata C. I. Koch, male.

Genital opening two or more times its width
distant from two tiny sclerites between
book-lung spiracles (Fig. 193); closer to
spinnerets than to book-lungs (Fig. 193) —
dorothea

Key to Male Pachygnatha

Spur on outside of chelicerae above base of

fang (Figs. 215, 227) 2

Chelicerae without such spur (Figs. 166,
203, 239) 4

52 Bulletin Musi'iini oJCoinixirative Zoology, Vol. 149, No. 1

Map 7. Distribution of Pachygnatha species.

Mecynogea, Metinae, Pachygnatha and Az/l;a North of Mexico • Levi 53

2(1). Embolus and conductor length 2 to 3 times

length of spherical tegulum (Fig. 223)

xanthostoma

Embolus and conductor at most 1.5 times
length of tegulum (Fig. 235) 3

3(2). Duct in tegulum undulating (Fig. 235);
paracymbium distally entire (Fig. 236); total
length more than 4.0 mm; Alaska, Canada

clercki

Duct in tegulum straight (Fig. 252); para-
cymbium distally with a notch (Fig. 253);

total length less than 2.0 mm; Florida

calusa

4(1). Posterior median eyes two to three times
diameter of posterior laterals (Fig. 203);
conductor pointed away from cymbium,
notched at tip (Fig. 213); paracymbium
straight and pointed (Fig. 212) __ autumnalis

- Posterior median eyes one and one-half
times diameter to subequal of posterior lat-
erals (Fig. 215); conductor not notched at
tip (Figs. 176, 189, 201); tip of paracymbium
curved (Figs. 175, 188, 200) 5

5(4). Conductor straight, embolus coiled behind
tegulum; length of conductor and embolus
two to three times length of spherical teg-
ulum (Fig. 247); abdomen with 2 curved
longitudinal lines not wavy behind (Fig.
244) tristhata

- Conductor lobed or twisted, embolus not
coiled behind tegulum; length of conductor
and embolus less than twice length of
spherical tegulum (Figs. 187, 199); two bro-
ken lines on dorsum of abdomen wa\ing to-
ward posterior (Figs. 184, 196) 6

6(5). Distal part of paracymbium without neck,
tip pointed toward cymbium (Fig. 175);
head with swelling on each side (Figs. 154,
163, 166) furcillata

- Distal part of paracymbium with narrow
neck, tip pointed away from cymbium (Figs.
188, 200); head without swelling (Fig. 184,
196) - 7

7(6). In ventral view of palpus, base of conductor
not visible above wide embolus (Fig. 189)

brevis

In ventral view of palpus, base of conductor
visible above embolus (Fig. 20\) - dorothea

Pachygnatha furcillata Keyserling

Figures 154, 156, 157, 162-176; Map 7

Pachiignatha furcillata Keyserling, 1884, X'erhandl.
Zool. Bot. Ges. Wien, 32: 662, pi. 21, fig. 11, 9.
Female syntypes from Philadelphia in the Marx
collection of the U.S. National Museum of Natu-
ral History, lost; one female in the Keyserling col-
lection of the British Museum, Natural History,
examined. McCook, 1894, American Spiders, 3:
271, pi. 28, fig. 3, 9. Bryant, 1933, Bull. Mus.

Comp. Zool., 74: 177, pi. 4, figs. 38, 39, 6. Kaston,
1948, Bull. Connecticut Geol. Natur. Hist. Surv.,

70: 267, figs. 842,858, 9, cJ.

Description. Female. Carapace or-
ange with darker lines. Head and eye
area yellow. Chelicerae yellow-orange.
Dorsum of folium widi angular lobes pos-
teriorly (Fig. 171). Head with a bulge on
each side (Figs. 154, 162, 165, 171); in
dorsal view, head outline not constricted
behind (Fig. 171). The genital opening is
narrow, a distance of about twice its
width from the two muscle sclerites be-
tween the book-lungs (Fig. 168), closer to
book-lungs than to spinnerets. Tricho-
bothrium on middle of third metatarsus,
none on foiuth metatarsus (Fig. 173). To-
tal length, 6.6 mm. Carapace, 3.2 mm
long, 2.2 mm wide. First femur, 3.2 mm
long; patella and tibia, 4.0 mm; metatar-
sus, 2.7 mm; tarsus, 1.4 mm. Second pa-
tella and tibia, 3.5 mm; third, 2.2 mm;
fourth, 3.0 mm.

Male. Three posterior cheliceral teeth
closely spaced, one close to base of fang
(Fig. 167). Total length, 5.5 mm. Cara-
pace, 2.7 mm long, 2.0 mm wide. First
femur, 3.3 mm; patella and tibia, 4T mm;
metatarsus, 2.7 mm; tarsus, 1.4 mm. Sec-
ond patella and tibia, 3.7 mm; third, 2.2
mm; fourth, 3.2 mm.

Variation. Female total length, 5.2 to

6.2 mm; carapace, 2.4 to 2.9 mm long.
Male total length, 4.6 to 5.9 mm; carapace,

2.3 to 2.9 mm long. The smallest speci-
mens come from the southeastern states.
Adult and juvenile females from the
southeastern states may have a spur
above the fang base on the chelicerae
(Fig. 165); the fang is minute in some,
absent in northern specimens. (The eyes
appear larger in Fig. 165 than in Fig. 162
because of the size difference: the small-
er southern specimen has relatively larg-
er eyes.)

Diagnosis. Females of this species
are readily identified by their light head
with bulges on each side (Figs. 154, 162,
171)., Females from the southeastern
states may have a spur on the chelicerae

54

Bulletin Museum oj Coiuparatiie Zuulu^y, \'ol. 149, No. 1

alK)\e the l)ase of the lanj:; (Fi,u;. 165).
Nhiles are identified h\ tlie kiK)hl)ed tip
and lateral lobe on the tip of eondnetor
(Figs. 174, 176) and by the tip oi the para-
cynibivun pointed toward the eymbium
(Fig. 175).

Natural History. ■ The speeies has
been colleeted from a cedar bog in Ohio,
a bog in Massachusetts, bottom land
hardwood in North Carolina.

Distribution. New England west to
Indiana, south to Florida (Map 7).

Pachygnatha brevis Keyserling
Figures 178-189; Map 7

Pacltt/gnatha tristriuta: — Keyserling, 1882, Ver-
handl. Zool. Bot. Ges. Wien, 32: 209 (not P. tris-
triuta C. L. Koch).

F. brevis Keyserling, 1883, Verhandl. Zool. Bot.
Ges. Wien, 33: 658. New name for P. tristriatci: —
Keyserling, 1882, misidentified. Male specimen
marked "Amer. N." in the Museum National
d'Histoire Naturelle, Paris, e.xamined and labeled
lectotvpe. Emerton, 1884, Trans. Connecticut
Acad.'Sci., 6: 336, pi. .34, fig. 21, pi. 40, figs. 8, 10,
9,6. PMcGook, 1894, American Spiders, 3: 267,
pi. 26, figs. 9, 10a, pi. 28, fig. 2, 9, 6. Comstock,
1912, The Spider Book, p. 407, figs. 101, 418, 6.
Kaston, 1948, Bull. Connecticut Geol. Natur.
Hist. Surv., 70: 267, figs. 841, 857, 9,6.

Note. A collection labeled in pencil Pachygna-
tha tristriata C. L. Koch, Amer. N., no. 2942 was
found within the collection of the Paris museum. In
the vial was one shriveled P. tristriata female, two
male P. clercki, and one pair of P. brevis. The male
of the two was marked lectotype, the female, para-
lectotype.

Description. Female. Carapace with
dark band enclosing median eyes (Fig.
184). Dorsum of folium lobed in posterior
part (Fig. 184). Carapace has a constric-
tion between head and thora.x (Fig. 184)

in dorsal \ iew. The genital opening is a
distance about ec^ual to its width or one
and one-half times its width from the two
sclerites between the book-lungs; it is
ecjuidistant or closer to sclerites than to
spinnerets (Fig. 181). Third metatarsus
has a trichobothrium one third distance
from proximal end in slightly posterior po-
sition; the fourth lacks a trichobothrium
(Fig. 186). Total length, 5.6 mm. Cara-
pace, 2.7 mm long, 1.9 mm wide. First
femur, 2.7 mm; patella and tibia, 3.5 mm;
metatarsus, 2.3 mm; tarsus, 1.2 mm. Sec-
ond patella and tibia, 3.2 mm; third, 1.9
mm; fourth, 2.7 mm.

Male. Posterior of the chelicerae with
four teeth, the distal three almost equalh
spaced (Fig. 180). Total length, 5.5 mm.
Carapace, 2.7 mm long, 2.0 mm wide.
First femur, 2.8 mm; patella and tibia, 3.6
mm; metatarsus, 2.4 mm; tarsus, 1.2 mm.
Second patella and tibia, 3.5 mm; third,
2.0 mm; fourth, 2.7 mm.

Variation. Female total length 4.6 to
5.5 mm; carapace 2.4 to 2.6 mm long.
Male total length 4.5 to 5.6 mm; carapace
2.2 to 2.6 mm long.

Diag.nosis. This species is easily con-
fused with P. dorothea; females are dif-
ficult to separate. The females can be dis-
tinguished by the wide genital opening,
equidistant from spinnerets and book-
lungs or closer to book-lungs (Fig. 181).
The male can be separated from P. do^
rothca by the heavy embolus (Fig. 189).

Natural History. Specimens have
been collected from swamps in New En-
gland, salt marsh in Maryland, and in an
asparagus bed in Massachusetts.

Distribution. Nova Scotia, New En-

Figures 178-189. Pactiygnatha brevis Keyserling. 178-180. Eye region and chelicerae. 178. Female. 179, 180. Male. 180.
Left chelicera, posterior. 181-186. Female. 181. Position of genital opening. 182, 183. Female genitalia. 182. Dorsal. 183.
Lateral. 184. Dorsal. 185. Abdomen, ventral. 186. Third metatarsus and tarsus, dorsal. 187-189. Left male palpus. 187.
Ventral. 188. Lateral. 189. Embolus and conductor, ventral (embolus black).

Figures 190-201. P. dorothea McCook. 190-192. Eye region and chelicerae. 190. Female. 191, 192. Male. 192. Left chelicera,
posterior. 193-198. Female; 193. Position of genital opening. 194, 195. Female genitalia. 194. Dorsal. 195. Lateral. 196.
Dorsal. 197. Abdomen, ventral. 198. Third metatarsus and tarsus, dorsal. 199-201. Male palpus. 199. Ventral. 200. Lateral.
201. Embolus and conductor, ventral.

Scale lines. 0.1 mm; Figures 178-181, 184-186, 190-193, 196-198, 1.0 mm.

Mecysogea, Metinae, Pachygsatha and Azilia North of Mexico • Levi 55

201

«r-i'

186

y

198

w

y

56 Bulletin Musciiin oj Comparative Zoology, Vol. 149, No. 1

gland to Virj^inia west to Indiana (Map posterior margin in two groups of two

7). The westennnost localities are P'ort (Fig. 192). Total length, 5.5 mm. Cara-

Wayne, Indiana, 9, 6,3 Ang. 1964 (W. pace, 2.4 mm long, 1.9 mm wide. First

Ivie) and Sable River, Ontario, 6, Ang. femnr, 2.6 mm; patella and tibia, 3.5 mm;

1963 (C]. B. Powell). metatarsus, 2.2 nnn; tarsns, 1.2 mm. Sec-
ond patella and tibia, 3.2 mm; third, 2.0

Pachygnatha dorothea McCook '"'"' i<)'ntb, 2.4 mm

Figures 190-201; Map 7 Variation. Females, total length, 5.0

to 6.2 mm; carapace, 2.5 to 2.7 mm long.

Pachy^mitha dorothea McCook, 1893, American Males, total length, 4.7 to 6.0 mm; cara-

Spulers 3: 270, pi 26, figs. 3 4, 9, d. Female 2.3 to 2.7 mm long,

and male svntypes trom neighborhood oi rhila- r-» • • t-i • i /•

delphia inthe Academy of Natural Sciences, Diagnosis. The species has oiten

Philadelphia, lost. Male neotype here designated been called P. hrevis; the two are very

from Ithaca, New York (N. Banks) in the Museum similar and females are difficult to sepa-

ot Comparative Zoology. rate. The female can best be identified

Pacluwiiatha kuratai Levi, 1951, Amer. Mus Nov- i .i ■. ^ • i • i • i j.

itates, 1501; 15, figs. 29-31, 9, 6. Male holotype \y ^^e genital openmg which IS closer to

from Toronto, Canada in the American Museum the spinnerets than to the sclerites be-

ofNatur;ilHistor>, examined. NEW SYNONYMY tween the book-lung spiracles (Fig. 193).

TSIote. McCook illustrated a right palpus, which The genital opening is narrower than that

shows by the base of the conductor that it is clearly of P. hrevis. The males have a more sleii-

this species. The new type locality is a place where der conductor and embolus than that of

this species is common and from which there are p /^^^.i;j,^.. i^ ventral view the base of

numerous collections. .i j . • • -i i /XT'- i nrv om \

the conductor is visible (Fig. 199, 201).

Description. Female. Carapace or- Natural History. This species has

ange with median longitudinal dark line been collected from lake shores, river

enclosing median ocular quadrangle banks, flood plains; in skunk cabbage,

(Fig. 196). Dorsum of abdomen with dark pitcher plants, and under the bark of

spots on each side (Fig. 196). Genital trees.

opening is the distance of two to three Knutson (1979) reports that 40 out of
times its width from the sclerites be- 150 skunk cdhhage {Symplocarpus foeti-
tween the book-lung spiracles; it is closer diis) flowers growing through a foot of
to spinnerets than to book-lungs (Fig. snow in April in Michigan were occupied
193). Trichobothrium of third metatarsus by one or more P. hrevis [P. dorothea,
one third its length from proximal end .s?r]. There were no webs and none of the
(Fig. 198). Total length, 5.0 mm. Cara- spiders was feeding, although there were
pace, 2.4 mm long, 1.7 mm wide. First insects on the warm plants that had melt-
femur, 2.4 mm; patella and tibia, 3.2 mm; ed their way through the snow,
metatarsus, 2.0 mm; tarsus, 1.1 mm. Sec- Distrihution. Quebec, Ontario, New
ond patella and tibia, 3.0 mm; third, 1.8 York to British Columbia, south to Ore-
mm; fourth, 2.6 mm. gon and New Mexico (Map 7). The south-
Male. Chelicerae with four teeth on ernmost record is San Fidel, Valencia

Figures 202-213. Pachygnathia autumnalis Keyserling. 202-204. Eye region and chelicerae. 202. Female. 203, 204. Male.
204. Left chelicera, posterior. 205-210. Female. 205. Position of genital opening. 206, 207. Female genitalia. 206. Dorsal.
207. Lateral. 208. Dorsal. 209. Abdomen, ventral. 210. Left third and fourth metatarsi, tarsi, dorsal. 211-213. Left male
palpus. 211. Ventral. 212. Lateral. 213. Embolus and conductor, ventral.

Figures 214-225. P. xanthostoma C. L. Koch. 214-216. Eye region and chelicerae. 214. Female. 215, 216. Male. 216.
Chelicera, posterior. 217-222. Female. 217. Position of genital opening. 218, 219. Female genitalia. 218. Dorsal. 219.
Lateral. 220. Dorsal. 221. Abdomen, ventral. 222. Third and fourth metatarsi and tarsi, posterodorsal. 223-225. Male
palpus. 223. Ventral. 224. Lateral. 225. Embolus and conductor, ventral.

Scale lines. 0.1 mm; Figures 202-205, 208-210, 214-217, 220-222, 1.0 mm.

Mecynogea, Metinae, PACHVc;iVA77/A AND AziuA NoKTH OF MEXICO • Levi 57

202

203 ^ ^0^
*^ O 0

205

^206 \ 207

217

218

209

208

210

221

220

224

58

Btillctiu Mtiscum of Coniixinitiie Zoology, Vol. 149, No. 1

Co., New Mexico, 9, 4 Sept. 1941 (W.
Ivie).

Pachygnatha autumnalis Keyserling
Plate?; Figures 155, 158. 159,202-213;
Map 7

Pachygnatlia autumnalis Keyserling, 1884, Ver-
handl. Zool. Bot. Gesell. Wien, 33: 660, pi. 21.
fig. 10, 6. Male holotype allegedly troni Harris-
biirg, Pennsylvania in die Marx collection of the
U.S. National Museum of" Natural History on per-
manent loan to the American Museum of Natural
History, e.\amined. Emerton, 1884, Trans. Con-
necticut Acad. Sci., 6: 337, pi. 34, fig. 22, pi. 40,
fig. 9, 9, 6. McCook, 1893, American Spiders, 3:
268, pi. 26, figs. 1,2,9,6. Roewer, 1942, Katalog
der Araneae, 1: 998. Kaston, 1948, Bull. Con-
necticut Geol. Natur. Hist. Surv., 70: 266, figs.
840, 849, 9,6.

Description. Female. Carapace dark
brown with a light patch on each side of
thorax posteriorly and a light rim around
thorax (Fig. 208). Folium very distinct
and posteriorly lohed, cardiac mark red-
dish (Fig. 208). Head region of carapace
not distinctly set off from thorax (Fig.
208). Eye area raised (Figs. 155, 202).
Posterior median eyes 1.7 diameters of
anterior medians. Laterals 0.8 diameters
of anterior medians. Anterior medians
their diameter apart, one diameter from
laterals. Posterior median eyes their di-
ameter apart, 0.8 diameters from laterals.
The genital slit is about its width distant
from the two sclerites between the lung
slits (Fig. 205). Chelicerae with only
three teeth, all small, on the posterior
margin; two proximally and one near
base of fang. Metatarsi 3 and 4 have a
trichobothrium close to the distal end
(Fig. 210). Total length, 4.7 mm. Cara-
pace, 2.2 mm long, 1.6 mm wide. First
femur, 2.0 mm; patella and tibia, 2.4 mm;
metatarsus, 1.6 mm; tarsus, 1.0 mm. Sec-
ond patella and tibia, 2.1 mm; third, 1.4
mm; fourth, 2.0 mm.

Male. Color like female. Chelicerae as
figured. Total length, 4.3 mm. Carapace,
1.9 mm long, 1.3 mm wide. First femur,
1.9 mm; patella and tibia, 2.3 mm; meta-
tarsus, 1.4 mm; tarsus, 0.8 mm. Second
patella and tibia, 1.9 mm; third, 1.2 mm;
fourth, 1.7 mm.

Variation. Female total length, 4.2 to
5.0 nun; carapace, 2.0 to 2.4 nnn long.
Male total length, 3.5 to 4.3 mm; carapace,
1.9 to 2.0 mm long.

Note. The large eyes have very small
rhabdom cells (Fig. 159), the lines of Fig.
159 show the dark pigment between the
rhalidom rows. The size is presumably a
functional adaptation.

Diagnosis. This species can be readi-
ly separated from others by its relatively
small size and the enormous posterior
median eyes, which are raised above the
carapace (Fig. 155, 208). Females have
only 3 teeth on the posterior margin of
the chelicerae; all other species have 4.
Males can be distinguished by the pres-
ence of a twisted embolus and conductor
of the palpus (Figs. 211, 213), and by a
very long pointed paracymbium (Fig.
212). Females are distinguished by the
very characteristic shape of the seminal
receptacle in side view (Fig. 207).

Natural History. This species has
been collected using pit-fall traps in low
woods, flat pine land in Florida, and in
old fields.

Distribution. Ontario, New England
south to Cuba, west to Missouri and Ar-
kansas (Map 7). The Cuban record is from
Soledad near Cienfuegos, Valles, Cuba,
9 (P. J. Darlington).

Pachygnattia xanthostoma C. L. Koch
Figures 160, 161, 214-225; Map 7

Fachygnatha xanthostoma C. L. Koch, 1845, Die
Arachniden, 12: 148, figs. 1068, 1069, 9,6. Orig-
inal specimens from Pennsylvania lost (M. Mor-
itz, pers. communication). Male neotype here
designated from Horseshoe Bend, northeast of
Jamison, Bucks County, Pennsylvania, 1959 (W.
Ivie) in the American Museum of Natural History.
Keyserling, 1884, Verhandl. Zool. Bot. Gesell.
Wien, 33: 659, pi. 21, fig. 9, 6. McCook, 1893,
American Spiders, 3: 269, pi. 26, figs. 7, 8, 9, 6.

P. f/orof/iea:— Bryant, 1933, Bull. Mus. Comp.
Zool., 74: 176, figs. 37, 41, 6 (misidentified, not
P. dorothea McCook).

Note. The type of P. xanthostoma has been lost
and there is no evidence that it ever was in the
Berlin Museum (M. Moritz, letter). Pachygnatha
xanthostoma came from Pennsylvania, was 2y4
lines long (4.8 mm, probably with chelicerae; Koch

Mecysogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi

59

had his specimens on pins). Koch described it as
having a punctate sternum and a carapace in part
simihuly sculptured; "thorax wider than head,
rounded, palpi and legs like Pachygnatha degeeri
(the smallest European species], and the chelicerae
and legs yellow." I consider this description most
applicable to the species described below, although
the head and chelicerae are often darker. The
species described below is the most similar to the
European P. degeeri, and the proportions of the pal-
pi of the two are similar. It has the most sculptured
carapace and sternum of the American species and
is also the most common.

The first reviser of Pachygnatha, Keyserling,
1884, considered this species to be xanthostoma
and illustrated the diagnostic chelicerae. In collec-
tions this species has been called xanthostoma on
labels by Bishop, Crosby and Gertsch, before 1933.
Miss Bryant, however, decided in 1933 that xan-
thostoma was an older name for P. brevis Keyser-
ling. She indicated as evidence that in 1912 N.
Banks examined the Koch types at the Berlin Mu-
seum and found that the species labeled P. xan-
thostoma was what had been called P. brevis. This
is an error. Miss Bryant also confused two species
under the name oi brevis: brevis and dorothea.

Description. Female. Dorsum of ab-
domen with an indistinct, lobed folium
enclosing a few white, paired spots and
a cardiac mark, as well as a broken me-
dian, longitudinal line; it is dusky to
black (Fig. 220). Punctate marks are vis-
ible on some areas of carapace and on
sternum. The genital opening is one-
third the distance from a line between
book-lungs to spinnerets (Fig. 217).
Third metatarsus has trichobothrium one
third its distance from pro.ximal end and
no trichobothrium on fourth metatarsus
(Fig. 222). Total length, 3.5 mm. Cara-
pace, 1.7 mm long, 1.3 mm wide. First
femur, 1.4 mm; patella and tibia, 1.9 mm;
metatarsus, 1.3 mm; tarsus, 0.7 mm. Sec-
ond patella and tibia, 1.7 mm; third, 1.2
mm; fourth, 1.6 mm.

Male. Total length, 3.3 mm. Carapace,
1.7 mm long, 1.2 mm wide. First femur,
1.5 mm; patella and tibia, 1.9 mm; meta-
tarsus, 1.3 mm; tarsus, 0.7 mm. Second
patella and tibia, 1.8 mm; third, 1.0 mm;
fourth, 1.7 mm.

Variation. Female total length, 3.0 to
4.2 mm; carapace, 1.6 to 1.9 mm long.
Male total length, 2.9 to 3.8 mm; carapace,
1.4 to 1.8 mm long.

Diagnosi.s. Females can be separated
from most American species by the small
size, from P. autumnaUs by small poste-
rior median eyes (Fig. 214) and by the
presence of a trichobothrium on the prox-
imal third of the third metatarsus, and by
the absence of trichobothrium on the
fourth metatarsus (Fig. 222). Males have
a very small tegulum, a very long and
complex conductor (Figs. 223, 225), and
a knobbed paracymbium (Fig. 224). Ex-
cept for P. clercki and P. cahisa it is the
only male with a spur on the chelicerae
above the origin of the fang (Fig. 215).

Natural History. This species has
been collected from pitfall traps, sifted
leaves in Michigan, lawn in Wisconsin,
under boards in Nebraska, in grass near
a stream in Utah, under a post by a pond
in a hay meadow in Colorado, in an al-
falfa field in Wyoming, and in the stom-
ach of a toad (Bufo) collected in Montana.

Distribution. Quebec to Alberta,
northern New England to Pennsylvania,
west to Washington, Idaho and Utah
(Map 7).

Pachiygnatha clercki Sundevall
Figures 226-237; Map 7

Pachygnatha clercki Sundevall, 1823, Specimen
Academicum Geneva, Araneidum Sueciae Exhi-
bens, p. 16. Specimens from Sweden. Roewer,
1942, Katalog der Araneae, 1: 996. Locket and
Millidge, 1953, British Spiders, 2: 106, figs. 71,
72, 73A, D, 9, 6. Bonnet, 1958, Bibliographia
Araneorum. 2: 3277. Wiehle, 1963, Tierwelt
Deutschlands, part 49: 61, figs. 99-106, 9,6.

Pachygnatha sewardi Ghamberlin and Ivie, 1947,
Bull. Univ. Utah, biol. ser., 10(3): 64, figs. 76-78,
9,6. Male holotype from Beluga Flats, Cook In-
let, Alaska in the American Museum of Natural
History, examined. NEW SYNONYMY.

Note. The illustrations were all made from
North American specimens, the palpi were made
from paratypes of P. sewardi.

Description. Female. Dorsum of ab-
domen with indistinct folium, only dark
outline present; posteriorly the dark
bands are broken, but are not wavy (Fig.
232). The sternum lacks connection be-
tween coxae 1 and 2 to the narrow dorsal
plate above. The genital slit is only one-

60

Htillctiii Museum of Compardtivc Zoology, Vol. 149, No. I

fourth tlie distance from a line l)etween
hook-lunus to the spinnerets (Fi^. 229).
One trichobothriinn on middle of meta-
tarsus 3, and one two-thirds the distance
from proximal end on metatarsus 4 (Fig.
234). Total length, 5.5 mm. Carapace, 2.6
mm long, 1.9 mm wide. First femur, 2.5
mm; patella and tibia, 3.2 mm; metatar-
sus, 2.0 mm; tarsus, 1.2 mm. Second pa-
tella and tibia, 2.9 mm; third, 1.9 mm;
foiuth, 2.6 mm.

Male. Coloration more distinct than fe-
male. Total length, 5.3 mm. Carapace, 2.4
mm wide, 1.8 mm long. First femur, 2.5
mm; patella and tibia, 3.3 mm; metatar-
sus, 2.0 mm; tarsus, 1.2 mm. Second pa-
tella and tibia, 3.1 mm; third, 1.9 mm;
fourth, 2.6 mm.

Variation. One male from the north
slope of Alaska had a relatively smaller
tegulum and a longer conductor and em-
bolus in the palpus. As a result of the
smaller tegulum, the loops of the duct
were shallower. The specimen belongs
to the Florida State Collection of Arthro-
pods.

Diagnosis. The female can be sepa-
rated from most larger North American
PachygnatJia by the presence of a tricho-
bothrium on the fourth metatarsus (Fig.
234); from P. tristriata by lack of a scler-
ite on the genital area (Fig. 229) and by
the posteriorly broken dorsal abdominal
markings (Fig. 232). The male is the only
large North American species with a spur
above the fang attachment on the chelic-
erae (Fig. 227), with a coiled duct in the
tegulum (Fig. 235), and with the shape
illustrated in Figs. 235, 237 of the palpal

conductor. Half grown specimens have
the trichobothriuiu on metatarsus 4.

Natural History. This species is found
in wet areas (Wiehle, 1963). The egg-sac
has been described by Wiehle (1963) and
pictured by Nielsen (1932, fig. 358); it is
attached to both bark and moss and is
guarded. The young leave after only 2
weeks.

Distribution. Eurasia, Alaska to Man-
itoba (Map 7). One male from Washing-
ton, D.C. in the N. Banks collection may
have been introduced.

Pachygnatha tristriata C. L. Koch
Plate 7; Figures 238-250; Map 7

Pachygnatha tristriata C. L. Koch, 1845, Arachni-
den, 12: 145, fig. 1066, 9. Female from Pennsyl-
vania in the Museum of the Humboldt Univer-
sity, Berlin, lost. Keyserling, 1884, Verhandl.
Zool. Bot. Ges. Wien, 33: 656, pi. 21, fig. 8, 9,
6. McCook, 1893, American Spiders, 3: 270, pi.
26, fig. 6, pi. 28, fig. 1, 9, 6. Emerton, 1909,
Trans. Connecticut Acad. Sci., 14: 202, pi. 5, fig.
6, 9, c5. Roewer, 1942, Katalog der Araneae, 1:
998. Kaston, 1948, Bull. Connecticut Geol. Natur.
Hist. Surv., 70: 266, figs. 838, 839, 847, 848, 9,
6.

Descriptioti. Female. Carapace or-
ange with longitudinal dark stripes and
a dark band from thoracic depression an-
terio-laterally (Fig. 244). Chelicerae, ster-
num brown. Dorsum of abdomen with a
gently curved black line on each side, far-
thest apart in middle (Fig. 244). Venter
of abdomen above the genital area has a
weakly sclerotized triangular patch,
pointing anteriorly (Figs. 241, 245). The
muscle impressions between the book-
lung slits are barely visible and may be

Figures 226-237. Pachygnatha clercki Sundevall. 226-228. Eye region and chelicerae. 226. Female. 227, 228. Male. 228.
Left male chelicera, posterior. 229-234. Female. 229. Position of genital opening. 230, 231. Female genitalia. 230. Dorsal.
231. Lateral. 232. Dorsal. 233. Abdomen, ventral. 234. Left third and fourth metatarsi and tarsi, dorsal. 235-237. Left male
palpus. 235. Ventral. 236. Lateral. 237. Embolus and conductor, ventral.

Figures 238-249. P. tristriata C. L. Koch. 238-240. Eye region and chelicerae. 238. Female. 239, 240. Male. 240. Left
male chelicera, posterior. 241-246. Female. 241. Position of genital opening. 242, 243. Female genitalia. 242. Dorsal.
243. Lateral. 244. Dorsal. 245. Abdomen, ventral. 246. Third and fourth metatarsi and tarsi, dorsal. 247-249. Male palpus.
247. Ventral. 248. Lateral. 249. Embolus and conductor, ventral.

Scale lines. 0.1 mm; except Figures 226-229, 232-234, 238-241, 244-246, 1.0 mm.

Mecynocea, Metinae, Pachvgnatha and Az//ja Nohih of Mexico • Levi 61

62

Bulletin Museum of Compdrcitivc Zoolo^ij. Vol. 149, \'o. 1

split into three little platelets. One trieli-
ohothiiuin on distal end of both nu^tatar-
sus 3 and 4 (Fi«. 246). Total length, 6.0
mm. Carapaee, 2.4 mm long, 1.7 mm
wide. First femnr, 2.3 mm; patella and
tibia, 3.0 mm; metatarsns, 1.9 mm; tarsns,
0.9 mm. Seeond patella and tibia, 2.7
mm; third, 1.8 nnn; fomth, 2.7 mm.

Male (from Massaehnsetts). There are
three teeth on the anterior margin of the
chelicerae (Fig. 239) and one blunt tooth
on the posterior margin (Figs. 239, 240).
Total length, 6.0 mm. Carapace, 2.7 mm
long, 1.9 mm wide. First femur, 3.3 mm;
patella and tibia, 4.1 mm; metatarsus, 2.7
mm; tarsus, 1.2 mm. Second patella and
tibia, 3.6 mm; third, 2.2 mm; fourth, 3.4
mm.

Variation. Specimens from the south-
ern part of the range are noticeably small-
er than northern ones. Female total
length, 4.9 to 6.5 mm; carapace, 2.3 to 2.9
mm long. Male total length, 4.5 to 6.2 mm;
carapace, 1.9 to 2.9 mm long.

Diagnosis. The females can be sepa-
rated from those of other species by the
weakly sclerotized triangle on the genital
area (Figs. 241, 245) and by the straight
lines on the dorsum of the abdomen (Fig.
244). This species seems to lack seminal
receptacles (Figs. 242, 243). The palpal
conductor and embolus are much longer
than those of other species (Figs. 247-
249). The embolus is coiled (Fig. 247),
and there is a unique blunt tooth on the
posterior margin of the chelicerae (Figs.
239, 240).

Natural History. This species has
been collected under a board in New
Hampshire, while sweeping through

xegetation, under stones, in an asparagus
bed in Massachusetts, in a marshy mead-
ow in New York, Bermuda grass pasture
(Cytiodoti dactijlon) in Louisiana, and in
pitfall traps in Ohio and in a soybean
field in Florida.

Distrihtitiou. Nova Scotia, Ontario,
New England south to northern Florida,
west to Wisconsin, Kansas and eastern
Texas (Map 7).

Pachygnatha calusa new species
Figures 251-254; Map 7

Holotypc. Male tiom Everglades, Florida, 27 De-
cember 1950 (A. M. Nadler) in the American Mu-
seum of Natural History. The name is a noun after
the Indians of the Everglades.

Note. The small size of the imique specimen
prevented studies of tracheae and trichohothria. It
is uncertain whether the specimen belongs to Pach-
ijfinatha or Gletw^natha. However, the description
permits recognition.

Description. Carapace, legs yellow.
Abdomen without markings. Posterior
median eyes 0.8 diameters of anterior
medians, laterals 0.6 diameters. Anterior
median eyes slightly more than their di-
ameter apart, slightly more than one di-
ameter from laterals. Posterior median
eyes slightly less than two diameters
apart, 1.5 from laterals. The laterals are
almost their diameter apart. The height
of the clypeus equals two diameters of
the anterior median eyes. The chelicerae
have 3 small teeth on the anterior margin,
2 on the posterior (Fig. 251). The endites
are widest distally. The third and fourth
metatarsi have (probably) one trichoboth-
riimi, one-third the article's length from
proximal end. The sternum connects to

Figure 250. Pachygnatha tristriata C. L. Koch, tracheae.

Figures 251-254. P. calusa new species, male. 251. Eye region and chelicerae. 252-254. Left palpus. 252. Ventral. 253.
Lateral. 254. Embolus and conductor, ventral.

Figures 255-271. Glenognatha emertoni Simon. 255-265. Female. 255. Eye region and chelicerae. 256. Left chelicerae.
posterior. 257. Sternum, labium and endites. 258. Venter of abdomen. 259. Female genitalia, dorsal view. 260. Tracheae.
261. Left lateral eyes. 262. Posterior median eyes. 263. Dorsal. 264. Abdomen, ventral. 265. Left third and fourth meta-
tarsus and tarsus. 266-271. Male. 266. Eye region and chelicerae. 267. Left chelicera, posterior. 268-271. Palpus. 268.
Ventral. 269. Lateral. 270. Embolus and conductor, ventral. 271. Embolus and conductor tip, mesal.

Scale lines. 0.1 mm; Figures 255, 256-258, 263-267, 1.0 mm.

Mecynogea, \\ET]>iAE, Pachygnatha and AziLiA North of Mexico • Levi 63

265

252

64

Bullet ni Miisctini of CoDijxinitirc 7ah)Io^\i, Vol. 149, No. 1

the narrow sclerite above the coxae be-
tween coxae 1 and 2, 2 and 3, 3 and 4.
The abdomen is qnite narrow; the spira-
cle is in acKance of spinnerets, about the
length oi One spinneret. Total length, 1.9
mm. Carapace, 0.82 nmi long, 0.62 nnn
wide. First femnr, 1.19 mm; patella and
tibia, 1.57 mm; metatarsns, 1.19 mm; tar-
sns, 0.52 mm. Second patella and tibia,
1.17 mm; third, 0.61 mm; fbnrth, 0.94
nnn.

Diagnosis. The small size and distal
notch on the paracymbinm (Fig. 253) sep-
arate the species from Pacluj^natha hrev-
is.

Glenognatha Simon

Glenognatha Simon, 1887, Ann. Soc. Entomol.
France, ser. 6, 7: 194. Type species by nionotypy
G. emertoni Simon. The name is feminine.

Mimognatha Banks, 1929, Bull. Mus. Comp. Zool.,
69: 90. Type species bv monotypv Mimognatha
foxi (McCook). NEW SYNONYMY.

?Hivaoa Berland, 1935, Bull. B. P. Bishop Mus.,
142: 50. Type species by original designation H.
argenteoguttata Berland from Marquesas Is-
lands. DOUBTFUL NEW SYNONYMY.

Didf^nosis. Glenognatha differs from
PacJu/gnatJia by having the posterior
\entral spiracle moved anterior (Figs.
258, 274), and by having an elaborate tra-
cheal system: two large trunk tracheae
that split into numerous fine tracheae
(Figs. 260, 275). Pachygnatlia, like most
Araneidae, has the spiracle close to the
spinnerets, and only four tracheae (Fig.
250). The abdomen of Glenognatha is
softer and more spherical (Figs. 263, 264,
273, 274) than that of Pachygnatha and
with paired light dorsal patches (Figs.
263, 273). Glenognatha differs from Tet-
ragnatha by having a spherical abdomen
and a canoe tapetum in the lateral eves
(Fig. 261).

Description. Carapace plain, usually
darker than legs. Legs not banded. Dor-
sum of abdomen with paired light and
dark marks, but no folium (Figs. 263,
273). Venter without distinct marks, dark-
er gray in a longitudinal band in center

(Fig. 264). Carapace smooth without se-
tae, no distinct thoracic depression. Eyes
closely grouped, median eyes their di-
ameter apart, 1.5 diameters from laterals.
Posterior median eyes may be slightly
larger than others. Glenognatha emer-
toni has tapetinn in the cells of the rhab-
doms of the posterior median eyes (Fig.
262). Height of clypeus 2 to 3 diameters
of anterior median eyes, more or less
swollen on each side above chelicerae
(Figs. 255, 266). Chelicerae large to very
large, always with 3 teeth on the anterior
margin, four on posterior (Figs. 255, 256,
266, 267). Those of male sometimes larg-
er than female, at times with spur above
base of fang (Figs. 276, 285). Labium
large, endites short to long (Figs. 257,
274). Sternum does not usually surround
coxae as in Pachijgnatha (Fig. 257). All
legs are thin and lack macrosetae (Fig.
273). The claws have long narrow teeth;
accessory setae are present. Dorsally the
base of tlie first and second femur has
one (G. foxi) to 3 (G. emertoni) tricho-
bothria, that of the third and fourth (G.
foxi) femur has no trichobothria or has
4 and 5 (G. emertoni). Besides the several
trichobothria on the tibiae, the metatarsi
have one trichobothrium longer than the
diameter of the article. The fourth meta-
tarsus lacks a trichobothrium in G. foxi.

Males are the same size as females, as
in Pachygnatha. The chelicerae are en-
larged. The palpus is similar to that of
Pachygnatha, with a spherical tegulum
and a long, coiled embolus supported by
a complex conductor (Figs. 270, 280,
289). The cymbium is narrow and the
paracymbiimi a free sclerite, quite simi-
lar in shape in all species (Figs. 269, 279,
288).

The female genitalia are similar to
Pachygnatha except that the median an-
terior chamber is spherical in G. emer-
toni.

Note. Banks (1929) placed G. foxi
into its own genus because it has shorter
chelicerae with only a few small teeth,
and because of the very large spherical

Mecynocea, Metinae, Pachygsatha and Azilia North of Mexico • Levi 65

TTTYl

G. erne

Map 8. Distribution of Glenognatha species.

palpal tegulum and smaller embolus and
conductor. I consider these specific dif-
ferences. GlenognatJid joxi males have
enlarged chelicerae, and those of the fe-
male are not small. Many characters of
G.foxi are associated with its small size:
the relatively large eyes (Fig. 272), the
convex sternum, and the great distance
between the posterior coxae (Fig. 274).
There are additional species in tropical
America, some as small in size as G. foxi.
Natural History. The species live in

more arid habitats than Pachygnatha
species. The habits of G. emertoni are
not known; G. foxi has been found in
meadows and on hot, dry wastelands in
Ohio. The unusual development of the
tracheae and the forward movement of
the spiracle is probably an adaptation
against water loss. Glenognatha foxi is
found on webs.

Distribution. The few Glenognatha
species named are American, but there
are specimens from the Pacific Islands in
American collections.

66

Bulletin Museum of CoDipanitiie Zooloi^y, Vol. 149, No. 1

Key to Clenoc.natha Species
NoHTH OF Mexico

1. Less than 3.0 nun total length; teniale with
chelicerae not enlarged (Fig. 272); male
with spnr on chelicerae (Figs. 276, 285);
emholns and conductor minute on huge
spherical tegulum (Figs. 278, 287); southern
Canada to Central America and West Indies
(Map 8) 2

- Total length more than 3.5 mm; female with
chelicerae enlarged (Fig. 255); male with-
out spur on chelicerae (Fig. 266); embolus
and conductor length greater than height of
spherical tegulum (Figs. 268); New Mexico,
Arizona (Map 8) emertoni

2(1). Female unknown; male with hooked tooth
on anterior margin of chelicerae (Fig. 285);
tip ot embolus coiled (Fig. 289); Mississippi
(Map 8) iviei

- Male without hooked tooth on anterior mar-
gin (Fig. 276); tip of embolus not coiled
(Fig. 280); southern Canada to Central
America, West Indies (Map 8) foxi

Glenognatha emertoni Simon
Figures 255-271; Map 8

Glenognatfxa emertoni Simon, 1887, Ann. Soc. Ent.
France, ser. 6, 7: 194. Male holotype from Arizona
in the Museum National d'Histoire Naturelle,
Paris, examined. 1894, Histoire Naturelle des Ar-
aignees, 1: 717, figs. 790, 791, 6. Banks, 1913,
Proc. Acad. Natur. Sci. Philadelphia, 65: 180, pi.
2, fig. 22, 6 . Roewer, 1942, Katalog der Araneae,
2: 995. Bonnet, 1957, Bibliographia Araneorum,
2: 1994.

Description. Female. Carapace, che-
licerae, legs orange-brown; sternum
brown. Dorsum of abdomen has paired
gray spots alternating with paired light
areas consisting of tiny white spots (Fig.
263). Clypeus swollen with two bulges
(Fig. 255). Eyes siibequal in size. Ante-
rior median eyes slightly more than their
diameter apart, two diameters from lat-

erals. Posterior median eyes slightly
more than their diameter apart, two from
laterals. Posterior median eyes have ta-
petum inside rhabdoms which are in
looped rows (Fig. 262). Endites are long
but not wider distally. Chelicerae very
large (Fig. 255). Abdomen spherical and
hairy. Metatarsus 3 has trichobothrium
one-third length of article from proximal
end. Metatarsus 4 with trichobothrium
one-third from proximal end. Total
length, 4.7 mm. Carapace, 2.6 mm long,
2.0 mm wide. First femur, 3.3 mm; pa-
tella and tibia, 4.3 mm; metatarsus, 2.9
mm; tarsus, 1.4 mm. Second patella and
tibia, 4.2 mm; third, 2.3 mm; fourth, 3.2
mm.

Male. Much like female. Chelicerae
smaller and less modified (Fig. 266).
Eyes slightly further apart. Total length,

4.0 mm. Carapace, 2.4 mm long, 1.9 mm
wide. First femur, 3.2 mm; patella and
tibia, 4.3 mm; metatarsus, 2.7 mm; tarsus,
1.5 mm. Second patella and tibia, 4.1
mm; third, 2.4 mm; fourth, 2.8 mm.

Variation. Female total length, 4.6 to
5.4 mm; carapace, 1.9 to 2.6 mm long.
Male total length, 4.0 to 5.0 mm; carapace,

2.1 to 2.5 mm long.

Diagnosis. Glenognatha emertoni has
a palpus (Figs. 268, 269) similar to that of
G. niinuta Banks found in Baja Califor-
nia; G. miniita is smaller in size, and the
cheliceral shape and teeth differ. The
larger size, large chelicerae readily sep-
arate G. emertoni from G.foxi. The sem-
inal receptacles are unusually thin-walled
and difficult to see (Fig. 259).

Natural History. Specimens have
been found under rocks near streams and
under rocks in nearly dry stream bed.

Figures 272-284. Glenognatha foxi (McCook). 272-275. Female. 272. Eye region and chelicerae. 273. Female. 274.
Sternum, coxae and venter of abdomen. 275. Female genitalia and tracheae, dorsal view. 276-284. Male. 276. Eye region
and chelicerae. 277. Left chelicera, posterior. 278-281, 283, 284. Left palpus. 278, 281-284. Ventral. 279. Lateral. 280.
Embolus and conductor, ventral. 278-280. (Pennsylvania). 281. (Jamaica). 282, 283. (Arkansas). 284. (Panama Canal
Zone). 282, 283. Right and left palpus from same individual.

Figures 285-289. G. iviei new species. 285. Eye region and chelicerae. 286. Left chelicera, posterior. 287-289. Palpus.
287. Ventral. 288. Lateral. 289. Embolus and conductor, ventral.

Scale lines. 0.1 mm; Figures 273, 274, 1.0 mm.

Mecynogea, Metinae, Pachygnatha and Azili a North of Mexico • Levi 67

272

277

68

Bullet ill MtisciiDi of Coinixirdtiic Zoology, Vol. 149, No. 1

Distribution. New Mexico and Ari-
zona (Map 8).

Glenognatha foxi (McCook)
Figures 272-284; Map 8

Theridiuni foxi McCook, 1893, American Spiders,
3, pi. 29, fig. 1.? Nomen nudum. Plate legend with
comment: "No \erbal description is made."

Mysincna hulhifera Banks, 1896, Trans. Amer. En-
tom. See, 23: 66. Three male, three female syn-
types from Washington, D.C. in the Mnsenm of
Comparative Zoology, examined. 1904, Proc.
Acad. Natiir. Sci. Philadelphia, 56: 127, pi. 7, figs.
12, 13, 6.

Glenognatha hulhifera: — Barrows, 1919, Ohio J.
Science, 19: 210, figs. 1, 2, 6 (mating, web).

Diploeephalus erumhi Petrnnkevitch, 1925, J. New
York Entomol. Soc. 33: 171, pi. 8, fig. 1, 2, 6.
Three male syntypes from Tennessee.

Glenognatha foxi: — Crosbv and Bishop, 1928,
Mem. Cornell Univ. Agric. Exp. Sta., 101: 1055.

Mimognatha foxi: — Banks, 1929, Bull. Mus. Comp.
Zool., 69: 90. Crosby and Bishop, 1936, J. New
York Ent. Soc, 44: 47. Kaston, 1948, Bull. Con-
necticut Geol. Natiir. Hist. Surv., 70: 264, figs.
834, 835, 9, c5. Bonnet, 1957, Bibliographia Ara-
neorum, 3: 2922.

Note. The name foxi has lieen used for this
species for fifty years, although the McCook name
lacked a description. His figure is adequate as an
indication for a new species (Int. Code Zool. No-
menclature, Art. 16, vii).

Description. Carapace orange, ster-
num dark orange. Coxae, legs yellow.
Dorsum of alidomen orange-white, some-
times with paired silver spots posteriorly
and darker patches (Fig. 273); also some
silver spots around anterior margin; ven-
ter without distinct markings. Eyes sub-
equal in size; anterior medians 1.3 di-
ameters apart, 1.5 diameters from laterals;
posterior medians 1.1 diameters apart,
1.2 from laterals. Chelicerae have three
teeth on the anterior margin, four on the
posterior (Fig. 272). A tooth on posterior
side of fang. Endites are not wider dis-
tally than at base. One trichol)othriiuu on
the metatarsi 1, 2 and 3, about one-quar-
ter the article's length from proximal end.
Total length, 1.6 mm. Carapace, 0.61 mm
long, 0.56 mm wide. First femiu', 0.67
mm; patella and tibia, 0.75 mm; metatar-
sus, 0.57 mm; tarsus, 0.36 mm. Second

patella and tibia, 0.66 mm; third, 0.45
nun; fourth, 0.65 mm.

Male. Much like female. Chelicerae
enlarged (Figs. 276, 277) and with spur
above base of" fang (Fig. 276). Total
length, 1.8 mm. Carapace, 0.70 mm long,
0.63 mm wide. First femur, 0.81 mm; pa-
tella and tibia, 1.01 mm; metatarsus, 0.65
mm; tarsus, 0.36 mm. Second patella and
tibia, 0.91 mm; third, 0.54 mm; fourth,
0.75 mm.

Variation. Some specimens have
hardly any abdominal markings, some
have scattered tiny white silver spots,
and some have a pair of ventral silver
patches between genital opening and
spiracle. The size of the male spherical
tegukmi varies in size (Figs. 281-284)
and left and right of the same individual
may even differ slighdy (Figs. 282, 283).
Total length of females, 1.6 to 2.6 mm;
carapace, 0.6 to 1.0 mm long. Total length
of males, 1.4 to 2.2 mm; carapace, 0.6 to
1.0 mm long. The largest specimens
came from Arizona and California.

Diafitiosis. The small size, relatively
small chelicerae (Figs. 272, 276) and the
very large spherical tegulum of the male
palpus separate this species from G.
emertoni and other species. Unlike some
similar tropical American species, the
male has a spur above the fang on the
chelicerae (Fig. 276). The tip of the em-
bolus is not twisted (Fig. 280) as is that of
G. iviei.

Natural History. Barrows (1919)
foimd the spider in meadows and waste-
lands, in hot, dry situations. The spider
makes a horizontal web 11 cm in diame-
ter, 5 cm above the ground. Grass may
grow through the web. The spiral strands
are close together, and viscid silk is pres-
ent. Spiders rest in the center of the web,
and when distiubed drop to the groimd
and rim away. Barrows observed mating
in June in the center of the web, both
indi\ iduals hanging down, the chelicerae
locked and legs clasped. The palpal or-
gans are inserted alternately every 5
minutes, copulation lasting 15 minutes.

Mecynogea, Metinae, Pachygnatha and AziLiA North of Mexico • Levi

69

Crosby and Bishop (1936) reported
males collected in March and April by
airplanes above Tallulah, Louisiana at
900 m and 60 m.

Other collections come from the edge
of pond and saltmarsh in iM as sachu setts,
from sweeping short grass in Virginia,
from pitfall traps in a cornfield, sweet po-
tato field, and a one-year-old abandoned
field in North Carolina; from a web over
a crawfish hole in Arkansas, a ground
crevice in a dry lake in Missouri, and be-
tween stones on the bank of a stream "but
did not notice an orb-web."

Distribution. The northernmost rec-
ords are from Nantucket Island, Massa-
chusetts, Wisconsin, Kansas, southern
Arizona, southern California, south to Ja-
maica and Panama (Map 8).

Glenognatha iviei new species
Figures 285-289; Map 8

Holotype. Male from Gulfport, Mississippi, 28 Au-
gust 1933 (W. Ivie, coll.) in the American Mu-
seum of Natural History. The species is named
after Wilton Ivie, who collected the specimen and
recognized that it was new.

Description. Carapace, sternum
brown, legs light brown. Dorsum of ab-
domen mostly gray to black, with paired
lighter patches lacking white or silver
pigment. Venter gray, darkest in the cen-
ter. Eyes subequal in size. Anterior me-
dian eyes slightly more than their diam-
eter from each other and from laterals.
Lateral eyes are on tubercles. One tricho-
bothrium on third metatarsus only, one-
third distance from proximal end. The
fourth metatarsus appears to lack a trich-
obothrium. Total length, L4 mm. Cara-
pace, 0.74 mm long, 0.65 mm wide. First
femur, 0.78 mm; patella and tibia, LOO
mm; metatarsus, 0.65 mm; tarsus, 0.41
mm. Second patella and tibia, 0.88 mm;
third, 0.47 mm; fourth, 0.67 mm.

Diagnosis. Glenognatha iviei differs
from G. foxi by the presence of an ante-
rior hooked tooth on each chelicera (Fig.
285) and by the twisted tip of the embolus
(Figs. 287, 289).

Azilia Keyserling

Azilia Keyserling, 1882, Verhandl. Zool. Bot. Ges.
Wien, 31: 270. Type species by monotypy A. for-
mosa from Peru. The name is feminine.
Note. According to Keyserling, the type speci-
mens of A. formosa are in Warsaw, Poland. The
staff of the Polish Academy of Sciences has not an-
swered requests, thus the type may he lost. A neo-
type will have to be designated for A. formosa
when the South American members of the genus
are revised, since there is doubt on the identity of
the species.

Diagnosis. Unlike all other Aranei-
dae, except Tetragnatha, Azilia have
the lateral eyes farther apart than the me-
dians (Figs. 295, 300). The eyes are
closely grouped and subequal in size.
Unlike most Araneidae but like Tetra-
gnatha, all eyes lack a tapetum (Figs.
299, 300). The oval abdomen (Figs. 295-
297) separates Azilia from those Tetrag-
natha species which have separated lat-
eral eyes. Unlike some other tetragnath-
ids, there are no trichobothria on the base
of the femora.

Description. Carapace streaked,
darker on sides and head region; narrow
in front. Eyes surrounded by black. Ster-
num dark, coxae light and legs banded.
Abdomen streaked, often with white
patch above spinnerets (Fig. 297). Eyes
closely grouped, subequal in size, or pos-
teriors slightly larger. Median eyes almost
their diameter apart; anteriors almost
touching laterals; posteriors about 1.5
from laterals; laterals about 2 diameters
from each other. Eyes without tapetum,
rhabdoms are in rows which loop (Figs.
299, 300). Clypeus height equal to or
slightly less than diameter of anterior
median eyes (Fig. 294). Chelicerae
strong but not enlarged as in Pachygna-
tha or Glenognatha. Endites long, wid-
est distally (Fig. 298); legs fairly long
with macrosetae (Fig. 295). No tricho-
bothria on base of femur. Abdomen
ovoid, widest anteriorly (Fig. 295), often
with a hump above spinnerets (Fig. 297).

Males slightly smaller than females
with stronger macrosetae on legs.

Genitalia. There is an epigynum

70 Bulletin Museum of Couiparativc Zor)/r)<,'f/, Vol. 149, Mo. 1

Plate 8. Azilia affinis O. P. -Cambridge. Web, diameter about 12 cm; note the closed hub (photo M. Stowe).

present (Figs. 291-293), but the openings
face posterior as in Meta. The seminal
receptacles (Fig. 290) are very thin-
walled like those of Leucauge. The pal-
pus is very simple (Figs. 301-305) with
tegulum (T) and embolus (E), but it lacks
a distinct conductor. The paracymbium
(P in Fig. 304) is only little modified.

Natural History. All species are found
in dark places and caves, like many

Meta. It is interesting that, despite the
crepuscular habits, the eye of the spider
lacks a tapetum. The web is vertical to
horizontal and loose. The hub, unlike
that o{ Meta, Leucauge and TetragtiatJia
wel)s, is closed (Plate 8). The spider
hangs on the underside of the web and
boimces when disturbed (S. Peck, per-
sonal communication).
Relationship. Like other genera in

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 71

Figures 290-308. Azilia affinis O. P. -Cambridge. 290-300. Female. 290. Female genitalia, dorsal view. 291-293. Epigy-
num. 291. Ventral. 292. Posterior. 293. Lateral. 294. Eye region and chelicerae. 295. Female. 296. Abdomen, ventral. 297.
Abdomen, lateral. 298. Labium and endites. 299. Left lateral eyes. 300. Posterior median eyes and left laterals. 301-308.
Left male palpus. 301. Ventral. 302. Lateral. 303-305. Expanded and cleared. 303. Mesal. 304. Ventral. 305. Lateral. 306-
308. Embolus. 306. (Florida). 307. (Georgia). 308. (Southern Texas).

Scale lines. 0.1 mm; Figures 294-298, 1.0 mm.

Abbreviations. C, conductor; E, embolus; P, paracymbium; S, subtegulum; T, tegulum.

72 Biillctiit Museum of Compdnitivc Z()olo<iy, Vol. 149, No. 1

if?^c2>'.

Azi lia affi ni s

<; I

Map 9. Distribution of Azilia affinis O. P. -Cambridge.

this paper, the plaeement oi Azilia is dif-
ficult. The palpus and the non-sclero-
tized seminal receptacles resemble Leu-
cauge superficially. The shape of the
abdomen and coloration resemble Meta.
Only Tetragnatha also lacks a tapetum in
all eyes. The long endites are tetragnath-
id-like, but in other genera long endites
are usually accompanied by enlarged
chelicerae. The solid hub of the web is
unlike any other web of Meta, Leucauge
or Tetragnatha relatives.

Species Differences. The species dif-
fer in the ventral marking of the abdo-
men, and the shape of the epigynum and
palpus. Unlike those of Leucauge and
Pachygnatha, the differences between
species are considerable, although all can
be recognized as Azilia by the unusual
eye arrangement.

Distribution. There are only 4 or 5
species in tropical America, of which
only one, A. affinis, extends its range into
temperate North America (Map 9).

Azilia affinis O. P. -Cambridge
Plate 8; Figures 290-308; Map 9

Azilia affinis O. P.-Cambridge, 1893, Biologia Cen-
trali-Americana, Aianeidea, 1: 115, pi. 15, fig. 4,
9. Female holotype from Teapa, Tabasco, Mexico
in British Museum, Natiual History, examined.

Azilia vagepicta Simon, 1895, Ann. Soc. Entomol.
France, 64: 153. Female specimens from "Amer.

sept, Georgia," in the Museum National d'Histoire
Naturelle, Paris, examined. Banks, 1913, Proc.
Acad. Natur. Sci. Philadelphia, 65: 180, pi. 11, fig.
14, 9. Roewer, 1942, Katalog der Araneae, 1: 923.
Bryant, 1945, Trans. Connecticut Acad. Arts Sci.
36: 208, fig. 10, 6. Bonnet, 1955, Bibliographia
Araneorum, 2: 840. NEW SYNONYMY.
Azilia mexicana Banks, 1898, Proc. California Acad.
Sciences, ser. 3, 1: 258, pi. 15, fig. 19. Adult syn-
types in California Academy of Sciences, de-
stroyed, one subadult male and one juvenile syn-
type from Tepic, Mexico in the Museum of
Comparative Zoology, examined. NEW SYN-
ONYMY.

Description. Female from Alachua
County, Florida. Dorsum of abdomen in-
distinctly patterned with dark chevrons
posteriorly and light areas consisting of
tiny white pigment spots. Sides with dark
anastomosing streaks. Venter with a
square central light patch consisting of
scattered tiny white pigment. Thorax
with an X-shaped depression (Fig. 295).
The epigynum has two openings and
very weakly sclerotized seminal recep-
tacles (Figs. 290-293). Total length, 8.9
mm. Carapace, 3.5 mm long, 2.5 mm
wide. First femur, 6.0 mm; patella and
tibia, 7.3 mm; metatarsus, 6.4 mm; tarsus,
L9 mm. Second patella and tibia, 5.3
mm; third, 3.2 mm; fourth, 4.7 mm.

Male from Alachua County, Florida.
Color like female. Smaller than female
but similar structure. Anterior part of X-
shaped thoracic mark less distinct than
posterior. Eye sizes and spacing as in fe-
male. Chelicerae like those of female.
Tooth on the distal end of the palpal fe-
mur. No tooth on the endite and no prox-
imal tooth on femur. None of the legs is
modified, but the macrosetae are relative-
ly longer than those of the female. Total
length, 6.8 mm. Carapace, 3.2 mm long,
2.4 mm wide. First femur, 8.0 mm; pa-
tella and tibia, 10.0 mm; metatarsus, 9.7
mm; tarsus, 2.3 mm. Second patella and
tibia, 6.7 mm; third, 3.0 mm; fourth, 4.9
mm.

Variation. The posterior hump is
more distinct in some specimens (Fig.
297) than others, almost absent in speci-
mens from Florida. No two populations
have similar genitalia, but the differences

Mecysogea, Metinae, Fachvcvatha and Azili a North of Mexico • Levi 73

in the epigynum are minor. The differ-
ences of Hie palpi are in the shape of tlie
embohis (Figs. 306-308) and the widtli of
the hiilb; also, no two males have a sim-
ilar duct coil within the bulb (Figs. 303-
305).

Females vary in total length from 6.9
to 9.9 mm; carapace, 2.9 to 3.8 mm long,
2.1 to 2.8 mm wide; first patella and tibia,
6.3 to 8.1 mm long. Males vary in total
length from 5.2 to 6.9 mm; carapace, 2.6
to 3.9 mm long, 1.9 to 2.4 mm wide; first
patella and tibia, 6.3 to 9.4 mm long.

Diagnosis. Azilia can be separated
from other Central and South American
species by its prominent square, light
patch on the venter of the abdomen (Fig.
296), the openings of the epigynum (Figs.
291-293), and its large size. The males of
related species are mostly unknown; one
has a long filiform embolus.

Natural History. Azilia vagepicta has
been collected in caves and wood rat
nests (Neotoma micropiis) and in other
presumably dark situations, including
mesic lush ravine; mesic woods, shady
recesses on the sides of a building struc-
ture, mesic hammock, in bottom of wet
ravine, and in dried-up cypress-hard-
wood swamp. Many were collected at
night with flashlight. In Florida, the
species occurs in twilight zones of caves,
where the spiders feed on fungus gnats.
Orb-webs are vertical to horizontal (Plate
8).

Distribution. Georgia, northern Flor-
ida, Gulf states, Mexico to Gosta Rica
(Map 9).

REFERENCES CITED

Anderson, J. F. 1978. Energy content of spider

eggs. Oecologia, 37: 41-57.
Balogh, J. I. 1934. Vorlaufige Mitteilung iiber

radnetzbauende Pachygnatlien. Folia zool. hy-

drobiol., 6: 94-96.
Benoit, p. L. G. 1962. Les Araneidae — Nephilinae

afVicains. Rev. Zool. Bot. Afr., 65: 217-2.31.
. 1964. La distribution geographique des

Araneidae — Nephilinae africano-malagaches des

genres Nephila Leach et NepJiilcn^i/s Koch.

Rev. Zool. Bot. Afr., 69: 311-326.

Christenson, T. E. and K. C. Goist, Jr. 1979.
Costs and benefits of male-male competition in
the orl) weaving spider Nephila clavipes, Be-
hav. Ecol. Sociobiol. 5: 87-92.

Dahl, F. 1912. Seidenspinne luid Spinnenseide.
Mitt. Zool. Mus. Berlin, 6: 1-90.

ExLlNE, H. 1948. Moriohology, habits and system-
atic position of Allepeira lemniscata (Walck-
enaer). Ann. Entom. Soc. Amer., 41: 309-325.

Farris, J. S. 1970. Methods for computing Wagner
trees. System. Zoo!., 19: 83-92.

Gerhardt, U. 1921. Vergleichende Studien iiber
die Morphologic des mannlichen Tasters und
die Biologic der Kopulation der Spinnen. Arch.
Naturgesch., 87: 78-245.

. 1926. Weitere Untersuchungen zur Biolo-
gic der Spinnen. Z. Morphol. Okol. Tiere, 6:
1-77.

. 1927. Neue biologische Untersuchungen

an einheimischen und auslandischen Spinnen.
Z. Morphol. Okol. Tiere, 8: 96-186.

. 1929. Uber Grossenxarianten der Mannch-

en von Nephila madagascariensis \'inson.
Zool. Anz., 86: 80-82.

Glatz, L. 1973. Der Spinnapparat der Orthogna-
tha (Arachnida, Araneae). Z. Morphol. Tiere,
75: 1-50.

K.A.STON, B. J. 1948. Spiders of Connecticut. Bull.
State Geol. Natur. Hist. Surv., 70: 1-874.

Knutson, R. M. 1979. Plants in heat. Natur. Hist.,
88: 42-47.

KovoOR, J. 1972. Etude histochimique et cytolo-
gique des glandes sericigenes de quelques Ar-
giopidae. Ann. Sci. Natur. Zool. Biol. Animate,
12 ser. 14: 1-40.

Lem, H. W. 1961. Exolutionary trends in the de-
velopment of palpal sclerites in the spider fam-
ily Theridiidae. J. Morphol., 108: 1-10.

. 1968. The spider genera Gea and Argiope

in America (Araneae, Araneidae). Bull. Mus.
Comp. Zool., 136: 319-352.

. 1975. The American orb-weaver genera

Larinia, Cercidiu and Matigora north of Mex-
ico (Araneae, Araneidae). Bull. Mus. Comp.
Zool., 147: 101-135.

. 1978a. Orb-weaving spiders and their

webs. Amer. Scient., 66: 734-742.

. 1978b. Orb-webs and phylogeny of orb-

wea\ers. Sxnip. Zool. Soc. London, 42: 1-15.
1978c. The American orb-weaver genera

Colplicpcira, Micnithena and Gasteracantha

north of Mexico (Araneae, Araneidae). Bull.

Mus. Comp. Zool., 148: 417-442.
Lem, H. W. AND L. R. Le\i. 1962. The genera of

the spider familv Theridiidae. Bull. Mus.

Comp. Zool., 127i 1-71.
Martin, D. 1978. Zum Radnetzbau der Gattung

Pachijgnathu Sund. (Araneae: Tetragnatliidae).

Mitteil. Zool. Mus. Berlin, 54: 83-95.
MlLLmCE, A. F. 1951. Key to the British genera of

the subfamiK Erigoninae (Famil> Lin>phiidae:

74

Bulletin Miisruni of Compdratiic 7a)oIo<^\i, Vol. 149, No. 1

Araneae). Ann. Maj^. Natur. Hist. (,ser. 12), 4:
545-562.

Nielsen, E. 1932. 'l'\\v Biology of Spiders. Levin
and Muiiks^aard, Copenhagen, 2: 1-724.

PalmgKEN, p. 1978. Taxonomic position of the ge-
nus Mcta (Araneida). Ann. Zool. Fennici, 15:
241-242.

. 1979. On the internal taxonomic characters

ot the genus Leucau^c White (Areneida). Bull.
Brit. Arachnol. Soc, 4: 394-395.

Robinson, M. H. .and H. Mirick. 1971. The pred-
atory behavior of the golden web spider Neph-
ila riavipe.s. Psyche, 78: 123-139.

Robinson, M. H.. AND B. Robinson. 1973. Ecology
and behavior of the giant wood spider Nephilu
mactihita (Fabricius) in New Guinea. Smith-
sonian Contrib. Zool., 149: 1-76.

. 1974. Adaptive complexity: the thermoreg-
ulatory postures of the golden-web spider,
Nephila ciaiipes, at low latitudes. Amer. Mid-
land Nat., 92: 386-396.

. 1976. The ecology and behavior of iVf'/)/) (7 r/

maculata: A supplement. Smithsonian Contrib.
Zool., 218: 1-22.

Senglet, a. 1972. Note sur les Pachygtiatha (Ara-

neae: Tetragnathidae) de la peninsule Ibericjue.

Mitteil. Schwei/erischen Fntomol. Cies., 45:

301 -.305.
SiMHKHLOFE, D. 1976. Species turnover and etjui-

librium island biogeography. Science, 194:

572-578.
SiMBERLOFF, D. AND E. (). WiLSON. 1970. Exper-
imental zoogeography of islands. A two-year

record of colonization. Ecology, 51: 934-937.
Simon, E. 1894-1895. Histoire Naturelie des Ar-

aignees, Librairie Encvclopedicjue de Roret,

Paris, 1: 489-1084.
Thorell, T. 1869. On European spiders. Nova

Acta Reg. Soc. Sci. Upsala ser., 3, 7: 1-242.
WiEHLE, H. 1963. Spinnentiere oder Arachnoidea

(Araneae). XII. Tetragnathidae-Streckspinnen

und Dickkiefer. Die Tierwelt Deutschlands,

49: 1-76.
. 1967. Meta — eine semienteleg\ne Gattiuig

der Araneae. Senckenbergiana Biol., 48: 183-

196.
Yaginuma, T. 1958. Re\ision of Japanese spiders

of the family Argiopidae. 1. Genus Meta and a

new species. Acta Arachnolog., 15: 24-31.

INDEX

\'ali(l names are printed in italics. Page numbers
refer to main references, starred page numbers to
illustrations.

affinis,Azilia. 70*, 71*, 72

affinitata, Allepeira, 14

Allepeira, 11

Araueiiiae, 8, 11

arg\ra, Argyroepeira, 28

argijra, Leucauge, 28, 30*, 31*

argyra, Meta, 28

argyra, Plesiometa, 29

argyra, Tetragnatha, 28

argijrohapta, Leucauge, 23

Argvroepeira, 23

aurulenta, Linyphia, 28

autumnalis, Paclujguatha. 49*, 51*, 57*, 58

Azilia, 69

basilica, Allepeira, 14

basilica, Cyrtophora, 14

basilica, Epeira, 14

basilica, Hentzia, 14

biologica, Pseudometa, 16, 43

hourneti, Meta, 43*

brevis, Pachijgnatha, .54, .55*

bidbifera, Glenognatha, 68

bulbifera, M>sniena, 68

calusa, Pacliygnatha, 62, 63*

clavipes, Aranea, 17

clavipes, Nephila, 17, 18*, 19*, 21*

clercki, Pacliygiuitha, .59, 61*

concolor, Nephila, 17

conferta, Allepeira, 14

conferta, Linyphia, 14

cintisi, Meta, 35

curtisi, Metellina, 33*, 35, 37*, 39*

curtisi, Pacliygnatha, 3.5

doll off, Meta, 43*

dorothea, Pachijgnatha, 55*, .56

dorothea, Pachygnatha, 58

eldorado, Metlcucauge, 45*, 46

einertoni, Glenognatha, 6.3*, 66

foxi, Glenognatha, 67*, 68

foxi, Mimognatha, 68

foxi, Theridium, 68

furcillata, Pachijgnatha, 49*, 53

Glenognatha, 64

Hentzia, 11

Hivaoa, 64

hortorimi, Argyroepeira, 2.5

hortorum, Epeira, 25

iviei, Glenognatha, 67*, 69

kuratai, Pachygnatha, 56

konipirensis, Metlcucauge, 45*

lemniscata, Allepeira, 14

lemniscata, Linyphia, 1.3

lemniscata, Mecynogea, 12*, 13, 15*

Leucauge, 23

longimana, Aranea, 17

mabelae, Leucauge (Argyroepeira), 25

maculata, Aranea, 17

maculata, Nephila, 17, 19*

Mecynogea, 11

menardii, Aranea, 42

menardi, Meta, 41*, 42

mengei, Metellina, 39*

Mecynogea, Metinae, Pachygnatha and Azilia North of Mexico • Levi 75

niericnidc, Mctclliua, 39*

Met a, 38

MetcIIiiui. 32

Metinae, 8, 16

Metleucaufie, 44

niexicana, Azilia, 72

mimetoicles, Metellina, 36, 37*

Mimounatha, 64

Nephila, 16

ornata, Linvphia, 28

Opas, 23

Pacliyf^iiatlui, 47, 49*

peckhaniii, Epeira, 35

pilipes, Aranea, 17

pilipes, Nephila, 17

Plesiometa, 23

plumipes, Nephila, 17

reticulata, Aranea, 36

reticulata, Meta, 36

segnientata, Meta, 36

segmentata, Metellina, 36, 39*

sewardi, Pachygnatha, 59

Tetragnatliinae, 9, 47

thomensis, Nephila, 18

tristriata, Pachygnatha, 51*, 60, 61*, 63*

tristriata, Pachygnatha, 54

vagepicta, Azilia, 72

venusta, Argyroepeira, 25

venusta, Epeira, 25

venusta, Leucauge, 24*, 25, 26*, 29*

wilderi, Nephila, 17

wistariana, Nephila, 17

x-notata, Zygiella, 16

xanthostoma, Pachygnatha, 49*, 57*, 58

u I let in

Museum of

omparative
Zoology

MUS. COMP. ZOOL..

L-IF*.'"-:' A r?v

SEP 1 2 tqno

ZOOGEOGRAPHY OF NORTH AkJIHct!!

UNIONACEA (MOLLUSCA: BIVALVIA)

NORTH OF THE MAXIMUM

PLEISTOCENE GLACIATION

RICHARD I. JOHNSON

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBER 2
16 SEPTEMBER 1980

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Breviora 1952-

BULLETIN 1863-

Memoirs 1864-1938

JOHNSONIA, Department of Mollusks, 1941-

OccASiONAL Papers on Mollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. I. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Sui-vey and Illustrated Catalogue of the Tere-
dinidae (Mollusca: Bivalvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts, 02138, U.S.A.

© The President and Fellows of Harvard College 1980.

ZOOGEOGRAPHY OF NORTH AMERICAN UNIONACEA
(MOLLUSCA: BIVALVIA) NORTH OF THE
MAXIMUM PLEISTOCENE GLACIATION

RICHARD I. JOHNSON'

CONTENTS

Abstract 77

Introduction 78

Zoogeography of North American Unionacea
North of the Maximum Pleistocene
Glaciation 78

Faunal Studies 83

Nomenclature 86

Distribution of Unionidae from the Rivers
South of the Maximum Pleistocene Glacia-
tion, Which Flow into the Missouri and Ohio
Rivers, That Have Not Extended Their
Range [Table 1] 86

Unionacea of the Mississippian and Ohioan
Regions with Refugia in Tributaries of the
Missouri and Ohio Rivers on Their South
Side [List 1] 86

Unionidae of the Ozarkian Region with Distri-
butions Not Extending into the Mississip-
pian or Ohioan Regions [List 2] 96

Unionacea of the Northern Atlantic Slope Re-
gion North of Maximum Pleistocene Gla-
ciation [List 3] 98

L^nionacea of the Pacific Coastal Region

[List 4] 99

Distribution of Unionacea in the Mississippian
and Ohioan Regions [Table 2] 103

Distribution of Unionacea in the St. Lawrence
River System [Table 3] 106

Distribution of Unionidae in the Canadian In-
terior Basin [Table 4] 106

Distribution of Unionacea in the Ozarkian Re-
gion [Table 5] 106

Systematic Section

Selection of Monographed Species 106

Acknowledgements 107

Pattern Used in Descriptions 110

Arrangement of Records 110

' Museum of Comparative Zoology, Harvard Uni-
versity, Cambridge, MA 02138.

Family Unionidae

Fusconaia uzarkensis (Call) 111

Anodonta (Utterbackia) suborhiculata Say 113

Arcidens confragosus (Say) 115

Arkansia wheeleh Ortmann and Walker __ 119

Actinonaias rafinesqueana (Frierson) 120

Lampsilis ahrupta (Say) 122

Lampsdis higginsi (Lea) 124

Lampsdis powelli (Lea) 127

Proptera capax (Green) 128

Vdlosa arkansasensis (Lea) 131

Vdlosa reeviana (Lea) 132

Cijprogenia aberti (Conrad) 134

Cyprogenia stegaria (Rafinesque) 136

Ptychobrcmchus occidentalis (Conrad) 138

Literature Cited 140

Index to Relevant Taxa 145

Abstract. The Unionacea (Mollusca: Bivalvia)
now found north of the maximum Pleistocene gla-
ciation in North America (which extended roughly
to the Missouri and Ohio rivers in the west and to
the New York-Pennsylvania boundary in the east),
excluding two species of western origin, migrated
there from southern refugia and are not autochtho-
nous. The probable origins and likely migration
routes of the unionids found in that northern region
are discussed with special emphasis on their move-
ments within the American Interior Basin.

In addition to a ubiquitous fauna widely dis-
persed in both the Mississippi and Ohio rivers, and
foimd on both the Ozark and Cumberland plateaus,
each plateau has a unique assemblage of species.
The post-glacial migration of some of these Ozar-
kian, Cumberlandian, and Ohioan species accounts
for the difierences in the faunas of the Mississippi
River above the Missouri River (Mississippian Re-
gion) and the Ohio River drainage (Ohioan Region).
Tables list the faunas of the several river systems
and the species are listed to indicate their probable
refugia. Fourteen species are monographed and in-

Bull. Mus. Comp. Zool., 149(2): 77-189, September, 1980 77

78 Bulletin Museum of Comparative Zoology, Vol. 149, No. 2

elude their known distritnitional records. Ei^ht ot
these are found only in the O/.arkian Region.
Among the additional six species, two other O/.ar-
kian specii's are included which extended their
ranges in post-glacial time, Lampsilis hi^^insi and
Propteru aiixix. Lampsilis hi^^insi and ahrupta, a
closely related Cumherlandian species, are com-
pared. They illustrate o\erlapping distribution in
the Mississippian and Ohioan regions. Cijpro^enia
stenaria, a Cumherlandian species which extended
its nmge into the Ohioan Region, is compared with
C. ahcrti, an Ozarkian species. Arcidens confra-
^osus and Anodonta suborbictdata, which extend
south to the Gulf Coastal Region, are examples of
species that entered the Ohioan Region from the
West. The repopulation of the St. Lawrence River
system by species from the Ozarkian, Mississippi-
an, Cumherlandian, Ohioan, and Northern Atlantic
Slope regions is discussed as is the repopulation of
the Canadian Interior Basin.

INTRODUCTION

This paper brings together the avail-
able data on the distribution of the
Unionacea found north of the maximum
Pleistocene glaciation in North America
(roughly to the Missouri and Ohio rivers
in the west and to the New York-Penn-
sylvania boundary in the east). Much of
this vast area was repopulated within ap-
proximately the past ten thousand years
during the retreat of the Wisconsin ice
sheet. Many of the dispersal routes have
been documented, but no previous paper
discusses the probable refugia of the
species, or lists the known faunas of the
relevant river systems.

It should be mentioned that the mobil-
ity, and thus the means of dispersal, of
the Unionacea is passive. It is dependent
on the temporary attachment of the larval
forms (glochidia) onto passing fish.

ZOOGEOGRAPHY OF NORTH
AMERICAN UNIONACEA NORTH
OF THE MAXIMUM PLEISTOCENE
GLACIATION

Introduction. Prior to the Cretaceous
Period the then Cumberland Plateau ex-
tended continuously from the Appala-
chian Mountains southwest into western
Texas. The Mississippi River did not ex-
ist. What is remarkable about the present

unionid faunas on the Ozark Plateau, on
both sides of the Ozark Crest, and on the
Cumberland Plateau, is their similarity,
even after the passage of some 200 mil-
lion years. The maximum Pleistocene
glaciation extended southward, west of
the Appalachian Mountains, roughly to
the present Missouri and Ohio rivers, and
to the Pennsylvania-New York boundary
in the east. During the Nebraskan glacial
stage, which represents the earliest and
most southerly extension of the ice sheet,
the unionid fauna was eliminated north
of this line. With the exception of two
species from the Pacific Coastal Region,
the Interior Basin (including the Cana-
dian Interior Basin), the St. Lawrence
River system, and the Northern Atlantic
Slope, have since been repopulated with
species from southern refugia.

Ozarkian Region. The upper portions
of the Red River system, Oklahoma and
Arkansas; Black River system, Arkansas;
Arkansas River system, Kansas, Okla-
homa, and Arkansas; White River system,
Missouri and Arkansas; St. Francis River
system, Missouri and Arkansas, all below
the Ozark Crest; and the Osage, Gasco-
nade, and Meramec river systems, all in
Missouri, and all above the Ozark Crest;
contain an assemblage of 8 species of
Unionidae found nowhere else (Tables 1,
5 E). These river systems comprise the
Ozarkian Region (H. and A. van der
Schalie, 1950: 450, map 1 IV).

In addition to these 8 species (Table 5
E) there are two others, Lampsilis hig-
ginsi and Proptera capax, which have
spread beyond into the Mississippian
and Ohioan regions, the latter also into
the Lake Erie drainage of the St. Law-
rence River system. Four species (Table
5 D) occur in the Ozarkian and Cumher-
landian regions which are not found in
the Mississippian Region. Nine species
(Table 5 C) found in the Ozarkian are not
found in the Mississippian or in the Cum-
herlandian regions but are found in the
Gulf Coastal Region. Nine species (Table
5 B) found in the Gulf Coastal and Ozark-

Zoogeography of North American VmoNACEA- Johnson 79

ian regions are also found in the Missis-
sippian and Ohioan regions. With the ex-
ception of Lampsilis radiiita siliquoidea
in the Cumberhuuhan Region, 40 species
(Table 5 A) occur in the Meramec River
system, which are generally distributed
in both the Mississippian and Ohioan re-
gions.

Cumberlandian Region. Ortmann
(1924a: 40) defined the Cumberlandian
Region to include: the drainages of the
Tennessee River system from the head-
waters to the vicinity of Muscle Shoals,
in Colbert and Lauderdale counties, Al-
abama; and the Cumberland River sys-
tem from the headwaters to the vicinity
of Clarksville, Montgomery County, Ten-
nessee (Ortmann, 1925: 366). Ortmann
(1924a: 40) discussed the unionid fauna
of the Duck River drainage, which is at
present, a tributary of the Tennessee Riv-
er system. The upper portion of the Duck
River has a fauna that is 38 percent Cum-
berlandian. Ortmann suggested that this
was the original fauna and that there once
had been stream confluence with it and
both the Tennessee and Cumberland riv-
er systems.

Originally Ortmann (1924a: 40) regard-
ed as Cumberlandian only those unionid
species currently confined to the areas of
the Tennessee and Cumberland river
systems as defined above. This Cumber-
landian fauna consists of 30 species (Ta-
ble 1). All but one of these occur in the
Tennessee River, and 21 are also found
in the Cumberland River. Later Ortmann
(1925: 370) suggested that there were
Cumberlandian species that had de-
scended the Tennessee and/or Cumber-
land rivers and invaded the Ohio drain-
age, though he never indicated which
they were. Table 2, C indicates that the
following are those additional Cumber-
landian species:

52 Lampsilis ahrupta
40 Carunciilina glans
20 Plethohasus cicatricosus
46 Plagiola personata

47 Plagiohi flexuosa

43 Plagiola sampsoni

44 Plagiola propinqua

See appropriate number under: Unionacea of the
Mississippian and Ohioan Regions With Ref'usia in
Tributaries of the Missouri and Ohio rivers on Their
South Side [List 1].

Thirty-seven of the 90 species of
unionids found in the Tennessee River
are Cumberlandian, as are 27 of the 78
found in the Cumberland River. These
t\yo assemblages are the largest number
of unionid species found in any of the
world's rivers.

Mississippian Region. H. and A. van

der Schalie (1950, map 1) regarded as the
Interior Basin (or Mississippian Region),
all of the river in the Gulf Coastal Region,
from the Rio Grande to and including the
Mobile-Alabama-Coosa River; the entire
Mississippi River system, including the
Ohio; the Canadian Interior Basin; and
the St. Lawrence River system west of
Lake Erie. Excepted were the Ozarkian
and Cumberlandian regions as defined
above.

The Mississippian Region is more nar-
rowly defined here to include: the Mis-
souri River and all those rivers draining
into its south side, excluding the Osage,
Gasconade, and Meramec river systems
which are in the Ozarkian Region; the
Blackwater River since it has no Ozarkian
species in it; and all of the Mississippi
River, above the Missouri, and the rivers
flowing into it west and north of the Ohio
River. In this area, as thus defined, occur
53 species of Unionacea; 40 are so uni-
formly distributed, in both the Mississip-
pian and Ohioan regions, that little can
be said of their general zoogeography. All
of them occur in the Meramec River sys-
tem, which may have been the major re-
fugium for these species west of the Ohio
River (Table 2 A). With the exception of
Margaritifera falcata, from the Pacific
coastal Region, 11 other species had re-
fugia below the Ozark Crest, and five also
had a refugium in the Meramec River.

80

Bulletin Muscuiii of (.oniixinitivi' Zoology, Vol. 149, No. 2

These 11 western species penetrated the
Ohioan Region to some extent, yet only
one Ohioan and one Cuniherhmdian
species occur in the Mississippian Re-
gion. During the Wisconsin glacial stage
(Plate 1), the ice sheet in the Mississip-
pian Region was much less extensive
than in the Ohioan Region, and perhaps,
the former fauna had achieved a stability
that prevented the newly available
Ohioan fauna from penetrating it.

Ohioan Region. This area consists of
all the rivers that flow into the Ohio Riv-
er, excluding the Tennessee and Cum-
berland rivers.

There are 72 species of Unionacea in
this area, 40 of which are found in the
Mississippian Region as well (Table 2,
A). The Green, Salt, and Kentucky rivers,
on the south side of the Ohio, are each
fairly well represented by this group of
species; but the Trade water. Licking,
and Big Sandy rivers are not, suggesting
that these were repopulated in post-gla-
cial time. In addition to the seven species
found in the Ohio River drainage, regard-
ed above as Cumberlandian, are 16 ad-
ditional species that appear to be of
Ohioan origin (Table 2, C). With the ex-
ception of Lasmig,ona coinpressa, miss-
ing in the Tennessee, Cumberland, and
Green rivers, all the other species occur
in the Tennessee River, and most are also
found in the Cumberland and Green
rivers.

Ortmann (1926: 186) concluded that
"there is a sharp line between the Cum-
berland and Green rivers in southern
Kentucky separating two apparently old
faunas, the Cumberlandian and Ohioan."
Ortmann (1913: 382) had previously in-
dicated that the Kentucky, Licking and
Big Sandy rivers contained Ohioan
species (Table 2, C) and suggested that
the uniformity of the fauna of the upper
Ohio basin was a character acquired in
post-glacial time. He further noted that
remnants of the pre-glacial Ohioan faima
may have had refugia in the Kanawha,
Monongahela, and [Allegheny] rivers.

Table 2, C indicates that the Tennes-
see and Cumberland rivers were major
refugia of both Cumberlandian and
Ohioan unionid species, and the Green
River was a major refugium for the
Ohioan species. These rivers appear to
be the most important refugia for those
species that repopidated the Wabash and
Greater Maumee rivers (Plate 1, D).

Table 2, B shows the distribution of 11
Mississippian unionids that penetrated
the Ohioan Region. Most of these species
penetrated at least to the Wabash River
drainage.

St. Lawrence River System. The St.

Lawrence River system has no natural
connections with the Mississippi or Ohio
river systems, and flows for some 2,000
miles east to the Atlantic Ocean. Walker
(1913: 18) studied the distribution of the
unionids found in the Great Lakes.

Lake Superior. Walker found Elliptio
complanata, an Atlantic Slope species, in
Lake Superior, at the western extremity
of the St. Lawrence River system. He
suggested that this species arrived in this
area by way of the Mohawk and Trent
outlets into Georgian Bay of Lake Huron
and from there into Lake Superior (Walk-
er, 1913: 58). Ortmann (1924b) confirmed
this, pointing out that toward the end of
glacial time. Lakes Superior, Michigan,
and Huron (then Lake Algonquin IV)
flowed through the Trent Valley from
Georgian Bay, Lake Huron, directly into
Lake Ontario (then Lake Iroquois). Later
the North Bay outlet opened, allowing
the water to flow from the end of Geor-
gian Bay through the Ottawa Valley to
the Ottawa Sea. After the weight of gla-
cial ice was removed by further melting,
the Ottawa Valley was sufficiently ele-
vated to eliminate it as an outlet.

Clarke (1973: 40) pointed out that E.
complanata has a continuous distribution
in the Hudson Bay drainage from the
Moose River system in northwestern
Quebec and northeastern Ontario to the
headwaters of the Albany River system in
northwestern Ontario. Since the head-

Zoogeography of North AMEmcAN Unioxace a •/o/inso/i 81

waters of the Albany River are still con- proper. On the basis of the unionid fan-

nected to Lake Superior through Long nas of rivers now flowing into Lake Erie,

Lake and Lake Nipigon, there can be lit- Ortmann concluded that the Huron and

tie doubt that this route also served in the Raisin rivers in Eastern Michigan; the

post-glacial migration of E. complanata Sandusky, Vermilion, Lower Cuyahoga

into Lake Superior. and Grand rivers in Ohio; as well as the

Lake Michigan. Van der Schalie Grand River in Ontario were once part of

(1963) established that a number of the Greater Maumee drainage. Van der

unionid species found in the Millecoquin Schalie (1938), also on the basis of unio-

River, in the eastern part of the Upper nid distribution, concluded that the Clin-

Peninsula of Michigan, the Carp and Oc- ton River, which flows into Lake St.

queoc rivers, in the northern part of the Clair, was also part of the Greater Mau-

Lower Peninsula, presently draining into mee drainage; but he determined that the

Lakes Michigan and Huron; indicate that Rouge River, between the Clinton and

these rivers were once connected to the Huron rivers, was of later origin (Figure

now submerged Mackinac River system. 1).

This system received its fauna from the Lake Ontario. This, the newest of the

Fox River which was then connected to Great Lakes, has a unionid fauna of seven

the Wisconsin River in the vicinity of species. Two are Atlantic Slope species.

Portage, Columbia Co., Wisconsin (Plate Clarke and Berg (1959: 3) monographed

1, B). He further established that the the species found in this lake and those

present distribution of the unionids in of the Genesee River and Finger Lakes

the rivers of western Michigan indicate Basin which flow into it on its south side,

that before the formation of Lake Michi- and suggested that the Atlantic Slope

gan, the rivers of western Michigan (e.g. species found there were derived from

the Muskegon, Grand, and St. Joseph) the Susquehanna Basin by way of glacial

connected with the Des Plaines River Lake Newberry. When the Wisconsin ice

through the Chicago Outlet and Illinois sheet melted, the Finger Lakes Basin

River to the Mississippi (Plate 1, C). He was inundated by Lake Newberry which,

also deduced that the Muskegon River in its earlier stages, drained southward

was at one time a tributary of the Grand into the Susquehanna drainage basin

River (of western Michigan) and also that (Plate 1, E). As the glacier continued to

the fauna of the Saginaw, now flowing recede a lower channel was uncovered

into Lake Huron, was derived from the and the water drained westward into the

previously mentioned Grand River dur- Erie Basin. Later, when an even lower

ing the confluence of the former drain- outlet was uncovered, the lake in the

ages (van der Schalie, 1961: 157). Erie Basin merged with the Finger Lakes

Lake Erie. Ortmann (1924b) pointed Basin and flowed eastward into the Mo-
out that during the Trent Stage of the hawk and Hudson river valleys (Plate 1,
Wisconsin, the Maumee River flowed F).

through the partially dry bed of Lake Lake Champlain. Simpson (1896: 381)

Erie into the Wabash River. The Man- suggested that sometime since the end of

mee and Wabash rivers were connected the Wisconsin, there had been a connec-

in the vicinity of Fort Wayne, Allen Co., tion between the Hudson River and Lake

Indiana (Plate 1, D). It was by this route Champlain when the Erie Basin merged

that 43 species of unionids (Table 3, A), with the Finger Lakes Basin and flowed

from the Mississippian, Cumberlandian, eastward into the Mohawk and Hudson

and Ohioan faunas, found their way into river valleys. He further suggested that

the Greater Maumee drainage. Many of the Lake Champlain favina may have

these species are not found in Lake Erie been derived from the St. Lawrence Riv-

82 Bulletin Mu.sciiin of C

oinixiid

live Zoolo^ij, Vol. 149, No. 2

ICE SHEET

Figure 1. Trent outlet stage of the Great Lakes. (After Ortmann, 1924, fig. 2.)

The darkened area represents the melt water which flowed through the Trent Outlet into the Mohawk River (arrow). The
Greater Maumee River system is postulated, and the rivers which formed part of it are numbered. 1. Clinton, 2. Huron,
3. Raisin, 4. Sandusky, 5. Vermilion, 6. lower Cuyahoga, 7. Grand River, Ohio, 8. Grand River, Ontario. The arrow between
3 and 4 represents the flow of the Maumee River system into the Wabash.

er by way of the Richelieu River. Since
the present fauna, with the exception of
EUiptio complanata, is from the west,
this second explanation is more probable.
Ottawa River. It is assumed that the
large number of Atlantic slope species in
the Ottawa River are derived from the
southern side of the lower St. Lawrence
River, but this is not documented here.

Canadian Interior Basin. (Plate 1, In-
sert.) North of the Mississippian Region
and the St. Lawrence River system, this
vast region comprises more than one
third of North America. It is divided into
the Arctic and Hudson Bay drainages.
The latter has a number of natural con-
nections with the St. Lawrence River sys-

tem. The Albany River system in Ontario
is joined to Lake Superior by Lake Nip-
igon and Long Lake, both of which have
north and south flowing outlets. The
Rainy River-Winnipeg River system is
similarly joined to Lake Superior at Sa-
ganaga [Seiganagah] Lake which drains
both east and west (Clarke, 1973: 6). Most
of the unionid fauna of the Canadian In-
terior Basin is of Mississippian origin
(Table 4). Thirteen of the 16 species in
the Hudson Bay drainage are derived
from a former confluence of the Minne-
sota River, of the upper Mississippi, and
the Red River of the North (Plate 1, A).
Prior to the recent erection of dams, these
rivers were periodicallv joined bv floods
(Dawley, 1947: 680). three of these 13

Zoogeography of North American Unionacea • Johnson

83

species were also derived from the St.
Lawrence River system, and an addition-
al species was described exclusively
from the latter system. One species is
from the Pacific Coastal region and one
appears to be speciating in the area. Only
three of the 16 species occur in the Arctic
Region.

Pacific Coastal Region. This consists
of rivers that flow into the Pacific Ocean
to, and including, Alaska which was un-
glaciated. The unionid fauna is com-
prised of eight species, two of which are
found outside of the region.

Northern Atlantic Slope Region. The

unionid fauna of the Northern Atlantic
Slope consists of 15 species, all of which
had refugia south of the Terminal Mo-
raine. This also includes one subspecies,
Anodonta cataract a fragilis found in
Newfoundland, New Brunswick and
Nova Scotia.

FAUNAL STUDIES

The tables in this paper are based
mostly on reliable published records, and
on specimens in the Museum of Com-
parative Zoology, or elsewhere, as noted.

Ozarkian Region. The unionid fauna
of the Red River system was enumerated
by Isely (1925), Strecker (1931), and by
Valentine and Stansbery (1971). The
unionids of the Ouachita River drainage
of the Black River system were listed by
Wheeler (1918). The unionids of the up-
per tributaries of the Arkansas River sys-
tem in Kansas were studied by Scammon
(1906), Clarke and Gillette (1911), and
again by Murray and Leonard (1962). The
unionids of the tributaries of the Neosho
River drainage in Missouri were listed by
Utterback (1917), and those of the Arkan-
sas River system in Oklahoma by Isely
(1925). The unionids of the White and
Black river drainages of the White River
system in Missouri were monographed
by Utterback (1915-1916), and those of
the White River in Arkansas were studied

by Call (1885). Meek and Clark (1912)
listed the unionids of the Big Buffalo
Fork, and Wheeler (1914) those of the
Cache River, both tributaries of the
White River in Arkansas. Utterback
(1917) listed the unionids of the St. Fran-
cis River system in Missouri, and Call
(1885) studied those in Arkansas. The
unionid faunas of the Osage, Gasconade,
and Meramec rivers were summarized by
Utterback (1917) in a table. A recent re-
port by Buchanan (1977) based on exten-
sive collections from the Meramec River
was very useful in establishing its fauna.
Buchanan (1980) later added Anodon-
toides ferussacianus and Simpsonicon-
cha ambigua to the Meramec River fau-
na. Gordon, et al. (1980), while primarily
interested in the unionids of Arkansas as
a political unit, added a number of rec-
ords to several of the Ozarkian river sys-
tems. The faunas of the Osage and Gas-
conade rivers, as well as some of the
drainages south of the Ozark Crest, were
augmented by records made by the 1964
Harvard-Ohio State Museum expedition
to this area.

Mississippian Region. The unionids
of the upper Missouri River were studied
by Coker and Southall (1915). West of the
Osage River, and like it flowing into the
Missouri River on its South side, is the
Blackwater River. Since, according to
Utterback's (1917) table its fauna in-
cludes no Ozarkian species, it properly
belongs in the Mississippian Region. The
distribution and number of unionid
species in the Mississippi River, dis-
cussed by H. and A. van der Schalie
(1915), was based on several mussel sur-
veys made in the Mississippi River, in-
cluding one by Crier and Mueller (1922-
23), and on the extensive collection made
by Max. Ellis during 1930-31.

Baker's (1928) monograph of the Wis-
consin mollusks was useful in establish-
ing the fauna of the Wisconsin and Fox
rivers, as was Dawley's (1947) report
which enumerated the fauna of the Min-
nesota River and mentioned its relation-

84

Bulletin Museum of Comparative Zoology, Vol. 149, No. 2

ship to the Red River of" the North. The
present (1973-77) distribution of the
unionids of the upper Mississippi River in
Wisconsin was made avaihible liy Ma-
thiak (1979). The fauna of the Illinois
River, past and present, was studied by
Starrett (1971). The unionids of its trib-
utaries in the Kanakee Basin had been
enumerated by Wilson and Clark (1912).
A popidar work by Parmalee (1967) on
the freshwater mussels of Illinois also
contained useful records.

Cumberlandian Region. Ortmann
(1918) reported on the unionid fauna of
the upper Tennessee River. Subsequent-
ly, additional collections have been re-
ported upon from its upper most tributar-
ies. The species in the Powell River were
listed by Ahlstedt and Brown (1980),
those of the Clincli River by Stansbery
(1973) and again by Bates and Dennis
(1978). Ortmann (1925) reported on the
lower Tennessee River, below Walden
Gorge, to Muscle Shoals in Lauderdale
and Colbert counties, Alabama, to which
van der Schalie (1939) added additional
notes. Ortmann (1924a) studied the
unionids of Duck River, a tributary of the
lower Tennessee. The Duck River unio-
nids were restudied by Isom and Yokley
(1968) and subsequently reexamined by
van der Schalie (1973). The unionid fau-
na of the Cumberland River was studied
by Wilson and Clark (1914), and that of
the upper part of the river, above the
Tennessee boundary, was examined
again between 1947-1949 by Neel and
Allen (1964) before the completion of the
Wolf Creek Dam. Blankenship and
Crockett (1972) listed the unionids found
in the Rockcastle River from Livingston,
Kentucky.

Ohioan Region. The first important
summary of the Ohioan unionid fauna
was made by Call (1897) in a paper on
the molluscan fauna of the hydrographic
basins of Indiana, in which he listed
those species found in the Wabash River
and its tributaries. Call's (1900) mono-

graph of Indiana mollusks was supple-
mented by both Blatchlev and Daniels
(1903) and Daniels (1915) and was re-
vised by Goodrich and van der Schalie
(1944). Baker (1922) studied the mollus-
can fauna of the Big Vermilion River, a
large tributary of the Wabash in Illinois.
Meyer (1974) investigated the fauna of
the Wabash and White rivers in Indiana,
and Clark (1976) reviewed the unionid
fauna, past and present, of the Wabash
River drainage. The unionids of the
Green River were extensively studied by
Ortmann (1926) and were listed again by
Clench and van der Schalie (1944). Stans-
bery (1965) added several formerly un-
reported species from the vicinity of
Munfordville, Kentucky. The Kanawha
(or New) River unionids were listed by
Ortmann (1913: 305-308) and again by
Morris and Taylor (1978). The fiiuna of
the Monongahela, Allegheny, and upper
Ohio rivers were discussed in Ortmann's
(1919) elegant monograph of Pennsylva-
nia naiades.

The unionids of the Tradewater and
Salt rivers were listed by Clench and van
der Schalie (1944), and Rosewater (1959)
again listed the species of the Salt. Those
of Floyd's Creek, a tributary of the Salt
River, were listed by Taylor (1980b). The
unionids of the Kentucky River were
studied by Danglade (1922). Ortmann
(1913: 308-310) listed those of the Lick-
ing and Big Sandy rivers.

In addition to the Wabash River, from
west to east, there are three other major
tributaries on the north side of the Ohio
River: the Miami, Scioto and Muskin-
gum. The Miami River is not included in
Table 2 because there is no published list
of its unionids. It is reasonalole to assume
that the fauna of the Miami River is, or
was, similar to that of the Scioto and Mus-
kingum rivers. A mimeographed list of
the Scioto River species was circulated
by Stansbery (1961). The species of Little
Darby Creek, one of its tributaries, were
listed by Stein (1965). Most of the species
of the Muskingum drainage are covered
by Sterki (1900, 1902) in papers on the

Zoogeography of North American UNioNACEA-yo/i/i.so/i

85

land and freshwater mollusca of the up-
per Tuscarawas Valley.

La Rocque's (1967) compilation on the
unionids of Ohio contained useful infor-
mation on the distribution of Ohio
species. Ortmann's (1909, 1912, 1919)
studies of the naiades of Pennsylvania re-
cord the unionid fauna of the upper Ohio
drainage.

A number of mimeographed reports by
state and federal agencies on the present
reduced naiad fauna of the Ohioan region
have appeared, and should be men-
tioned, although they did not include
data required here. Williams (1969) re-
ported on the mussel fisheries in the Ten-
nessee, Ohio and Green rivers. Bates
(1970) on the Ohio mussel fisheries, Den-
nis (1970) on the mussels of Pennsylva-
nia, and Bates (1971) and Taylor (1980a)
on the mussels of the West Virginia por-
tion of the Ohio River.

St. Lawrence River System.

Lake Superior. The Lake Superior
unionids were listed by Goodrich and
van der Schalie (1932).

Lake Michigan. The unionids of Lake
Michigan, and the tributaries on its west-
ern side in Wisconsin, were studied by
Baker (1928). The Wisconsin tributaries
were extensively collected during 1973-
77 and reported on by Mathiak (1979),
who added five new records. Van der
Schalie (1961) enumerated the fauna of
the Fox River in Wisconsin that flows
into western Lake Michigan, and listed
the species from a number of western
Michigan rivers that flow into Lake Mich-
igan on its eastern shore. Van der Schalie
(1936, 1941, 1963) also listed the unio-
nids of the St. Joseph, Grand and Mus-
kegon rivers, on the lower Michigan Pen-
insula. The unionids of Lake Michigan
proper were listed by Goodrich and van
der Schalie (1932).

Lakes Huron and St. Clair. The
unionids of these lakes were listed by
Goodrich and van der Schalie (1932).

Lake Erie. The Lake Erie unionids
were listed by Walker (1913), Ortmann

(1924b), and by Goodrich and van der
Schalie (1932). Stansbery and Stein
(1962) circulated a useful mimeographed
list of the unionids of the Island Region
in the western part of Lake Erie. Robert-
son and Blakeslee (1948) listed the
unionids found in the easternmost part of
the Lake and of the rivers which flow into
it in the Niagara Frontier Region. The
naiades of the Huron River in southern
Michigan, that flows into western Lake
Erie, were studied by van der Schalie
(1938). The fauna of the Maumee River,
which also flows into western Lake Erie,
were enumerated by Call (1897), and
again by Wilson and Clark (1912) based
on new collections. Clark (1977) listed
the unionids of the St. Joseph River, a
tributary of the Maumee. Strayer (1979)
again listed the naiads of the St. Joseph
River, as well as those of the Raisin and
Huron rivers.

Lake Ontario. The Lake Ontario
unionids were listed by Goodrich and
van der Schalie (1932). Clarke and Berg
(1959) monographed the unionids of
Lake Ontario and those species found in
the Finger Lakes Region, draining into
the Lake on its south side.

Lake Champlain. The Lake Cham-
plain unionids were listed bv C. W. John-
son (1914).

Ottawa River. The unionids of Ottawa
River, a major tributary of the St. Law-
rence River, are based on the lists of
Latchford (1882) and Whiteaves (1895),
and especially on La Rocque and Ough-
ton's (1937) account of the Unionidae of
Ontario.

Canadian Interior Basin. The union-
id fauna of the Arctic and Hudson Bay
drainages is based on the summary by
Clarke (1969) and on the distributional
maps in Clarke's (1973) monograph of the
Canadian Interior Basin mollusks.

Pacific Coastal Region. The list of
Unionacea from the Pacific Coastal Re-
gion is from Burch (1973, 1975).

Northern Atlantic Slope Region. The

86

Bulletin Museum of Comparative Zoology, Vol. 149, No. 2

list of Unionacea from the Northern At-
lantic Slope is from Johnson (1970). Pa-
pers by Athearn and Clarke (1962) and
Clarke and Rick (1963) were also used.

NOMENCLATURE

The nomenclature and arrangement of
the species covered in this paper are es-
sentially those of Burch (1973, 1975). E.x-
ceptions are mentioned in the notes and
under the several monographed species.
The species concepts and arrangement oi
Phigiola are those of Johnson (1978), as
are the Atlantic Slope species (1970). The
unionid species from the Canadian Inte-
rior Basin are from Clarke (1973).

Simpson, and some later authors, es-
pecially F. C. Baker, gave nomenclatorial
standing to a number of ecophenotypic
variants; while the names applied to
these have no standing as species or sub-
species in modern systematics, some still
continue to appear in the literature.

Included as notes under the species in
the several faunal lists are names which
have been promulgated in the literature
since Simpson's (1914) monographic
work on the Unionacea. Not listed are the
taxa introduced by the Sicilian noble-
man, Marchese A. De Gregorio (1914),
who redescribed many well known North
American unionids with abandon. All of
his taxa were synonymized by Walker
(1918) and have since been forgotten.

Morrison (1969) asserted that a number
of Rafinesque's taxa, previously thought
unrecognizable, had priority over a num-
ber of well known ones. Since he gave
no support for his assertions, Rafines-
que's names are placed under their sup-
posed synonyms.

DISTRIBUTION OF UNIONIDAE FROM
THE RIVERS SOUTH OF THE MAXIMUM
PLEISTOCENE GLACIATION, WHICH
FLOW INTO THE MISSOURI AND OHIO
RIVERS, THAT HAVE NOT EXTENDED
THEIR RANGE [Table 1J

UNIONACEA OF THE MISSISSIPPIAN
AND OHIOAN REGIONS WITH REFUGIA
IN TRIBUTARIES OF THE MISSOURI
AND OHIO RIVERS ON THEIR SOUTH
SIDE [List 11

In addition to the 72 species listed be-
low, the fauna also includes two species
from elsewhere: a. Margaritifera falcata
from the North Pacific Coastal Region
found in the Mississippian Region, and
B. Elliptio complanata from the North
Atlantic Slope Region found in the
Ohioan Region.

FAMILY MARGARITIFERIDAE

1. Cumberlandia monodonta (Say 1829)
Found below the Ozark Crest only in
the:

Black River System

Ouachita River Drainage. Arkan-
sas: Ouachita River, above Skillern's
Shoals, Arkadelphia, Clark Co.
(Wheeler, 1918: 121).

FAMILY UNIONIDAE

2. Amblema plicata (Say 1817)

Unto peruviana Lamarck 1819; Am-
blema costata Rafinesque 1820;
Unio undulatus Barnes 1823; Vnio
rariplicata Deshayes 1830; Unio
perplicata Conrad 1841; Quadrula
perplicata quintardii Cragin 1887;
Unio pihbnji Marsh 1891, Nautilus
5: 1 (Little Red River, Arkansas; ho-
lotype ANSP 41589a figured by Pils-
bry, 1892, Proc. Acad. Nat. Sci. Phila.
44: 131, pi. 8, figs. 7-8; also 1892,
Nautilus 6: 1, pi. 1, figs. 7-8); Coke-
ria southalli Marshall 1916.

3. Fiisconaia ebena (Lea 1831)
Obliquaria pusilla Rafinesque 1820
(Morrison, 1969: 24).

4. Fiisconaia flava (Rafinesque 1820)
Unio rubiginosus Lea 1829; Unio
cerinus Conrad 1838; Unio hebeta-
tus Conrad 1834; Unio rubidus Lea
1861; Fiisconaia flava parvula Crier
1918.

Zoogeography of North American \J:sio^acea' Johnson 87

OZARKIAN REGION

CUMBERLANDIAN

TABLE I

DISTRIBUTION OF UNIONIDAE OF THE RIVERS
SOUTH OF MAXIMUM PLEISTOCENE GLACIATION
THAT FLOW INTO THE MISSOURI AND OHIO
RIVERS, THAT HAVE NOT EXTENDED THEIR

RANGE

Lampsilis brevicula (Call 183?)
Ptychobranchus occidental is (Conrad 1
Cyprogenia aberti (Conrad I85O)
Fusconaia ozarkensis (Call I887)
Fusconaia barnesiana (Lea I838)
Quadrula tuberosa (Lea 1840)
Pleurobema oviforme (Conrad 183^*)
Alasmidonta raveneliana (Lea 183^)
Pegias fabula (Lea I836)
Actinonaias pectorosa (Conrad 183^)
Plagiola interrupta (Rafinesque 1820)
Plagiola arcaeformis (Lea I83I)
Plagiola lenior (Lea 1842)
Plagiola biemarg inata (Lea 1857)
Plagiola capsaeformi s (Lea 183'*)
Plagiola florentina (Lea 1857)
Plagiola turgidula (Lea I858)
Plagiola haysiana (Lea 183^*)
Plagiola stewardsoni (Lea 1852)
Medionidus conradicus (Lea 183^*)
Villosa taeniata (Conrad 183'*)
Villosa trabalis (Conrad 183'*)
Villosa vanuxemens is (Lea I838)
Dromus dromas (Lea 1834)
Ptychobranchus subtentum (Say 1825)
Fusconaia cuneolus (Lea 18'40)
Fusconaia edgariana (Lea 1840)
Quadrula intermedia (Conrad I836)
Lexingtonia dolabel loides (Lea \8kO)
Lasmigonia holstonia (Lea I838)
Lampsilis virescens (Lea I858)
Lemiox rimosus (Rafinesque 1831)
Carunculina cylindrella (Lea 1868)
T'leurobema qibbcrur? (Lea I838)
Villosa ortmanni (Walker 1925)
Lasmigona subviridis (Conrad 1835)

836)

REGION

E
4)
■M
U)
>-

to

E

B

1U

l-

(U

V

4-)

(1>

■I-'

E

4-i

in

>

l/l

(U

in

(U

>-

E

>-

4->

>-

en

to

E

a:

^-J

<U

oo

in

to

rn

OJ

t/1

4-»

^

c

u

■w

, ^

<u

in

u

(/I

i_

.—

0)

Ol

3

1-

>-

<u

<u

(T3

>

:^

a>

0

CO

>

u

>

L.

—

to

z

.—

0)

Q

a:

^

L_

0:

>

CC

i_

i-

1-

<D

—

l-

-D

0)

__0

n:

>

a>

oc

1)

0)

C

>

N

•—

-0

0)

>

ID

0

oc

m

0

Ul

—

cr:

ID

c

a>

U1

cc

1-

^

3

a>

0

E

OJ

m

c

3

0

ai

0

<D

c

-^

^

<u

TO

—

n}

t/i

1-

c

u

E

<u

C

0)

i/i

ID

0)

(U

3

3

\-

(T3

JD

0

0

2:

H-

Q

tJ

t3

:^

X

X

X

X

X

X

X

X

X

X

X
X
X
X
X
X
X
X

X
X
X

X

X
X
X
X
X
X
X
X

X
X
X

X
X

X
X
X

X

X

X

X

X

X
X
X
X
X
X
X
X
X
X
X

X

X
X

X

X

X
X
X
X
X
X
X
X
X

X
X
X
X
X
X

X
X

X

X

X

X

6

1

2

3

29

15

22

1

1

NOTES

(1)
(1)
(1)
(2)
(3)

(5)

(6)

(7)
(8)
(9)

(10)

(11)
(12)

(13)
(14)
(15)

(16)

(17)
(18)

(19)

Notes

(1) These four species are part of the Ozarkian fauna.

(2) Fusconaia barnesiana (Lea 1838). Ortmann (1918: 536, 537) recognized two additional forms: big-
byensis (Lea 1841) and tumescens (Lea 1845).

(3) Quadrula tuberosa (Lea 1840). Not in Burch (1975). Unio tuberosus Lea 1840, Proc. Amer. Philos.
Soc. 1 : 286 (Caney Fork [of the Cumberland River] and Cumberland River, Middle Tennessee); 1842,
Trans. Amer. Philos. Soc. 8: 210, pi. 14, fig. 25, figured holotype USNM 84219; 1842, Obs. Unio 3:
48. Ortmann (1918: 540) regarded Unio intermedius Conrad 1834 (Tennessee River system), as being

88 Bulletin Museum oj Coinpiiiiilive Zoology, Vol. 149, No. 2

the same as Unia luhcrostis Lea 1840 (C]innl)iTlaiKl Hi\er system), and Ihiio spdrstis Lea 1841 (Ten-
nessee River system). An examination of inore recently collected material indicates that while .vprtr.sus
is a sNiionvm oi tiilycrosa, as is Quddnila hidiiauhita Morrison 1942 (Tennessee River, Tnscinnhia,
Colbert Co., Tennessee, holotype I'.SWl 84221 figured hy Johnson ( 1975a: 2.5, pi. 3, fi«. .'3), (Juadnda
intermedia (Conrad 18.36), is a distinct species restricted to the Tennessee River system. Stanshery
(1970: 13), writing on species in the Tennessee River system, in addition to intermedia, recognized
Quadrula sparsa (Lea) as valid, and mentioned that, "this form [sp(Osa\ stands between Q. metaiievra
and Q. intermedia and merges with neither."

(4) Pleurohema oviforme (Conrad 1834). Ortmann (1918: 552, 524 recognized two additional forms:
ar^enteum (Lea 1841) and holstonense (Lea 1840).

(5) Alasmidonta raveneliana (Lea 1834). Alas))iudon atropurpuremn Rafinesque 1831 (Morrison, 1969:
24). Ortmann (1918: 562) recognized Rafinesijue's taxa, bnt later Ortmann and Walker (1922: 39)
declared atroptirpureum an imidentified form. A. H. Clarke (pers. comm.) regards raveneliana, from
the upper Tennessee River system, and atropurpurenm, from the tipper Cumberland River system,
as distinct species, possibly since some specimens from the latter river system show a slight rugosity
on the posterior slope.

(6) Actiononais peetorusa (Conrad 18.34). Lampsilis suwerhijana Frierson 1927, Check list N American
naiades, p. 69 (Duck River, Tennessee; based on the figure in Sowerbv, 1868, Conch. Iconica 16,
Unio, pi. 63, fig. 316, holotype [lost] teste Johnson, 1972, Occ. Pap. Moll. 3: 149, pi. 28, fig. 3 [from
Sowerbyl).

(7) Medionidus eonradieus (Lea 18.34). Ortmann (1918: 575) as plateolus (Rafinesque 18.31). Mediunidus
was monographed (Johnson 1977, [eonradieus] p. 165, pi. 18, figs. 1, 2).

(8) Villosa taeniata (Conrad 18.34). Ohliquaria teneltus Rafinesque 1820 (Morrison, 1969: 24).

(9) Villosa trahalis (Conrad 1834). Stansbery (1971: 18d, fig. 37) separated trahalis, from the upper
Cumberland River system, Kentucky and perpurpurea (Lea 1861) (fig. 38), from the Clinch drainage
of the upper Tennessee River system, Virginia and Tennessee, possibly because specimens from the
latter locality sometimes have violet nacre.

(10) Dromtis dromas (Lea 1834). Ortmann (1918: 566) recognized an additional form: eaperatus (Lea
1845).

(11) Fusconaia euneolus (Lea 1840). Ortmann (1918: 531) recognized an additional form: appressa (Lea
1871).

(12) Fuseonaia edp.ariana (Lea 1840). Not in Burch (1975). Klon Unio eor Conrad 1834, teste Ortmann
(1925: 330). Ortmann (1918: 533) recognized an additional form: analo^a Ortmann 1918. The holotype
ofFusconaia edgariana analoga Ortmann British Museum (Nat. Hist.) 1964117 was figured by Johnson
(1977b: 2.36, pi. 27, fig. 1).

(13) Lexingtonia dolahelloides (Lea 1840). Ortmann (1918: 546) recognized an additional form: eonradi
(Vannata 1915).

(14) Lasmigona holstonia (Lea 1838). Alasmidon hadium Rafinescjue 1831 (Morrison 1969: 23).

(15) Lampsilis vireseens (Lea 1858). Not in Burch (1975). Figured by Stansbery (1971: 183, fig. 40).

(16) Carunculina cijlindrella (Lea 1868). Not in Burch (1975). Ortmann (1918: 573) did not recognize this
ta.xon, but later, Ortmann (1925: 353) accepted it as a form of C. maestus (Lea 1841). Included, here,
on the authoritv of Stansbery (1971: 18e, fig. 39), and Ahlstedt and Brown (1980, pp. 41, 43).

(17) Pleurohema gihberum (Lea 1838). Not in Burch (1975). Unio gibber Lea 1838, Trans. Amer. Philos.
Soc. 6: 35, pi. 10, fig. 30 (Caney Fork [of Cumberland River], Tennessee, figured holotype USNM
84544); 18.38, Obs. Unio 2: 35. Apparently restricted to the type locality.

(18) Villosa ortmanni (Walker 1925). This species is very close to vanuxemensis (Lea 1838), if, indeed,
it is distinct.

(19) Lasmigona subviridis (Conrad 18.35). This is the dominant species in the Greenbrier and New rivers
of the Kanawha River system, West Virginia. Ortmann (1913: 371) suggested that his species origi-
nated in this system. With this exception, its distribution is entirely on the Atlantic slope and in the
St. Lawrence River system.

Additional Notes

Alasmidonta ntinor (Lea 1845) was recognized by Ortmann (1918: 560) as a Cumberlandian species.
Ortmann (1926: 167) later declared that it was, "an absolute synonym of A. ealeeolus (Lea 1829)." See
under: 24. Alasmidonta ealeeolus in List 1. Villosa nebitlosa (Conrad 18.34) was recognized by Ortmann
(1924a: 42) as a Cumberlandian species. Ortmann (1926: 180) later conceded that nebulosa and iris (Lea
1829) were the same species, but mistakenly thought that Conrad's name was the earlier. See under 68.
Villosa iris in List 1.

Zoogeography of North American UNiONACEA-yo/i/iw/j

89

5. Fuscoiuiid suhrotimda (Lea 1831)
Unio kirtlandidnus Lea 1834; Unio
lesueurianus Lea 1840; Unio pilaris
Lea 1840; Unio bursa pastori.s Wright
1896; Quadrula kirtlandiana minor
Simpson 1900, lectotype USNM
150155, selected by Johnson (1975a:
15, pi. 3, fig. 5); Pleurohema mis-
souriensis Marsh 1901, Nautilus 15:
74 (Black River, near Poplar Bluff,
Missouri, holotvpe MZUM 38855
figured bv Walker, 1915, Nautilus
28: 140, pi. 5, figs. 1-2). Walker also
determined that the shell was an
ecophenotype of suhrotimda, which
occurs only in the Cumberlandian
and Ohioan regions; Fusconaja [sic]
subrotunda leucogona Ortmann
1913, lectotype CM 61.5239 selected
by Johnson (1977: 238, pi. 27, fig. 2).
Fusconaia maculata mac ul at a (Ra-
finesque 1820). So listed by Tay-
lor (1980a) on the authority of Stans-
bery who, apparently by some undis-
closed means, has been able to
identify EUiptio nigra maculata Ra-
finesque as Fusconaia subrotunda
(Lea).

6. Fusconaia undata (Barnes 1823)
Obliquaria lateralis Rafinesque 1820
(Morrison, 1969: 24); Unio trigonus
Lea 1831; Unio friersoni B. H.
Wright 1896, Nautilus 9: 134, pi. 3
(Bayou Pierre, an arm of the Red Riv-
er, in De Soto Parish, Louisiana.
"The type locality- has been drained,"
teste Frierson, 1927, Check list N
American naiades, p. 54. Lectot>'pe
USNM 133432 selected bv Johnson,
1967, Occ. Papers on Moll. 3: 6, pi.
5, fig. 4); Unio askewi Marsh 1896,
Nautilus 10: 91, pi. 1, figs. 3, 4 (Vil-
lage Creek, Hardin Co.; Sabine Riv-
er; both Texas. Figured holotvpe
ANSP 70448a from the latter locali-
ty). Placed with undata by Ortmann
(1914, Nautilus 28: 20); Fusconaia
selecta Wheeler, 1914, Nautilus 28:
76, pi. 4 (Cache River, Nemo, Craig-
head Co., Arkansas; holotype [re-

tained by Wheeler, possibly now in
the Alabama Mus. Nat. Hist. J; para-
types MZUM 38300); Fusconaia un-
data trigonoides 'Frierson' Utter-
back 1915; Fusconaia undata
wagneri Baker 1928.

7. Quadrula cylindrica (Say 1817)

Unio strigillatus Wright 1898. Known
from the Cumberlandian and Ohioan
regions, and from the Ozarkian Re-
gion below the Ozark Crest.

Specimens Examined
Ozarkian Region

Black River System

Ouachita River Drainage. Arkan-
sas: Ouachita River, 3 mi. SE Pencil
Bluff, Montgomery Co.; Ouachita
River, Arkadelphia Clark Co.; (both
MCZ).

Saline River Drainage. Arkansas:
Saline River, Benton, Saline Co.
(MCZ, CM).

Arkansas River System

Fall River Drainage. Kansas: Fall
River, Wilson Co. (MCZ).

Verdigris River Drainage. Kansas:
Verdigris River, Neodesha, Wilson
Co. (MCZ).

Neosho River Drainage. Kansas:
Neosho River, 3 mi. E Oswego, La-
bette Co. (Murray and Leonard);
Neosho River, near state line (CM).
Oklahoma: Neosho River, Miami,
Ottawa Co. (CM).

Spring River Drainage. Missouri:
Spring River, Carthage (MCZ); Cen-
ter Creek, Webb City (CM); both
Jasper Co.

White River System

White River Drainage. Missouri:
White River, Hollister, Taney Co.
(Utterback). Arkansas: White River,
Cotter (MCZ, CM); North Fork
White River, Norfolk (CM); both
Baxter Co.

Big Buffalo Fork Drainage. Arkan-
sas: Big Buffalo Fork, Newton, Sear-

90

Bulletin Museum of CU)ini)(ir(itii-c Zoo/o^'f/, Vol. 149, Mo. 2

cy, and Marion Cos. {(ill Meek and
Clark).

Black River Drainage. Missouri:
Black River, Willianisville, Wayne
Co. (MCZ). Arkansas: Black River,
Pocahontas, Randolph Co. (MCZ);
Black River, Black Rock, Lawrence
Co. (MCZ, CM).

Spring River Drainage. Arkansas:
Spring River, Ravenden (Mnrray and
Leonard); Spring River, Black Rock
(MCZ); hotJi Lawrence Co.

St. Francis River System

St. Francis River Drainage. Mis-
souri: St. Francis River, Greenville,
Wayne Co. (Utterback).

8. Quadrula metanevra (Rafinesque
1820)

Unio wardii Lea 186L

9. Quadrula nodulata (Rafinesque 1820)
Unio pustulatus Lea 1831.

10. Quadrula pustulosa (Lea 1831)
Obliquaria huUata Rafinesque 1820
(Morrison 1969: 24); Unio school-
craftensis Lea 1834; Unio kieneri-
anus Lea 1852; Unio asperatiis Lea
1861; Unio refulgens Lea 1868.

11. Quadrula quadrula (Rafinesque
1820)

Studied by Neel (1941) who includ-
ed as synonyms: Unio rugosus
Barnes 1823; Unio lachrymosus Lea
1828; Unio asperrimus Lea 1831;
Unio fragosus Conrad 1836. Also in-
cluded here as synonyms are Unio
asper Lea 1831; Unio prasinus Con-
rad 1834; Unio nobilis Conrad 1854;
Quadrula quadrula contrayensis Ut-
terback 1915; Quadrula quadrida
hullocki Baker 1928.

12. Tritogonia verrucosa (Rafinesque
1820)

Unio tuberculatus Barnes 1823.

13. Megalonaias gigantea (Barnes 1823)
Elliptio nervosa Rafinesque 1820
(Morrison 1969: 24); Unio heros Say
1829.

14. Ci/clonaias tuhcrctilata (Rafinesque
1820)

Unio graniferus Lea 1838; Quadrula
granifera pusilla Simpson 1900, lec-
totype MCZ 20176, selected by
Johnson (1966: 131, pi. 2, fig. 4);
Quadrula tuherculata utterback-
iana Frierson 1927.

15. Elliptio crass id ens crassidens (La-
marck 1819)

Elliptio nigra Rafinesque 1820.

16. Elliptio dilatata Rafinesque 1820
Unio gibbosus Barnes 1823; Unio
subgibbosus Lea 1837; Unio gibbo-
sus delicatus Simpson 1900; Elliptio
dilatatus sterkii Grier 1918.

17. Hemistena lata (Rafinesque 1820).

18. Plethobasus cooperianus (Lea 1834)
Obovaria striata Rafinesque 1820
(Morrison, 1969: 24).

19. Plethobasus cyphyus (Rafinesque
1820)

Unio aesopus Green 1827; Unio
compertus Frierson 1911. Reported
below the Ozark Crest from the Ver-
digris River, Coffeyville, Montgom-
ery Co., Kansas. This record from the
Arkansas River system was based on
3 specimens. The one in the R. E.
Call collection is neither at Wash-
burn University, Topeka, Kansas nor
in the Museum of Comparative Zo-
ology. The single valve, thought to
be extant, figured by Murray and
Leonard (1962: pi. 4), was sent to the
National Museum of Natural History
(Leonard, pers. comm., 1979), but it
is not there now (Clarke, pers.
comm., 1979). Branson (1967: 282)
reported P. cyphyus from 3 localities
in the Spring River of the Neosho
River drainage of the Arkansas River
system. These specimens could not
be located at the Pittsburg [Kansasl
State University (Kelting and Bran-
son, pers. comms., 1979).

20. Plethobasus cicatricosus (Say 1829)
Obovaria pachosteus Rafinesque

Zoogeography of North American Unionacea • Johnson 91

1820 (Morrison, 1969: 24); Not in
Burch (1975). See: Simpson (1914, 2:
807), figured bv Stansbery (1971, fig.

24).

21. Pleurohema clava (Lamarck 1819)
Not found on either side of the Ozark
Crest. The several records from the
Mississippian Region could not be
verified, and are probably spurious.
This is an Ohioan species.

22. Pleurohema cordatum (Rafinesque
1820)

Stansbery (1967) sent 2 pages of
mimeographed ephemera on the
Pleurohema cordatum complex to
interested parties. Since he later
(1970: 21) included this work in his
bibliography, it is essentially quoted
and discussed as if it were validly
published, with additional com-
ments in brackets. He recognized 4
species, one possibly undescribed.

(1). Pleurohema cordatum (Rafines-
que 1820). Outline: equilateral
triangle.

Ujiio ohliqua Lamarck 1819.
Unidentifiable [=cordatum Ra-
finesque (Morrison, 1969: 24).
Johnson (1969, Nautilus 83: 54)
substantiated Stansbery's opin-
ion that ohliqua is unrecogniz-
able].

Ohovaria cordata Rafinesque
1820

Unio plenum Lea 1840. Speci-
men with shortened posterior [is
Pleurohema premorsa Rafines-
que 1831 (Morrison, 1969: 24)].

(2). Pleurohema coccineum (Conrad
1836). Outline: subcircular.
Unio coccineus Conrad 1836.
Compressed with red nacre.
Unio catillus Conrad 1836.
Moderately wide specimen —
medium rivers.

Unio coccineum Lea 1838. Mod-
erately wide specimen — medi-
um rivers.
Unio solidus Lea 1838. Very

wide specimen — large rivers [is
sintoxia (Rafinesqvie 1820)
(Morrison, 1969: 24)].
Unio cuneus Conrad 1838 (?) A
wide "dwarfed" specimen.
Unio gouldianus Ward 1839.
Specimen with white nacre.
Unio fidgidus Lea 1845. Mod-
erately wide specimen — quite
young.

Quadrula coccinea paupercula
Simpson 1900. Stunted "Lake
Erie" specimen [Lectotype
USNM 134834, selected by
Johnson, 1975a; 17, pi. 3, fig. 1,
non Unio paupercula Lea 1861,
unnecessarily changed bv Simp-
son (1914, Cat. Naiades 2: 884)
to magnalacustris].
Pleurohema coccineum missis-
sippiensis Baker 1928 (in part).
Very wide specimen(s) from
Miss. River at Lake Pepin. [Hav-
lik and Stansbery (1978: 11) rec-
ognized P. coccineum as a valid
species.]

(3). Pleurohema pijramidatum (Lea
1834). Outline: scaline triangle.
Unio pijramidatus Lea 1834 [Is
P. ohliquata (Rafinesque 1820)
(Morrison, 1969: 24)]; Unio mijt-
iloides Rafinesque 1820. Conrad
1836. Actually P. clava (La-
marck 1819).

Unio cardiacea Guerin [1829]
Pleurohema coccineum missis-
sippiensis Baker 1928 (in part).
Well developed specimen(s)
from Miss. River at Lake Pepin.
[Havlik and Stansbery (1978: 11)
recognized P. ruhrum (Rafines-
que 1820) = pijramidatus Lea
1831, possibly on the basis of
Ortmann's (1918: 550) recogni-
tion of Pleurohema ohliquum
ruhrum (Rafinesque) and the
subsequent selection of a lecto-
type for it, ANSP 20237, figured
bv Johnson and Baker (1973:
169, pi. 3, fig. 2)].

92

Bulletin Museum oj Comfxtrative Zoology, Vol. 149, No. 2

(4). Pleurohema . . . (undescribed).
Outline: isosceles triangle.

The taxa under cordatuni and
pyrdiniddttnn are usually asso-
ciated with specimens froin
large rivers, while specimens
that lose their sinus and become
rounded with an elongated pos-
terior end, from smaller streams
and headwaters, have been in-
correctly called P. cordatum
coccineum.

[Pleurobema cordatum is a
species with a nimiber of eco-
phenotypic forms which are not
species or subspecies in the
modern sense. Note: Pleurohe-
ma missouriensis Marsh 1901.
See under no. 5, Fusconaia sub-
rotunda (Lea 1831)].

23. Uniomerus tetralasmus (Say 1831)
Unio camptodon Say 1832; Unio sayi
Ward 1839. For the complete syn-
onymy see: Johnson (1970: 339).

White River System

White River Drainage. Arkansas:
Cache River, Nemo, Craighead Co.
(Wheeler, 1914, Nautilus 28: 76).

Previously unreported from the
Tennessee River system. Found in
Kentucky Lake, Little Eagle Creek
embayment, Benton Co. in 1971 and
Hurricane Creek, Henderson and
Hardin Cos. in 1977; both Tennes-
see (Brown and Pardue, in press).
Like Arcidens confragosus, no. 32,
it is suggested here that tetralasmus
was accidentally introduced behind
the impoundment with fishes.

24. Alasmidonta calceolus (Lea 1830)
Anodonta viridis Rafinesque 1820
(Morrison, 1969: 23); Margaritana
minor Lea 1845; Alasmidonta cal-
ceolus danielsi Baker 1928; Alasmi-
donta calceolus magnalacustris Bak-
er 1928.

25. Alasmidonta marginata Say 1819

Margaritana marginata truncata
Wright 1898; Alasmidonta margina-
ta variabilis Baker 1928.

26. Anodonta grandis grandis Say 1829
Anodonta plana Lea 1834; Anodonta
stewartiana Lea 1834; Anodonta gi-
gantea Lea 1838; Anodonta footiana
Lea 1840; Anodonta virens Lea
1852; Anodonta opacea Lea 1856;
Anodonta leonensis Lea 1857; Ano-
donta danielsii Lea 1858; Anodonta
texasensis Lea 1859; Anodonta
healii Lea 1863; Anodonta dakota
Frierson 1910.

27. Anodonta corpulenta Cooper 1834
Regarded as a subspecies by Burch
(1973: 17; 1975: 15) and Havlik and
Stansbery (1978: 15). A. corpulenta
has not been regarded as a subspe-
cies in the modern sense, as a geo-
graphically isolated population. It is
either a synonym of grandis, or more
probably, it is a valid species.
Found in the Meramec Basin, Mis-
souri. (Buchanan, pers. comm.).

28. Anodonta imbecillis Say 1829
Anodonta ohiensis Rafinesque 1820
[of authors]; Utterbackia imbecillis

fusca Baker 1928. In spite of the
statements by various authors, the
original orthography is with the dou-
ble "1".

Range extension on the North Atlan-
tic Slope:

Delaware River System

Schuylkill River Drainage. Penn-
sylvania: Pickering Creek, Chester
Co. (Fuller and Hartenstine, 1980).
[most probably a recent introduction].

29. Anodonta suborbiculata Say 1831
See under species monographed.

30. Anodontoides ferussacianus (Lea
1834)

Anodonta subcylindracea Lea 1838;
Anodonta modesta Lea 1857; Ano-
dontoides bergei Baker 1928.

Zoogeography of North American Vnionacea' Johnson 93

Meramec River System

Boiirbeuse River Drainage. Mis-
souri: Dry Fork Creek [Maries and
Gasconade Cos.]; Brush Creek [Gas-
conade and Crawford Cos.]; {both
Oliio State Museum, teste, Buchan-
an, pers. comm.).

White River System

Little Red River Drainage. Arkan-
sas: Little Red River, Clinton, Van
Buren Co. (E. Pleas, MZUM, teste
Gordon, pers. comm.) Listed by Gor-
don, et al. (1980: 35).

St. Francis River System

St. Francis River Drainage. Ar/can-
sas: St. Francis River, Marked Tree,
Poinsett Co. (M. Ellis, MZUM, teste
Gordon, pers. comm.) Listed by Gor-
don, ef. al. (1980: 35).

31. Arcidens confragosus (Say 1829)
Previously unreported from the Ten-
nessee River system. Found in: Ken-
tucky Lake, 6 mi. above New John-
sonville, Benton Co., Tennessee,
collected in 1967 by Stansbery and
Jenkinson. It is suggested that it was
accidentally introduced behind the
impoundment with fishes. See under
species monographed.

32. Lasmigona complanata (Barnes 1823)
Unio katherinae Lea 1838
Previously unreported from the
Ozarkian Region below the Ozark
Crest.

White River System

Black River Drainage. Missouri:
Black River, Hendrickson, Butler
Co. (MCZ 260973).

33. Lasmigona compressa (Lea 1829)
Not found on either side of the Ozark
Crest.

Not found in the Illinois River, prop-
er, but recorded from its tributaries.
Illinois: Vermilion River; Panther

Creek, Woodford Co. {both Parma-
lee, 1967: 53, pi. 17, A).

34. Lasmigona costata (Rafinesque 1820)
Lasmigona costata pepinensis Baker
1928; Lasmigona costata nuda Ba-
ker 1928.

35. Simpsoniconcha ambigua (Say 1825)

White River System

Little Red River Drainage. Arkan-
sas: Little Red River, Clinton, Van
Buren Co. (E. Pleas, USNM, teste,
Gordon, pers. comm.). Listed by
Gordon, et al. (1980: 35).

Meramec River System

Bourbeuse River Drainage. Mis-
souri: Bourbeuse River, 0.1 mi. S.
Tea (Buchanan, Ohio State Mu-
seum). Buchanan allowed the author
to examine an authentic specimen
from this lot.

St. Lawrence River System

Great Lakes Drainage (Lake Mi'^b-
igan). Wisconsin: Wolf River, Sha-
wanao Co. (Mathiak, 1979, p. 62, pi.
9, A, single specimen).

36. Strophitus undulatus (Say 1817)
Anodon rugosus Swainson 1822;
Alasmidonta edentula Say 1829; An-
odonta parvonia Lea 1836; Strophi-
tus undulatus tennesseensis Frier-
son 1927; Strophitus undulatus
ovatus Frierson 1927; Strophitus ru-
gosus pepinensis Baker 1928; Stro-
phitus rugosus winnebagoensis Ba-
ker 1928; Strophitus rugosus
lacustris Baker 1928.

37. Actinonaias ligamentina (Lamarck
1819)

Unio carinatus Barnes 1823, teste
Morrison (1969: 24). Lamarck's ho-
lotype of Unio ligamentina was fig-
ured by Johnson (1969, Nautilus 83:
53; fig. 8); Lampsilis ligamentinus

94 Bulletin Museum of Comparative Zoology, Vol. 149, No. 2

gibhus Simpson 1900, figured holo-
type British Museum (Nat. Hist.)
107.10.28.262; Lampsilis ligamen-
tinus nigrescens Simpson 1914, lec-
totype USNM 150410, selected by
Johnson (1975a: 13, pi. 1, fig. 2); Ac-
tinonaias carinata orbis Morrison
1942, hoh)type USNM 85998, fig-
ured by Johnson (1975a: 32, pi. 1, fig.
3).

38. Actinonaias ellip.siformis (Conrad
1836)

Unio venustus Lea 1838; Vnio pleasi
Marsh 1891, Nautilus 5: 2 (Little
Red River, Arkansas, lectotype ANSP
56480a selected by Johnson and Ba-
ker, 1973, Proc. Acad. Nat. Sci. Phila.
125: 166, pi. 9, fig. 5). Found in the
streams of the southern half of the
Lower Michigan Peninsula which
flow into Lake Michigan. During pe-
riods of confluence, it entered the
Saginaw River drainage, but it has
not migrated into any of the streams
of southeastern Michigan draining
into Lake St. Clair or Lake Erie (van
der Schalie, H. and A., 1963: 9).

39. Carunculina parva (Barnes 1823)
Unio texasensis Lea 1859; Unio
cromweUi Lea 1865; Carunculina
parva cahni Baker 1927.

40. Carunculina glans (Lea Dec. 1831)
Toxolasma livida Rafinesque 1831
(Morrison, 1969: 24). Not in Burch
(1975). See: Simpson (1914, 1: 153).
Stansbery (1970): 18) recognized
moaesta (Lea) as a form or subspe-
cies, which "exists in . . . several
headwater streams of the Cumber-
land Plateau and the Southern Ap-
palachians." Reported by Call (1885a:
31) as from the Green and Kentucky
rivers, based on MCZ 5410, two
.specimens, so labeled. These are re-
garded here as erroneous records.
Known from the Cumberlandian and
Ohioan regions, and from the Ozark-
ian Region below the Ozark Crest.

Specimens Examined
Ozarkian Region

Black River System

Ouachita River Drainage. Arkan-
■sa.s: Caddo River, near Arkadelphia;
Ouachita River and Old River, near
Arkadelphia; Terra Noire Creek, Mt.
Zion; all Clark Co. {all Wheeler,
1918: 119).

Arkansas River System

Spring River Drainage. Missouri:
Spring River, Carthage, Jasper Co.
(MCZ).

Elk River Drainage. Missouri: Elk
River, [McDonald Co.] (Utterback).

White River System

White River Drainage. Arkansas:
White River, Carroll Co.; James Fork
of White River, Galena, Stone Co.;
{both MCZ).

Big Buffalo Fork Drainage. Arkan-
sas: Big Buffalo Fork, Searcy Co.
(Meek and Clark).

Little Red River Drainage. Arkan-
sas: Little Red River (MCZ).

41. Plagiola triquetra (Rafinesque 1820)
Found below the Ozark Crest only in
the

White River System

Black River Drainage. Missouri:
[Black Riverl. Poplar Bluff, [Buder
Co.], (MZUM). Based on a single
specimen, the locality written on the
shell, with the additional data [W. A.]
Marsh, March 3, 1891. This record
may be spurious as is that of Pleu-
robema missouriensis Marsh 1901.
See also under: no. 5. Fusconaia
subrotunda (Lea 1831). P. triquetra
was not listed from the Kentucky
River system by Johnson (1978).
Now recorded from: South Fork Ken-
tucky River, Booneville, Ousley Co.,

Zoogeography of North Ameriga.n Unionacea • Johnson 95

Kentucky (MCZ). Found in the Mus-
kegon, Grand and St. Joseph rivers
on the eastern side of Lake Michi-
gan, but not in the Lake proper
(Johnson, 1978: 292).

42. Plagiola torulosa (Rafinesque 1820)
Found in the Grand River on the
eastern side of Lake Michigan, but
not in the Lake proper (Johnson,
1978: 296).

43. Plagiola sampsoni (Lea 1861)

See: (Johnson, 1978, pi. 6, fig. B, tri-
angles).

44. Plagiola propinqua (Lea 1857)

See: (Johnson 1978, pi. 6, fig. B,
dots).

45. Plagiola personata (Say 1829)
See: (Johnson, 1978, pi. 6, fig. A).

46. Plagiola obliquata (Rafinesque 1820)
Unio sulcatus Lea 1829.

47. Plagiola flexuosa (Rafinesque 1820)
Rafinesque's records: Green, Salt,
and Kentucky rivers, are spurious.
See: (Johnson, 1978, pi. 5).

48. Ellipsaria lineolata (Rafinesque 1820)
Unio securis Lea 1829.

49. Lampsilis teres (Rafinesque 1820)
The availability of teres over ano-
dontoides Lea 1834 and fallaciosiis
Smith 1899 was discussed by John-
son (1972: 244). This discussion was
ignored by Morrison (1979: 61) who
regarded teres and anodontoides as
separate species.

50. Lampsilis fasciola (Rafinesque 1820)

51. Lampsilis higginsi (Lea 1857)

Not in Burch (1975). See under
species monographed.

52. Lampsilis ahrupta (Say 1831)
Listed by Burch (1975) as orhiculata
(Lea 1836) non Hildreth 1828. See
under species monographed.

53. Lampsilis ovata (Say 1817)
Lampsilis cardium Rafinesque 1820;

Unio ventricosus Barnes 1823; Unio
occidens Lea 1829; Unio canadensis
Lea 1857; Lampsilis ventricosa lu-
rida Simpson 1914; Lampsilis ven-
tricosa perglobosa Baker 1928;
Lampsilis ventricosa winnehagoen-
sis Baker 1928. L. ventricosa Barnes
is regarded as a valid species by Hav-
lik and Stansbery (1978: 11). For a
discussion of ecophenotypic varia-
tion in L. ovata see: Johnson (1970:
388).

54. Lampsilis radiata siliquoidea (Barnes
1823)

Unio luteola Lamarck 1819 (Morri-
son, 1969: 24); Unio rosacea DeKay
1843; Unio superiorensis Marsh
1897; Lampsilis siliquoidea pepi-
nensis Baker 1927; Lampsilis sili-
quoidea chadwicki Baker 1928.
Johnson (1969, Nautilus 83: 54, fig.
11) figured the type of Unio luteola
and restricted the type locality to the
Susquehanna River, Columbia, York
Co., Pennsylvania. This choice from
the two localities given by Lamarck,
placed the equivocal type in the At-
lantic Slope Region where only the
typical form is found. See: Clarke
and Berg (1959: 60; 80) who define
siliquoidea as a subspecies in the
modern sense. L. radiata siliquoidea
has never been reported from the
Tennessee or Cumberland River sys-
tems.

55. Leptodea fragilis (Rafinesque 1820)
Lampsilis simpsoni Ferriss 1900,
Nautilus 14: 38, fig. (Spring River,
Hardy [Sharp Co.l, Arkansas, figured
holotype Field Mus. Nat. Hist, [not
seenl); such specimens, without a
dorsal "wing" are found throughout
the range of L. fragilis; Leptodea
fragilis lacustris Baker 1922.

56. Leptodea laevissima (Lea 1829)
Potamilus ohiensis (Rafinesque 1820)
(Morrison, 1969: 24). Originally de-
scribed as Anodonta ohiensis {Last-
ena ohiensis).

96

Bullet i)t Mit.scuin of (Umifxinitiic Zoolofj,!/, VdI. IW, No. 2

57. Leptodcd Icptodon (Rafinesciiie 1820)
Lauipsilis blatchlciii Daniels 1902,
Nautilus 16: 13, pi." 2 (Wabash River,
Section 32, Lynn Township, Posey
Co., Indiana; holotype MZUM
82660). Known only from the type
lot. It was suggested hy Goodrich
and van der Schalie (1944: 317) to be
a form of leptodon.

58. Li^iiniid recta (Lamarck 1819)
Elliptio latissima Rafinesque 1820;
Unio sageri Conrad 1831.

59. Ligumia suhrostrata (Say 1831)
Lampsilis suhrostrata fiirva Simp-
son 1914, lectotype USNM 126921,
selected bv Johnson (1975a: 13, pi.
1, figs. 4, 6).

60. Ohovaria olivaria (Rafinesque 1820)
Unio ellipsis Lea 1828.

61. Ohovaria retusa (Lamarck 1819)

62. Ohovaria suhrotunda (Rafinesque
1820)

Elliptio levigata Rafinesque 1820;
Unio lens Lea 1831; Ohovaria lens
parva Simpson 1914, lectotype
USNM 150457 selected bv Johnson
(1975a: 16, pi. 1, fig. 8); Ohovaria
lens elongata Simpson 1914, lecto-
tvpe USNM 35734, selected by John-
son (1975a: 12, pi. 1, fig. 9; Quadrula
suhrotunda glohula Morrison 1942,
holotype USNM 85789, figured bv
Johnson (1975a: 29, pi. 1, fig. 7).

63. Proptera alata (Say 1817)
Metaptera megaptera Rafinesque
1820. Found below the Ozark Crest
only in the Neosho River drainage of
the Arkansas River system. (Scam-
mon, 1906: 300; Murray and Leon-
ard, 1962: 129).

64. Proptera capax (Green 1832)
Published records from the Arkansas
River system are in error. See under
species monographed.

65. Truncilla truncata Rafinesque 1820
Non Unio truncatus Spengler 1793;

Unio elegans Lea 18.31; Truncilla
truncata lacustris Baker 1928.

66. Truncilla donacifonnis (Lea 1828)

67. Villosa fahUis (Lea 1831)

68. VUlosa iris (Lea 1829)

Unio nehulosus Conrad 1834, the
records under this name from the
Cumberlandian Region are iris. The
type locality of nehulosus is: Black
Warrior River [Mobile-Alabama-
Coosa River system], Alabama, and
it may be a valid species; Unio ellip-
siformis Conrad 1836; Unio novie-
boraci Lea 1838.

69. Villosa lienosa (Conrad 18.34)
Range extended into Ohio and West
Virginia. Based on records in Jenkin-
son and Kokai (1978).

70. Cyprogenia stegaria (Rafinesque
1820)

See under species monographed.

71. Ohliquaria reflexa Rafinesque 1820

72. Ptijchohranchus fasciolare (Rafines-
que 1820)

Not found in the Mississippian Re-
gion. Valentine and Stansbery (1971:
23) claim that there are specimens of
this species in the Ohio State Mu-
seum from Missouri. These are prob-
ably P. occidentalis (Conrad).

UNIONIDAE OF THE OZARKIAN REGION
WITH DISTRIBUTIONS NOT EXTENDING
INTO THE MISSISSIPPIAN OR OHIOAN
REGIONS [List 2]

I. Fusconaia ozarkensis (Call 1888)
Not in Burch (1975).
See under species monographed.

II. Plectomerus domheijanus (Va-
lenciennes 1827)
Unio trapezoides Lea 1831.

111. Quadrula apiculata (Say 1829)
Not in Burch (1975). Studied by
Neel (1941: 4, pi. 1, figs. ,5-12)
who regarded apiculata as a sub-
species of Quadrula quadrida.

Zoogeography of North American UNiONACEA-yo/i;j.so»

97

They appear to l)e separate
species with an overlapping dis-
tribution in the Gulf" Coastal Re-
gion.

IV. Arkansia wheeleri Ortmann and
Walker 1912
See under species monographed.

V. Actinonaias rafinesqueana
(Frierson 1927)
Not in Burch (1975).
See under species monographed.

VI. Plagiola florentina (Lea 1857)
Truncilla curtisii Frierson and
Utterback 1916, Amer. Mid. Nat.
4: 453 [190], pi. 6, figs. 14 a-d,
pi. 28, figs. 109 A-D (White Riv-
er, Hollister, [Taney Co.], Mis-
souri, lectotype MZUM 90748,
selected by Johnson, 1978, Bull.
Mus. Comp. Zool. 148: 272, pi.

13, fig. 10).

See: (Johnson, 1978, pi. 2, dots).

VII. Plagiola turgidiila (Lea 1858)

Truncilla lefevrei Utterback 1916,
Amer. Mid. Nat. 4: 455 [192], pi.
6, figs. 13 a-d, pi. 28, figs. 108 A-
D (Black River, Williamsville,
[Wayne Co.], Missouri, figured
syntypes [lost]).

See: (Johnson, 1978, pi. 2, trian-
gles).

VIII. Glebula rotundata (Lamarck
1819)

Unio suhorbiculata Lamarck
1819.

IX. Lampsilis excavata (Lea 1857)
Plate 15, figures 1, 2.
Several lots collected in the
Black River system by the joint
Harvard-Ohio State Museum ex-
pedition, 1964, confirmed the au-
thenticity of older records from
the Saline River, Benton, Saline
Co., Arkansas.

X. Lampsilis satur (Lea 1852) Plate

14, figures 1, 2.

Unio satur Lea 1852, Trans.

Amer. Philos. Soc. 10: 205, pi.
17, fig. 19 ([Red River], Alexan-
dria, [Rapides Parish], Louisiana;
holotype USNM 84505); 1852,
Obs. Unio 5: 21.
Lampsilis ventricosa satur (Lea).
Simpson, 1900, Proc. U. S. Nad.
Mus. 22: 527; 1914, Cat. Naiades
1: 41.

Not in Burch (1975). Ortmann
(1916, Nautilus 30: 56) correcUy
suspected tliat satur was a valid
species, distinct from Lampsilis
ovata (Say). They occur together
in the White and St. Francis river
systems, but ovata is missing fur-
ther south.

XL Lampsilis hijdiana (Lea 1838)
Plate 15, figures 5-7.
Unio hydianus Lea 1838, Trans.
Amer. Philos. Soc. 6: 14, pi. 6,
fig. 14 (Teche River, [St. Landry
Parish], holotype USNM 85010;
vicinity of New Orleans [Orleans
Parish]; both Louisiana); 1838,
Obs. Unio 2: 14.
Unio placitus Lea 1852, Trans.
Amer. Philos. Soc. 10: 279, pi.
23, fig. 38 (Alabama; holotype
USNM 85152) 1852, Obs. Unio
5: 35.

Lampsilis hijdiana (Lea). Simp-
son, 1900, Proc. U.S. Nad. Mus.
22: 536; Cat. Naiades 1: 66.
Simpson correctly regarded hij-
diana as a valid species distinct
from Lampsilis radiata siliquoi-
dea (Barnes), pointing out that
hijdiana has a more inflated
shell, a more waxy texture, and
that the rays are bolder, more dis-
tinct, and more brilliant. Both
species occur together in the up-
per Arkansas, White, and St.
Francis river systems, Ijut sili-
quoidea is missing further south.

XII. Lampsdis powelli (Lea 1852)
Not in Burch (1975).
See under species monographed.

98

Bulletin Mu\(uni of Compardtivc Zoology, Vol. 149, No. 2

XIII. Ohovarid jdcksonidfui (Friersoii
1912)

As pointed out by Stansbery
(1976: 47) Unio castaneus Lea
December 1831 is preoccupied
by LJnio castaneus Rafinesque
1831.

XIV. Proptera purpurata (Lamarck
1819).

XV. Villosa arkansasensis (Lea 1862)
Not in Burch (1975).
See under species monographed.

XVI. Villosa reeviana (Lea 1852)

See under species monographed.

XVII. Villosa vibex (Conrad 1834) Plate
17, figures 6-8.

The synonymy was listed by
Johnson (1970: 373) with the ex-
ception of:

Lampsilis streckeri Frierson
1927, Check list N American
naiades, p. 74 (Little Red River,
Arkansas; holotype MZUM
91075, figured by Frierson, 1928,
Nautilus 41: 139, pi. 2, fig. 1);
Strecker, 1931, Baylor Univ.
Mus., Special Pub. no. 2, p. 40.
Actinonaias streckeri Valentine
and Stansbery 1972, Sterkiana,
no. 42, p. 32, non Frierson. Is
Actinonaias rafinesqueana
(Frierson 1927).

Frierson, in the original descrip-
tion, mentioned that the types of
streckeri were found together
with specimens of vibex. An ex-
amination of the type of the for-
mer with the specimens also
mentioned there as from Onion
Creek [of the Colorado River
system], Travis Co., Texas,
MZUM 79921, as streckeri; in-
dicates that the "chain-like rays,
sometimes becoming a ribbon or
band of cross lines," which char-
acterize streckeri, occur in widely
separated populations oi vibex. It
is found in most coastal river sys-
tems from the Guadeloupe River

system, Texas, east and north to
the Cape Fear River system.
North Carolina. Ozarkian region,
known only from:

White River System

Little Red River Drainage. Ar-
kansas: Archies Fork, Clinton
(MZUM); Little Red River, Clin-
ton (MCZ, MZUM); both Van
Buren Co.

XVIII. Cyprogenia aberti (Conrad 1850)
See under species monographed.

XIX. Ptychobranchus occidentalis
(Conrad 1836)
See under species monographed.

UNIONACEA OF THE NORTHERN AT-
LANTIC SLOPE REGION NORTH OF
MAXIMUM PLEISTOCENE GLACIATION
[LIST 3]

FAMILY MARGARITIFERIDAE

A. Margaritifera margaritifera (Lin-
naeus 1758)

Reported from the Lower Saskatch-
ewan River by Dall, but shown to be
a spurious record by Clarke (1973:
130).

FAMILY UNIONIDAE

B. Elliptio complanata (Lightfoot 1786)
Also found in the St. Lawrence River
system and in the Ohio River system,
Tuscarawas River, New Philadel-
phia, Tuscarawas Co., Ohio (CM).
Sterki (1900: 10) suggested that it
migrated from the east to Lake Erie
by canals [Erie Canal] and by the
Ohio Canal to the Tuscarawas River.
St. Lawrence River system, not found
in Lake Erie proper, but found in the
Grand River, Eagleville, Ashtabula
Co., Ohio (CM) (Ortmann, 1919:
110). Found in an Indian Midden ca.
1350 A.D. on the shore of the Little
Miami River, Warren Co., Ohio (Rus-
sell Barber, pers. comm.). Canadian

Interior Basin, reported from the Sas-
katchewan River by Dall, but shown

Zoogeography of North American Unionacea -yo^nson

99

to be a spurious record by Clarke
(1973: 130).

C. Lasmigona suhviridis (Conrad 1835)
Ortmann (1913: 371) found this to be
the dominant species in the Green-
brier and New rivers of the Kanawha
River system, which extends across
the divide but drains to the west, into
the Ohio. Ortmann thought that it
spread into the Atlantic Slope region
by former stream capture. With this
exception, it is found only on the At-
lantic slope and in the St. Lawrence
drainage. It is arbitrarily regarded as
an Atlantic Slope species, since it
spread from there into the St. Law-
rence drainage.

D. Alasmidonta undulata (Say 1817)

St. Lawrence River System

Lake Champlain Drainage. Ver-
mont: Otter Creek, Addison Co., (D.
Smith, pers. comm.).

E. Alasmidonta varicosa (Lamarck 1819)

F. Alasmidonta heterodon (Lea 1830)
This species was placed in the mono-
typic subgenus Prolasmidonta by
Ortmann (1914, Nautilus 28: 44).
Prolasmidonta was raised to generic
rank by Fuller (1977: 169), based on
the same data available to Ortmann.
Johnson (1970: 347) maintained the
subgeneric status of Prolasmidonta.

C. Anodonta cataracta cataracta Say
1817

H. Anodonta cataracta fragilis Lamarck
1819

marginata of authors, non Say 1817;
brooksiana van der Schalie 1938.
The concept oi fragilis as a subspe-
cies is based on the work of Clarke
and Rick (1963: 22). The use of this
combination for certain populations
of Anodonta in the St. Lawrence Riv-
er system may be suspect. The Lake
Superior population, at least, might
be another geographic subspecies, A.
grandis simpsoniana Lea.

I. Anodonta implicata Say 1829

Morrison (1975: 36) stated that Mijt-

ilus fucata Dillwyn 1817 is this
species, and suggested that after a
century and a half of usage. Say's
name should be replaced by that of
Dillwyn.

J. Strophitns undulatus (Say 1817)

This species is ubiquitous, and is in-
cluded on the list of Unionacea found
in the area north of the maximum gla-
ciation as no. 36.

K. Ligumia nasuta (Say 1817)

L. Lampsilis cariosa (Say 1817)

M. Lampsilis ochracea (Say 1817)

Morrison (1975: 38) stated that Myt-
ilus fluviatilis Gmelin 1791 is this
species, and suggested that after a
century and a half of usage. Say's
name should be replaced by that of
Gmelin. He also removed this species
from Lampsilis, placing it in Lep-
todea.

N. Lampsilis radiata radiata (Gmelin
1791)

UNIONACEA OF THE PACIFIC COASTAL
REGION [List 4]

FAMILY MARGARITIFERIDAE

a. Margaritifera falcata (Gould 1850)
Reported from the Madison River,
Montana by Stober (1972: 343) as
Marga ri tife ra m argari t if era .

FAMILY UNIONIDAE

b. Gonidea angulata (Lea 1838)

c. Anodonta heringiana Middendorff
1851

d. Anodonta californiensis Lea 1852

e. Anodonta dejecta Lewis 1875

f. Anodonta kennerhji Lea 1860

g. Anodonta oregonensis Lea 1838

h. Anodonta nuttalliana Lea 1838

Anodonta wahlamatensis Lea 1838.
The first name is chosen on the basis
of the page precedence.

100 Bitdcfin Mu.scuin of Compamthc Zoology, Vol. 149, No. 2

ujseg joijaiui ueipeueo

J3Ajy uin6ui>(snn

jSAjy oaops

jSAiy Apues Big

jaAjy 6ui>)3n

jSAiy A>(Dn3U3x
g J3A!a lies

LU

"^ JSAjy jaiewspeji

sjaAjy AugqSai IV 5 e(3L|e6uouow

jaAiy (M3N JO) eL|Meue»

<
o

o

j3Aiy 4SBqef\

ui3}sAs JSAjy 0140

<

Q JSAiy pueiJaqujn^

jaAjy 93ss3uuax

jaAi y s loui 1 I I

1/, CD

— ^ lua^sAs JSAiy iddississiw

jaAjy osLuejaw

jSAiy apeuooseg

j3Ajy 36eso

,,,, JSAjy ja^eft >|DBig

1S3J3 >|Jezo MOiaq

X

X

X X

X

X

X

X

X

X

X

+

+

X X

+

X

X

X

X

X

+

X

X X X X +

X

X

X

X

X

X X

X

X X

X

X

X

X

X

X

X

X

X X

XXX

X

X

+

X

X

X

X

X

+

X

X X

+

+

+

+

+

+ +

+

+

X

X X

X

X X

X

X

X

X

X

X

X

X

X X

X X

+

X

+

X

X

X X

X

X X

X

+

X

X

+

X

X

+

X X

X

X

X

X

< o

_1 —

LU UJ

CO CC

CC

<

X

X X

X

X

X X

X

X

+

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

LTv

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

-3-

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

+

X

X

X

X

X

X

X

+

X

X

+

X

X

X

X

X

X X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X

X

X

X

X

X

X

XXXXXXXXX+ XXXXXXXXXXXXXXXX +

«

LU

uo

•^ 0

3:

■z.

a, -t^

1-

0

a.

—

•vi ■«

Z

0

to K

—

UJ

CO ,s <a

a:

•tJ -w «

<

to -t^ 0

LU

z

to 3 •'-a

0

<:

•^ "«

<

0

§: CO a

^

0

:i

tB 0 «

0

■« -v^ -t^

2

+^ (32 Cn

0

Q

UJ s

2

^ Sh^

Li_

<

0 <i) •

<r>

s; Sh CO

2

Ci) b CS

T'

<

« 0 la 3)

0

'^•^ ^ J,

Q-

•^ ^ ^ a

t-

0.

sr 0 •<-■ „

->

0 CO ^

CNJ

cn

l/l

•t^ T3 to +^

UJ

a:

_I

(—

GO

ca

UO

UO

<

—

"iC

1-

0

■s.

■0

TJ

0

—

in

3

in

CT

TO

TO

ra

• —

c

i_

i-i

r—

—

TO

0

m

m

If)

w

4-»

in

TO

TO

lyi

3

tn

c

(-

TO

in

3

4-*

+-»

0

m

D

ni

3

C

c

TO

>.

TO

TO

TO

TO

TO

0

in

\f\

tj

TO

m

h

m

4-1

.—

0)

>

x:

TO

X

u

-—

4-1

C

-i

0

—

1-

i/i

4-J

TO

4-»

TO

cn -0

u

a.

4->

1)

4-1

0

TO

TO

m

1-

ro

nj

.—

a>

• —

T

ns

m

m

X>

I-

• —

UJ

>•

TO

•—

0)

in

0

• —

-M

\-

m

-i

>

4-j

■ ~

SI

TI

TJ

—

4-'

1-

m

in

c

i.)

*-t

14-

3

c

0)

C

01

01

4->

4-1

01

m

m

3

m

C

C

TJ

. —

TO

0

i-

in

TO

in

OJ

CJ'

•—

3

b

1)

>

m

in

>

m

4-1

i-t

TO

D

m

0

TO

4-1

in

0

in

1_

•—

>—

0)

0

-i

u-

0

u

u

l_

u

in

.—

m

i_

OJ

3

0

• —

l-

4-'

4-)

^_

m

m

n

<4-

i/i

(i>

m

t/1

m

-n

TO

4-1

0

h

-C

L_

TO

4-4

TO

a

ro

01

m

1-

3

t/j

(-

c

4-»

TO

—

—

—

in

TO

lU

r

<n

m

• —

ra

• —

(I)

XJ

• —

TO

0

<U

0

0

ra

TO

c;

I.

TO

• —

—

n

1-

m

r—

OJ

m

ro

4J

. —

t~-

r

-O

TJ

1 —

XI

0

TO

TO

—~

TO

.—

■ —

0

F

(1)

T

r

. —

TJ

c

c

r

, —

n

4-1

0

4-1

J

0

cn

in

3

0

in

• —

• —

cn

(1)

n

4-1

U

0

l/l

0

0

(-

0

n

(/J

c

Q.

L.

1-

a.

j_

J-

—

0

CT

• —

a.

in

0

—

4-J

0. -a

u

(\

3

"O

n

n.

.—

. —

3

l/l

—

TJ

t-t

1-

—

cn

.—

h

n

.—

0

m

i/i

h

Q-

0

cn

0

1-

h

4-1

—

ID

TO

—

TO

(U

in

' —

• —

TO

' —

h

in

■1

I-

u

3

-1

m

(1)

>-

c;

*j

rn

0

3

ni

-U

TO

TO

1—

Q-

O-Li.

_i

_l

0

-J

<

t/j

-J

<

LU

Q.

<

UJ

0' Cl-

_i

>

0

CL

UJ

_I

1—

CN csj mo-^ r^ ir\ -^ CO \0 \D -^ r^o <n ursLnooa^-roo — -co o^ltxcn

, — vX), — LALTN' — LACNlr^LArO — CNICN' — — fV^vOf^-J'-^'-Tv^f^

Zoogeography of North American Unionacea • yo/mson 101

X

X

CM

+ X

X X

+

X X

en

XXX

X X +

X X

f^

u

+ + +

u

X X X X X X

X

X

+

X

X

X

CM

-I

X

CM

CM

CM

X

X

X

X

X

X

X

X

X

X

X

X

X.

X

X

X

X

X

X

X

X

X

X

o

-3-

X

X

X

X

X

X

X

X

X

X

X

X

X

o

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

no

X

X

X

X

X

X

X

X

X

X

X

X

X

o

X

X

X

X

X

X

X

X

X

X

X

X

X.

o

X

X

X

X

X

X

CM

X

xxxxxxxx

r^

X

X XX

r-

+

X++ xxx^xxx

X

X

c

X X X X

XX X X X X X

xxxx xxxxxx

XX XXX XX X

xxxxxxxxxxx ro

X X X X X

XX X

X XX X

xxxxxxxxxx +

ID

D

OJ E

C

OJ

. —

V l-

(/)

o

ro

3

0

4-. O

• —

■o

H

ru

(U

c u-

—

n>

0

ra

c

—

o

TO —

-^

>

t.

• —

i/i

o

3

• —

Ol O

• —

\-

01

o

i-

i/i

"O

1_

nj

— ro

o

ra

c

£

fO

■ —

o

^

T3

o

CTl C

(U

a.

a)

>

>

■^

m

0

XI

JD

m

c)

n

( }

3

ID

t/1 T3

1-

m

(U

. —

in

m

O"

•4-J

13
— (D

c

ra

T3

C

o

' —

(B

o

<D —

n>

. —

(D

t>

ra

ra

O}

—

u

C —

i~»

D

ro

1 —

• —

<u

03

13

- —

o —

c

o

c

l_

u

T3

T

W

i-

1-

— o

C)

c

o

(U

m

o

n

-o

Ifi

13 C

X)

-J

u

n

>

4-J

, i

TO

ro

cn 3

o

i_

in

F

o

Q.

Q.

<

=J

0) 1_

r

m

■T

-I

n

HI

ni

cy <

S 1-

<

tj

Ll_

o

o

_j

_j

^-

-3-

(^ v^ oo

o^

rr\

,

o

-.o

r^

' —

CSI

— ^D

CNJ

ro

VD

un

LP*

— — "D

4-i -t-t ~0

s

s

^

lU

E

s

w

4-1

tn

O

03

3

01

T

V4-

4-1

•j

in

< —

01

h

. —

03

01

&

fr

n

D

*J

l/l

.

in

4->

I)

\_

O

03

03

t/1

Q.

c

n>

01

CT

•—

i-

.—

c

. —

1)

01

4-1

0!

01

-Q

4J

03

—

^~-

14-

n3

-1

tn

1_

1-

V

</>

I-

m

—

T

XI

O

M-

o

01

0

4-1

m

o

Q.

03

c

c

r

-O

a.

1-

<U

.—

^

1-

=j

r

CU

()

13

03

-Q

4-«

x:

l/l

O

91

u

tn

o

D

03

o

M-

5;

0!

—

l/l

(n

tn

.—

• —

<D

U)

03

03

D

• —

01

03

^^

CD

ns

. —

ra

c:

4-1

\-

03

L.

. —

r

4~J

,—

c

-J

tn

(1)

r

(1)

0)

._

n

t-

3-

o

i-

0

-o

o

4-1

F

F

i/i

c

o

03

o

■o

^~

•—

T3

a.

-1

()

tl

•—

Tl

m

t/1

fU

< —

o

O

f)

m

F

4->

n

\-

CQ

=!

->

• —

l_

r

3-

c

03

o

r

03

U-

c/

>

<

<

Q-

_J

=)

-J

<

<

\0 O^ <T\ • — (T\ -^ <T\ c<\ ■ — COf'^

^Or^rsivOLTscsiLnrocM ro

102

Bulletin Miisiuni of Coinixinitirc Zoology, Vol. 149, No. 2

uiseg jo|ja;u| ueipeueo

jaAiy iun6u|->jsnn

jaAiy oaops

jaAiy Apues 6!9

jaAjy 6ui>(0|-i

jSAjy A>(on5U3X

^ JSAiy lies

3
b

^ j3Aiy jaieMapeJi

SJ3Aiy Au3L|63l [v 5 e[3L|e6uouow

jaA]y (M3N JO) eqMeue>|

z

X
D

3

j9Aiy qseqett
luajsAs JSAiy oiljo
jSA|y pue[j3qujn3
jaAiy aassauuai
jaAiy siou I 1 I I

X

CN

XXX + XXXXXX +

X

C-i

XXXXXXXX + XXXX

fv

X XX

r*

X X

o

XXXX XXX X

a

XX X

c

<

o

z z

< o

_I —

ai o

LU LlJ

CD Oa

H

1^ S LuajsAs JSAiy iddississiw

jaAiy Daiuejaw

jaAjy apeuooseg

jaAiy a6eso

,,jaAiy ja^e^ •>\^e[q

isaj] >(jezo Moiaq

o
o

z
<

<

M
O

o

X X

XXXX

X

X

X X

o

XXXX

X

X

\£>

XXXX

X

X

X X

X

X

X

X

X

X

-cr

X X

XXXX

X

X

X X

X

X

X

X

X

X

X

X

X

0-1

X X

XXXX

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

CNJ

X

XXXX

X

X

X

X

X

X

X

X

X

X

X

X

r^

X

XXXX

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

o

CN

■^x

CNJ

X

-

o

o

o

o

X

X

rg

r<^

-^

-3-

co

CO

CO

-T

■J-

en

CN

o

LP*

LA

r--

LA

o

o

en

oo

LTV

o

LPv

r*~\

OA
CO

CsJ

LTN

LPl

o

0

ui

en

-n

on

D

c

u

ro

c

Q.

01 "^

(U
3
C

c
o

U

<:

n

3

0

C(

a)

0)

C

o

r

01

()

01

u

T)

01

01

(U

•S

l/l

U1

T3

m

r

4-1

o

0)

■ —

u

OJ

03

■t-J

^

in

m

m

r

-3

m

.—

01

*->

tn

1-

in

■M

*-t

•—

'U

^.

w

H-

o

-u

I-

l-

10

03

in

Q.

c

OJ

OJ

c

U

c

u

n

■M

>

n

o

■n

o

i/l

D

C)

C)

03

03

o

c

o

c

(TJ

^

n

-1

m

n

01

o

1-

I-

r.

■—

01

rr

3

o

in

—

.—

o

in

■—

■ —

fn

u

-J

L.

i/i

n

1-

■—

-J

^

■M

X

u

D

Ol

u

in

d.

o

Q

n

r

n

o

01

n

.—

. —

1-

—

01

. —

<u

4-)

1-

i-

o

b

•^

o

m

1 >

-1

u-

ifi

>.

<)

f)

Xl

Wl

<1>

01

in

OJ

OJ

.i:;

c

m

01

01

u

4-t

0)

O

l+-

D

l_

c

D

G.

in

a.

O

^.^

m

01

F

in

01

u~

01

in

•—

in

r

L-

(0

(1)

.--

r

OJ

01

c

01

o:

0)

03

•^

OJ

(0

03

03

O

t.M

n

r—

n

.—

.n

^—

01

fl)

^

01

a>

■ —

—

-U

•—

D

^

<—

p—

■ —

• —

T^

,_

o

u

o

.—

r

m

-J

o

l/l

4-J

o

C)

O

l_

o

O

o

O

o

bi

in

r

01

I-

t/i

o

o

u

f^

lO

-C

OJ

c

SI

•—

•—

• —

CL

F

n

o

>

T

o

o

u

-n

m

• —

ZT.

f.T

4-»

>

D

i->

Ol

Ul

U)

■ —

>

tn

F

>.

o

Q)

f-

l/l

Q.

(U

OJ

. —

F

(U

OJ

OJ

0

I-

QJ

0)

OJ

OJ

' —

•-1

m

m

*j

n

01

-1

^

D

(1)

-Q

OJ

_j

_i

Q_

o

Q-

—I

Ll

<_J

O*

Q-

>

3:

a.

Q-

O-

O

c_>

Q_

Q.

a.

o.

LU

X X X X X X

o

a:

<
^-
o

Z()Oc;e()GRAphy op^ North Ameiucan Untonacea •^(^/jn.son 103

DISTRIBUTION OF UNIONACEA IN THE Mississippian and Ohioan regions from

MISSISSIPPIAN AND OHIOAN REGIONS either or both areas.

[Table 2J B. Eleven species found in both the

Mississippian and Ohioan regions, most-

The number or letter preceeding the ly missing in the Tennessee and Cum-

name of the species in the table refers to berland river systems, appear to have

the designation on the several faunal lists reached the Ohioan Region from the west,

already given. Seven of these species appear to have

An "X" indicates that the record was had a refugium in the Meramec River or

found in published works and a " + " in- to have come from the Ozarkian Region

dicates that the record is supplemental to south of the Ozark Crest, or from even

these, based on specimens in the Mu- farther south.

seum of Comparative Zoology. A small With the exception of Margaritifera
letter, capital letter, or arabic numeral in falcata, of the Pacific Coastal Region,
the body of the table indicates that there which is limited in the Mississippian Re-
are additional data in the several lists that gion to tributaries of the upper Missouri
should be consulted concerning the dis- River in Montana, all of the species from
tribution of the species under the cate- the west are also found in the Ohioan
gory in which it occurs. The rivers Region. As noted by H. and A. van der
thovight to be possible refugia for the Schalie (1950: 456) certain species do not
Mississippian fauna are listed from west usually inhabit the large rivers, but are
to east. The arrangement of the Cumber- found in smaller streams, sloughs, or
landian and Ohioan rivers continues in lakes. Among these species are Villosa
the same direction but in an arbitrary lienosa, Anodonta suhorhiculatu, A. cor-
fashion. The Wabash follows the Cum- pulenta, Ligumia subrostrata, and
berlandian rivers, since these appear to Uniomerus tetralasmus. Anodonta sub-
he its major source of species. The or/;?'<:'u/«f<'/ is monographed as a represen-
Green, Kahawha, and Monongahela and tative of this group of species. Lampsilis
Allegheny rivers then follow as the most higginsi, Proptera capax, and Arcidens
likely refugia of some of the oldest parts confragosus are also monographed to il-
of the Ohioan fauna. These are followed lustrate the Ohioan penetration by sev-
by other possible refugia, and then by the eral other species from the Mississippian
Scioto and Muskingum rivers which, af- Region.

ter the Wabash, are among the major trib- C. The Ohioan Region has 21 species

utaries of the Ohio on its north side. As which had refugia in at least the Tennes-

mentioned elsewhere, there is no list of see and Cumberland or Monongahela

the Miami River Unionidae. This river, and Allegheny rivers. An additional

between the Wabash and Scioto, proba- species, Elliptio comphinata from the

bly has a fauna similar to that of the Scio- Northern Atlantic Slope, reached the

to and Muskingum rivers. Ohioan Region by way of the St. Law-
rence River system.

A. Of the 74 species found in the Mis- Only two, possibly three, of these 22

sissippian and Ohioan regions, with the species are found in the Mississippian

exception of Lampsilis radiata siliqiioi- Region, although a number of them are

dea, which is unexplainably missing from found in the St. Lawrence River system,

both the Tennessee and Cumberland Two of these Ohioan species: Quadrula

rivers, 40 species are sufficiently ubiq- cijlindrica and Carunculina glans occur

uitous on both sides of the Ozark Plateau, in the Ozarkian Region, south of the

and on the Cumberland Plateau, to indi- Ozark Crest, but are entirely missing

cate that they may have spread into the north of it. A Cumberlandian species,

104 Bulletin Museum oj CUmtfxirdtice ZooId^i/, Wil. 149, \'(). 2

JSAiy OIJEIUQ

uieidmeqo 3^e-\

uiseg sa>je"| ja6uij

oijeiUQ 3>|ei

3IJ3 a>|ei

jaAiy aauinew

jiB[3 -IS 3^ei

uojriH 3'>|e"i

ue6iL|3!w 3>|e-i

SJSAiy ueBiqoiw ujsjsa/^

sjsAiy uisuoosiM uj33se3

joijadns 3>|ei

X xxxxxxxx

X

X

X X

XX X XX XXX

X

X

XX xxxxxxxx

X X

X X

X

X

X

X

X X

XX X >< X >< ;' ;<

X

X

^c

xxxxxxxxxxx

X

X X

X X

X

X

X

X

X

X

X

X X

X X

X X

X

X X

xxxxxxxxxxx

X

X X

X X

X

X

X

X

X

X

X X

XX xxxxxx

X X

X

X

X

X

X

X

X

X X

X X

xxxxxxxx XX

X X

X

X

X

X

X X

XX XX

X

X

-3-

X

-3-

X

xxxxxxxxxxx

X

X X

X X

X

X

X

X

X

X

X

X X

X

X

xxxxxxxxxxx

X

X X

X X

X

X

X

X

X

X

X

X X

X X

X X

XXX

<

1—

GO

>-

tt)
-a

(0

LU

CO

n

13

<

□::

3
D"

C
01

OJ

^

01

r

in

o

:>

s ^

—

o

OJ

4-<

in

—

cd

« s

>—

m

01

4-1

1/1

03

4->

-)

b

•v^ ti

in

-t-j

1/1

05

3

c

in

F

■ —

1-

03

ftj

01

1/1

r

4-1

D

0)

•—

03

OJ

3

OJ

03

O

O

OJ

in

Si- K

fO

m

c

■D

03

1/1

.—

0)

O

01

F

f—

03

l-

4-t

4-*

in

X

03

03

11-

03

• —

in

01

•^ «

t~<

■u

in

m

\-

4->

—

m

\-

4->

03

<—

>

4-*

OJ

0)

O

03

03

U

—

•—

in

u

<U

CO -iJ.

m

01

fl)

01

. —

L.

3

0)

0)

01

01

m

—

L-

0)

u

"O

■ —

—

4-1

' —

-J

u

o

—

i-

u

to '"Q

s

-n

n

4-1

u-

01

4-1

>—

01

X)

JD

4-1

4-»

>

.—

u

03

D

c

u

3

u-

03

03

L-

CD

•—

o

O-

01

<

• tj s

ri

-n

r

F

01

4-1

01

01

m

F

c

3

in

03

01

1 —

<u

Q.

cr

3

o

4-*

0)

-D

O

"D

L

3

in

b

>

CO Q

&

(D

m

n

>

1/1

01

4-1

01

D

4-t

o

O

in

—

J3

•—

1-

u

in

l-

C

03

O

1-

OJ

o

CO t-i

1.

1-

(>

n

(1)

o

•—

i_

01

o

—

.—

M-

in

F

03

1-

4->

3

3

OJ

3

"O

O

it-

u

.—

1—

•ri Sh

■r^

m

T)

01

(1)

■n

14-

4-t

1/1

1/1

—

1-

03

C

4-)

03

Q-

03

4-1

CJ

4-1

n

§: S

fi

{/)

(11

m

l-

c

3

OJ

01

Q-

.—

03

.—

.—

03

F

—

03

c

03

in

<u

ro

rQ

r>

lU

c

n

01

o

01

n

4-1

c

03

03

—

03

. —

(U

OJ

1-

•—

O

03

f—

03

—

c

01

ca p

•^^

p—

4-J

n

, —

4~t

I-

01

<1)

-n

.—

01

o

01

03

03

C

4-t

3

. —

>—

J3

■ —

03

01

T3

■ —

—

■^

>--

u

1:J

fs; s

,v

._

r

m

.«

r

(1)

4-»

-n

x:

c

m

F

in

C

O

c

U

o

—

O

3

3

c

•—

3

• —

o

'—

CTI

I.

— I

tn lo O

l/l

n

1/1

n

4-J

F

Q.

O

F

Q.

o

<U

O

o

c

O

c

•—

u

i_

1-

CT

o

h

U

o

•—

in

•—

03

CD

Q- -D

F

Q- -a

n

-1

.—

4-1

ifl

(■>

I —

(-

r—

. —

o

•—

TD

::

CD

t-

-3

-o

•—

u

in

-D

c

C71

LL

b

>

cf

1-

,

F

o

in

F

n

n

m

. —

a

01

\-

u

in

-Q

. —

in

4-1

O

I-

03

3

i)

03

■ —

in

OJ

03

3

03

b

in

n

1—

'=■:

IT]

r

m

m

r

1-

,—

m

+J

>-

01

F

.—

3

o

c

03

i —

1-

i —

3

J3

3

■ —

3

i-

■ — '

03

03

-Q

_l

<

_i

_i

<

Q-

_i

UJ

_i

<

l/l

c_)

_l

<

>

U.

<

<

o

Q-

h-

Q.

O' O

U-

<

O 1—

Q-

_l

_i

o

-a- \D

C-)

rr\

o

r^co

^O

U\ IS\\0

-3-

-a-

CM

(X)

-3-

r^c»

en

—

u-\

CSI

o

^

vT)

-3-

^

vD

CM

O

ra

o

LTV

CM

r^

LA

P0\0

La

< —

LA

CM

pa

■ —

ra

vO

oa

CM

tv-(-3-

v£)

CM

' —

r--.

CM

' —

vU

^3-

u\

c\

<x>

ZOOGEOGFIAPHV OF NORTH AMERICAX UnIONACEA ■ yo/inso/j 105

xxxxxxx:xxx:x

X X X X

X X X X X X X

oo

LA

-a-

OA

LTV
CM

CM

XX XX XXX

1^

X ;< + X

<r

xxxxxxxxxx

c

XXX

rv-1

X X

CM

o

X

-

X X

CM

o

o

o

X X

CM

o

CM

-3*

(NJ

-3-

LA

-3-

LA
CA

CM

CM

o

CM

CM
PA

LA

T3
CD

C
O

U

<

in

OJ

(U

(/)

0

•—

i_

1-

r-o.

■ —

03

0)

(D

E
1-
0

CO

•—

4J

14-

»—

m

O)

03

—

O

m

n

0

01

•—

4-)

•—

m

D

en

fU

•r^

1-

XI

—

o

01

"O

H

•—

01

03

OJ

S^

(D

03

in

01

t; i/i

XI

c

i/i

-Q

F

O

0

u

3

^

'D

03

*—

in

4-1 (D

c

D

m

F

01

r

CL

<3

03

4-»

03

cn

03

x>

U

03

fa 4-

3

4~>

. —

01

4-1

in

n

1- 0!

r

o

r

t-j

4-J

03

OJ

—

I-

4-1

•—

01

0

<U

3

4-1

in

o

01

Q.

in

T3

XI >

01

<1l

-P

U

C

4-1

3

03

0

•—

U

D" (/I

o

—

1-

\-

01

3

—

o

X C 01

r

^

"=1;

03

lU

lU

U

t

03

'~

u

(U

— 3

1_

.—

XI

4-J

-T

L.

>

4-J

01 — ^

0

D

(L)

>

a>

^—

fO

•—

1_

>

'—

'U

— x:

-Q

.—

D

a)

O

XI

(1)

n

Q. — O

^^

rC

03

u.

4-1

X)

3

OJ

03

XI

L.

>

-C

-Q O

J

j:i

Ul

*->

r

01

01

01

03 >-

01

C)

1_

HJ

4-1

t

_)

01

i-

4-1

3

03

u

O C

(/)

03

o

O U 03

r

01

in

-\

01

?^

03

0

in

l_

OJ

(U

03

in

0

03

0

(D

4-

01

in

u

in

E

0

0

0

03

14-

0

03 1-

TO

•—

3

<—

01

01

01 03 (L)

.^

1-

fU

^

CI

c

in

c

—

03

in

C

in

— JD

•—

(D

OJ

l_

c

^—

H)

<t)

1_ — XI

-I

01

0
C71

s-

01

u

'~

u

(U

L.

*~

u

*~

O O

1_

l/l

c

0)

o

.—

■n

X)

0) 3 O

( 1

m

r

\

0

c

Q)

,3

•~"

(U

' —

u

"~"

4-1

U

• —

u

'~~

— s:

fU

C)

<)

H

in

m

n

n

4-1 1_ l_

r

n

o

E

03

rC

C^

c

•—

—

—

0-1

—

•—

'—

en o

>

o

O

Q.

n

4-'

4-J

CX XJ 3

-I

t^

kj

0

Q.

t

in

t

03

0

—

in

b

in

fD >~

o

-^

in

F

F

n

n

o 01 0)

I

,

rn

CT)

0

■0

• —

-J

u.

in

Ol X3

fc

LL

in

u.

' 4-1

X)

•—

3

c

OJ

(U

4)

1- 3 ^

01

,

01

•

u

'—

01

fc

01

U

b

03

b

Cl_ CL

o

>

Li_

Z5

CO

_j

_j

_i

Q- O" O.

o

UJ

2:

—i

0

h-

<

OQ

<

LU

_1

(U

<

2:

<

03

<

0)

_1

LACM CM r^LAOALACMvO r^-3- r-~— 000 C^ CT\ ~ CMOO
-a-p^vX)^XI CslOALALALAvXl CM-a--3- — LAvO — tv^

31 QQ i<i

Q < C3 O _J

CD

(•ft

<

o

o

I-

106

Bulletin Miisciini of Conipdrativc ZooId^ij, Vol. 149, No. 2

Lampsilis ahrupta, which also occurs in
the Ohioan and Mississippiaii regions, is
monographed.

DISTRIBUTION OF UNIONACEA IN THE
ST. LAWRENCE RIVER SYSTEM [Table
3J

A. Fifty-two of the species in the St.
Lawrence River system entered there
from the Mississippian, Cumberlandian,
and Ohioan regions.

B. Eleven species entered the St. Law-
rence River system from the Northern
Athmtic Slope Region.

DISTRIBUTION OF UNIONIDAE IN THE
CANADIAN INTERIOR BASIN [Table 4]

Thirteen of the species found in the
Hudson Bay drainage are Mississippian
in origin, having entered that region from
a former confluence of the Minnesota
River, the upper Mississippi River and
the Red River. Lasmigona compressa,
Lampsilis radiata siliquoidea, and Ano-
dontoides ferussacianns appear to have
also invaded the Hudson Bay drainage by
way of the St. Lawrence River system.
Elliptio complanata entered only by way
of the latter system. Anodonta kenneriyi
is derived from the Pacific Coastal re-
gion. Anodonta grandis simpsoniana ap-
pears to be speciating within the Cana-
dian Interior Basin. Only three unionids
occur in the Arctic drainage.

DISTRIBUTION OF UNIONACEA IN THE
OZARKIAN REGION [Table 5]

Fifty-three of the 74 species found in
the Mississippian and Ohioan regions
occur below the Ozark Crest (Plate 1, 1),
and have the same arabic numerals as
those in List L An arabic numeral in the
body of the table indicates that there are
additional data of general interest in the
list, while one at the end of a line indi-
cates additional data specifically on the
Ozarkian Region. The uniquely Ozarkian
unionacean fauna, as well as those

species also found in the Gulf Coastal
Region are represented by Roman nu-
merals (List 2).

The Ozarkian rivers below the Ozark
Crest are arranged on the Table from
south to north, while those above the
Crest are listed from east to west.

A. With the exception of Lampsilis ra-
diata siliquoidea, which is missing from
the Tennessee and Cumberland rivers,
40 species are found on the Ozark Pla-
teau, above the Ozark Crest, and on the
Cumberland Plateau.

B. Nine species found in the Gulf
Coastal and Ozarkian regions, have ex-
tended their range into the Mississippian
and Ohioan regions.

C. Nine species appear to be of south-
ern origin.

D. Four species afford evidence of the
uniformity of the fauna of the Ozarkian
and Cumberland plateaus before the for-
mation of the Mississippi River. Qtiad-
riila cylindrica and Carunculina glans
are also found in the Ohioan Region, but
not in the Mississippian Region.

E. With the exception oi Lampsilis Jiig-
ginsi and Proptera capax, two Ozarkian
species which have spread into the Mis-
sissippian and Ohioan regions, 8 species
are unique to the Ozarkian Region.

SYSTEMATIC SECTION

Selection of the Monographed Species.
Fourteen species are monographed. Eight
of these are found only in the Ozarkian
Region. Six others were selected to illus-
trate certain distributional pecularities
previously discussed. The detailed dis-
tribution of the species, based on the
published record, and an examination of
the specimens in the major collections, is
reported for the first time. Among the ad-
ditional six species, are two other Ozark-
ian species which extended their ranges
in post-glacial time, Lampsilis Iiigginsi
and Proptera capax. Lampsilis Iiigginsi
and abrupta, a closely related Cumber-
landian species, are compared. They il-
lustrate overlapping distribution in the

Zoogeography of North American Unionacea • Johnson 107

ARCTIC

DRAINAGE

HUDSON BAY DRAINAGE

TABLE li

DISTRIBUTION OF THE UNIONIDAE
IN THE CANADIAN INTERIOR BASIN

4 Fusconala flava (Rafinesque)
11 Quadrula quadrula (Rafinesque)

2 Amblema plicata (Rafinesque)
3^ Lasmigona costata (Rafinesque)
36 Strophitus undulatus (Say)
63 Proptera alata (Say)
58 Ligumia recta (Lamarck)

53 Lamps i 1 i s ovata (Say)

26 Anodonta grandis grandis Say
33 Lasmigona compressa (Lea)

Anodonta grandis simpsoniana Lea

54 Lampsilis radiata siliquoidea (Barnes)
32 Lasmigona complanata (Barnes)

30 Anodontoides ferussacianus (Lea)
B Elliptio complanata (Lightfoot)
f. Anodonta kennerlyl Lea

F

m

O)

1_

i_

s_

<u

0)

>

>

in

• ^

• —

S-

u

s_

i-

(U

<U

>

c

>

c

—

»_

• ^

i_

u

(U

ai

0)

4-1

■M

—

tn

-0

1/1

—

0)

0)

fD

ra

2

a:

0)

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X
X

X

X

3

6

14

6

Mississippian and Ohioan regions. Their
specific status, nomenclature, and distri-
bution have been matters of some dis-
pute which are resolved here. Cijprog,e-
nia stegaria, a Cumberlandian species
which extended its range into the Ohioan
Region, is compared with C. aherti, an
Ozarkian species. Arcidens confragosus
and Anodonta suhorhiculata are includ-
ed as examples of species which entered
the Ohioan Region from the west in post-
glacial time, that are widely distributed
in the Gulf Coastal Region. A. suhorhic-
ulata was chosen to show the distribution
of that group of species which do not
usually inhabit the large rivers, but are
usually found in smaller streams, sloughs,
or lakes. Among these are also Villosa
lienosa, Anodonta corptdenta, Ligumia
suhrostrata and Uniomerus tetralasmus.

Acknowledgements. Thanks are ex-
tended to Drs. Joseph Rosewater and Ar-
thur H. Clarke, of the National Museum
of Natural History, for allowing me ac-
cess to the collection under their care and
for photographs of types. Dr. George M.
Davis, of the Academy of Natural Sci-
ences of Philadelphia, and Dr. Juan J.
Parodiz, of the Carnegie Museum, Pitts-
burgh, kindly made the collections in
their charge available for study as did Dr.
John B. Burch, of the Museum of Zoolo-
gy, University of Michigan, Ann Arbor.
Dr. Alex S. Tompa, also of the latter in-
stitution, kindly arranged for the loan of
material. Dr. H. W. Charnley, Jr., of the
Buffalo Museum of Science, made two
critical lots of L. capax available, and Mr.
Don Manning, McKenzie, Tennessee,
kindly loaned material. Thanks are also

108

Bulletin Museiun of Comparative Zoologij, Vol. 149, No. 2

OZARKIAN REGION

below Ozark Crest

above Ozark Crest

TABLE 5

DISTRIBUTION OF UNIONACEA IN THE
OZARKIAN REGION

A. Unionaaea oornmon to both

the Ozarkian and Cumberlandian
Regions.

2 Amblema pi icata

12 Tritigonia verrucosa
16 El 1 ipt io di latata

53 Lamps! 1 is ovata

36 Strophitus undulatus
71 Obliquaria reflexa

65 Truncilla truncate
kS Lamps ills teres

55 Leptodea frag Mis

10 Quadrula pustulosa
k Fusconaia flava

11 Quadrula quadrula
28 Anodonta imbecillis

25 Alasmidonta marginata

37 Actinonaias ligamentina
58 Ligumia recta

3') Lasmigona costata
22 Pleurobema cordatum

3 Fusconaia ebena

8 Quadrula metanerva
H Cyclonaias tuberculata

54 Lampsilis radiata siliquoidea
m Alasmidonta calcedus

48 Ellipsaria lineolata

57 Leptodea leptodon

60 Obovaria olivaria

68 Vi 1 losa iris

39 CaruncuHna parva

15 Elliptio crassidens crassidens

26 Anodonta grandis grandis

66 Truncilla donaciformis
32 Lasmigonia complanata
41 Plagiola triquetra

56 Leptodea laevissima

13 Megalonaias gigantea
63 Proptera alata

19 Plethobasus cyphyus

1 Cumberlandia monodonta
30 Anodondoites f erussac i anus

35 S impson i concha ambigua

E

<u

1_

E

■M

01

<U

in

>

u

4-i

>~

*

•i;

QJ

1/1

on

H:

cc

1-

>

>~

I-

dJ

CO

i~

0)

1_

u

l/l

>

a:

1-

(U

>

QJ

0)

.—

(U

U

>

—

>

>

u

ce

(U

>

0)

•—

oc

.—

.—

c

X)

>

cd

oc

cc

fD

u

nj

a:

• —

0

i_

(U

c

QC

^

JZ

(V

J^

Ll.

E

0

<v

0

1/1

+J

u

ra

0

CD

■a

(D

0

—

n}

.

L.

(/I

ra

(D

■ —

(U

J=

' —

w

0)

ID

Ul

al

CO

z

3

CD

1/1

X

<J

0

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

+

X

X

+

+

X

X

X

X

X

X

X

+

X

X

X

X

X

X

X

X

+

+

X

X

X

X

X

X

X

+

+

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

+

+

X

X

X

X

X

+

+

+

X

X

X

X

+

X

X

X

+

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

+

X

X

X

X

X

X

X

X

X

+

+

X

+

+

X

X

X

X

X

X

+

X

X

X

X

X

X

X

+

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

+

X

)'

X

X

X

+

X

X

X

+

+

X

X

X

X

X

X

X

+

X

X
X

X

X
X

X

X

X

X

X

X

+

X

X

X

X

X

X

+

+

X

X

X

X

X

X

+

+

X
X

X

X

X

+

+

X

X

X

X

X

X

X
X

X
X

+
X

X
X
X
X

X

/,

X
X
X
X

X
X

X

(2)

(55)

(37)

(32)

(63)
(19)
(1)

(^0)

(35)

20 28 32 32 31 31

2D

31

extended to my colleagues, Dr. Kenneth manuscript and made a number of signif-

J. Boss who read the manuscript; and to icant suggestions. Dr. David H. Stans-

Dr. Ruth D. Turner who thankfully read bery, of the Ohio State Museum, help-

the manuscript with a very critical eye. fully, offered several opinions. Mr.

Dr. Arthur H. Clarke of the National Mu- Douglas G. Smith, of the Museum of Zo-

seum of Natiual History kindly read the ology, University of Massachusetts, Am-

Zoogeography of North American Unionacea •/()/»! vo/;

109

TABLE 5. Continiu'd

OZARKIAN REGION

B. Unionidae common to both the
Gulf Coastal and Ozavkian
Regions, also found in the
Mississippian and Ohioan
Regions.

6 Fusconaia undata

31 Arcidens confragrosus

23 Uniomerus tetralasmus

59 Ligumia subrostrata

38 Actinonaias el 1 i ps i formis

9 Q.uadrula nodulata

69 V i 1 losa I ienosa

29 Anodonta suborb iculata

27 Anodonta corpulenta

C. Unionidae common to both the
Gulf Coastal Region and the
Ozarkian Region, below the
Ozark Crest.

XIV Proptera purpurata

XI Lamps ills hydiana

XI 11 Obovaria jacksoniana

11 Plectomerus dombeyanus

X Lamps i 1 is satur

XVI I Vi 1 losa vibex

III Quadrula apiculata
VI I I Glebula rotundata
IX Lamps! lis excavata

D. Unionidae common to both
the Ozarkian Region, below
the Ozark Crest, and the
Cumberlandian Region.

7 Quadrula cylindrica
^0 Carunculina glans
VI Plagiola florentina

VII Plag iol a turg idula

below Ozark Crest

above Ozark Crest

E
(U

4-1
t/1
:^

i_
HI

>

x>

(D

E
0)
■w
1/1

i_
a)

>

u

■!;
OJ

>

a:
O

-C

in
O
0)

z

(U

>

en

CD
+-»

SI

3

^;

1-

>

O

TO

CD

>

1/1

o

c
nj

V-
U-

4->

1-
>

o

(U

E

(D

i_
01

3:

(1)

>

a:

a)

-D
ID

c
o
o

Ul
(D

>
a)

C7)
TO

1/1
o

X
X
X
X

X
X

+
X

X
X

X

+
X
X

+

X
X

+
X

X

X
X
X

X
X
X

X

+

X

+
X
+
X
X

X
X

X

+
X
X

+

X
X
X
X
X

X

X

X
X
X
X
X

X X X + + +

XX X + X

XX X + X

XX X + +

XX X X +

X
X
X

+

u

o

0)
0)

I/)

(6)

(23)

(38)

(X)
(IX)

(VIM)

X X X X X

X + X

X X

X X

E. The Ozarkian Fauna

XIX Ptychobranchus occidental is

XVIII Cyprogenia aberti

51 Lamps ills higg ins i

XI I Lamps i 1 is powel 1 i

I Fusconaia ozarkens's

V Actinonaias raf i nesqueana
XVI Villosa reeviara
6^* Proptera capax
IV Arkans ia wheeleri
XV Villosa arkansasensis

Arkansas River System

All the species are found

ioi the Neosho River drainage, -r , „

^ Total A-E

"•'• White River System

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X
X

1

2

6

1

5 8

5 K ■'.

"T^ , '

37

h3

^5 Ise

51 l^g 150

33

1

37 1

■^■: To

tal 6

3

110 Bulletin Museum <>J C^oniixmitive Zoology, Vol. 149, No. 2

heist, supplied certain data on the Lake
Cvhamphiin fauna. Most of the photo-
graphs were taken by Mr. David L.
MacHenry. The maps and tables were
carehilly prepared by Miss Catliarine G.
Kessler wlio also helped with other nu-
merous details, including proofreading.
At my last moment recpiest, Steven
Ahlstedt, Tennessee Valley Authority;
Alan C. Buchanan, Missouri Department
of Conservation; Mark E. Gordon; Uni-
versity of Arkansas; and R. D. Oesch,
Glendale, Missouri; all kindly supplied
me with data and specimens.

Pattern Vscd in Descriptions.

Abbreviations. The following abbrevi-
ations have been used in the text and on
the plate captions:

ANSP Academy of Natural Sciences of Phila-
delphia, Pennsylvania

CM Carnegie Museum, Pittsburgh, Pennsyl-

vania

MCZ Museum of Comparative Zoology, Cam-
bridge, Massachusetts

MZUM Museum of Zoology, University of Mich-
igan, Ann Arbor, Michigan

OSM Ohio State Museum, Columbus, Ohio

USNM National Museum of Natural History,
Washington, D. C.

Synonymy. For ease of reference, fidl
citations are included for each taxon, in-
cluding the type locality and the location
of the type when known. References to
plates and figiues are not included under
Lea's Observations on the Genus Unio
since they are always the same as in the
preceding entry. In some instances, lec-
totypes are selected. Elsewhere in the
text, references are abbreviated and re-
quire the use of the bibliography. Except
for the original references, only relevant
citations since 1914 are included here
since earlier ones are available in Simp-
son (1914).

Descriptions. The measurements are
only intended to convey the general size
of specimens from a given station, and to
indicate sexual differences.

Anatomy and Breeding Season. The
available data are cited.

Habitat. Included when known.

Remarks. These are designed to elu-
cidate the differences between the sexes
and the species, and include comments
on distribution and taxonomy.

Range. The distribution is summa-
rized.

Abundance. The former abundance of
the species is based on the ninnber and
size of the lots found in the several col-
lections studied. Their present abun-
dance is based on the opinions of Stans-
bery (1970, 1971, 1976).

Specimens Examined. Most of the rec-
ords are based on the specimens in the
collections mentioned above. These col-
lections contain most of the specimens
available for study. With the exception of
those in the Ohio State Museum, most of
those in the other institutions have been
personally examined. The records are fol-
lowed by the initials of the institution in
which they were observed. In most in-
stances when records were duplicated
only references to the specimens in the
Museum of Comparative Zoology were
included, though in certain critical cases
several references are mentioned. Spec-
imens not seen are credited to the re-
sponsible individual or published refer-
ence and to the associated institution, if
it is known. The published references
may be found either under Synonymy or
under Literature Cited, or sometimes
under both headings.

Distribution. The various river sys-
tems are listed from west to east. East of
the Mississippi River this arrangement
approximates the direction of post-glacial
dispersal.

Arrang,ement of the Records. The river
systems of the Gulf Coastal Region are
listed from west to east. The drainages
below the Ozark Crest are listed from
south to north. In the Ozarkian Region,
above the Crest, the rivers flowing into
the Missouri from the south are listed
from west to east. The records for all of
the above mentioned rivers are arranged

Zoor.EOGRAPHV OF XORTII AMKHK:A.\ UnIOXACEA • /r)/i//.S7)/( 111

from the headwaters to the mouth. Al)ove
the Missouri River the Mississippian
drainages are listed hom south to north
and the records are arranged from mouth
to headwaters, approximating the direc-
tion of post-glacial dispersal. In the Cum-
berlandian and Ohioan regions the rivers
flowing into the Ohio are listed from west
to east. The records of those on the south
side of the Ohio River are arranged from
headwaters to mouth, while those of the
Ohio itself and the tributaries above it are
arranged from mouth to headwaters, ap-
proximating the direction of post-glacial
dispersal.

The cumbersome term, Mobile-Ala-
bama-Coosa river system, is used be-
cause it reflects the main channel of the
river, which was given a different name
at each important confluence.

Figures. When available, holotypes, al-
lotypes and lectotypes are generally used
to illustrate each species. Some of the
data on the plate captions, such as the
measurements, are not repeated else-
where.

The distributional maps are based on
Rand McNally and Company's Commer-
cial Atlas of America 1912 Edition, plate
21, which indicates United States Inland
Waterways.

Superfamily UNIONACEA Thiele 1935

Family UNIONIDAE (Fleming 1828) Ort-

mann 1911

Subfamily AMBLEMINAE

Tribe AMBLEMINI

Genus Fusconaia Simpson

Fusconaio Simpson 1900, Proc. U. S. Natl. Mus. 22:
784. Type species, Unio trifionus Lea, original
designation.

Fusconaia ozarkensis (Call)
Plate 9, figures 1-4
Distribution: Plate 8 A

Unio ozarkensis Call 1887, Proc. U. S. Natl. Mus.
10: 498, pi. 27 (Currant [Current] River, Shannon
Co., Missouri, and in lacks Fork and Big Creek,
tributaries to it; lectotype MCZ 5707 selected by
lohnson, 197.5, Occ. Papers on Moll. 4: 14.3, spec-

imen portraved in figs. 1-3; figured paralectotype
MCZ 5705;'hoth from lacks Fork [Shannon Co.]).

Pleurohema hrevis suhelliptica Simpson 1900,
Proc. Acad. Nat. Sci. Phila. 52: 80 (Spring River,
Hardy [Fulton Co.], Arkansas; lectotype USNM
1599.52 selected by Johnson, 1975, Special Occ.
Pub. no. 4, Dept. Moll., Mus. Comp. Zool., p. 20,
pi. 2, fig. 7).

Pleurohema pannosa Simpson 1900, Proc. Acad.
Nat. Sci. Phila. 52: 82 (White River; [Ouachita
River], Hot Springs, [Garland Co.]; both Arkan-
sas); 1900, Proc. U. S. Natl. Mus. 22: 799. The
measured type, from the latter locality, was not
located by Johnson (1975a: 16). Included in this
synonymv on the authoritv of Ortmann (1917,
Nautilus 31: 62).

Pleurohema utterhaekii Frierson 1915 [in] Utter-
back, Amer. Midi. Nat. 4: 197 [86], pi. 5, figs. 12
a, b; pi. 20, figs. 63 A-D (White River, Hollister,
[Taney Co.], Missouri; holot\'pe MCZ 271448, re-
figured by lohnson, 1972, Occ. Papers on Moll.
3: 151, pi. 24, fig. .5).

Quadrula flava sampsoniana P'rierson 1927, Check
list N. American naiades, p. 55 (Elk River,
[McDonald Co.], Missouri; holotype MZUM
92226, refigured bv Johnson, 1972, Occ. Papers
on Moll. 3: 149, pi'. 25, fig. 5).

Fuseonaia ozarkensis (Call). Ortmann, 1917, Nau-
tilus 31: 62.

Description. Shell medium in size, sel-
dom exceeding 70 mm in length. Outline
elliptical to subtrapezoid. Valves usually
somewhat compressed, solid, inecjuila-
teral. Anterior end regularly rounded.
Posterior end slightly biangulate below
the medial line. Ventral margin straight
or regularly curved. Dorsal margin
straight forming an indistinct angle with
the obliquely descending posterior mar-
gin. Hinge ligament long. Posterior ridge
double, but very faint. Umbos small and
low, not much projected above the hinge
line, their sculpture consisting of several
coarse undulations. Surface of the shell
rather smooth, rest stops usually clearly
marked. Periostracum yellowish-brown
or olivaceous, often marked with numer-
ous obscure narrow green rays, arranged
on the central portion of the disk.

Left valve with two triangular pseu-
docardinal teeth of about equal height,
two curved lateral ones. Right valve with
one high, triangular pseudocardinal, a
smaller one before and behind it, and one
strong lateral tooth. Anterior and poste-

112

BitUctin Museum of ('oinixirdtiic '/.oolof^i/. \'()l. 1 49. Xo. 2

rior muscle scars, and pallial line all well
impressed. Nacre usually white or bluish,
iridescent posteriorly.

Length Heiulit W'idtli
mm mm mm

68 42 26 Whik' River, Hollister,

Taney Co., Missouri. Holo-
of P. utterhdckii.

59 38 20 Big Fork, Texas Co.,

Missouri.

55 36 20 Jacks Fork, Shannon Co.,

Missouri. Leetotxpe of [7.
ozarkensis.

Anatomy and Breeding Season. The
anatomy was discussed by Ortmann
(1917: 63), and on the basis of it, he rec-
ognized that ozarkensis was a Fusconaia.
He further noted that the glochidia are
subelliptical, slightly higher than long;
0.15 X 0.18, and that the species is tachy-
tictic.

HaJntat. Found by Call in shallow
mountain streams, "being swift, and lim-
pid, with rocky bottoms."

Remarks. Fiiseonaia ozarkensis (Call)
does not much resemble any other Ozark-
ian species. It is, however, very close to
Fiiseonaia harnsiana (Lea) of the Cum-
berlandian Region. F. ozarkensis has a
more elongated, subtrapezoidal shell,
more anterior umbos, and the rays on the
disk are faint, at best, and generally en-
tireK' wanting, where those of harnsiana
are well developed.

Ortmann (1917: 63) indicated that the
syntype, subsequently selected as the
lectotype, is less typical of the general
shape of the shell than is that of the fig-
ured paralectotype, and that the differ-
ence in shell morphology is not a sexual
one as it was thought to be by Call.

Ran^e. Found only in the Ozarkian Re-
gion: below the Ozark Crest in the Black,
Arkansas, White, and St. Francis river
systems; and above it in the Gasconade
River system.

Abundance. Apparently not taken in
large numbers, except in the Current Riv-
er drainage of the White River system
and in the Big Piney River of the Gas-
conade River system.

Specimens Examined

Black River System

Ouachita Rixer Drainage. Arkansas:
[Ouachita RixerJ Hot Springs, Garland
Co. (Simpson).

Ark.\nsas River System

Neosho River Drainage. Missouri:
Neosho River (Utterback, 1917, Table [as
utterhackii]); Elk River, McDonald Co.
(MZUM).

White River System

White River Drainage. Missouri: James
River, Galena, Stone Co.; White River,
Hollister, Taney Co. (both MCZ). Arkan-
sas: White River, Cotter (MZUM), and
Norfolk (CM), both Baxter Co.; Big Buf-
falo Fork of White River, below Gilbert,
Searcy Co. (Meek and Clark, 1912: 10).

Current River Drainage. Missouri: Big
Creek, Texas Co.; Jacks Fork; Current
River, 5.8 mi. NE Eminence; both Shan-
non Co.; Current River 1 mi. SW Doni-
phan, Ripley Co. (all MCZ).

Spring River Drainage. Arkansas:
Spring River, Hardy, Fulton Co. (MCZ).

St. Francis River System

St. Francis River Drainage. Missouri:
St. Francis River (Utterback, 1917, Table
[as utterhackii]).

Gasconade River System

Gasconade River Drainage. Missouri:
Big Piney River, 10 mi. W Licking, Texas
Co.; Gasconade River, 6 mi. NW Vichy,
Maries Co. (both MCZ).

Subfamily UNIONINAE
Tribe ANODONTINI
Genus Anodonta Lamarck

Aiuxlonta Lamarek 1799, Memoires de la Soe.
d'Hist. Nat. de Paris, p. 87. Type species, Mytihts
cy^iu'tis Linnaeus. Monohpic. Placed on tlie Of-
ficial List ot Generic Names in Zoology, in 1926,
Opinion 94. Reconfirmed, 1959, Opinions and
Declarations rendered hv Int. Comm. Zool. .No-
men., 20 (28): 303-310, Opinion 561.

Zo()CEO(;r\phy of \okth American \J'SIO\aceA' Johnson 113

Subgenus Utterbackia F. C. Baker

Uttcrhackia F. C. Baker 1927, Amer. Midi. Nat. 10:
221, 222 (misspelled as Vttcrharhia on p. 221).
Type species, Anodonta imhccillis Say, original
desiKiiation.

Uttcrl)ackia>ta Frierson 1927, Cheek list \ Ameri-
can naiades, p. 17. Type species, Anodonta suh-
orhiculatu Sav, monotypic. Discussed by John-
son, 1970. Bull. Mus. Comp. Zool. 140: 362.

Sayadina Haas 1969, Das Tierreich, pt. 88, p. .359,
nan Gregorio 1914. Walker (1918: 176) correctly
pointed out that the type species of Sdycidina,
Anodonta lenustu Gregorio is a synonym ot An-
odonta grandis Say.

Anodonta {Utterbackia) suborbiculata
Say

Plate 9, figure 5

Distribution: Plate 5

Anodonta suborbiculata Say 1831 [Jan. 29] New
Harmony [Indiana] Disseminator [no pagination]
(Ponds near the Wahash [River, near New Har-
mony, Posey Co., Indiana], type ANSP [lost],
teste Johnson and Baker, 1973, Proc. Acad. Nat.
Sci. Phila. 125: 146). Say, 1831, American Conch,
no. 2 [letter press], pi. 11. Simpson, 1914, Cat.
Naiades 1: 400. Parmalee, 1967, Illinois State
Mus. Pop. Sci. Ser. 8: 48, pi. 1.5 B. Starrett, 1971,
Illinois Nat. Hist. Surv. Bull. 30: 320, pi. 3, fig.
19.

Anodonta (Utterbackiana) suborl)icuIata Say.
Frierson, 1927, Check list N American naiades,
p. 17.

Anodonta (yai/adina) suborbiculata Sav. Haas,
1969, Das Tierreich, pt. 88, p. 360.

Description. Shell medium to large,
reaching over 130 mm in length. Outline
suborbiculate. Valves subinflated, thin
but strong, gaping in front and behind.
Anterior end regularly rounded, posterior
end broadly pointed at about the medial
line. Ventral margin regularly curved.
Dorsal margin straight, ending in a wing
like angle before and behind. Posterior
slope slightly incurved above the low
posterior ridge. Umbos very low and
compressed in young specimens, becom-
ing somewhat fuller in adults, but not
much if at all; raised above the hinge
line, umbonal sculpture consisting of a
few irregular, feeble undulations, each of
which has on it, small sharp, tubercles
which form two imperfectly radiating
rows, located somewhat anterior of the

center. Periostraciun smooth and shining
in \'oimg shells, pale yellow green, with
numerous very fine green rays. Older
shells are concentrically sculptured and
the periostracum is smokey or yellowish
and often nearly rayless.

Hinge without any trace of denture.
Umbonal cavities shallow. Muscle scars
indistinct. Nacre silvery and iridescent,
sometimes tinted bluish or purplish.

Length Height Width
mm mm mm

170 126 58 Keokuk Lake, Muscatine,

Muscatine Co., Iowa.

130 100 43 Spoon River, Illinois.

Anatomy and Breeding Season. The
anatomy and breeding season of A. sub-
orbiculata are not known. Sexuality and
other aspects of reproduction of all the
other members of the subgenus Utter-
backia are discussed by Heard (1975). A
glochidium was figured by Surber (1915:
4, pi. 1, fig. 1).

Habitat. Found in a sand or mud bot-
tom, in water with an average depth of
five to eight feet, usually in lakes or
sloughs where the current is not strong.

Remarks. Anodonta suborbiculata Say
is very easily identified. Since the hinge
teeth are completely absent it is easily
recognized as an Anodonta, and as the
umbos are flat, and not raised above the
hinge line it may be recognized as be-
longing to the subgenus Utterbackia. Its
orbiculate outline separates it from the
other members of the subgenus.

The records of A. suborbiculata from
Nebraska (Aughey, 1877) have never
been verified. This species was not found
in the Tennessee River system by Ort-
mann (1918, 1924, 1925) which suggests
that its occurrence behind the Wheeler
Dam may be the result of the subsequent
introduction of fishes. In any event, the
dam created an environment the species
prefers.

Range. Gulf Coastal Region: Trinity
River system, Texas and the Escambia
River system, Florida. Lower Mississippi
River drainages. Ozarkian Region: Red

114 Bulletin Musciini of Ctiiiipdnilivc ZoolouU- ^ "/■ ^~^^^^ ^'O- -

and Black ri\er systems, Arkansas; Ar-
kansas River system, Kansas; White and
St. Francis river systems, Missomi and
Arkansas; all below the Ozark Crest;
abo\e it, the Meramec River system, Mis-
souri. Mississippian Region: Missouri
and Illinois rivers; Mississippi River and
drainages from Illinois north to Wiscon-
sin. Cimiherlandian Region: Tennessee
River, Alabama. Ohioan Region: Wabash
River drainage, Indiana; Green River,
Kentucky.

A})undance. This species appears to he
rare, though occasionally very abundant
locally. Taken in great numbers by
Strode (1891) in Thompson's Lake, Illi-
nois.

Specimens Examined

Gulf Coastal Region

Trinity River System

Trinit\' River Drainage. Texas: Wards
Prairie Lake, nr. Romavor, Libertv Co.

(MCZ).

EscAMRiA River System

Escambia River Drainage. Ahil^ama:
Escambia River, nr. Pollard, Escambia
Co. (MZUM). Florida: Chumucklee
Spring Lake, nr. McDavid, Escambia Co.
(MZUM).

Conecuh River Drainage. Alabama:
Gantt Lake, Clearview, Covington Co.
(MCZ).

Ozarkian Region

Red River System

Red River Drainage. Arkansas: Little
River (MZUM). Mercer Bayou, N Black
Diamond, Miller Co. (MCZ). Louisiana:
Red River (MZUM); Frierson (USNM);
Clear Lake; Edwards Lake (both
MZUM); all De Soto Parish. Black Lake
Bayou, Lebanon, Bienville Parish. [Red
River] Alexandria, Rapides Parish
(USNM).

Black River System

Ouachita River Drainage. Arkansas:
Old River, Arkadelphia, Clarke Co.
(Wheeler, 1918: 121).

Aricansas River System

Neosho River Drainage. Kansas: Neo-
sho River, Woodson Co. (Call, 1885b:
96); pond E Neosho River, 4.5 mi S
Hiunboldt, Allen Co. (Miuray and Leon-
ard, 1962: 160).

White River System

Black River Drainage. Missouri: Black
River, Hendrickson, Butler Co. (MCZ).
Arkansas: Black River, Black Rock
(MCZ); Spring River, Black Rock
(MZUM); both Lawrence Co.

St. Francis River System

Tyronza River Drainage. Arkansas:
Pond, nr. Tvronza River, Tyronza, Poin-
sett Co. (MZUM).

Mer\mec River System

Meramec River Drainage. Missouri:
Creve Coeur Lake, St. Louis Co. (MCZ).

Mississippi River System

Mississippi River Drainage. Tennes-
see: Mississippi River, Presidents Island,
nr. Memphis, Shelby Co. (MCZ).

Mississippian Region
Missouri River System

Elkhorn River Drainage. Nebraska:
Elkhorn River (Aughey 1877: 704).

Missouri River Drainage. Missouri:
Lake Contrarv, St. Joseph, Buchanan Co.
(MZUM).

Kansas River Drainage. Nebraska: Big
Blue River (Aughey 1877: 704).

Mississippi River System

Mississippi River Drainage. Illinois:
Mississippi River, Alton, Madison Co.

(MCZ).

Zoogeography of North Amerigan Unionagea •yo/in.son 115

Sangamon River Drainage. Illinois:
Clear Lake nr. Springfield, Sangamon
Co. (MCZ).

Spoon River Drainage. Illinois: Spoon
River; Simpsons Lake; Thompsons Lake;
{all MCZ) all Fnlton Co.

Illinois River Drainage. Illinois: Illi-
nois River, Mossville, Peoria Co. (Star-
rett, 1971: 320).

Kaskaskia River Drainage. Illinois:
Kaskaskia River (Parmalee, 1967: 48).

Mississippi River Drainage. Iowa:
Keokuk Lake, Museatine, Muscatine Co.
(MCZ). Wisconsin: Mississippi River, La
Crosse, La Crosse Co. (Mathiak, 1979: 7).

Cumberlandian Region

Tennessee River System

Tennessee River Drainage. Alabama:
Tennessee River, Wheeler Reservoir
[Lawrence Co.l (MZUM).

Ohioan Region
Ohio River System

Wabash River Drainage. Indiana:
Ponds nr. the Wabash River, New Har-
mony, Posey Co. (Say); Wabash River
(Call, 1900: 533).

Creen River Drainage. Kentucky: Bar-
ren River; Pond nr. Bowling Green, War-
ren Co. {both MZUM).

Genus Arcidens Simpson

Arcidens Simpson 1900, Proc. United States Natl.
Mus. 22: 661. Type species, Alasmodonta con-
franosa Say, original designation. This genus is
monotypic.

Arcidens confragosus (Say)
Plate 10, figures 1, 2
Distribution: Plate 4

Alasmodonta confra^osa Say 1829, New Harmony
[Indiana] Disseminator 2 (22): .339 (Fo.\ River, of
the [Little] Wabash [River, Richland Co., Illinois],
type ANSP [lost], teste Johnson and Baker, 1973,
Proc. Acad. Nat. Sci. Phila. 125: 146); 1831,
American Conch, no. 3 [letter press], pi. 21; 1832,
no. 4, p. [4] of wrapper.

Arcidens confra^osa jacintoensis Strecker 1931,
Baylor Univ. Mus. [Waco, Texas], Spec. Pub. 2,

p. 13 (San Jacinto River, Liberty Co., Texas; syn-
types presumed to be in the Strecker Museum
[not available]; topotype MCZ 2.3317).
Arcidens confragosus (Say). Simpson, 1914, Cat.
Naiades, 1: 475. Utterback, 1915, Amer. Midi.
Nat. 4: 254 [103], pi. 22, figs. 72 A and B. Coker,
1919, Bull. U. S. Bur. Fisheries for 1917-18, 36:
34, pi. 21, lower pair. Frierson, 1927, Check list
N American naiades, p. 21. Baker, 1928, Wiscon-
sin Geol. and Nat. Hist. Surv. Bull. 70 (2): 196,
pi. 63, figs. 4-7. Murray and Leonard, 1962, Univ.
Kansas, Mus. Nat. Hist. Misc. Pub. 28: 97. Par-
malee, 1967, Illinois State Mus. Pop. Sci. Ser. 8:
51, pis. 16 A; 33 A. La Rocque, 1967, Ohio Geol.
Surv. Bull. no. 62 (2): 208, fig. 94; fig. 95 distri-
butional map. Haas, 1969, Das Tierreich, pt. 88,
p. 396. Starrett, 1971, Illinois Nat. Hist. Surv.
Bull. 30: 312, pi. 3, fig. 13.

Description. Shell medinm to large,
often exceeding 100 mm in length. Ont-
line rhomboid. Valves inflated, subsolid,
or solid, slightly inequilateral. Anterior
end regularly rounded. Posterior end
somewhat produced below the medial
line. Ventral margin regularly curved.
Dorsal margin straight, forming a sharp
wing like angle with the very straight,
obliquely descending posterior margin.
Hinge ligament long. Posterior ridge
rather high, but indistinct. Posterior
slope with mmierous heavy, radial cor-
rugations. Umbos high and full, much
projected above the hinge line, their
sculpture consisting of irregular, doubly-
looped ridges, the bases of the loops
being developed into strong, pinched-up
nodules, which e.xtend out on to the disk
of the shell; in front of, and behind, the
loops there are a number of radiating,
wavy, subnodulous lirae or small ridges;
from the anterior row of knobs extending
backwards and downwards there is a se-
ries of strong folds and these are crossed
by wrinkled, radiating, sometimes zig-
zagged threads. The anterior base usually
has merely concentric striae. Periostra-
cum subshiny or dull, brownish green, or
olive; occasionally (luite greenish and
banded.

Left valve with an arched, somewhat
elongated pseudocardinal tooth under
the umbo which curves upward, in front
of which, there is a lower compressed

116

Bulletin Mtiscuiii oj Coinixmitive Zoology, Vol. I U), No. 2

tooth. Hiiiht valve with a strong sul)c()in-
pressed tooth in front of the umbo. The
hinge phite is cut away from the umbo for
the reception of the tooth in the left
valve. The lateral teeth reduced to ves-
tigial blurs. Unibonal cavities deep, an-
terior muscle scars shallow, posterior
ones and pallial line quite indistinct. Na-
cre bluish-white or dull white, with a
wide prismatic border.

Length Height Width
nun mm mm

115 78 48 Rough River, Dundee,

Ohio Co., Kentucky.

105 77 51 Fox River, IlHnois. Topo-

t>pe MCZ 53004.

84 57 32 San Jacinto River, Texas.

Topotype MCZ 23317.

Anatomy and Breeding Season. The
anatomy was discussed by Ortmann
(1912: 284, fig. 12). A glochidium was fig-
ured by Surber (1912, pi. 1, fig. 5) 0.355 x
0.350. This species is bradytictic, proba-
bly from September to Jvuie (Baker, 1928:
197). Host fish not recorded.

Habitat. Found in a sand or mud bot-
tom in sluggish water a few feet deep.

Remarks. Arcidens confragosus (Say),
a widely distributed species, is not very
closely related to any other unionid, ex-
cept Arkansia wheeleri Ortmann and
Walker which is found only in the Ozark-
ian Region in the upper Red and Black
river systems. The sculpture of Arcidens
is much more complicated than that of
Arkansia and covers most of the surface
of the shell, while that oi Arkansia is re-
stricted to the posterior half of the disk.
The radiating wrinkles and curved folds
on the dorsal slope are much alike in
both, but the folds on the posterior disk
of Arkansia more closely resemble those
of Am})lema plicata (Say). The umbonal
sculpture of Arcidens somewhat resem-
bles that of Qiiadrula qiiadnila Rafines-
que, but Arcidens does not much resem-
ble the latter in any other way. While the
umbonal scidpture of Arkansia is com-
paratively simple and confined to the up-

per umbo, that o{ Arcidens is developed
into strong pitched up nodides, at the
base of the loops, and extends out onto
the disk. The lateral teeth in Arcidens are
not as well developed as those oi Arkan-
sia, but the presence of the interdental
projection in the left valve and the cut-
ting away of the interdentum in the right
valve opposite it, relates Arkansia, Arci-
dens, and Lasmigona.

This species was not found in the Ten-
nessee River by Ortmann (1918, 1924,
1925), which strongly suggests that its
present occurrence in Kentucky Lake be-
hind an impoundment on the lower Ten-
nessee River is the result of the recent
introduction of fishes.

Range. Gulf Coastal Region: from the
Colorado River system, Texas, east to the
Tombigbee River drainage of the Mo-
bile-Alabama-Coosa River system, Mis-
sissippi. Lower Mississippi River drain-
ages. Ozarkian Region: Red, White and
St. Francis river systems below the Ozark
Crest; above it the Osage and Meramec
river systems. Mississippian Region:
Missouri River system; Illinois River;
Mississippi River and drainages from Il-
linois north to Minnesota. Cumberland-
ian Region: lower Tennessee River, Ten-
nessee. Ohioan Region: Wabash River
drainage, Indiana; Green River drainage,
Kentucky; Ohio River, east to Cincinnati,
Hamilton Co., Ohio.

Reported from the St. Lawrence River
System: Great Lakes Drainage (Lake
Michigan) Illinois: Chicago, Cooke Co.
by Baker (1906: 74) as in the State Lab-
oratory, now the Illinois Natural History
Survey, Urbana, Illinois. This record
could not be verified there (Smith, pers.
comm.) or at the Museum of Natural His-
tory of the University of Illinois, also at
Urbana (Hoffmeister, pers. comm.). It is
regarded as an erroneous record, since
like other anodontine species Arcidens
prefers the oozy mud of river margins
where there is little current.

Abundance. "The species is rare but
widely distributed" (Coker, 1919: 34).

Zoogeography of North American Unionacea • Johnson 117

Specimens Examined

Gulf Coastal Region

Colorado River System

Colorado River Drainage. Texas: Col-
orado Rixer, Austin, Travis Co. (Streck-
er); Skull Creek, Colorado Co. (USNM).

Brazos River System

Brazos River Drainage: Texas: West
Yegua Creek, Lee Co. (MZUM).

Buffalo Bayou System

Buffalo Bayou Drainage. Texas: Buf-
falo Bavou, near Eureka (MCZ), Houston
(USNM); both Harris Co.

San Jacinto River System

San Jacinto River Drainage. Texas: San
Jacinto River, Liberty Co. (MCZ,
MZUM).

Trinity River System

Trinity River Drainage. Texas: Trinity
River, Dallas (MCZ); Elm fork of Trinity
River (Strecker); both Dallas Co. Trinity
River (MZUM); Chambers Creek
(Strecker); both Navarro Co.

Neches River System

Neches River Drainage. Texas: Neches
River, Smith Co. (Strecker). Neches Riv-
er, Big Eddy, Henderson Co. (Strecker).
Poe Lake, Nacogdoches Co. (MZUM).

Sabine River System

Sabine River Drainage. Texas: Sabine
River, Gladewater, Gregg Co. (Strecker).
Louisiana: Sabine River, Logansport, De
Soto Parish (MCZ).

Atchafalaya River System

Atchafalaya River Drainage. Louisi-
ana: Bayou Waukasha, 3 mi. S of Lebeau,
arm of Atchafalaya River, 10 mi. SE of
Lebeau {both MCZ); both St. Landry

Parish. Bavou Teche, St. Marv Parish

(USNM).

Lake Maurepas-Pontchartrain-
BoRGNE System

Louisiana: Baton Rouge, East Baton
Rouge Parish (MCZ).

Pearl River System

Pearl River Drainage. Mississippi:
Pearl River, Jackson, Hinds Co. (MCZ,

USNM, MZUM).

Pascagoula River System

Chickasawhay River Drainage. [Missis-
sippi: Chunkey Creek], Enterprise,
[Clarke Co.] (Simpson, 1914: 476, as En-
terprise, Alabama).

Mobile-Alabama-Coosa River Sys-
tem

Tombigbee River Drainage. Mississip-
pi: Tombigbee River, Columbus,
Lowndes Co. (MZUM).

Mobile River Drainage. Alabama: Mo-
bile River, 30 mi. N Mobile, Mobile Co.
(MCZ).

Lower Mississippi River Drainages

Tensas River System

Tensas River Drainge. Louisiana:
branch of Tensas River, 15 mi. W Tallu-
lah, Madison Co. (MZUM).

Big Black River System

Big Black River Drainage. Mississippi:
Big Black River (Hinkley, 1906, Nautilus
20: 53).

Yazoo River System

Yalabusha River Drainage. Mississip-
Yalabusha River, Grenada Co.

pi.

(MZUM).

118 Bulletin Museum of CU)iii})(tr(itiic Zoology, Vol. 1 U), So. 2

Ozarkian Region

Red Riveh System

Red River Drainage. Arkansas: Mercer
Bavou, N Black Diamond, Miller Co.
(MCZ). Louisiana: Red River (MZUM)
Caddo Lake, Chicot Bavou; Wallace Bay-
ou; {all MZUM) all Caddo Parish. Frier-
son (USNM); Bayou Pierre (MZUM;
CM); both De Soto Parish.

White River System

Black River Drainage. Missouri: Black
River, Hendrickson, Butler Co. (MCZ).

Cache River Drainage. Arkansas:
Cache River, Nemo, Craighead Co.
(MZUM).

St. Francis River System

St. Francis River drainage. Arkansas:
St. Francis River, Marked Tree, Poinsett

Co. (MZUM).

Meramec River System

Meramec River Drainage. Missouri:
Meramec River, Crawford Co. (Buchan-
an); Meramec River, Fern Glen, St. Louis
Co. (MZUM).

Bourbeuse River Drainage. Missouri:
Bourbeuse River, Franklin Co. (Buchan-
an).

Osage River System

Osage River Drainage. Kansas: Marais
des Cygnes River, 3 mi. E Ottawa, Frank-
lin Co. (Murray and Leonard). Missouri:
Marais des Cygnes River, Rich Hill
(MZUM), Papinsville (Utterback); both
Bates Co.; Grand River, Henry Co.
(MZUM); Grand River, Warsaw, Benton
Co.; Osage River {both Utterback).

Mississippian Region

Missouri River System

James River Drainage. South Dakota:
James River, near Frankfort, Spink Co.;
James River, Riverside, Hanson Co.
(both Coker and Southall).

Missouri River Drainage. Missouri:
Lake Contrary, St. Joseph, Buchanan Co.;
Platte River (both Utterback).

Mississippi River System

Big Muddy River Drainage. Illinois:
Big Muddy River, DeSoto, Jackson Co.;
Blairsville, Williamson Co.; Waltonville,
Jefferson Co. {all Mus. Nat. Hist., Univ.
Illinois, pers. comm.).

Kaskaskia River Drainage. Illinois:
Kaskaskia River, Washington Co. (MCZ).

Mississippi River Drainage. Illinois:
Mississippi River, Alton, Madison Co.
(MCZ).

Illinois River Drainage. Illinois: Illi-
nois River, 10 mi. below Hardin, Cal-
houn Co. (Bartsch; USNM). Illinois Riv-
er, Meredosia, Morgan Co. (MCZ).
Illinois River, Havana, Mason Co. (Par-
malee). Illinois River, Peru (MCZ); Utica
(Baker); both LaSalle Co.

Sangamon River Drainage. Illinois:
Sangamon River, Springfield, Sangamon
Co. (MCZ). Sangamon River, 4 mi. above
Mahomet, Champaign Co. (Baker, 1922:
143).

Mississippi River Drainage. Illinois:
Mississippi River, Quincy, Adams Co.
(MCZ). Mississippi River, Warsaw, Han-
cock Co. (MCZ). Iowa: Mississippi River
below Burlington, Des Moines Co.
(MCZ). Illinois: Benton Slough, Hender-
son Co. (MCZ).

Rock River Drainage. Illinois: Peca-
tonica River (MCZ).

Mississippi River Drainage. Illinois:
Mississippi River, Moline, Rock Island
Co. (MCZ). Iowa: Mississippi River,
Clinton, Clinton Co. (MCZ). Wisconsin:
Mississippi River, Prairie du Chien,
Crawford Co. (Baker). Mississippi River,
Pierce Co. (Grier).

Minnesota River Drainage. Minnesota:
Minnesota River, 2.8 mi. SW Blooming-
ton, Hennepin Co. (Stansbery, OSM).

Cumberlandian Region
Tennessee River System

Tennessee River Drainage. Tennessee:
Kentucky Lake (of Tennessee River) 6

Zoogeography of North Amehicax U\io^acea • Jolin.son 1 19

mi. above New Johnsonville, Benton Co.

(MCZ).

Ohioan Region
Ohio River System

Wabash River Drainage. Indiana: Wa-
bash River, Posey Co. (MCZ). Illinois:
Fox River, Richhind Co.; Little Wabash
River, Carmi, White Co.; (both MCZ).

Green River Drainage. Kentucky: Mud
River, Rectors Bridge, 3 mi. E Dunmore,
Muhlenberg Co.; Rough River, Dundee,
Ohio Co.; (both MCZ). Pond River
(MZUM).

Ohio River Drainage. Ohio: Ohio Riv-
er, Cincinnati, Hamilton Co. (Sterki,
1907: 393).

Genus Arkansia Ortmann and Walker

Arkansia Ortmann and Walker 1912, Nautilus 25:
97. Type species, Arkansia uJiecleri Ortmann
and Walker, original designation. This genus is
monotypic.

Arkansia wheeler! Ortmann and Walker

Plate 11, figure 1
Distribution: Plate 8 A

Arkansia wheeleri Ortmann and Walker 1912, Nau-
tilus 25: 98, pi. 8 (Old River [a bayou of the
Ouachita River], Arkadelphia, [Clark Co.], Arkan-
sas; holotype MZUM 105514) Simpson, 1914,
Cat. Naiades 1: 478; Ortmann, 1921, Nautilus 34:
141.

Description. Shell usually of mediiun
size, reaching 80 mm in length, though
occasionally exceeding 100 mm. Shell
subrotund to subovate or subrhomboidal.
Valves inflated, rather thick and solid.
Anterior end regularly rounded. Poste-
rior end somewhat truncate. Ventral mar-
gin regularly curved. Dorsal margin
curved, forming a distinct angle with the
obliquely descending posterior margin.
Hinge ligament short. Anterior margin
straight. Posterior ridge not prominent,
usually rounded, but sometimes slightly
biangulate. Umbos very prominent, pro-
jecting anteriorly and incurved over a
large lunule, their sculpture consisting of

two or three double-looped bars, the
loops slightly swollen or tubercular;
sculpture restricted to the extremity of
the umbos. Posterior half of the disk
sculptiued with irregular, oblique folds,
sometimes very faint, which curve up-
wards on the dorsal slope, and in front of
the posterior ridge are crossed by nu-
merous, irregular, radiating, small folds
or wrinkles at right angles to the lines of
growth; anterior portion of the disk
smooth. Periostracum with a silky luster,
dark reddish-brown or black, usually
lighter toward the umbos, which in
young shells are chestnut colored.

Left valve with two not strongly differ-
entiated pseudocardinal teeth coalescing
along the hinge line, the anterior one nar-
row and parallel with the hinge line, the
posterior one somewhat wider and heavi-
er, slightly separated from a strong pro-
jection of the interdentum, which is con-
tinuous with the lower lateral tooth and
slopes gradually to its extremity; upper
lateral low, the groove between deep, ex-
tending nearly to the umbo. Right valve
with a single, strong pseudocardinal
tooth, with a deep pit behind it to receive
the anterior pseudocardinal tooth of the
left valve, interdentum cut away to re-
ceive the interdental projection of the left
valve; one short strong, lateral tooth.
Muscle scars not much impressed. Nacre
usually salmon colored above the pallial
line, bluish-white, or entirely white be-
low, rather thin, slightly iridescent with
a wide, dark, prismatic border. Sexual
differences not marked.

Length Height Width

87 73 48 Old River, Arkadel-

phia, Clark Co., Arkansas.
Parat\'pe. Wheeler colln.

73.5 62 41 As above. Holot\pe

MZUM.

35 33 23 As above. Parat>pe.

Wheeler colln.

Anatomy and Breeding Season. The
anatomv was discussed bv Ortmann and
Walker' (1912, Nautilus 25: 97-99) and
on the basis of it, they recognized that
this species belonged to the subfamily

120 Hiilldiii Miiscuin of Coiiiixiniticc Zoolo^ij, \'()l. 149, .Vo. 2

Anodontinacv The species is l)raclvtictic
(Wheeler, 1918: 113).

Habitat. "Found in the shallow waters
both on the sand bars and muddy bot-
toms, but like other anodontine species
the\' prefer the oozy mud ot the river mar-
gins where there is little or no current"
(Wheeler, 1918: 113).

Remarks. Arkansia wheeleri Ortmann
and Walker is not very closely related to
any other species, except Arcidens con-
fra^osus (Say), a widely distributed
species. In general, their external ap-
pearance is somewhat similar. In both
the smaller series of radiating wrinkles
and the cui^ved folds on the dorsal slope
are very much alike. Both have strong,
oblique folds on the posterior half of the
disk, but those oi Arkansia more closely
resemble those of Amblema plicata
(Say). Arkansia has a heavier more in-
flated shell; the umbos are fuller, more
projecting, and more anterior. The um-
bonal sculpture of Arkansia is compara-
tively simple and confined to the upper
umbo, whereas that of Arcidens is devel-
oped into strong pitched up nodules, at
the base of the loops, and extends out
onto the disk. The lateral teeth are better
developed in Arkansia than in Arcidens,
but the presence of the interdental pro-
jection in the left valve and the cutting
away of the interdentum in the right
valve opposite it, relates Arkansia, Arci-
dens and Lasmigona.

The Wheeler collection is in the Ala-
bama Museum of Natural History, Uni-
versity of Alabama, University, Alabama,
but was not available for study.

Range. Found only in the Ozarkian Re-
gion: below the Ozark Crest in the upper
Red and Black river systems.

Abundance. "Has apparently never
been found in numbers. The only recent
record is from the Kiamichi River in
Oklahoma" (Stansbery, 1970: 18).

Specimens Examined
Red River System

Kiamichi River Drainage. Oklahoma:
Kiamichi River, Antlers, Pushmataha Co.

(CM); Kiamichi River, 8.5 mi. NE Hugo,
Choctaw Co. (USNM).

Little River Drainage. Arkansas: Little
River, White Cliffs, Little River Co.

(MZUM).

Black River System

Ouachita River Drainage. Arkansas:
Ouachita River and Old River, both Ar-
kadelphia, Clark Co. {both MCZ and

MZUM).

Tribe LAMPSILINI

Genus Actinonaias Crosse and Fishcher

ActinoiKiid.s Crosse and Fischer 1893, Recherches
Zooloyicjiies . . . de I'Anierique Centrale et du
Mexique, Pt. 7, 2: 600. Type species, Unio sa-
putalensis Lea, monotypic. Frierson, 1917, Nau-
tilus 31: 48. Ortmann, 1919, Mein. Carnegie Mus.
8: 232.

Ortmanniana Frierson 1927, Check listN American
naiades, pp. 11, 79. Type species, Unio carinatiis
Barnes, original designation.

Venustaconchu Frierson 1927, Check list N Amer-
ican naiades, pp. 11, 81, Errata et corrigenda.
Type species, Unio venustus Lea, original des-
ignation.

Actinonaias rafinesqueana (Frierson)
Plate 11, figures 2-4
Distribution: Plate 8 B

Lumpsilis rafinesqueana Frierson 1927, Check list
N American naiades, p. 69 ([Illinois River],
Moodvs, [Cherokee Co.], Oklahoma; holotvpe
MZUM 87576 figured bv Frierson, 1928, Nautilus
41: 1.38, pi. 1, figs. 1,2).

Actinonaias streckeri Valentine and Stansbery
1972, Sterkiana no. 42, p. .32, non Frierson 1927.

Description. Shell medium to large,
often exceeding 100 mm in length. Out-
line long, elliptical. Valves subinflated,
subsolid, slightly inequilateral. Anterior
end regularly rounded. Posterior end
ending in a point or biangulation below
the medial line. Ventral margin regularly
curved. Dorsal margin straight, forming
an angle with the obliquely descending
posterior margin. Hinge ligament long.
Posterior ridge generally indistinct and
faintly double. Umbos rather full, not
much projected above the hinge line,
their sculpture not observed, located con-

Z00GE()GR\PHV OK NoRTH AMERICAN UnIONACEA • /o/lMSO/l 121

sidenihly forward ol the middle of the
shell. Surface of the shell smooth, except
for wide growth rests. Periostraciim dirty
yellowish-green, or tawny, often marked
by faint, broad, broken, greenish rays.

Left valve with two triangular pseu-
docardinal teeth of about equal height,
and two lateral teeth. Right valve with a
strong tooth in front of the umbo, with a
smaller one before it. Interdentum nar-
row and flat. One lateral tooth. Anterior
and posterior muscle scars, and pallial
line, all well impressed. Nacre bluish-
white to creamy, with a wide prismatic
border.

Male shells are long elliptical, some-
what biangulate, posteriorly. Female
shells are more rounded and expanded
posteriorly, having a slight, wide marsu-
pial swelling. Both male and female
shells gape slightly at the anterior base.

Length Height Width
mm mm mm

104 62 35 White River, Cotter,

Baxter Co., Arkansas. Male.

86 60 35 [Illinois River], Moodys,

Cherokee Co., Oklahoma.
Holotype. Female.

Anatomy and Breeding Season. Gor-
don (pers. comm.) has studied the anat-
omy, and in a work in preparation, on the
basis of the "mantle flaps" will show that
this species belongs in the genus Lamp-
silis. If this is so, then arguing heuristi-
cally, so will A. pectorosa.

Remarks. Actinonaias rafinesqueana
(Frierson) bears a very close resemblance
to the Cumberlandian species, A. pecto-
rosa (Conrad). Females, of the former
species, are more fan shaped posteriorly,
and while both species generally have a
similar dirty yellowish green periostra-
cum, most specimens of rafinesqueana
are generally slightly rayed or rayless,
whereas specimens of pectorosa are al-
most always rayed. In both species the
rays, when present, are characteristically
broken.

In the Ozarkian Region, rafinesqueana
can be confused with A. ligamentina

(Lamarck), but the latter has a much
heavier shell, a stronger posterior ridge,
is more heavily rayed, and usually has a
more greenish or brownish periostracum.
A. rafinesqueana has also been confused
with Lampsilis powelli (Lea), but the lat-
ter shell is proportionally longer, brighter
yellow, and is without rays.

Range. Found only in the Ozarkian Re-
gion, below the Ozark Crest, in the upper
Arkansas and White river systems.

Abundance. Apparently never taken in
large numbers.

Specimens Examined
Arkansas River System

Verdigris River Drainage. Oklahoma:
Verdigris River (Gordon, pers. comm.).

Neosho River Drainage. Missouri:
Spring River [Jasper Co.]; Elk River,
[McDonald Co.] {both Nordstrom et al.
1977: 19); Indian Creek [of Elk River];
McDonald Co. (figured allotype, MZUM
[lost]).

Illinois River Drainage. Oklahoma:
[Illinois River], Moodvs, Cherokee Co.
(Holotype MZUM 87576; paratypes
MZUM 90665; paratype ANSP 145238).

White River System

White River Drainage. Arkansas: White
River, Monte Ne, Benton Co.; White Riv-
er, Cotter, Baxter Co. {both MCZ).

Black River Drainage. Arkansas: Black
River (MZUM).

Genus Lampsilis Rafinesque

Lampsilis Rafinesque 1820, Ann. Gen. des Sci. Phy-
siques (Bruxelles) 5: 298. Type species. Unto
ovatiis Say, subsequent designation. Herrmann-
sen, 1847, Indicis Generum Malacozoorum 1:
575. Frierson (1927: 11, 79) introduced under
Lampsilis, the subgenus Ortmanniana. Type
species, Unio carinatus Barnes, original desig-
nation. He included L. ahrupta and hi^^insi in
this subgenus. Ortmann (1919: 2.33) had already
placed carinatus [=lig,amentina (Lamarck)] un-
der Actinonaias Crosse and Fisher 1893 because
its anatomy was not that of a Lampsilis.

•">•■)

Bulletin MusciiDi of C.Dniixinitivi' 'Zoolo^ij, Vol. 149, No. 2

Lampsilis abrupta (Say)
Plate 12, figures 1, 2
Distribution: Plate 2

I'nio iihniptits S;i> 1831, Aint'iican Conch, no. 2
[Ii'tti'i iiress], pi. 17 (Wahash |I{i\er, New Har-
Mi()n\, P()st'\ ('()., Indiana), t\i)(' ANSP |l()st|,
teste Johnson antl Hakt'r, 1973, Proc. Atad. Nat.
Sci. Phila. 125: 146).

Ihiio cyctips Hafinescjne 18.31. Cont. Monot?. Bi-
\al\e Shells Hi\er Ohio, p. 2 (ri\er[s] Ohio and
Mississippi, type ANSP [lost], teste lohnson and
Raker, 1973, Proe. Acad. Nat. Sci. "Phiia. 125:
146).

Unio orhieulatus Lea 1836, Synopsis Naiades, p.
25, nan Hildreth 1827.

Vnio erassus Conrad 1836, Monography Unionidae,
no. 2, p. 34, pi. 16 (rivers of Ohio, Indiana and
Illinois, type .\NSP [lost], teste Johnson and Bak-
er, 1973,'Proc. Acad. Nat. Sci. Phiia. 125: 14,5)
non Say 1817.

Lampsilis orhienlata Simpson 1914, Cat. Naiades
1: 76. Coker, 1919, Bull. U. S. Bur. Fisheries for
1917-18, 36: 29, pi. 14, lower fig. Ortmann, 1919,
Mem. Carnegie Mus. 8: .320 [description partiui],
pi. 20, fig. 8, pi. 21, figs. 1, 2. Neel and Allen,
1964, Malacologia 1: 448, fig. 53. Parmalee, 1967,
Illinois State Mus. Pop. Sci. Ser. 8: 67, pi. 23 B.
La Rociiue, 1967, Ohio Geol. Sury. Bull. no. 62
(2): 217, fig. 103, fig. 104 distributional map
[jHirtiml All non Hildreth 1827.

L(U)ij)silis (Ortnianniana) abrupta (Say). Frierson,
1927, Check list N American naiades, p. 80. Haas,
1969, Das Tierreich, pt. 88, p. 461.

Lampsilis orhieulata forma orhieulata. Starrett,
1971, Illinois Nat. Hist. Surv. Bull. 30 (5): .338.

Description. Shell medium to large,
sometimes e.xceeding 100 mm in length.
Outline subelliptical, suhovate or
suhcjuadrate; gaping at the anterior base.
Valves somewhat inflated; e.xtremely
thick and heavy. Anterior end regularly
rounded; posterior end pointed in males,
abruptly truncated in females. Ventral
margin regularly rounded in males, al-
most straight in females. Dorsal margin
short and slightly curved, forming an in-
distinct angle with the obliquely de-
scending posterior margin. Hinge liga-
ment long and full. Posterior ridge
visible in males and yoiniger specimens,
indistinct in older females. Posterior
slope slightly convex or flattened, nar-
row. Umbos moderately inflated, project-
ing a little above the hinge line, their
sculptiue not distinct, more or less in-

clined forward, h)cated in front of the
middle of the shell. Surface of the shell
smooth, except for wide, low, concentric
ridges. The rest periods are often marked
by a sulcus. Periostracum subshiny or
dull, light or dark, yellow, yellowish
brown. Rays, seldom present, except in
younger shells; when visible they are
grayish green or brownish and are usu-
ally fine and widely spaced.

Left valve with two triangular pseu-
docardinal teeth of about eciual height,
two curved lateral ones. Right valve with
one high triangular pseudocardinal, with
a smaller low one before and behind it.
Interdentum narrow and flat. One strong
lateral tooth. Anterior and posterior ad-
ductor muscle scars, and pallial line, all
well impressed. Nacre generally silvery-
white, iridescent, sometimes salmon tint-
ed.

Male shells are subelliptical or sub-
ovate, the regularly curved ventral and
posterior margins meeting in a blunt
point mid-way from the base. Female
shells develop a post-basal swelling
which renders them posteriorly subtrim-
cate or subquadrate in outline.

Length Height Width

mm

mm

mm

103

75

61

Ohio Riyer, Portsmouth,
Scioto Co., Ohio. Male.

(CM).

93

65

43

Ohio Riyer. Male.

92

83

57

Ohio Riyer, Industry,
Beayer Co., Pennsylvania
Female. (CM).

73

68

42

Ohio Ri\er. Female.

Anatomy and Breeding Season. The
anatomy was discussed by Ortmann
(1912: 353) and Simpson (1914: 76). The
glochidia were figured by Ortmann
(1911, pi. 89, fig. 22) 0.19 X 0.21 and
0.20 X 0.25 mm. This species is bradvtic-
tic (Ortmann, 1919: 323).

Habitat. Found on riffles in the strong
currents of large rivers (Ortmann, 1919:
324).

Remarks. Lampsilis abrupta (Say) in

Zoogeography of North American Unionacea -yo/inson 123

the Cumherlandian and Ohioan regions
is easily recognized, except for young
shells, by the thickness of the shell,
which for its size is probably the heaviest
shell in those regions. Also characteristic
is its yellowish brown periostracum
which may have narrow, widely spaced
green rays. Young males resemble those
of Actinonaias ligamentina (Lamarck)
but the latter usually have wide green
rays which are close together. The female
of L. ahrupta does not resemble A. lig,-
amentina. In the Mississippian and the
Ohioan regions L. ahrupta may be mis-
taken for L. higginsi (Lea), see: Remarks.

Frierson (1924: 135) showed that Lea
(1836: 25) had mischievously misidenti-
fied Unio orhiculatus Hildreth 1827 as
Unio ahriiptus Say 1831. Hildreth's type
measured 100 mm in length, 100 mm in
height, and had a purple [orl violet nacre.
These data clearly indicate that orhicu-
latus is not ahruptus Say which is never
as long as wide, and does not have a pur-
ple nacre. In a sample of 74 specimens of
L. ahrupta. Ball (1922: 113) was unable
to reconcile Hildreth's measurements
with those of his specimens. Unio orhic-
ulatus Hildreth is Ohovaria retusa (La-
marck 1819).

Range. Cumherlandian Region: Ten-
nessee and Cumberland river systems.
Ohioan Region: Monongahela and Alle-
gheny rivers above maximum glaciation;
Ohio River proper; and the Wabash, Scio-
to, and Muskingum rivers; St. Lawrence
River System, western Lake Erie. Missis-
sippian Region: Illinois River drainage,
Illinois.

Ahundance. Apparently never taken in
large quantities. Ortmann (1919: 323)
found it most abundant in the Ohio River,
Beaver Co., Pennsylvania, and said that
he had not collected more than two doz-
en specimens below Pennsylvania.
Stansbery (1970: 19) said that "This
species is still taken occasionally in the
Tennessee River below Wilson and Gun-
tersville Dams, and rarely in the Muskin-
gum River, Ohio. "

Specimens Examined

Mississippian Region

Mississippi River System

Illinois River Drainage. Illinois: Illi-
nois River, 1 mi. below Hardin, Calhoun
Co. (Bartsch, USNM); Illinois River, Peo-
ria Lake, Chillicothe, {hoth Danglade),
hoth Peoria Co.; Illinois River, La Salle,
La Salle Co. (MZUM).

Ohioan Region

Ohio River System

Ohio River Drainage. Illinois: Ohio
River, Brookport, Massac Co. (Parmelee).
Kentucky: Ohio River, Paducah, Mc-
Cracken Co. (Bartsch, USNM).

Cumherlandian Region
Tennessee River System

Clinch River Drainage. Tennessee:
Clinch River, Offutt (CM); Clinch River,
.5 mi. below dam at Norris (MZUM);
hoth Anderson Co.; Clinch River, Sol-
way, Knox Co. (CM).

Holston River Drainage. Tennessee:
Holston River (MZUM).

French Broad River Drainage. Tennes-
see: French Broad River (MCZ).

Tennessee River Drainage. Tennessee:
Tennessee River, Washington Ferry,
Meigs Co. (MZUM).

Flint River Drainage. Tennessee: Flint
River (MZUM).

Tennessee River Drainage. Alahama:
Tennessee River, 11 mi. S Huntsville,
Madison Co. (MZUM).

Limestone Creek Drainage. Alahama:
Limestone Creek, The Points, Moores-
ville. Limestone Co. (MZUM).

Tennessee River Drainage. Alahama:
Tennessee River, Decatur, Morgan Co.
(MZUM); Tennessee River, Muscle
Shoals (MZUM), Florence (MCZ,
MZUM), hoth Lauderdale Co.; Tennes-
see River, Tuscumbia, Colbert Co.
(MCZ).

124 Bulletin Museum of Coniparative Zoology, Vol. 149, No. 2

Duck River Drainage: Tennessee:
Duck River, Columbia, Maury Co.
(MZUM).

Tennessee River Drainage. Tennessee:
Tennessee River, Trotter's Landing,
Humphreys Co. (MZUM). Tennessee
River, 1 mi. above Fort Henry, Steward
Co. (MZUM).

Cumberland River System

Cumberland River Drainage. Ken-
tucky: Cumberland River, Robertsport;
Monticello; (both MZUM) both Wayne
Co. Cumberland River, below Wolf
Creek Dam; Horse Shoe Bottom Ferry;
below Rowena; {all MZUM) all Russell
Co. Cumberland River, Burkesville;
Neely's Ferry, 4 mi. S Burkesville; (both
MZUM) both Cumberland Co.

Obey River Drainage. Tennessee: Obey
River, mouth of Jolly Creek, Picket Co.
(MZUM). Obey River, mouth of Horse
Creek, Clay Co. (MZUM).

Cumberland River Drainage. Tennes-
see: Cumberland River, Hartsville,
Trousdale Co. (Manning coll'n.). Nash-
ville, Davidson Co. (MCZ). Cumberland
River, Cowers Island [between Nashville
and Dover] (MZUM).

Ohioan Region
Ohio River System

Wabash River Drainage. Indiana: Wa-
bash River, Lafayette, Tippecanoe Co.
(MZUM). White River (Call, 1898: 493).

Ohio River Drainage. Indiana: Ohio
River, Vevay, Switzerland Co. (MZUM).
Ohio: Ohio River, Constance, Boone Co.
(MCZ). Ohio River, Cincinnati, Hamilton
Co. (MCZ; MZUM). Ohio River, below
New Richmond, Clermont Co. (MZUM).

Scioto River Drainage. Ohio: Scioto
River, Chillicothe, Ross Co. (MZUM).

Ohio River Drainage. Ohio: Ohio Riv-
er, Portsmouth, Scioto Co. (CM). Ohio
River, Portland, Meigs Co. (CM). West
Virginia: Ohio River, Parkersburg, Wood
Co. (CM). Ohio: Ohio River, Marietta,
Washington Co. (MCZ).

Muskingum River Drainage. Ohio:
Muskingum River (Stansbery, 1970: 19).

Ohio River Drainage. West Virginia:
Ohio River, St. Marys, Pleasants Co.
(CM). Ohio: Ohio River, Toronto (CM);
Shanghae [Knox Township] (MCZ); both
Jefferson Co. Pennsylvania: Ohio River,
Industry, Cooks Ferry, Shippingport; (all
CM) all Beaver Co.

Allegheny River Drainage. Pennsylva-
nia: Allegheny River, Kelly, Godfrey;
(both CM) both Armstrong Co.

Monongahela River Drainage. Penn-
sylvania: Monongahela River, Charleroi,
Washington Co. (CM).

St. Lawrence River System

Great Lakes Drainage (Lake Erie): Ni-
agara River, Buffalo, [Erie Co.l. The sin-
gle specimen collected by Miss Walker
in 1906, reported by Robertson and
Blakeslee (1948: 111), as in the Buffalo
Museum of Science, could not be located
(H. W. Charnley, pers. comm.). Since
Robertson and Blakeslee describe A. li-
garnentina, and the above specimen sep-
arately, the record is accepted as authen-
tic.

Lampsilis higginsi (Lea)
Plate 13, figures 1, 2
Distribution: Plate 2

Unio hifi^insii Lea 1857, Proc. Acad. Nat. Sci. Phila.

9: 84 ([Mississippi River], Muscatine, [Muscatine

Co.], Iowa); 1862, Jour. Acad. Nat. Sci. Phila. (2)

5: 188, pi. 24, fig. 258, figured holotype USNM

8482.3; 186.3, Ohs. Unio 9: 10.
Latnpsilis hi^^insii ^randis Simpson 1914, Cat.

Naiades 1: 78 (Illinois River, near Utica, [Lasalle

Co.], Illinois, type USNM [lost]).
Lampsilis hi^^.insii (Lea). Simpson 1914, Cat.

Naiades 1: 78. Utterback, 1916, Amer. Midi. Nat.

4: 445 [182], pi. 28, figs. 105 A and B. Coker,

1919, Bull. U. S. Bur. Fisheries for 1917-18, 36:

29, pi. 14, middle pair. Baker, 1928, Wisconsin

Ceol. and Nat. Hist. Surv., Bull. 70: (2): 293, pi.

95, Pannalee, 1967, Illinois State Mus. Pop. Sci.

Ser. 8: 67, pi. 23: A.
Lampsilis (Ortmanniana) hi^,fiinsii (Lea). Frierson,

1927, Check list N American naiades, p. 80. Haas,

1969, Das Tierreich, pt. 88, p. 461.
Lampsilis orhiculata hrmahi^ginsii (Lea). Starrett,

1971, Illinois Nat. Hist. Surv. Bull. 30 (5): 339.

Zoogeography of North American Vnionacea- Johnson 12=

Description. Shell niediuni to large,
sometimes exceeding 100 mm in length.'
Outline elliptical, oval, or rhomboid, gap-
ing at the anterior base. Valves much in-
flated, thick and heavy. Anterior end reg-
ularly rounded; posterior end somewhat
pointed in males; truncated or subtrun-
cated in females. Ventral and dorsal mar-
gins slightly curved, the latter forming an
indistinct angle with the obliquely de-
scending posterior margin. Hinge liga-
ment long and full. Posterior ridge rather
sharp in some males and younger speci-
mens, indistinct in older females. Poste-
rior slope slighdy convex or flattened,
narrow. Umbos swollen, inflated, consid-
erably raised above the hinge line, their
sculpture consisting of a few, slightly
looped, feeble ridges, located consider-
ably forward of the middle of the shell.
Surface of the shell smooth, except for
wide, low, concentric ridges. The rest
periods are often marked by a sulcus.
Periostracum usually dull and dark, yel-
lowish to olive. Rays frequendy present,
green and often wide.

Left valve with two triangular pseu-
docardinal teeth of about equal height,
two curved lateral ones. Right valve with
one high triangular pseudocardinal, with
a smaller low one before and behind it.
Interdentimi narrow and flat. One strong
lateral tooth. Anterior and posterior ad-
ductor muscle scars, and pallial line, all
well impressed. Nacre generally silvery-
white, iridescent, sometimes salmon tint-
ed, yellowish or orange.

Male shells are elliptical or oval, the
regularly curved, ventral dorsal margins
meeting, often imperceptibly, mid-way
from the base. Female shells develop a
post-basal swelling which renders them
posteriorly subtruncate and rhomboidal
in oudine. They also become greatly in-
flated.

Length Height Width
mm mm mm

9'^ 68 56 Mississippi River, Mercer

Co., Illinois. Male.
45 Illinois River, Illinois.
Male.

100

91

96

78

76

69

72

62

69

63

59

45

Illinois River, Meredosia,
Morgan Co., Illinois.
Female.

Mississippi River, Daven-
port, Scott Co., Iowa.
Female (MZUM).
Spring River, Lawrence
Co., Arkansas. Male.
(MZUM).

Old River, Arkadelphia,
Clark Co., Arkansas. Fe-
male (MZUM).

88

63

Anatomy and Breeding Season. The
anatomy of higginsi and abrupta have
not been separately discussed. A glochid-
ium of the fonner was figured by Surber
(1912: 9, pi. 2, fig. 23) which measured
0.21 X 0.26 mm. He also (1912: 7) indi-
cated that this species is bradytictic, and
(1913: 107) that Stizostedion canadense
(Smith) is [at least one of| its natural host
fish.

Habitat. Found on riffles in the strong
currents of large rivers.

Remarks. In the Mississippian region
Lampsilis higginsi can be confused with
Obovaria olivaria (Rafinesque), but the
latter has a more regularly ovate shell,
fine, rather than broad green rays, and
more especially it has only a slight inter-
dentum between the most posterior
pseudocardinal tooth and the lateral
teeth in the right valve and none in the
left valve. L. abrupta and higginsi over-
lap in their distribution in the Illinois
and lower Ohio rivers.

Simpson (1914: 76, 78) correctly de-
scribed L. abrupta and higginsi Ort-
mann (1919: 320) included both in his
description, and suggested there was
only one species, as did H. and A. van
der Schalie (1950: 456). An examination
of the specimens available to them re-
vealed misidentifications which led to
their erroneous conclusion. Morrison in
writing to Starrett (1971: 338) noted that,
"The geographic range of abruptus over-
laps that of higginsi in the region of the
Ohio-Mississippi confluence," and that
they were distinct species. Starrett (1971:
338), however, followed the classification

126 Bulletin Muscuiii of Comparative Z00I01J.1J, Vol. 149, No. 2

suggested by Stanshery whicli implies
they are a single species, since "forms"
have no status in modern nomenclature.
Stanshery and Kokai (1979: 57) later sug-
gested, "that there are several distinct
taxa within this [Lampsili.s orhiculata]
complex," without any further elucida-
tion.

L. hi^^insi differs from ahnipta in that
the former has umbos which are much
higher and fuller. The shells of both
sexes in L. higginsi are more inflated, the
females often being very bulbous, while
those o{ ah nipt us are rather flat sided. L.
higgin.si has a darker greenish periostra-
cum in contrast to the yellowish perios-
tracum o{ abrupt a. The rays of L. higgin-
si are broader and green rather than
grayish green or brownish as in abrupta.

Range. Ozarkian Region: below the
Ozark Crest in the Black and White river
systems; and above it in the Meramec
and Gasconade river systems. Mississip-
pian Region: Illinois River; Mississippi
River from Illinois north to Wisconsin.
Ohioan Region: lower Ohio River, Illi-
nois. Listed by Gordon, et al. (1980: 35)
as occurring in the Little River Drainage
of the Red River System, Arkansas. These
specimens, in the MZUM, were exam-
ined by the author and are Lampsilis sat-
ur (Lea), a species not included in their
list.

Abundance. "Is a rather uncommon
species, but a few may be found in almost
any carload of mixed shells" (Coker 1919:
29). "Is known living today only in the
upper Mississippi River" (Stanshery
1970: 19). Listed as a rare and endan-
gered species by Stanshery (1971: 15). In
the area between Pairie du Chien, Craw-
ford Co., Wisconsin and Harpers Ferry,
Alamakee Co., Iowa, Mathiak (1979: 57,
pi. 3, 10 C) collected 40 live specimens
in five days during 1975. He said that,
"in any calculation taking into account
the number of tons of shells being pro-
cessed [by fisheries] daily, the number of
Higgin's Eye mussels I found in a few
hours would lead to the conclusion that
hundreds had actually been harvested in

1975." He further noted, "no one work-
ing with the endangered species program
took the opportunity to search for the
Higgin's Eye or other rare mussels dur-
ing [1974-1975] when all the species
were being harvested from the Prairie du
Chien area." The collecting by Fuller
(1978: 3) offers further evidence that L.
higginsi is probably about as abundant in
the upper Mississippi River as it was ear-
ly in the century. It woidd appear that
this species was placed on the endan-
gered list by over enthusiastic bureau-
crats. This would be of little importance
if it did not lead directly to the unnec-
essary harassment of commercial fisher-
men.

Specimens Examined
Ozarkian Region
Black River System

Ouachita River Drainage. Arkansas:
Ouachita River (MZUM), Old River
(MZUM, CM), both Arkadelphia, Clark
Co.

Saline River Drainage. Arkansas: Sa-
bine [sic] = Saline River (Stanshery,
1970: 19).

White River System

White River Drainage. Arkansas: North
Fork White River, nr. Norfork, Baxter Co.
(MCZ).

Black River Drainage. Arkansas: Black
River, above Pocahontas, Randolph Co.
(MZUM, CM): Black River, Black Rock,
Lawrence Co. (MCZ, MZUM, USNM).

Spring River Drainage. Spring River
[Lawrence Co.] (MZUM).

Meramec River System

Meramec River Drainage. Missouri:
Meramec River, Crawford, Franklin, and
St. Louis Cos. (Buchanan).

Gasconade River System

Gasconade River Drainage. Missouri:
Gasconade River (Stanshery, 1970: 19).

Zoogeography of North American llNioNACEA-yo/inso/j 127

Mississippian Region

Missouri River System

Missouri River Drainage. Ne])r(isk(i:
Nemeha River (Aughey, 1877: 702).

Mississippi River System

Illinois River Drainage. Illinois: Illi-
nois River, 1 mi. below Hardin (Bartsch,
USNM; MZUM); Kampsville (Dan-
glade); both Calhoun Co. Illinois River,
Florence, Pike Co. (Danglade). Illinois
River, Meredosia, Morgan Co. (MCZ). Il-
linois River, Beardstown, Cass Co.
(Danglade). Illinois River, Bath (Dan-
glade); Havana (MZUM); both Mason
Co. Illinois River, Peoria (Danglade; Par-
malee); Chillicothe (Danglade); ])oth
Peoria Co. Illinois River, Marquette, Bu-
reau Co. (Shimek, USNM). Illinois River,
nr. Utica, La Salle Co. (Simpson). Illinois
River, Morris, Grundy Co. (Shimek;
USNM). Kankakee River (MZUM).

Mississippi River Drainage. Missouri:
Mississippi River, Louisiana, Pike Co.
(Utterback). Mississippi River, Hannibal,
Marion Co. (Utterback). Iowa: Mississip-
pi River, Montrose, Lee Co. (Bartsch,
USNM). Iowa River, Iowa City, Johnson
Co. (Shimek, USNM). Mississippi River,
Muscatine (USNM); Fairport (Morrison,
USNM); Montpelier (Bartsch, USNM);
all Muscatine Co. Mississippi River, Buf-
falo (MZUM); Davenport (MCZ); Prince-
ton (Bartsch, USNM); all Scott Co.,
Crooked Slough, 1 mi. N Bellevue, Jack-
son Co. (Bartsch, USNM). Mississippi
River, Dubuque, Dubuque Co. (Sharpe,
USNM). Wisconsi7i: Wisconsin River,
Arena, Iowa Co. (MCZ). Mississippi Riv-
er, Prairie du Chien (Stansbery, Ohio
State Museum 38489, 21 fresh speci-
mens, 1976); Lynxville (Bartsch USNM);
both Crawford Co. Mississippi River,
Victorv; Genoa; Stoddard; (all Bartsch,
USNM); all Vernon Co. Black River, be-
tween R. R. Ijridge and La Crosse, La
Crosse Co. (Bartsch, USNM). Minnesota:
Mississippi River Dresback, Winona Co.
(Bartsch, USNM). Wisconsin: Mississip-

pi River, Buffalo, Buffalo Co. (Grier).
Minnesota: Lake Pepin, Red Wing,
Goodhue Co. (MCZ; USNM). Wisconsin:
Mississippi River, above Prescott, Pierce
Co. (Bartsch, USNM). St. Croix River,
Hudson, St. Croix Co. (MZUM).

Ohioan Region

Ohio River System

Ohio River Drainage. Illinois: Ohio
River, Hillerman, Massac Co. (Bartsch,

USNM).

Lampsilis powelli (Lea)
Plate 15, figures 3, 4
Distribution: Plate 8 B

Unio poweUii Lea 1852, Trans. Amer. Philos. Soc.
10: 270, pi. 19, fitr. 25 (Saline River, Arkansas;
figured holotvpe USNM 85042); 1852, Obs. Unio
5: 26.

Lampsilis poweUii (Lea). Simpson, 1914, Cat.
Naiades 1: 75. Scammon, 1906, Kansas Univ. Sci.
Bull. 3: 288. Utterback, 1916, Amer. Midi. Nat.
4: 446 [183], pi. 28, fig. 103 A, B [figured as L.
hiteola], Black River, Williamsville, Wavne Co.,
Missouri, MCZ 271445. Isely, 1925, Proc. Okla-
homa Acad. Sci. 4: 111.

Description. Shell generally of medi-
um size, occasionally large, exceeding
100 mm in length. Outline elliptical or
long obovate. Valves subinflated, rather
thin to subsolid. Anterior end regularly
rounded. Posterior end somewhat point-
ed. Ventral margin regularly curved, or
straight. Dorsal margin straight forming
a barely perceptable angle with the
obliquely descending posterior margin.
Hinge ligament long. Posterior ridge in-
distinct. Umbos moderately full, slightly
projected above the hinge line, their
sculpture consisting of double looped
corrugations, located considerably for-
ward of the middle of the shell. Surface
of the shell smooth except for growth
rests. Periostracum generally shining, ol-
ive-brown, tawny, and always without
rays.

Left valve with two triangular pseu-
docardinal teeth (the anterior one higher),
and with two lateral teeth. Right valve

128

Bulletin Museum oj Coinparutiic Zoolo'^iii, Vol. 149, So. 2

with one strong pseuclocarcliiial tooth in
front of tlie nniho and with a smaller one
before it. Interdentnm very narrow. One
lateral tooth. Anterior adduetor niuscle
scars well impressed, posterior ones less
so. Pallial line visible anteriorly. Nacre
bluish white and iridescent.

Male shells are feebly biangnlate pos-
teriorly, near the medial line. Female
shells are wider and ronnder posteriorly.

Lenjith Height Width
mm mm mm

80 56 29 Saline River, Arkansas.

Holotype. Male.

77 39 24 As above. Allotype. Fe-

male.

Anatomy and Breeding Season. Un-
known.

Remarks. Lampsilis powelli (Lea)
bears a very close resemblance to the
Cumberlandian species, L. virescens
(Lea) which is found only in the Tennes-
see River system, mainly in the Paint
Rock River drainage. Shells of the former
are always without rays, while those of
virescens usually have green rays above
the sometimes marked posterior ridge. In
the Ozarkian Region, powelli has been
confused with L. radiata siliquoidea, hij-
diana, teres, and Villosa reeviana. With
the exception of L. teres, which is some-
times without rays, all of these other
species are rayed to some degree. L.
powelli differs from teres, in that the lat-
ter is more elongate, and pointed, much
more inflated, and the females exhibit
more sexual dimorphism.

Range. Found only in the Ozarkian Re-
gion, below the Ozark Crest, in the Sa-
line River drainage of the Black River
system; the upper Arkansas, and upper
White River systems.

Abundance. Apparently never taken in
large numbers.

Specimens Examined

Black River System

Saline River Drainage. Arkansas: Sa-
line River, Benton, Saline Co. (MCZ,

MZUM).

Arkansas River System

Neosho River Drainage. Kansas: Neo-
sho River, Oswego (Scammon, 1906:
288); Chetopa (Isely, 1925, Table 2); both
Labette Co. Spring River, Baxter Springs,
Cherokee Co. (MCZ). Missouri: Elk Riv-
er, McDonald Co. (Utterback, 1916: 183).

Illinois River Drainage. Oklahoma: Il-
linois River, Moodys, Cherokee Co. (Ise-
ly, 1925: 67).

White River System

White River Drainage. Missouri: White
River (Utterback, 1917, Table).

Black River Drainage. Missouri: Black
River, Williamsville, Wayne Co. (MCZ).

Genus Proptera Rafinesque

Potamilis Rafinesque 1818, Amer. Monthly Mag.
and Critical Rev. [New York] 3: 355. Type
species, Unio alatus Say, monotypic. In 1818, 23
of the 24 taxa listed under Potamilis were nomina
niida. Morrison (1969: 24) correctly stated that no.
12, P. alatus is Unio alatus Say 1817. Say had
sent Rafinesque a copy of his article from the first
edition of Nicholson's Encyclopedia (Johnson,
1975: 265).

Proptera Rafinesque 1819, Jour. Phys. Chim. Hist.
Nat. [Paris] 88: 426. Type species, LJnio alatus
Say. Subsequent designation, Herrmannsen,
1847, Indicis Generiun Malacozoonun 2: 41, un-
der Metaptera Rafinesque, an objective synonym
of Proptera. Metaptera Rafinesque 1820, Ann.
Gen. Sci. Physiques, Bruxelles 5: 299. Rafines-
que deemed Proptera inappropriate. Potamilis
was overlooked from 1818 until its availability
was indicated by Morrison (1969: 24). It was
adopted by Valentine and Stansbery (1971: 25),
and has since been promulgated by the latter.
Proptera has been in general usage since 1900
(Simpson, 566). As no question of priority of au-
thorship is involved, the resurrection of Potamilis
appears nugatorv. The most recent revision of the
Rules (1974, Bull. Zool. Nomencl. 31 (2): 80) un-
der Article 23, states: "A zoologist who considers
that the application of the Law of Priority would
in his judgment disturb stability or iniiversality
or cause confusion is to maintain the existing
usage and must refer the case to the Commission
for a decision under the plenary powers [Art.
79]." This author maintains existing usage, and
suggests that those who would promulgate Po-
tamilus seek the riding.

Proptera capax (Green)
Plate 16, figures 1, 2
Distribution: Plate 3

Unio capax Green 1832, Cabinet of Nat. Hist, and
Amer. Rural Sports 2: 290 ([Mississippi River],

Zoogeography of North Amerigax Unionacea • Johiison 129

Falls of St. Anthony [Minneapolis, Hennepin Co.,
Minnesota], type [lost]; Bayou Teche [error]).

StimpJn/nota ^lohosa Lea 1832, Trans. Amer. Phi-
"los. Soc. 5: 41, pi. 4, fitr. 12 (River Ohio, 150 miles
below Louisville [Kentuckv], fi,u;ured type [lost];
syntypes ANSP 56609); 1834, Ohs. Unio 1: 153.
Lea (1852, Synopsis of Naiades, third edit., p. 27)
acknowledged that capax had priority over ^lo-
hosa.

Lampsilis capax (Green). Smith, 1899, Bull. U. S.
Bur. Fisheries for 1898, pi. 84. Simpson 1914, Cat.
Naiades 1: 76; Coker, 1919, Bull. U. S. Bur. Fish-
eries for 1917-18, 36: 31, pi. 17, lower fig.

Proptera capax (Green). Ortmann, 1914, Nautilus
28: 67. Utterhack, 1916, Amer. Midi. Nat. 4: 394
[162], pi. 26, figs. 93 A and B. Frierson, 1927,
Check list N American naiades, p. 87. Baker,
1928, Wisconsin Geol. and Nat. Hist. Sur., Bull.
70 (2): 248, pi. 84, figs. 1-4. Parmalee, 1967, Il-
linois State Mus. Pop. Sci. Ser. 8: 83, pi. 28: B.
La Rocque, 1967, Ohio Geol. Surv. Bull. no. 62
(2): 260, fig. 149, fig. 150 distributional map. Star-
rett, 1971, Illinois Nat. Hist. Surv. Bull. 30 (5):
330.

Description. "The valves of this shell
are much more convex or globose than
any of the uniones which I have seen,
and as they are quite thin compared with
most of the western species, the cavity in
which the animal is lodged is exceeding-
ly capacious — hence its name. The ante-
rior end is broad, rounded, and slightly
angular near the hinge; the posterior mar-
gin is very narrow, and also rounded;
these valves do not close perfectly on
each other, but gape at the opposite mar-
gins; this is more remarkable in old than
in young individuals. The epidermis is
smooth, yellowish, and frequently cloud-
ed with brown. The nacre is bluish
white, and often very beautifully irides-
cent. The beaks are recurved over the
tegument. The teeth resemble very much
those of the U. ovatits of Mr. Say, but
they are much thinner. These characters,
I think, will be sufficient to distinguish
the Unio capax from every other shell."
The original description is quoted since
it is quite unavailable, and because it is
lucid.

Length Height Width
mm mm mm

128 90 74 Wabash River, New

Harmony. Posey Co.,
Indiana.
92 77 61 As above.

Anatomy and Breeding Season. The
anatomy was discussed by Ortmann
(1914, Nautilus 24: 67). The "axe-head"
glochidiimi was figured bv Coker and
Surber (1911: 179, pi. 1, fig. 4, 4a). Re-
ported gravid in June, July, August, and
October (Surber, 1912: 7 and Ortmann,
loc. cit.). Duration of the season not in-
dicated.

Habitat. "Has been taken on both a
sand and mud bottom, in flowing water,
and at depths of only a few inches to
eight feet or more" (Parmalee, 1967:83).

Remarks. Proptera capax (Green) may
be easily recognized by its thin, globose
shell with inflated umbos, and by its peri-
ostracum which is always rayless. It may
be confused with Lampsilis ovata (Say)
but the latter has a heavy shell, is not as
globose, is generally rayed, and exhibits
sexual dimorphism, which capax does
not.

Coker and Surber (1911: 20) indicated
that capax was not a Lampsilis, but a
Proptera, where it was placed by Ort-
mann (1914, Nautilus 24: 67).

Range. Found in the Ozarkian Region,
only below the Ozark Crest, in the White
and St. Francis river systems. Mississip-
pian Region: Illinois River; Mississippi
River from Illinois north to Minnesota.
Ohioan Region; Wabash River; St. Law-
rence River system, western Lake Erie
and eastern Lake Ontario drainages.

Reported from the Arkansas River sys-
tem, Neosho River, Lyon Co., Kansas, by
Murray (1962, Nautilus 75: 95). The sin-
gle specimen found was figured by Mur-
ray and Leonard (1962: 132, pi. 36) and
is a female of Proptera purpurata (La-
marck), with a heavy shell and character-
istic purple nacre. The single male shell
[no longer available] reported from the
Verdigris River, Nowata Co., Oklahoma,
by Branson (1963: 510), as capax, was
probably also P. purpurata. P. capax
does not exhibit sexual dimorphism.

Abundance. "Most numerous in the
lower Ohio and Wabash rivers — espe-
cially in the latter" (Parmalee, 1967: 83).
"Largely, if not entirely, gone from the

130 Bulletin Miiscuin oj CDmfxirdtiic '/a)iAo<1[i, Vol. 149, \(>. 2

entire Ohio Rixer clrainaj^e. This species
still survives in the White and St. Francis
rivers of Arkansas" (Stansi)ery, 1970: 18).

Specimens Examined

Ozarkian Region

White River System

White River Drainage. Arkcnisas: White
River, 1 mi. NE Devall Bluff, Prairie Co.

(MZUM).

St. Francis River System

St. Francis River Drainage. Arkansas:
St. Francis River, Parkin (MCZ; MZUM);
Wittshurg(Call); both Cross Co. St. Fran-
cis River, Madison, St. Francis Co.
(MCZ; CM).

Mississippian Region

Missouri River System

Missouri River Drainage. Nebraska:
Blue River; Elkhorn River {both Aughey,
1877: 702).

Mississippi River System

Mississippi River Drainage. Illinois:
Mississippi River, Alton, Madison Co.
(MCZ; MZUM).

Illinois River Drainage. Illinois: Illi-
nois River, Grafton, Jersey Co. (Dan-
glade). Illinois River, Kampsville, Cal-
houn Co. (Danglade). Illinois River,
Pearl; Florence {both Danglade); both
Pike Co. Illinois River, Beardstown, Cass
Co. (Danglade). Sangamon River (Dan-
glade). Illinois River, Havana, Mason Co.
(Danglade). Spoon River, Fulton Co.
(MZUM). Illinois River, Pekin, Tazewell
Co. (Danglade). Illinois River, La Salle
Co. (Calkins).

Mississippi River Drainage. Illinois:
Mississippi River, Hamburg, Calhoun
Co. (MCZ). Missouri: Mississippi River,
Hannibal, Marion Co. (MCZ). Mississip-
pi River, La Grange, Lewis Co. (Utter-
back). Illinois: Mississippi River, Ham-

ilton (MZUM), Warsaw (MCZ); both
Hancock Co. Iowa: Des Moines River,
Keokuk, Lee Co. (MZUM). Mississippi
River, Burlington, Des Moines Co.
(MCZ). Illinois: Mississippi River, Mer-
cer Co. (MCZ). Illinois: Mississippi Riv-
er, Martins Landing, Rock Island Co.
(MZUM). Iowa: Mississippi River,
McGregor, Clayton Co. (MZUM). Wis-
consin: Mississippi River, Alma, Buffalo
Co. (MZUM). Minnesota: Mississippi
River, Falls of St. Anthony, [Minneapolis,
Hennepin Co.] (Green).

Ohio River System

Ohio River Drainage. Illinois: Ohio
River, Hillerman, Massac Co. (USNM).
Ohio River, Shawneetown, Gallatin Co.
(MCZ).

Wabash River Drainage. Illinois: Wa-
bash River, 2 mi. W Maunie, White Co.
(Parmalee). Indiana: Wabash River, New
Harmony, Posey Co. (MCZ). Illinois:
Wabash River, Gravville, White Co.
(USNM). Indiana: White River, Hazle-
ton, Gibson Co. (MZUM). West Fork,
White River, Strawtown, Hamilton Co.
(MZUM).

St. Lawrence River System

Great Lakes Drainage (Lake Erie).
New York: Niagara River, Buffalo, Erie
Co. (Lake Ontario). New York: Wilson's
Creek [= Twelvemile Creek, near
Wilson], Niagara Co. These records (Rob-
ertson and Blakeslee 1948: 106) are
based on specimens in the Buffalo Mu-
seum of Science, which have been ex-
amined and are authentic.

Genus Villosa Frierson

Micromyd Agassiz 1852, Archiv fiir Naturge-
schiclitf, 18 (1): 47. Type species, Unio lapillus
Say, subsequent designation, Herrniannsen, 1852,
Indicis Generum Malacozooruni, Supp. et Corr.,
p. 83, nan Mivromija Rondani 1840 (Insecta).

Villosa Frierson 1927, Check list N American
naiades, pp. 11, 80. Type species, Unio villosus
Wright, original designation.

Zoogeography of North American Unionacea'/o/jmsom 131

Villosa arkansasensis (Lea)
Plate 16, figures 3, 4
Distribution: Plate 8 A

Unio arkansasensis Lea 1862, Proc. Acad. Nat. Sci.

Phila. 14: 169 ([Ouachita River] near Hot

Springs, [Garland Co.], Arkansas); 1862, Jour.

Acad. Nat. Sci. Phila. (2) 5: 206, pi. 30, fig. 275,

figured holotvpe USNM 25710; Ohs. Unio 9: 28.

Call, 1895, Trans. Acad. Sci. St. Louis 7: 6.
Lampsilis arkansasensis (Lea). Simpson, 1900,

Proc. U. S. Natl. Mus. 22: 557; 1914, Cat. Naiades

1: 1.30.
Eurynia {Microniya) arkansasensis (Lea). Ort-

mann, 1916, Nautilus 30: 54. Wheeler, 1918,

Nautilus 31: 118.

Description. Shell small in size, sel-
dom exceeding 40 mm in length. Outline
ovate, or short obovate. Valves subinflat-
ed, solid. Anterior end regularly round-
ed; posterior end somewhat pointed.
Ventral margin regularly curved. Dorsal
margin rather rounded, forming a more or
less distinct angle with the obliquely de-
scending posterior margin. Posterior
ridge, low, almost double, ending a slight
biangulation below the medial line.
Hinge ligament short. Umbos not much
swollen, slightly projected above the
hinge line, the sculpture not observed,
located in the anterior third of the shell.
Surface of the shell with irregularly con-
centrical striations. Periostracum dull to
subshiny, yellowish to brownish, usually
with fine green rays over the entire sur-
face.

Left valve with two rather heavy, tri-
angular, pseudocardinal teeth of about
equal size, and two short lateral teeth.
Right valve with two pseudocardinals,
the posterior one chunky, the more an-
terior tooth vestigial. No interdentum.
One lateral tooth. Anterior adductor mus-
cle scars well impressed, posterior ones
faint. Pallial line visible anteriorly. Nacre
silvery white, bluish, iridescent poste-
riorly.

Male shells are somewhat pointed. Fe-
male shells are more broadly roimded,
and truncated below the medial line. In
mature females there is a distinct "con-
striction" in the middle of this truncation.

Length Height Width
mm mm mm

40 28 18 [Ouachita River], Hot

Springs, [Garland Co.],
Arkansas. Holotvpe. Male.

38 27 18 Saline River, Benton,

Saline Co., Arkansas.
Female.

Anatomy and Breeding Season. The
anatomy was discussed by Ortmann
(1916, Nautilus 30: 54).

Remarks. Villosa arkansasensis (Lea)
is most closely related to the Cumber-
landian species, V. vannxemensis vaniix-
emensis (Lea). Shells of the former are
more swollen, less elongate, and the um-
bos are located more anteriorly. Further,
in females the enlarged and truncated
posterior end is more evenly rounded,
and not as produced as it is V. v. vannx-
emensis. The periostracum of arkansas-
ensis is usually yellowish to brownish
and rayed. The nacre is always white.
The periostracum of vannxemensis is
from dirty tawny through olive to nearly
black, and is not always rayed. The nacre
is very rarely white. It is usually dirty
purplish-white, through salmon-tinted to
dark purple. The most important similar-
ity between the two species is the char-
acteristic "constriction" on the posterior
truncation of the female shells.

Range. Found only in the Ozarkian Re-
gion, below the Ozark Crest, in the upper
Black River System.

Abundance. Never taken in large num-
bers.

Specimens Examined
Black River System

Ouachita River Drainage. Arkansas:
Ouachita River, 3 mi. S.E. Pencil Bluff,
Montgomery Co. (MCZ); Ouachita River,
near Hot Springs, Garland Co. (USNM);
Ouachita River, S Arkadelphia, Clark Co.
(Wheeler).

Little Missouri River Drainage. Arkan-
sas: Little Missouri River, Murfreesboro,
Pike Co. (MZUM).

132 Bulletin Museum of Conifxirutivc 7A)()l(><iti, Vol. I4^j, .Vo. 2

Caddo Ri\er Drainage. Arkansas: Cad-
do River (MZUM).

Saline River Drainage. Arkansas: Sa-
line River, Benton (MCZ, MZUM); 3.5
mi. SE Traskwood (MCZ); both Saline
Co.

Villosa reeviana (Lea)
Plate 17, figures 1-5
Distribution: Plate 7 B

Unio reevianus Lea 1852, Trans. Ainer. Philos. Soc.
10: 272, pi. 20, fig. 28 (Alexandria, Louisiana
[erroneous]; holotype USNM 85025 male; para-
type MCZ 178918 male, labeled, "Arkansas. Lea
so decided from this specimen and another, Pow-
ell."; allotype USNM 85025a, labeled, "White
River, Arkansas." The type locality is here re-
stricted to: White River, Elkins, Washington Co.,
Arkansas); 1852, Obs. Unio 5: 28.

Unio brciicuht.s Call 1887, Proc. U. S. Natl, fvlus.
10: 499, pi. 28 (Currant [Current] River, Shannon
Co., Missouri and in Jacks Fork and Big Creek,
tributaries to it; lectotype selected by Johnson,
1975, Occ. Papers on Moll. 4 (54): 141, MCZ
.5020, male specimen portrayed in figs. 1, la, lb,
from Jacks Fork [Shannon Co.]; figured allotype
MCZ 5023 portrayed in figs. 2, 2a, 2b, from Big
Creek [Fork], Texas Co.); 1895, Trans. Acad. Sci.
St. Louis 7: 6, pi. 17.

Lampsilis brittsi Simpson 1900, Proc. Acad. Nat.
Sci. Phila. 52: 76, pi. 5, figs. 1-2 (no locality [Ni-
anugua River, Camden Co., Missouri]; holotype
USNM 152700).

Lampsilis reeviana (Lea). Simpson, 1914, Cat.
Naiades 1: 74. Utterback, 1916, Amer. Midi. Nat.
4: 449 [186]. Frierson, 1927, Check list N Amer-
ican naiades, p. 73.

Lampsilis hrevieula (Call). Simpson, 1914, Cat.
Naiades, 1: .57. Ortmann, 1918, Nautilus 32: 14.

Eurynia iMieronu/a) hrevieula (Call). Utterback,
1916, Amer. Midi. Nat. 4: 434 [171], pi. 27, figs.
98 A-D.

Lampsilis reeviana hrevieula (Call). Frierson 1927,
Check list N American naiades, p. 7.3.

Lampsilis hreviculus hrittsi Simpson. 1900. Proc.
U. S. Natl. Mus. 22: 5.33. 1914, Cat. Naiades 1:
58.

Eunpiia (Mieronii/a) hrevieula hrittsi (Simpson).
Utterback, 1916, Amer. Midi. Nat. 4: 4.35 [172],
pi. 27, figs. 99 A, B.

Description. Shell of medium .size, not
exceeding 80 mm in length. Outline
ovate, elliptical or obovate. Valves sub-
inflated, rather thin to subsolid, inequi-
lateral. Anterior end regularly roimded,
posterior end more broadly rounded.
Ventral margin slightly curved. Dorsal
margin somewhat curved, forming a bare-

ly perceptible angle with the ()l)li(iuely
descending posterior margin. Hinge lig-
ament is long, passing under the umbos
and appearing anteriorly in front of a
small lunule. Posterior ridge indistinct.
Umbos are not very full, but slightly el-
evated above the hinge line, their scidp-
ture not observed, located very anterior-
ly. Surface of the shell is smooth.
Periostracum is generally shiny or sub-
shiny, greenish straw color, having a
smoky tint, or yellowish horn-color, with
dark green rays, generally over the entire
surface. The rays are usually delicate,
and generally broken.

Left valve with two delicate, subcom-
pressed pseudocardinal teeth. These are
located anterior of the umbo, with the
more anterior tooth being the higher. Of
the two lateral teeth present, the inner
tooth is the higher. Right valve with a
somewhat compressed pseudocardinal
tooth, with a smaller one above it sepa-
rated by a deep, parallel-sided pit, often
with the vestige of another tooth behind.
Interdentum is very narrow and curved.
One lateral tooth is truncate posteriorly.
Muscle scars and umbonal cavities are
shallow, pallial line only visible anterior-
ly. Nacre is whitish, pinkish, bluish
white.

Male shells are quite elliptical, and
faintly biangulate posteriorly. Female
shells are shorter and higher than those
of the male and strongly inflated post-ba-
sally. The low posterior ridge ends in a
decided point about three-fifths of the
way up from the base. The post-l)asal
margin appears truncated. This trunca-
tion may be slightly to considerably
emarginate.

Length Height Width
mm mm mm
75 41 28 [Arkansas] Holotype of

U. reevianus. Male.
62 39 23 White River, Arkansas.

Allotype of L'. reevianus.
Female.

64 40 24 Jacks Fork, Current River,

Shannon Co., Missouri.
Lectotvpe of U. hrevieulus.
Male. '

Zoogeography of North American UNiONACEA'7o/in.son 133

58 40 25 Big Fork, Current River,

Texas Co., Missouri. Allo-
type ofU. hrevicithis.
Female.

Anatomy and Breeding Season. The
anatomy was discussed bv Call (1887,
Proc. U. S. Nad. Miis. 10:' 499) and by
Ortmann (1918, Nautilus 32: 15), on the
basis of which, the latter placed U. brev-
iciilus under Lampsilis. A glochidium
was figured by Surber (1915, pi. 1, fig. 14)
as L. b. brittsi, 0.23 x 0.29. Breeding sea-
son not recorded.

Habitat. Found by Call in shallow
mountain streams, "being swift and lim-
pid, with rocky bottoms."

Remarks. Villosa reeviana (Lea), as
pointed out by Call, does not much re-
semble any other Ozarkian species.
There is some variation within the sev-
eral populations as to the thickness of the
shell, width of the rays and degree of sex-
ual dimorphism in the female. The male
shell of V. reeviana has been confused
with specimens of V. vibex (Conrad). The
latter has a consistently thinner shell,
broader green rays, and the female differs
from the male only in that the posterior
end of the shell is more broadly rounded,
the post-basal margin is not truncated.
The specimens mentioned by Frierson
(1927: 74) from Onion Creek, Travis Co.,
Texas MZUM 79921 are vibex.

Range. Found only in the Ozarkian Re-
gion: below the Ozark Crest in the White
River system; and above it in the Mera-
mec. Gasconade and Osage River sys-
tems.

Abundance. Taken in large numbers
from several localities in the Meramec
River system by the Harvard-Ohio State
Museum Expedition in 1965.

Specimens Examined

White River System

White River Drainage. Arkansas: Main
Fork White River, Elkins, Washington
Co. (MCZ). Missouri: James River, Ca-
lena. Stone Co. (MCZ, MZUM); White
River, Hollister and Forsyth {both

MZUM), both Taney Co. Arkansas:
White River, Cotter (MZUM) and Nor-
folk (MCZ), both Baxter Co.

Big Buffalo Fork Drainage. Arkansas:
Big Buffalo Fork, near Gilbert, Searcy
Co.; Big Buffalo River, Buffalo River
State Park, Marion Co.; {both MZUM).

Current River Drainage. Missouri: Big
Creek, Texas Co.; Jacks Fork; Current
River between Eminence and Ellsinore;
both Shannon Co.; Current River, Big
Spring State Park, Carter Co.; Current
River, 1 mi. SW Doniphan, Ripley Co.;
Eleven Point River, Riverton, Oregon
Co.; {all MCZ).

Black River Drainage. Missouri: Mill
Creek, 1 mi. E Lesterville, Reynolds Co.
(MZUM).

Spring River Drainage. Arkansas:
Spring River, Salem, Fulton Co. (MCZ).

Meramec River System

Big River Drainage. Missouri: Cedar
Creek (MZUM), Mineral Fork, both
Washington Co.; Big River, Jefferson Co.
{both Buchanan, 1977, p. 31).

Meramec River Drainage. Missouri:
Huzzah Creek, Crawford Co. (MZUM);
Meramec River, Dent Co.; Meramec Riv-
er, Steelville, Crawford Co. {both MCZ);
Meramec River, at Franklin and St. Lew-
is Co. line (Buchanan, 1977, p. 31).

Bourbeuse River Drainage: Missouri:
Bourbeuse River, Franklin Co. (Buchan-
an, 1977, p. 31).

Gasconade River System

Gasconade River Drainage. Missouri:
Big Piney River, 7 mi. W Licking, Texas
Co.; Gasconade River, 6 mi. NW Vichy,
Maries Co. {both MCZ).

Osage River System

Osage River Drainage. Missouri: Little
Niangua River (MCZ, MZUM); Niangua
River, Hahatonka (Utterback); both Cam-
den Co.

134 Bulletin Musi'iini of CUniiixinitiiv 'Aoolo^y, Vol. 149, No. 2

Genus Cyprogenia Agassiz

Cypro^ciiid At^assi/. 1852, Aicliiv fiir Natuijjt>-
scliichtf, 18 (1): 47. Type s]ieties, lUiio irronitiis
Lea, original desij^nation. Ortinaiin, 1912, Aim.
Carneunr Mus. 8: 312.

Cyprogenia aberti (Conrad)
Plate 19, figure 3
Plate 20, figures 1-4
Distribution: Plate 6

Unio (iherti Conrad 1850, Proc. Acad. Nat. Sci. Phi-
la. 5: 10 (Rapid.s of Verdigris River, Chambers'
Ford [not located, Oklahoma] Arkansas [River
system]; 1854, Jonr. Acad. Nat. Sci. Phila. (2) 2:
295, pi. 26, fig. 1, figured type not located by
Johnson and Baker, 1973, Proc. Acad. Nat. Sci.
Phila. 125: 147. Call, 1895, Trans. Acad. Sci. St.
Louis 7: 4.

Unio himarckiatius Lea 1852, Trans. Amer. Philos.
Soc. 10: 266, pi. 17, fig. 20 (Caddo River; Washita
[Ouachita] River, near the Hot Springs, [Garland
Co.]; both Arkansas; figured holotype USNM
84306, labeled, "White River, Arkansas."); 1852,
Obs. Unio 5: 22.

Unio popenoi Call 1885, Bull. Washburn College
Lab. Nat. Hist. 1: 49, pi. 2 (Fall River, Wilson
Co., figured holotype MCZ 4934; Verdigris River;
both Kansas); is Unio aberti Conrad, teste Call,
1887, Amer. Nat. 2 1 : 860.

Cyprogenia aberti (Conrad). Scammon, 1906, Kan-
sas Univ. Sci. Bull. 3: 315, pi. 71, fig. 2. Simpson,
1914, Cat. Naiades 1: 328. Utterback, 1916, Amer.
Midi. Nat. 4: 322 [134], pi. 25, figs. 83 A, B. Iselv,
1925, Proc. Oklahoma Acad. Sci. 4: 103. Murray
and Leonard, 1962, Univ. Kansas, Mus. Nat. Hist.,
Pub. no. 28, p. 105, pi. 27, figs. 1^, text fig. 23.
Branson, 1966, Sterkiana no. 23, p. 7.

Cyprogenia aberti laiuarckiana (Lea). Simpson,
1914, Cat. Naiades, 1: 329. Utterback, 1916,
Amer. Midi. Nat. 4: 323 [135].

Unio irroratus Call. 1895, Trans. Acad. Sci. St.
Louis 7: 20, non Lea 1828.

Description. Shell of medium size, .sel-
dom exceeding 80 mm in length. Outline
usually rhomboidal, occasionally round-
ed triangular. Valves slightly inflated,
thick and heavy. Anterior end regularly
rounded; posterior end subtrimcate. Ven-
tral margin more or less roiuided, tending
to be straight or slightly concave poste-
riorly. Dorsal margin slightly convex
forming a more or less distinct angle with
the almost vertical, sometimes concave,
posterior margin. Hinge ligament rather
long. Posterior ridge well developed,

high and narrow at the umbonal region,
becoming flattened and somewhat dou-
ble, with a wide shallow groove before it.
Posterior slope with a radial furrow, very
narrow. Umbos low and compressed,
curved forward over the hnnile, located
at the middle of the shell or somewhat
anteriorly. Surface of the shell with
strong, low, more or less numerous, con-
centric ridges; in addition it is sometimes
subvertically wrinkled and nodulous.
Periostracum scarcely shining, yellowish
green, with dots and flecks of dark green
which fall into broad broken rays. Hinge
very broad and flat. Left valve with two
heavy, triangular, blunt and ragged,
pseudocardinal teeth of about equal
height; also with two short, very low lat-
eral teeth. Right valve with one large
pseudocardinal, sometimes with a feeble
one on each side of it. Interdentum is
very broad. One low double or triple lat-
eral tooth. Umbonal cavities are very
shallow and compressed. Anterior and
posterior adductor muscle scars and pal-
lial line are all well impressed. Nacre is
white, thinner and somewhat iridescent
anteriorly.

Length Height Width
mm mm mm

80 75 44 Fall River, Wilson Co.,

Kansas. Paratype oi Unio
popenoi.

48 41 26 Ouachita River, Arkadel-

phia, Clark Co., Arkansas.

Anatomy and Breeding, Season. Un-
known.

Habitat. Reported from Kansas as liv-
ing in rocky, gravel or soft mud substrates
(Murray and Leonard, 1962: 107).

Remarks. Cyprogenia aberti (Conrad)
of the Ozarkian Region is readily distin-
guishable from all other unionids there
by the peculiar, mottled color of the peri-
ostracum; its generally compressed, oval
or rhoml)oidal shell; and by the subvert-
ical wrinkles or nodules which occur at
least on the upper part of the disk. While
most specimens of aberti are oval or
rhomboidal, some specimens become

Z(K)c.p:ogr\phy of North American Unionacea -./o/i/i.son 135

rounded triangular, much resembling C.
stc^aria of the Cumherlandian and
Ohioan regions. Call (1895: 20) in a col-
lection made in the St. Francis River,
Wittshurg, Cross Co., Arkansas, noted
that many of the young appeared to he
(ihcrti, hut that, "the triangular outline is
lost with age and the circular form he-
comes more and more marked." This lot,
MCZ 5492, now consists of only two
adult specimens. He also mentioned
specimens from the Saline River of the
Black River system, now lost. He regard-
ed both of these lots as heing irroratus =
stegaria. Frierson (1927: 66) mentioned
that stegaria and ahcrti "almost merge
into one unbroken chain in the state of
Arkansas." Having studied virtually all of
the available material, it is admitted that
occasional specimens of both aberti and
stegaria closely resemble one another,
but the shell of the former species always
has a much narrower, and more com-
pressed posterior slope.

Range. Found only in the Ozarkian Re-
gion: below the Ozark Crest in the Black,
Arkansas, White, and St. Francis river
systems; and above it in the Meramec
River system.

Abundance. Formerly abundant in the
Fall River, Kansas. Murray and Leonard
(1962: 107) claimed that this species has
not been collected in Kansas since 1906.
Collected in some numbers by the Har-
vard-Ohio State Museum Expedition,
1965, in the Ouachita River, of the Black
River system, near Pencil Bluff, Mont-
gomery Co., Arkansas.

Specimens Examined

Ozarkian Region

Black River System

Ouachita River Drainage. Arkansas:
Ouachita River, 3 mi. SE Pencil Bluff,
Montgomery Co. (MCZ); Ouachita River,
nr. Hot Springs, Carland Co. (Lea);
Ouachita River, N Cove Creek, Hot
Springs Co. (CM); Caddo Creek (USNM);

Ouachita River, Arkadelphia, Clark Co.
(MCZ).

Saline River Drainage. Arkansas: Sa-
line River, Benton, Saline Co. (MCZ).

Arkansas River System

Verdigris River Drainage. Kansas: Fall
River, 3.5 mi. S Fredonia, Wilson Co.
(MZUM); Verdigris River (MCZ). Okla-
homa: Verdigris River, Oologa (USNM),
Catoosa, NW Inola (both Isely), all Rog-
ers Co. Verdigris River, Chambers Ford
[not located] (Conrad).

Neosho River Drainage. Kansas: Neo-
sho River, Lyon Co. (Murray and Leon-
ard); Spring River, 2 mi. E Pittsburg
(Branson).

White River System

White River Drainage. Arkansas: White
River, Elkins, Washington Co. (MCZ).
Missouri: White River, Hollister (Utter-
back), Forsyth (MZUM), both Taney Co.
Arkansas: White River, Cotter, Norfork
(both MCZ), both Baxter Co.

Big Buffalo Fork Drainage. Arkansas:
Big Buffalo Fork, Newton Co. (Meek and
Clark); Big Buffalo Fork, 2.5 mi. below
Red Cloud Mine, Marion Co. (MZUM,
from Meek and Clark, Station 25).

Little Red River Drainage. Arkansas:
Little Red River, Clinton, Van Buren Co.
(MCZ).

Black River Drainage. Missouri: Black
River, .5 mi. W Hendrickson (MCZ), Pop-
lar Bluff (MZUM), both Buder Co. Ar-
kansas: Black River, Pocahontas, Ran-
dolph Co. (MCZ).

Spring River Drainage. Arkansas:
South Fork Spring River nr. Salem, Ful-
ton Co. (MZUM); Spring River, Hardy,
Sharp Co.; Spring River, Black Rock,
Lawrence Co. {both MCZ).

St. Francis River System

St. Francis River Drainage. Missouri:
Little St. Francis River, 9 mi. E Ironton,
Madison Co. (MZUM); St. Francis River,
Greenville, Wayne Co. (MZUM); St.

136 Bulletin Museum oj Coniputativc Zoolomi, Vol. 149, No. 2

sas: Tyronza Hi\er, Tyroiiza, Poinsett Co.
(MZUM); St. Francis River, Wittsburj^,
Cross Co. (MCZ).

Mehamec River System

Merainec River Drainage. Missouri:
Meraniec Ri\er, Crawford Co. (see Plate
20 fig. 4)

Cyprogenia stegaria (Rafinesque)
Plate 18, figures 1, 2
Plate 19, figures 1, 2
Distribution: Plate 6

Ohovcirid stegaria Rafinesque 1820, Ann. Gen. Sci.
Physiciues, Bruxelles 5: 308 (I'Ohio [River]; lee-
totype ANSP 20215, selected by Johnson and
Baker, 1973, Proc. Acad. Nat. Sci. Phila. 125: 173,
pi. 4, fiK^ 3).

Unio verrucosus alhus Hildreth 1828, Anier. Jour.
Sci. 14: 281 (Muskingum [River, near Marietta,
Washin^iton Co., Ohio]; type [lost]).

Vnio irroratus Lea 1828, Trans. Amer. Philos. Soc.
3: 269, pi. 5, fijjr. 5 (Ohio; fisured holotvpe USNM
84637 [lost]. Reeve, 1864, Conch. Iconica 16:
Unio, pi. 12, species 44).

Cypro^cnid irrorata pusilla Simpson 1900, Proc. U.
S. Natl. Mus. 22: 610 (Green River [Mammoth
Cave, Edmonson Co.] Kentucky; lectotype USNM
152004, selected by Johnson, 1975, Special Occ.
Pub., Dept. Moll., Mus. Comp. Zool. no. 4: 17;
pl. 2, fig. 2).

Unio stegarius (Rafinesque). Say, 1834, Amer.
Conch, no. 6 [no pagination] Unio irroratus Lea
is listed as a synonym. Conrad, 18.38, Monogra-
phy Unionidae of North America, no. 10, p. 83,
pl. 46, fig. 1. Reeve, 1864, Conch. Iconica 16:
Unio, pl. 12, species 45.

Cyprogenia stegaria (Rafinesque). Ortmann, 1919,
Mem. Carnegie Mus. 8: 218, pl. 13, fig. 5. Frier-
son, 1927, Check list N American naiades, p. 66.

Cyprogenia irrorata (Lea). Simpson, 1900, Proc. U.
S. Nat. xMus. 22: 610. Ortmann, 1912, Ann. Car-
negie Mus. 8: 312, fig. 17. Simpson, 1914, Cat.
Naiades 1: 327. Ortmann and Walker, 1922, Occ.
Papers, Mus. Zool., Univ. Michigan no. 112, p.
43. La Rocciue, 1967, Geol. Surv. Ohio, Bull. No.
62 (2): 232, figs. 118, 119. Parmalee, 1967, Illinois
State Mus., Popular Sci. Ser. 8: 61, pl. 20: A,
Frontispiece.

Description. Shell of medium size, sel-
dom e.xceeding 80 mm in length. Outline
suhcircular, suhtrapezoidal, or roiuided
triangular. Valves somewhat inflated,
thick and heavy. Anterior end broadly
rounded; posterior end subtruncate. Ven-

tral margin more or less rounded, tending
to be straight or slightly concave poste-
riorly. Dorsal margin slightly convex,
fornu'ng an angle with the almost vertical
posterior margin. Hinge ligament rather
long. Posterior ridge well developed,
high and narrowly rounded at the um-
bonal region; with a radial groove ante-
rior to it and another shallower one pos-
terior to it. Posterior slope slightly
concave or flattened, rather broad. Um-
bos moderately high and full, but flat-
tened, curved forward over the lunule,
their sculpture consisting of a few slight-
ly double-looped bars, located at the mid-
dle of the shell or somewhat anteriorly.
Surface of the shell with strong, low,
more or less numerous concentric ridges.
The anterior end may be smooth, but
there are usually nodules elsewhere
where they assume a radial pattern. The
nodules may be missing in older speci-
mens. Periostracum light green or yel-
low, to light brown, ornamented with
green mottlings, which fall into fine or
broad green rays.

Hinge well developed. Left valve with
two heavy, triangular, blunt and ragged,
pseudocardinal teeth of about equal
height, two short heavy lateral ones.
Right valve with one large pseudocardi-
nal. Interdentum very broad and short.
One strong double lateral tooth. Umbo-
nal cavities shallow, and compressed.
Anterior and posterior adductor muscle
scars and pallial line all well impressed.
Nacre silvery white, seldom bluish or
pinkish.

Length Height Width
mm mm mm

64 64 44 Tuscarawas Ri\er, New

Philadelphia, Tuscarawas
Co., Ohio.

49 49 35 Clinch River, Clinton,

Anderst)n Co., Tennessee.

Anatomij and Breeding Season. Ort-
mann (1919: 219) discussed the sexeral
anatomical descriptions. A glochidium
was figured bv Ortmann (1912: pl. 19, fig.
6) and by Suri3er (1912: pl. 1, fig. 11). Ac-

ZOOCEOGRAPHV OF NoRTH AMERICAN Unionacea • Johnson 137

cording to Ortiiiann it measured 0.18 x
0.15. He also (1919: 219) indicated that
this species is hradytictic.

Hahitdt. Found in firmly packed grav-
el, in strongly flowing water in the Clinch
and Holston rivers and in mussel beds on
gravelly bottoms with steady currents in
the deep channel of the Ohio River be-
tween Cincinnati and Pittsburgh (Ort-
mann, 1919: 220).

Remarks. In the Cumberlandian and
Ohioan regions, Cijpro^enia stc^aria
(Rafinesque) is distinguishable from all
other unionids by the peculiar, mottled
character of the color of the periostracum,
and by its generally subglobular shape
and nodular surface, although the shape
and development of the nodules is vari-
able. Occasional specimens resemble C.
aberti (Conrad) of the Ozarkian Region
(see Remarks under this section).

Specimens from the Cumberlandian
Region are, in general, not as large as
those from the Ohioan Region. C. stega-
ria appears to have found a more favor-
able environment in the Ohioan Region
in post-glacial time, as apparently did
Plagiohi flexuosa (Rafinesque) (Johnson
1978: 284).

As indicated in the synonymy, Rafines-
que's name for this species has been used
by Say, Conrad, Ortmann, Frierson and
others. Ortmann and Walker (1922: 43)
argued that this species could not be
identified from Rafinesque's description
alone. Frierson (1927: 66) indicated that
it could be so identified. The present
Rules of International Nomenclature
(1964) emphasize the identification of the
type, so Rafinesque's name is clearly
available.

Range. Cumberlandian Region: Ten-
nessee and Cumberland river systems.
Ohioan Region: Ohio River and its trib-
utaries.

Abundance. This species appears to be
locally abundant.

Specimens Examined
Cumberlandian Region

Tennessee River System

Clinch River Drainage. Virginia:
Clinch River, Clinchport, above mouth of
Copper Creek, 1.5 mi. below Speers Fer-
ry bridge, all Scott Co.; Clinch River, 1.5
mi. S Dona, Lee Co. (all MCZ). Tennes-
see: Clinch River, below Kvles Ford
bridge, Hancock Co. (MCZ); Clinch Riv-
er, Clinch River Station, Claiborne Co.
(CM); Clinch River, 4 mi. NW Thorn
Hill, Grainger Co.; Clinch River, Clinton
and Edgemoor, both Anderson Co.;
Clinch River, 3 mi. S Wheat, Roane Co.
(all MCZ).

Holston River Drainage. Tennessee:
Holston River, Turley Mill, Grainger Co.
(CM); Holston River, Mascot and near
Knoxville, both Knox Co. (both MCZ).
Tennessee River Drainage. Tennessee:
Tennessee River, Knoxville, Knox Co.
(MCZ); Tennessee River, Rathburn,
Hamilton Co. (MZUM). Alabama: Ten-
nessee River, Bridgeport, Jackson Co.
(MCZ).

Flint River Drainage. Alabama: Flint
River [Madison Co.] (MZUM).

Tennessee River Drainage. Alabama:
Tennessee River, Florence, Lauderdale
Co.; Tennessee River, Tuscumbia, Col-
bert Co. {both MCZ); Pickwick Basin,
mounds between Barton, Colbert Co. and
Waterloo, Lauderdale Co. (Morrison).

Duck River Drainage. Tennessee: Duck
River, Columbia, Maury Co. (Hinkley
and Marsh, single specimen).

Cumberland River System

Cumberland River Drainage. Ken-
tucky: Cumberland River, below Wolf
Creek Dam, Russell Co. (MCZ); Cum-
berland River, Albany Landing; Cloyds
Landing; both Cumberland Co.; (both
CM).

Obey River Drainage. Tennessee: Obey
River, Duncan Ford, 4 mi. SE Lilvdale,
Pickett Co. (MCZ).

Cumberland River Drainage. Tennes-
see: Cumberland River, Goodall Island,
Smith Co. (Wilson and Clark); Cumber-

138 Bulletin Museum of ('oinixirdtive yAiolo^j,//, Vol. 140, \o. 2

Francis River, Butler Co. (MCZ). Arkan-
land River, Nashville, Davidson Co.

(MCZ).

Ohioan Region
Ohio River System

Wabash Ri\'er Drainat^e. Indiana: Wa-
bash River, New Harniony, Posey Co.;
Wabash River, Mt. Carmel, Wabash Co.;
Wabash River, Terre Haute, Vigo Co.;
Wai)ash River, Lafayette, Tippecanoe
Co.; Wabash River, Delphi, Carroll Co.;
Wabash River, Tippecanoe River, both
White Co.; {all MCZ).

White River Drainage. Indiana: White
River (MCZ).

Green River Drainage. Kentiickij:
Green River, Rio, and Munfordville, both
Hart Co. {both MCZ); Green River,
Mammoth Cave (MCZ); Onyx Cave
(CM); both Edmonson Co.; Barren River,
Bowling Green, Warren Co.; Green Riv-
er, below dam at Rochester, Muhlenberg
Co. {both MCZ).

Salt River Drainage. Kentucky: Rolling
Fork, Salt River, .5 mi. S New Haven,
Nelson Co. (MCZ).

Kentucky River Drainage. Kentucky:
Kentucky River (Danglade).

Licking River Drainage. Kentucky:
Licking River, Butler, Pendleton Co.
(MCZ).

Ohio River Drainage. Ohio: Great
Miami River (Sterki); Ohio River, Cincin-
nati, Hamilton Co. (MCZ); Litde Miami
River, Xenia, Green Co. (CM); Scioto
River (MCZ); Ohio River, Portsmouth,
Scioto Co. (CM): Virginia: Big Sandy
River (MCZ). Ohio: Ohio River, Port-
land, Meigs Co. (CM). West Virg,inia:
Ohio River, Parkersburg, Wood Co.
(CM).

Muskingum River Drainage. Ohio:
Muskingum River, Lowell, Washington
Co.; Tuscarawas River, New Philadel-
phia, Tuscarawas Co.; {both MCZ).

Ohio River Drainage. West Virginia:
Ohio River, St. Marys, Pleasants Co.
(CM); Ohio River, Clarington, Monroe
Co. (MCZ). Pennsylvania: Ohio River,
Industry, Beaver Co. (CM).

Big Beaver River Drainage. Ohio: Ma-
honing River (Sterki).

Allegheny River Drainage. Pennsylva-
nia: Allegheny River, Godfrey and Al la-
din, both Armstrong Co. {both CM); Al-
legheny River, Natrona, Allegheny Co.
(CM).

Monongahela River Drainage. Penn-
sylvania: IMonongahela River] Indian
kitchen-midden opposite Point Marion,
Fayette Co. (CM).

Genus Ptychiobranchius Simpson

Ptychobrdiuhiis Simpson 1900, Proc. Acad. Nat.
Sci. Phila. 52: 79. Type species, Uuio phaseolus
Hildretli, oritfinal desiy;nation. Oitiiiaiiii, 1912,
Ann. Carnegie Mus. 8: 305.

Subtentus Frierson 1927, Check list N American
naiades, pp. 10, 65. T\pe species, Unio stihfentus
Say, original designation. Frierson created tliis
monotypic subgenus tor P. stihtcntiDU (Say), a
Cumherlandian species, prohahK' liecause of the
corrugations on its posterior slope. Its use seems
unnecessary in a genus consisting of 4 species. In
addition to P. suhtcntum., there is fasciohire (Ra-
ftnescjue) from the Ciuiilierland and (Jhioan Re-
gions; occidentalis (Conrad) from the Ozarkian
Region; and f^reeni (Conrad) which is restricted
to tlie Mohile-Ahihania-Coosa River system.

Ptychobranchus occidentalis (Conrad)
Plate 20, figures 5, 6
Distribution: Plate 7 A

Unio occidentalis Conrad 1836, Monography
Lhiionidae, no. 7, p. 64, pi. 36, fig. 1 (Current Riv-
er [Randolph Co.], Arkansas, figured type not lo-
cated, teste, Johnson and Baker, 1973, Proc. Acad.
Nat. Sci. Phila. 125: 163).

Ptt/chohi'dnchtis clintoncnsis Simpson 1900, Proc.
Acad. Nat. Sci. Phila. 52: 79, pi. 5, fig. 3 (Archies
Fork of Little Red Ri\er, near Clinton [\'an Buren
Co.], Arkansas; holotvpe USNM 124615); 1900,
Proc. U. S. Natl. Mus. 22: 613; 1914, Cat. Naiades
1: 335. Iselv, 1925, Proc. Oklahoma Acad. Sci. 4:
102.

Ellipsaria clintoncnsis (Simpson). Utterhack, 1916,
Amer. Midi. Nat. 4: 316 [128], pi. 25, figs. 81 A, B.

Ptychobranchus occidentalis (Conrad). Valentine
and Stansbery 1971, Sterkiana, no. 42, p. 23.

Ptychobranchus jyhascolus Scammon. 1906, Univ.
Kansas Sci. Bull. 3: 319, pi. 72, fig. 2. Isely, 1925,
Proc. Oklahoma Acad. Sci. 4: 102; non Hildreth
1828.

Ptychobranchus fasciohire Murray and Leonard.
1962, Univ. Kansas, Mus. Nat. Hist., Pub. no. 28,
p. 162, pi. 45, fig. 1, non Rafinesqtie 1820.

Zoogeography of North American UNiONACEA'yo^n.soM 139

Description. Shell medium to large,
occasionally exceeding 100 mm in length.
Outline oblong, ovate, elliptical or sub-
rhomboid, sometimes subtriangular.
Valves subinflated, very solid. Anterior
end regularly rounded. Posterior end
more broadly rounded. Ventral margin
straight or slightly curved. Dorsal margin
curved imperceptably joining the
obliquely descending posterior margin.
Hinge ligament long. Posterior ridge usu-
ally indistinct, rounded, often somewhat
double, ending in a slight biangulation
near the base. Umbos not inflated, or
much raised above the hinge line, their
sculpture not seen, located considerably
forward of the middle of the shell. Sur-
face of the shell smooth. Periostracum
yellowish-green, brownish, with very
delicate, wavy, continuous green rays,
often over the entire surface.

Left valve with two low, small, solid,
pseudocardinal teeth, and two low lateral
teeth. Right valve with one large tooth
directly under the umbo, with a vestigial
one on either side of it. Interdentum
short but very wide. One lateral tooth
which is sometimes more or less double.
Umbonal cavities very shallow; anterior
and posterior muscle scars well im-
pressed. Pallial line impressed anterior-
ly. Nacre bluish white, often with lurid
blotches.

Length Height Width
mm mm mm

116 67 40 Verdigris River, Neodesha,

Wilson Co., Kansas.

91 45 27 Archies Fork, of Little

Red Ri\er, nr. Clinton, \'an
Buren Co., Arkansas.
Paratvpe of P. clintonensis
MCZ 192371.

Anatomy and Breeding Season. Dis-
cussed in part under P. pJiaseolus (Hil-
dreth), by Ortmann (1912: 306). P. occi-
dentalis is bradytictic.

Remarks. Ftijchohranchus occidental-
is (Conrad) is restricted to the Ozarkian
Region. It is most closely related to P.
fasciolare (Rafinesque) of the Cumber-
landian and Ohioan regions. Valentine

and Stansbery (1971: 23) claimed to have
specimens of fasciolare from Missouri,
and while their specimens were not ex-
amined, all other specimens studied from
Missouri proved to be occidentalis.
While occidentalis and fasciolare are
very similar, most of the rays on the shell
of the former are always very fine, uni-
form, and generally continuous, while
those of the latter are always wide, and
often broken into spots.

In the Ozarkian Region occidentalis
has been confused with Elliptio dilatata
(Rafinesque). It does not much resemble
the latter which has a longer, less high
shell, often with an arcuate ventral mar-
gin. It is seldom rayed, and usually has
a violet nacre.

Range. Foimd only in the Ozarkian Re-
gion: below the Ozark Crest in the Red,
Black, Arkansas, White, and St. Francis
river systems; and above it in the Mera-
mec River system.

Abundance. The size of the lots exam-
ined tend to indicate this species is rather
abundant.

Specimens Examined
Red River System

Blue River Drainage. Oklahoma: Blue
River, Milburn, Johnson Co.; Blue River,
Durant, Bryan Co.; {both Valentine and
Stansbery, 1971: 23).

Kiamichi River Drainage. Oklahoma:
Kiamichi River, Tuskahoma, Pushmataka
Co. (MZUM).

Little River Drainage. Oklahoma:
Glover Creek, Glover, McCurtain Co.
(MCZ). Arkansas: Cossatot River, 5.5 mi.
ESE Gillham; Rolling Fork River, S Gill-
ham; both Sevier Co. (both MCZ).

Black River System

Ouachita River Drainage. Arkansas:
Ouachita River, 3 mi. SE Pencil Bluff,
Montgomery Co.; Ouachita River, Arka-
delphia, Clark Co. {both MCZ); Ouachita
River, Camden, Ouachita Co. (xMZUM).

Caddo Creek Drainage. Arkansas:

140 Hullitiii Muscitiii of Coiiiixiidtivc Zoolo^ij, Vol. 149, So. 2

Caddo Creek, Caddo Ciap, Montgomery
Co.; Caddo Creek, 4 mi. above mouth,
Caddo Valley, Clark Co.; {both MCZ).
Saline River Drainage. Arkdusas: Sa-
line Ri\er, Benton, Saline Co.; Saline
River, lenkins Ferry, Grant Co.; {hoth
MZUM).

Meramec River System

Big River Drainage. Missotiri: Mineral
Fork, Washington Co. (Buchanan, 1977,
fig. 48).

Meramec River Drainage. Missouri:
Meramec River, Meramec State Park,
Franklin Co. (MCZ).

Arkansas River System

Verdigris River Drainage. Kansas: Fall
River, Elk Co.; Verdigris River, Neode-
sha, Wilson Co.; {both MCZ).

Neosho River Drainage. Kansas: Neo-
sho River, lola, Allen Co. (MCZ, MZUM).
Missouri: Spring River, Carthage (MCZ)
and Webb City (Utterback); both Jasper
Co. Kansas: Spring River, Cherokee Co.
(Branson, 1966, Nautilus 80: 22).

White River System

White River Drainage. Arkansas: White
River, Sulphur City and N Elkins {both
MZUM), both Washington Co. Missouri:
James River, Galena, Stone Co. (MCZ);
White River, Hollister (Utterback), For-
syth (MZUM); both Taney Co. Arkansas:
White River, Cotter (MCZ) and Norfolk
(MZUM), both Baxter Co.; Archie's Fork,
Little Red River, near Clinton, Van Bur-
en Co. (MCZ).

Black River Drainage. Missouri: Black
River, Poplar Bluff, Butler Co. (MCZ).

Spring River Drainage. Arkansas:
South Fork Spring River, near Salem,
Fulton Co.; Spring River, Hardy and
Willford; both Sharp Co.; Spring River,
Black Rock, Lawrence Co.; {all MZUM).

Current River Drainage. Missouri: Big
Creek, Texas Co. (MCZ, MZUM); Jacks
Fork, Shannon Co. (Utterback); Current
River, Big Spring State Park, Carter Co.
(MZUM)J Current River, 1 mi SW Doni-
phan, Ripley Co. (MCZ).

St. Francis River System

Little St. Francis River Drainage. Mis-
souri: Little St. Francis River, 9 mi. E
Ironton, Madison Co. (MZUM).

LITERATURE CITED

Ahlstedt, S. a. and S. R. Brown. 1980. The
naiad fauna of the Powell River in Virginia and
Tennessee. Bull. Amer. Mai. Union for 1979,
pp. 40-43.

AthearN, H. D. 1952. Some new records of
naiades from eastern North America. Ann.
Repts. Amer. Mai. Union for 1952, pp. 8-9.

Athearn, H. D. and a. H. Clarke. 1962. The
freshwater mussels of Nova Scotia. Natl. Mus.
Canada Bull. no. 193, pp. 11^1, pis. 1-4.

.\ughey, S. 1877. Catalogue of the land and fresh-
water shells of Nebraska. Bull. U. S. Geol. and
Geog. Surv'. of the Territories 3: 697-704.

Baker, F. C. 1909. Mollusks from Kansas and
Oklahoma. Nautilus 23 (7): 91-94.

. 1922. The nioUuscan fauna of the Big Ver-
milion River, Illinois. Illinois Biol. Monogr. 7
(2): 1-126, pis. 1-15.

. 1928. The fresh-water Mollusca of Wiscon-

sin. Pt. II. Pelecvpoda. Wisconsin Geol. and
Nat. Hist. Surv. Bull. 70 (2): iv, 1-495, pis. 29-
105.

Ball, G. H. 1922. Variation in freshwater mussels.
Ecology 3: 93-121.

B.\tes,J. 1970. Ohio mussel fisheries investigation
. . . Pt. 1, Mussel studies. Center for Aquatic
Biology, Eastern Michigan University', Ypsilan-
ti, Michigan, pp. 108 [mimeographed].

. 1971. Mussel investigations, state of West

Virginia. June 1, 1969-June 15, 1971. Final re-
port Part 1 — section 1. Mussel distribution and
ecology. Center for Aquatic Biology, Eastern
Michigan Univesity, Ypsilanti, Michigan, pp.
iv, 91 [mimeographed].

B.a^tes, J. AND S. D. Dennis. 1978. The mussel
fauna of the Clinch River, Tennessee and Vir-
ginia. Sterkiana nos. 69-70, pp, 3-23.

Blankenship, S. and D. R. Crockett. 1972.
Changes in the freshwater mussel fauna of the
Rockcastle River at Livingston, Kentucky.
Trans. Kentucky Acad. Sci 33: 37-39.

Blatchley, W. S. AND L. E. Daniels. 1903. On
some Mollusca known to occur in Indiana. A
supplementary paper to Call's Catalogue.
Twenty-seventh Ann. Rept. Indiana Dept.
Geol. and Nat. Resources for 1902, Indianapo-
lis, pp. .577-628, 3 pis.

Bra.nsoN, B. a. 1963. New mollusk records from

Zoogeography of North American Unionacea 'Johnson 141

Oklahoma and their zooyeoj^raphical sij^nifi-
cance. Trans. Kansas Acad. Sci. 66 (3): 501-
512.

. 1966a. Alasmiclonta inar^indta and Ptij-

chuhranchiis fdsciolari.s in Kansas. Nautilus 80
(1): 21-24, map 1.

. 1966h. L'nionid records from Kansas, Ar-
kansas and Louisiana. Sterkiana, no. 23, pp.
7-8.

. 1967. A partial hiolotjical survey of the

Spring River drainage in Kansas, Oklahoma and
Missouri. Part 1. Collecting sites, basic limno-
logical data, mollusks. Trans. Kansas Acad. Sci.
for 1966, 69 (3-4): 242-293, figs. 1-7, tables 1-
10.

1969. Glehula in Oklahoma. Sterkiana, no.

36: 22.

Buchanan, A. C. 1977. Statvis of knowledge re-
port, naiades of the Meramec River Basin. Un-
published report for the U. S. Army Corps of
Engineers, St. Louis Dist. Pt. 1. Text, pp. [4] 1-
46; Pt. 2. pp. [2] 1-51.

. 1980. Mussels (naiades) of the Meramec

River basin, Missouri. Bull. Amer. Mai. Union
for 1978, p. 64 [Abstract only].

BURCH, J. B. 1973. Freshwater Unionacean clams
(Mollusca: Pelecypoda) of North America. Bio-
ta of Freshwater Ecosystems. Identification
Manual no. 11. U. S. Environmental Protection
Agency, pp. xi, 1-176, 154 te.xt figs.

. 1975. Freshwater Unionacean clams (Mol-
lusca: Pelecypoda) of North America. Malacol-
ogical Publications, Hamburg, Michigan, pp.
xviii, 1-204, 252 text figs.

C.-\LL, R. E. 1885a. A geographic catalogue of the
Unionidae of the Mississippi Valley. Bull. Des
Moines [Iowa] Acad. Sci. 1 (1): 5-57.

. 1885b. Contribution to a knowledge of the

fresh-water Mollusca of Kansas. III. Fresh-
water bivaKes. Bull. Washburn College Lab.
Nat. Hist. 1 (3): 93-97.

. 1895. A study of the Unionidae of Arkansas,

with incidental references to their distribution
in the Mississippi Valley. Trans. Acad. Sci. St.
Louis 7 (1): 1-65, pis. 1-21.

. 1897. The hydrographic basins of Indiana

and their Molluscan fauna. Proc. Indiana Acad.
Sci. for 1896, 6: 247-258, 1 map.

1900. A descriptive illustrated catalogue of

the Mollusca of Indiana. 24th Annual Rept.
Dept. Geol. and Nat. Resources, Indianapolis,
Indiana, for 1899, pp. 337-535, pis. 1-78; In-
dex, pp. 1013-1017.

Carlander, H. B. 1954. A history offish and fish-
ing in the Upper Mississippi River. Upper Mis-
sissippi River Conservation Commission, Des
Moines, Iowa, pp. 1-96.

Clark, C. F. 1976. The freshwater naiades of the
lower end of the Wabash River, Mt. Carmel,
Illinois to the south. Sterkiana, no. 61, pp. 1-
14.

. 1977. The freshwater naiades of Ohio. Pt.

I: St. Joseph River of the Maumee. Sterkiana,
nos. 65-66, pp. 14-36.

Clark, H. W. and G. H. Gillette. 1911. Some
observations made on Little River near Wichi-
ta, Kansas, with reference to the Unionidae.
Proc. Biol. Soc. Washington 24: 63-68.

Clarke, A. H. 1959. Unionidae from upper St.
Lawrence River. Nautilus 72: 98-99.

. 1969. Freshwater mussels of the Canadian

Interior Basin. Ann. Repts. Amer. Mai. Union
for 1969, pp. 27-28.

— . 1973. The freshwater mollusks of the Ca-

nadian Interior Basin. Malacologia 13: pp. i-
iv, [1-12], 1-509, maps, 1-87, charts, 1-35, pis.
1-28.

Clarke, A. H. and C. O. Berg. 1959. The fresh-
water mussels of Central New York. Cornell
Univ. Agr. Expt. Sta., Ithaca, N. Y., Mem. no.
367: 1-79, 7 pis.

Clarke, A. H. and A. M. Rick. 1963. Supplemen-
tary records of Unionacea from Nova Scotia
with a discussion of the identity q{ Anodonta
fragilis Lamarck. Natl. Mus. Canada, Ottawa.
Bull. no. 199, pp. 15-27.

Clench, W. J. and H. van der Schalie. 1944.
Notes on naiades from the Green, Salt, and
Tradewater Rivers in Kentucky. Papers Mich.
Acad. Sci., Arts and Letters 29: 223-229.

Coker, R. E. 1919. Fresh-water mussels and mus-
sel industries of the United States. Bull. U. S.
Bur. Fisheries, 1917-18, 36: 1-89, 46 pis., fold-
ing map.

Coker, R. E. and J. B. Southall. 1915. Mussel
resources in tributaries of the upper Missouri
River. U. S. Bur. Fisheries, appendix 4 to rept.
of U. S. Commissioner of Fisheries for 1914,
pp. 1-17, pi. 1.

Coker, R. E. and T. Surber. 1911. A note on the
metamorphosis of the mussel Lciinpsilis laevis-
simus. Biol. Bull. 20: 179-182.

Coon, T. G., et al. 1977. Relative abundance and
growth of mussels (Mollusca: Eulamellibran-
chia) in pools 8, 9 and 10 of the Mississippi
River. Freshwater Biology 7: 279-285.

Danglade, E. 1914. The mussel resources of the
Illinois River. U. S. Bur. Fisheries, appendix 6
to rept. of U. S. Commissioner of Fisheries for
1913, pp. 1-48.

. 1922. The Kentucky River and its mussel

resources. U. S. Bur. Fisheries, Doc. no. 934,
pp. 1-8.

Daniels, L. E. 1903. A check list of Indiana mol-
lusca with localities. Twenty-seventh Ann.
Rept. Indiana Dept. Geol. and Nat. Resources
for 1902, Indianapolis, pp. 629-652.

. 1914. A supplemental check list of Indiana

mollusca, with localities and notes. Thirty-
ninth Ann. Rept. Indiana Dept. Geol. and Nat.
Resources for 1902, Indianapolis, pp. 318-326.

Dawley, C. M. 1947. Distribution of aquatic mol-
lusks in Minnesota. Amer. Midi. Nat. 38: 671-
697.

1 42 Bullt'lin Mii.sciini of C'oiiiixircitii c '/aioIo^ij, Vol. I W, Xo. 2

Dennis, S. 1). 1970. l\inis\ K.mia mussel studies
. . . Center lot A(iuatie Biology, Eastern Mitli-
i^iiu Lhii\t'rsit> , ^'j^silaiiti, Michiuaii, pj). 13S.

Kl.INT, R. K. 1971. Glacial and Quaternary yeolo^y.
John Wile\ and Sons, New York, pp. ix, 1-892.

Fhiehson, L. S. 1924. Interestinti facts in the liis-
tory of" Vnio orhiculcitus Hildretli and (/.
ahrtiptiis Say. Nantilns 37: 135-137.

. 1927. A classified and annotated check list

of the North .\inerican naiades. Baylor Univ.
Press, Waco, Texas, pp. 1-111, Errata et (Cor-
rigenda, [separate sheetj.

Fui.LEH, S. L. H. 1977. Freshwater and terrestrial
inolhisks. pp. 143-194. In: Cooper, J. E., ct ill.
[Ed. J Endangered and threatened plants and
animals of North Carolina. North Carolina State
Mus. Nat. Hist., Raleigh, N.C. pp i-xvi, 1-444.

[Fuller, S. L. H.] 1978. Corps/Service cooperate
to protect endangered mussels. Endangered
Species Technical Bull. 3 (9): 3-6.

Fuller, S. L. H. .and R. H. 41.\rtenstine. 1980.
Anodonta imhecillis Say in the Delaware River
hasin. Nantilns 94 (1): 4.

Goodrich, C. and H. van der Schalie. 1932. 1.
On the increase in the naiad fauna of Saginaw
Bay, Michigan. 11. The naiad species of the
Great Lakes. Occ. Papers Mns. Zool. Univ.
Mich., 238: 1-14.

. 1939. A(jnatic moUusks of the npi^er pen-
insula of Micliigan. Uni\. Mich. Mns. Zool.
Misc. Pnhl. 43: 1-45.

Gordon, M. E., L. R. Kraemer, and A. V.
Brown. 1980. Unionaceae of Arkansas: his-
torical review, checklist, and observations on
distributional patterns. Bnll. Amer. Mai. Union
for 1979, pp. 31-37.

. 1944. A revision of the Mollnsca of Indiana.

Amer. Midi. Nat. 32 (2): 257-326.

Gregorio, a. De. 1914. Sn taluni molluschi di
acqua dolce di America (American fresh water
shells of America). 11 Natnralista Siciliano 22
(2-3): 31-72, pis. 3-12.

Crier, N. M. 1922. Final report on tlie stutly and
appraisal of mussel resources in selected areas
of the Upper Mississippi River. Amer. Midi.
Nat. 8: 1-33.

. 1926. Report on the study and appraisal of

nmssel resources in selected areas of the Upper
Mississippi River, 1920-1925. Amer. Midi. Nat.
10: 89-110, 113-134.

Crier, N. M. and J. F. Mueller. 1922-1923.
Notes on the naiad fauna of the Upper Missis-
sippi. II. The naiades of the Uj^per Mississii:)pi
drainage. Nantilns 36: 46-49; 1923, 96-103.

Haas, F. 1969. Snperfamilia Unionacea. [ln\ Das
Tierreich. Berlin. Pt. 88, pp. 1-X, 1-663.

Havlik, Marian E. and D. H. Stansbery. 1978.
The naiad mollnsks of the Mississsippi River
in the vicinity of Prairie dn Chien, Wisconsin.
Bnll. Amer. Mai. Union for 1977, pp. 9-12.

Heard, W. H. 1975. Sexnality and other aspects of

reproduction in .Xnodontd ( Ptdecs poda:

Unionidae). Malacologia 15 (1): 81-103.
IhNKLEY, A. A. 1906. SoTne shells of Mississiiii)i

and Alabama. Nautilus 20: 52-55.
l.SEI.V, F. B. 1925. The fresh-water nuissel lauua

of eastern Oklahoma. Proc. Oklahoma .\cad.

Sci. 4: 43-118, 2 tables.
IsoM, B. G. AND p. YOKLEY, Jh. 1968. 4"he mussel

fauna of Duck River in Tennessee, 1965. Amer.

Midi. Nat. 80 (1): 34-42.

JENKINSON, J. J. AND F. L. KOKAI. 1978. Villosd
lienosa (Conrad, 1834) in Ohio. Bnll. Amer.
Mai. Union for 1977, pp. 82-83.

Johnson, C. W. 1914. Fauna of New England. List
of the Mollnsca. Occ. Pap. Boston Soc. Nat.
Hist. 7 (13): 1-231.

Johnson, R. 1. 1956. The types of naiades (Mol-
lnsca: Unionidae) in the Museum of Compar-
ative Zoologv. Bnll. Mns. Comp. Zool. 115 (4):
102-142; pis'. 1-2.

. 1970. The systematics and zoogeography

of the Unionidae (Mollnsca: Bivalvia) of the
Sonthern Atlantic Slope Region. Bnll. Mns.
Comp. Zool. 140 (6): 263-450.

. 1972. The Unionidae (Mollnsca: Bivalvia)

of Peninsular Florida. Bnll. Florida State Mns.
16 (4): 181-249, figs. 1-10.

. 1975a. Simpson's nnionid types and mis-

cellaneons nnionid types in the National Mu-
senm of Natural History. Special Occ. Pub. 4.
Dept. Moll., Mns. Comp. Zool., Cambridge,
Mass., pp. 1-56, 3 pis.

. 1975b. First paper on the Conchology of

the United States by an American author,
Thomas Sav, 1817. Jour. Soc. Biblphy. Nat.
Hist. 7 (3): 265-267. "

. 1977a. Monograph of the genns Mcdioiii-

(lu.s (Bi\al\i;i: Unionidae) mostK from the Apa-
lachicolan region southeastern United States.
Occ. Pap. on Moll. 4 (56): 161-187.

. 1977b. Arnold Edward Ortmann, a bibli-
ography of his work on mollnsks, with a cata-
logue of his recent molluscan taxa. Occ. P;ip.
on Moll. 4 (58): 229-247, pi. 7.

. 1978. S\stematics and zoogeograph\' of

Plu'^iohi { = Dy.siwniid = Epiohhi.snia), an al-
most extinct genus of freshwater mussels (Bi-
\ al\ ia: Unionidae) from middle North America.
Bnll. Mns. Comp. Zool. 148 (6): 239-321.

1979. The types of Unionacea (Mollnsca:

Bi\alvia) in the Museum of Zoology, Lfni\ersit\
of Michigan. Malacological Review 12 (1-2):
29-36.

Johnson, R. 1. and H. B. Baker. 1973. The types
of Unionacea (Mollnsca: Bi\;il\ia) in the Acad-
em\- of Natinal Sciences of Philadeli-)hia. Proc.
Acad. Nat. Sci. Phila. 125 (9): 145-186, pis. 1-
10.

La RocyUE, A. 1967. Pleistocene Mollnsca of
Ohio. State of Ohio, Ck'ol. Survev Bnll. no. 62
(2) [Chapter 4— Naiades], pp. vii-xiv, 113-285.

Zoogeography of North American UNioNACEA'/o/in.son 143

Selected References and Index, 1970 (4): 772-
800.

La RocyuE, A. and J. ()uc;ht()X. 1937. .\ prelim-
inai\' account of the Unionidae of Ontario. Ca-
nadian Jour. Res., Sec. D, 15: 147-155.

Latchfohd, F. R. 1882. Notes on the Ottawa
Lhiionidae. Trans. Ottawa Field Nat. Cluh, pp.
48-57.

Lea, Isaac. 1836. A s\nopsis of the fauiiK' of
naiades. Philadelpliia: Care\', Lea and Blan-
chard, pp. i-\ iii, 1-59 [2], 1 col. pi.

Mathiak, H. a. 1979. A river survey of the unionid
mussels of Wisconsin 1973-1977. Sand Shell
Press, Horicon, Wisconsin, pp. 75, 11 col. pis.,
distributional maps.

Meek, S. E. 1896. A list of fishes and niolhisks
collected in Arkansas and Indian Territory in
1894. Bull. U. S. Fish Comm. for 1895, pp. 341-
349.

Meek, S. E. and H. W. Clahk. 1912. The mussels
of the Big Buffalo Fork of White River, Arkan-
sas. U. S. Bur. Fisheries, Doc. 759, pp. 1-20, 2
tables.

Meyer, E. R. 1974. Unionid mussels of the Wa-
bash, White and East Fork of the White rivers,
Indiana. Viriiinia Jour. Sci. 25 (1): 20-25.

Morris, J. S. and R. W. Taylor. 1978. A survey
of the freshwater mussels (Bi\al\ia: Unionidae)
of the Kanawha Ri\er of West Virginia. Nauti-
lus 92 (4): 153-155.

Morrison, J. P. E. 1929. A preliminary list of the
mollusca of Dane County, Wisconsin. Trans.
Wisconsin Acad. Sci., Arts and Letters 24: 405-
425.

. 1932. A report on the mollusca of the north-
eastern Wisconsin lake district. Trans. Wiscon-
sin Acad. Sci., Arts and Letters 27: 359-396.

. 1942. Preliminary report on mollusks foinid

in shell mounds of the Pickwick Landing Basin
in the Tennessee River Valley. Bineau Amer.
Ethnologv (Smithsonian Inst., Washington)
Bull. no. i29: 339-392.

. 1969. The earliest names for North Amer-
ican naiades. Ann. Repts. Amer. Mai. Union for
1969, pp. 22-24.

. 1975. Maryland and Virginia mussels of

Lister. Bull. Amer. Mai. Union for 1974, pp. 36-

39, 1 pi.
Murray, H. A. and B. Leonard. 1962. Unionid

mussels in Kansas. Uni\ . of Kansas, Mus. Nat.

Hist. Misc. Pub. no. 28, pp. 1-184, 45 pis., 42

figs.
Neel, J. K. 1941. A taxonomic study of Quadridu

(fuadrula (Rafinesque). Occ. Papers Mus. Zool.

Univ. Mich., no. 448, pp. 1-8, pi. 1.
Neel, J. K. and W. R. Allen. 1964. The mussel

fauna of the upper Cimiberland basin before its

impoundment. Malacologia 1 (3): 427-459.
Nordstrom, R., ct al. 1977. Rare and endangered

species of Missouri. Missouri Dept. Conserva-
tion, pp. 1-129.

Ortmann, a. E. 1909. A preliminary list of the
Unionidae of western Pennsylvania, with new
localities for species from eastern Pennsylva-
nia. Ann. Carnegie Mus. 5 (2-3): 179-21o!

. 1911. A monograph of the naiades of Penn-

syKania. Part 1. Anatomical investigations; Part
2. The system of North American naiades.
Mem. Carnegie Mus. 4 (6): 279-347, pis. 86-
89.

. 1912. The geological origin of the fresh-
water fainia of Pennsylvania. Pennsylvania To-
pog. and Geol. Survey 1910-1912, pp. 130-149.

. 1913. The Alleghenian Divide, and its in-
fluence upon the fresh water faima. Proc. Amer.
Philos. Soc. 52: 287-390, pis. 12-14.

. 1918. The Nayades (freshwater mussels) of

the upper Tennessee drainage. With notes on
svnonvmv and distribution. Proc. Amer. Philos.
Soc. 57 (6); 521-625.

. 1919. A monograph of the naiades of Penn-
sylvania. Part 3. Systematic account of the gen-
era and species. Mem. Carnegie Mus. 8 (1): 1-
384, pis. 1-21.

. 1920. Correlation of shape and station in

freshwater mussels (Naiades). Proc. Amer. Phi-
los. Soc. 59 (4): 269-312.

. 1924a. The naiad-fauna of Duck River in

Tennessee. Amer. Midi. Nat. 9 (1): 18-62.

. 1924b. Distributional features of naiades in

tributaries of Lake Erie. Amer. Midi. Nat. 9:

101-117.
— . 1925. The naiad-fauna of the Tennessee

River system below Walden Gorge. Amer.

Midi. Nat. 9 (8): 321-373.
— . 1926. The naiades of the Green River

drainage in Kentucky. Ann. Carnegie Mus. 17
(1): 167-188, pi. 8.

Ort.mann, a. E. and B. Walker. 1922. On the
nomenclature of certain North American
naiades. Occ. Papers, Mus. Zool., Univ. Mich-
igan, no. 112, pp. 1-75.

Par\l\lee, p. W. 1967. The fresh-water mussels
of Illinois. Illinois State Mus. Popular Sci. Ser.
8, pp. i-ix, 1-108, 35 pis.

RiGGS, C. D. and G. R. Werr. 1956. The mussel
population of an area loamv-sand bottom of
Lake Texoma. Amer. Midi. Nat. 56 (1): 197-
203, tables 1-2.

Rorertson, I. C. S. and C. L. Blakeslee. 1948.
The xMollusca of the Niagara frontier region.
Bull. Buffalo Soc. Nat. Sci. 19 (3): i-xi, 1-191,
14 pis., 1 folding map.

Rosewater, T- 1959. Mollusks of the Salt F^iver,
Kentucky. Nautilus 73: 57-63.

Say, T. 1830-34. American Conchology, or de-
scriptions of the shells of North America, illus-
trated by colored figures. New Harmony, In-
diana. 68 col. pis., with letter press. Pts. 1-7.

Scammon, R. E. 1906. The Unionidae of Kansas,
Part I. Univ. Kansas Sci. Bull. 3: 279-373, pis.
52-86.

144 Bulletin Museum of Coinixinitivc /.ooId^ij. Vol. 149, No. 2

Simpson, C. T. 1893. On sonif fossil miios and
other fresh-water shells from the chift at Toron-
to, Canada: with a re\ iew of the distrihntion of
the Unionidae of northeastern North America.
Proc. U. S. Natl. Mus. 16: 591-595.

. 1896. On the Mississippi Valley Unionidae

fonnd in the St. Lawrence and Atlantic drai li-
aise areas. Amer. Nat. 30: ;379-:384.

. 1900. Synopsis of the naiades, or pearly

fresh-water mnssels. Proc. U. S. Natl. .Vlns. 22:
501-1044, pi. 18.

1914. A descriptive catalogue of the

naiades, or pearly fresh-water mussels. Detroit,
Michitjan. Parts 1-3, pp. i-xi, 1-1.540.

Smith, H. M. 1899. The mussel fishery and pearl
button industry of the Mississippi River. U. S.
Fish Comm. Bull, for 1898, pp. 289-314, pis.
65-85.

St.anSBERY, D. H. 1961. Species of naiades re-
corded for central Ohio (Scioto Drainage above
Circle\ille) 2 pp. [mimeographed].

. 1965. Naiad fainia of the Green River at

MunforcK ille, Kentuckv. Ann. Repts. Amer.
Mai. Union for 1965, pp. 13-14.

- — ■ . 1967. A provisional classification of the

Pleurnhema cordatiim complex in the Missis-
sippi drainage l)asin of North America. Ohio
State Museum 2 pp. [mimeographed].

. 1970. American Malacological Union Sym-
posium: rare and endangered mollusks. (2).
Eastern freshwater mollusks (1). The Mississip-
pi and St. Lawrence River systems. Malacolo-
gia 10: 9-31, pis. 1-2.

. 1971. Rare and endangered mollusks in

eastern United States, pp. 5-18. [//(] S. E. Jor-
gensen and R. Sharp (Editors) Proc. of a sym-
posium on rare and endangered mollusks
(naiads) of the U. S. Sport Fish and Wild., Dept.
Interior, Region 3, Twin Cities, Minn., pp. 1-
79.

. 1973. A preliminary report on the naiad

fauna of the Clinch River in the southern Ap-
palachian mountains of Virginia and Tennessee
(Mollnsca: Rival via: Unionidae). Bull. Amer.
Mai. Union for 1972: 20-22.

1976. [in] Boschung, H., Endangered and

threatened plants and animals of Alabama.

Bull. Alabama Mus. Nat. Hist., no. 2, Naiad

mollusks, pp. 42-52.
Stansbery, D. H. .\xd F. Kok.\i. 1979. A revision

of the Lamp^ilis orhiculata (Hildreth, 1828)

comple.x. Bidl. Amer. Mai. Union for 1978, p.

57 [Abstract only].
Stansbery, D. H. and C. B. Stein. 1962. The

Unionidae of the Island region of western Lake

Erie. A check list [2 pp. mimeographed].
Starrett, \V. C. 1971. A survey of the mussels

(Unionacea) of the Illinois River: a polluted

stream. Illinois Nat. Hist. Sur., Bull. 30 (5):

267-403, 4 col. pis., 17 figs.
Stein, C. B. 1965. Naiad fauna of Little Darin-

Creek in C^'utral Ohio. Ann. Repts. Amer. Mai.
Union for 1965, pp. 22-23.

S tehki, V. 1900. List of the land and fresh water
.Mollnsca of Tuscarawas County, Ohio. Eighth
Ann. Rept. Ohio State Acad. Sci., pp. 1-13.

. 1902. Some additions and corrections to the

list of land and fresh water Mollnsca. Ohio Nat.
2 (8): 286-287.

Stober, y. J. 1972. Distribution and age of Mar-
^aritiffru imir^aritifera (L.) in a Madison River
(Montana, U. S. A.) mussel bed. Malacologia 1 1
(2): .343-350.

Strayer, D. 1979. Some recent collections of
mussels from southeastern Michigan. Malaco-
logical Review 12 (1-2): 93-95.

Strecker, J. K. 1931. The distribution of the
naiades or pearly fresh-water mussels of Texas.
Baylor Univ. Mus. Spec, Bull. pp. 1-71.

Strode, W. S. 1891. Mollusks of Thompson's
Lake, Illinois. Nautilus 4 (2): 133-134.

Surber, T. 1912. Identification of the glochidia of
freshwater mussels. U. S. Bur. Fisheries Doc.
771, pp. 1-10, pis. 1-3.

. 1914. Notes on the natural hosts of fresh-
water mussels. Bull. U. S. Bur. Fisheries for
1912, 32: 101-115, pis. 29-31.

1915. Identification of the glochidia of

freshwater mussels. Appendix 5 to Rept. U. S.
Commissioner of Fisheries for 1914, U. S. Bur.
Fisheries Doc. 813, pp. 1-9, pi. 1.

Taylor, R. W. 1980a. A survey of the freshwater
inussels of the Oliio River from Greenup locks
and dam to Pittsburgh, Pa. A report submitted
to the U. S. Army Corps of Engineers. Hun-
tington [West Virginia] District. [Mimeo-
graphed] pp. iv, 1-71, 20 pis.

. 1980b. Mussels of Floyd's Fork, a small

northcentral Kentuckv stream. Nautilus 94 (1):
13-15.

1980c. Freshwater bivalves of Tygart

Creek, northeastern Kentuckv. Nautilus 94 (2)
89-91.

Utterback, W. 1. 1915-1916. The naiades of .Mis-
souri. Amer. Midi. Nat. 4: 41-53; 97-152; 181-
204; 244-273. 1916, pp. 311-327; 339-3.54;
387-400; 432-464, pis. 1-29. Reprinted and re-
paged, 1916, pp. 1-200, pis. 1-29.

. 1917. Naiadgeographv of Missouri. Amer.

Midi. Nat. 5: 26-30, 1 table.

Valentine, B. D. and D. H. Stan.sbery. 1971.
An introduction to the naiades of the Lake Tex-
oma Region, Oklahoma, with notes on the Red
River fauna. Sterkiana, no. 42, pp. 1—40.

Vanatta, E. G. 1910. Unionidae from southeast-
ern Arkansas and N. E. Louisiana. Nautilus 23
(8): 102-104.

VAN DER SCHALIE, H. 1936. The naiad fauna of the
St. Joseph River drainage in southwestern
Michigan. Amer. Midi. Nat. 17: 523-527.

. 1938. The naiad fauna of the Huron River,

in southeastern Michigan. Misc. Pub. Mus.

Zoogeography of North American VJsio'sacea • JoJinson 145

Zool., Univ. Michigan, no. 40, pp. 83, 12 pis.,

1 folding map.
— . 1939. Additional notes on the naiades of

the lower Tennessee Ri\ er. Amer. Mid. Nat. 22

(2): 452-457.
. 1941. Zoogeography of naiades in the

Grand and Muskegon Rivers of Michigan as re-
lated to glacial history. Papers Michigan .Acad.

Sci., Arts and Letters 26: 297-310, 2 figs.
— . 1945. The valne of mnssel distribution in

tracing stream confluence. Papers Michigan

Acad. Sci., Arts and Letters 30: 355-373.
— . 1961. The naiad (fresh-water mussel) fauna

of the Great Lakes. Great Lakes Res. Div., Inst.

Sci. and Tech., Univ. Michigan, Pub. no. 7:

156-157.
. 1963. Mussel distribution in relation to

stream confluence in northern Michigan, USA.

Malacologia 1: 227-2.36.

1973. The mollusks of the Duck River

drainage in central Tennessee. Sterkiana, no.

52, pp. 45-55.
VAN DER SCHALIE, H. AND A. 1950. The mussels

of the Mississippi River. Amer. Midi. Nat. 44:

448-i66.
. 1963. The distribution, ecology, and life

history of the mussel, Actinonaias ellipsiformis

(Conrad), in Michigan. Occ. Papers Mus. Zool.

Univ. Mich., no. 633, pp. 1-17, 3 pis.

Walker, B. 1913. The Unione fauna of the Great
Lakes. Nautilus 27: 18-23, 29-34, 40-47, 56-
59.

. 1918. A synopsis of the classification of the

fresh water Mollusca of North America, north
of Mexico and a catalogue of the more recently
described species, with notes. Mus. Zool.,
Univ. Michigan, Mis. Pub. no. 6, pp. 1-213.

Wheeler, H. E. 1914. The unione fauna of Cache
Ri\'er, with description of a new Ftisconaia
from Arkansas. Nautilus 28 (7): 73-78, pi. 4.

. 1918. The Mollusca of Clark County, Ar-
kansas. Nautilus 31 (4): 109-125.

WiLLLAMS, J. C. 1969. Mussel fishery investigation
Tennessee, Ohio and Green rivers final report.
Kentucky Department of Fish and Wildlife Re-
sources, Frankfort, Kentucky, pp. 107.

Whiteaves, J. F. 1895. Notes on recent Canadian
Unionidae. Canadian Record of Sci. 6: 250-
263; Additional notes . . . pp. 365-366.

Wilson, C. B. and H. W. Clark. 1912. The mus-
sel fauna of the Maumee River. U. S. Bur. Fish-
eries Doc. 757, pp. 1-72.

- — . 1912. The mussel fauna of the Kankakee

Basin. U. S. Bur. Fisheries Doc. 758, pp. 1-52,
1 map.

. 1914. The mussels of the Cumberland Riv-
er and its tributaries. U. S. Bur. Fisheries Doc.
no. 781, pp. 1-67, pi. 1.

INDEX TO TAXA

All molluscan ta.xa mentioned in this paper are in-
cluded. References to monographed species are in
bold face type.

aberti (Conrad), Cyprogenia 98, 134

aherti Conrad, Unio 134

abrupta (Say), Lampsilis 122

abruptus Say, Unio 122

Actinonaias Crosse and Fischer 120

aesopus Green, Unio 90

alata (Say), Proptera 96, 107

alatiis Say, Unio 128

albus Hildreth, Unio verrucosus 136

ambigua (Say), Simpsoniconcha 93

Ambleminae 111

Amblemini 111

analoga Ortmann, Fusconaia edgariana 88

angulata (Lea), Gonidea 99

Anodonta Lamarck 112

Anodontinae 112

anodontoides Lea, Unio 95

apiculata (Say), Quadrula 96

appressa (Lea), Fusconaia cuneolus 88

arcaeformis (Lea), Plagiola 87

Arcidens Simpson 115

argenteum (Lea), Pleurobema oviforme 88

arkansasensis (Lea), Eurynia 131

arkansasensis (Lea), Lampsilis 131
arkansasensis Lea, Unio 131
arkansasensis (Lea), Villosa 98, 131
Arkansia Ortmann and Walker 119
askewi Marsh, Unio 89
asper Lea, Unio 90
asperatus Lea, Unio 90
asperrimus Lea, Unio 90
atropurpureum Rafinesque, Alasmidon 88
badium (Rafinescjue), Alasmidon 88
barnsiana (Lea), Fusconaia 87, 112
bealii Lea, Anodonta 92
bergei Baker, Anodontoides 92
beringiana Middendorff, Anodonta 99
biangulata Morrison, Quadrula 88
bimarginata (Lea), Plagiola 87
bigbyensis (Lea), Fusconaia barnesiana 87
blatchleyi Daniels, Lampsilis 96
breviculus Call, Unio 132
britsi Simpson, Lampsilis 132
britsii Simpson, Lampsilis breviculus 132
brooksiana van der Schalie, Anodonta 99
bullata Rafinescjue, Obliquaria 90
bullocki Baker, Quadrula quadrula 90
bursapastoris Wright, Unio 89
cahni Baker, Carunculina parva 94
calceolus (Lea), Alasmidonta 88, 92

146 Bulletin Museutn of Comixiratiic Zooloi^y, Vol. 149, No. 2

ctiliforniciisis Lea, .\in)d(>nt(i 99

cuinptadon Say, Unio 92

canadensis Lea, Viiio 95

capax (Creen), Lanipsilis 129

capax (Green), Proptcra 96, 128

capax Green, Ibiio 128

capcranis (Lea), Dronitis drunias 88

capsaeformis Lea, Phifiiola 87

cardiacca Gnerin, Vnio 91

cardiuni Rafinesqne, Lanipsilis 95

carinatus Barnes, (7nio 93, 120

cariosa (Say), Lanipsilis 99

castanciis Lea, C'/n'o 98

castancus Rafinesqne, Unio 98

cataracta Say, Anodonta cataracta 99

c<itilltis Conrad, C'n/o 91

ccrintis Conrad, C'n/'o 86

chadicicki Baker, Lanipsilis siliquoidea 95

cicatricosus (Say) Plethobasus 90

clava (Lamarck), Pleurohenia 91

clintonensis Simpson, Pfychohranchtis 138

coccincuni (Conrad), Pleurohenia 91

coccineum Lea, Pleurohema 91

coccineum Lea, C^/uo 91

coccineus Conrad, C/;ho 91

Cokeria Marshall 86

complanata (Barnes), Lasmifiona 107

compel tits Frierson, t//u'() 90

complanata (Lightfoot), Elliptio 98, 107

comprcssa (Lea), Lasmi<iona 93, 107

confra^osus Say, Alasniidonta 115

confragosus (Say), Arcidens 93, 115

conradi (Vanatta), Lexingtonia dolahelloides 88

conradicus (Lea), Medionidus 88

contraryensis Utterhack, Quadrula quadrula 90

cooperianus (Lea), Plethobasus 90

cor Conrad, C/»io 88

cordatum Rafinesqne, Ohovaria 91

cordatum (Rafinesqne), Pleurohema 91

corpulenta Cooper, Anodonta 92

costata Rafinesqne, Amhlema 86

costata (Rafinesqne), Lasmigona 93, 107

crassidens (Lamarck), Elliptio crassidens 90

crassus Conrad, (7nto 122

cromwelli Lea, Unio 94

cuneolus (Lea), Fusconaia 88

cuneus Conrad, C/^/o 91

curtisii Frierson and Utterhack, Tnincilla 97

cyclips Rafinesqne, C/nfo 122

cygneus Linnaens, Mytilus 112

cylindrella (Lea), Carunculina 88

cylindrella (Lea), Carunculina maesta 88

cylindrica (Say), Quadrula 89

cyphyus (Rafinesqne), Plethobasus 90

Cyprogenia Agassiz 134

dakota Frierson, Anodonta 92

danielsi Baker, Alasniidonta calceolus 92

danielsii Lea, Anodonta 92

dejecta Lewis, Anodonta 99

delicatus Simpson, L^;uo gihhosus 90

dilatata Rafinesqne, Elliptio 90

dolahelloides (Lea), Lexingtonia 88

domheyanus (\'alenciennes), Plectoinerus 96

donaciformis (Lea), Truncilla 96

droinas (Lea), Dronius 88

ebena (Lea), Fusconaia 86

edentula Say, Alasniidonta 93

edgariana (Lea), Fusconaia 88

elegans Lea, C^»i/(» 96

ellipsiformis Conrad, Actinonaias 94

ellipsiformis Conrad, t'o/o 96

ellipsis Lea, LZ/i/'o 96

elongata Simpson, Ohovaria lens 96

excavata (Lea), Lanipsilis 97

/«/;/7/.v (Lea), V(7/().v« 96

fabula Lea, Pegias 87

fallaciosus Smith, Lanipsilis 95

falcata (Gonid), Margaritifera 99

fasciola Rafinesqne, Lanipsilis 95

fasciolare Mnrrav and Leonard, Ptychobranchus

138
fasciolare (Rafinesqne), Ptychobranchus 96, 138
ferussacianus Lea, Anodontoides 92, 107
//<;l<; Rafinesqne, Fusconaia 86, 107
flexuosa (Rafinesqne), Plagiola 95, 137
florentina (Lea), Plagiola 87, 97
fluviatilis Gmelin, Mytilus 99
footiana Lea, Arioc/onf^ 92
fragilis Lamarck, Anodonta cataracta 99
fragilis (Rafinesqne), Leptodea 95
fragostis Conrad, L/riio 90
friersoni B. H. Wright, L';i/o 89
fucata Dillwyn, Mytilus 99
fulgidus Lea, C/rn'o 91
furva Simpson, Lampsilis subrostrata 96
Fusconaia Simpson 111
gibber Lea, (./h/o 88
gibherum Lea, Pleurohema 88
gibhosus Barnes, tZ/uo 90
gibbus Simpson, Lampsilis ligamentinus 94
gigantea (Barnes) Megalonaias 90
gigantea Lea, Anodonta 92
gouldianus Ward, L^/i/o 91
glans (Lea), Carunculina 94
globosa Lea, Syniphynota 129
globula Morrison, Quadrula subrotunda 96
grandis Say, Anodonta grandis 92, 107
grandis Simpson, Lam))silis higginsii 124
graniferus Lea, (7n/o 90
greeni (Conrad), Ptychobranchus 138
haysiana (Lea), Plagiola HI
hebetattis Conrad, C^n/'o 86
Jieros Say, C/juo 90
heterodon (Lea), Alasniidonta 99
higginsi Lea, Lampsilis 95, 121, 124
higginsii Lea, C/fiio 124
holstonense (Lea), Pleurohema oviforme 88
holstonia (Lea), Lasmigona 88
hydiana (Lea), Lanipsilis 97
imhecillis Say, Anodonta 92
implicata Say, Anodonta 99
intermedia (Conrad), Quadrula 87
intermedins Conrad, (7nio 87
interrupta (Rafinescine), Plagiola 87

Zoogeography of North Amerigan \Jnio\acea- Johnson 147

iris (Lea), Villosa 88, 96

irroratus Lea, Unio L34, 136

jdcintocnsis Strecker, Arciclcns c()tifr(i<^()siis 115

jacksotiiami (Frierson), Ohuidiici 98

katherinae Lea, C^nfo 93

kcnncrlyi Lea, Anodonta 99, 107

kicncrianus Lea, C/nio 90

kirtlandianus Lea, t/nfo 89

hicfirytnosus Lea, L'n/o 90

lacustris Baker, Leptodea fr(i<j.ilis 95

lacustris Baker, Strophitus rugosus 93

Idcustris Baker, TrunciUa tnowata 96

laei'issima (Lea), Leptodca 95

lamarckianus Lea, t^/nO 134

lamarckiana (Lea), CZ/i/'o ahcrti 134

Lanipsilini 120

Lampsilis Rafinesque 121

hipillus Say, C^n/o 130

/«^« (Rafinesque), Hemistena 90

latissima Rafinesque, EUiptio 96

lateralis Rafinesque, Ohliquaria 89

Icfevrei Utterlxick, TrunciUa 97

/ens Lea, t./nio 96

lenior (Lea), Plagiola 87

Iconensis Lea, Anodonta 92

leptodon (Rafinesque), Leptodca 96

lesucurianus Lea, L'^/no 89

leucogona Ortmann, Fusconaja suhrotunda 89

levigata Rafinesque, EUiptio 96

lienosa (Conrad), Villosa 96

ligamcntina (Lamarck), Actinonaias 93, 121

lineolata (Rafinesque), Ellipsaria 95

livida Rafinesque, Toxolasma 94

litrida Simpson, Lampsilis icntricosa 95

luteola Lamarck, C/nio 95

maculata Rafinesque, EUiptio nigra 89

maculata (Rafinesque), Fusconaia maculata 89

magnalacustris Baker, Alasmidonta calccolus 92

magnalacustris Simpson, Quadrula coccinca 91

margaritifera (Linnaeus), Margaritijcra 98, 99

marginata non Say, Anodonta 99

marginata Say, Alasmidonta 92

megaptera Rafinesque, Metaptcra 96

metanevra (Rafinesque), Quadrula 88, 90

Metaptcra Rafinesque 128

Micromya Agassiz 130

minor (Lea), Alasmidonta 88

minor Lea, Margaritana 92

minor Simpson, Quadrula kirtlandicnia 89

mississippiensis Baker, Pleurobema coccincum 91

missouriensis Marsh, Plcurohcma 89

moaesta (Leu), Carunculina glans 94

modesta (Lea), Anodonta 92

monodonta (Say), Cumberlandia 86

mytiloidcs Rafinesque, C//h'o 91

nasuta (Say), Ligumia 99

Sayadina Haas 113

nehulosus Conrad, L'/u'o 96

nchulosa (Conrad), Villosa 88

nervosa Rafinesciue, EUiptio 90

nigra Rafinesque, EUiptio 90

nigrescens Simpson, Liunpsilis ligamcntinus 94

nohilis Comad, l';u'(; 90

nodulata (Rafinesque), Quadrula 90

noviehoraci Lea, t^Mi'o 96

»n<(/(/ Baker, Lasmigona costata 93

nuttalliana Lea, Anodonta 99

ohliqua Lamarck, C'n/o 91

obliquata (Rafinesciue), Plagiola 95

obliquata (Rafinesque), Pleurobema 91

occidens Lea, C'/uo 95

occidentalis (Conrad), Ptychobranchus 138

occidentalis Conrad, L/nio 138

ochracea (Say), Lampsilis 99

oliiensis Rafinescjue, Anodonta 92

ohicnsis Rafinesque, Lastena 95

ohiensis (Rafinesque), Potamilis 95

oliiaria (Rafinesque), Obovaria 96

opacea Lea, Anodonta 92

orbiculata Simpson, LampsUis 122

orbiculatus non Hildreth, C7nio 122

or/j/s Morrison, Acfinonaias carinata 94

oregonensis Lea, Anodonta 99

Ortmanniana Frierson 120

ortmanni (Walker), V!7/o.s« 87, 88

OL«^« (Say), Lampsilis 95, 97, 107

ovatus Say, L'/nO 121

ovifonne (Conrad), Pleurobema 88

ozarkensis (Call), Fusconaia 96, 111

ozarkensis Call, (.//no 111

pachosteus Rafinesque, Obovaria 90

pannosa Simpson, Pleurobema 111

parva (Barnes), Carunculina 94

parva Simpson, Obovaria lens 96

parvonia Lea, Anodonta 93

parvula Grier, Fusconaia flava 86

paupercula Simpson, Quadrula coccinca 91

pectorosa (Conrad), Actinonaias 121

pepinensis Baker, Lampsilis siliquoidea 95

pepinensis Baker, Lasmigona costata 93

pepinensis Baker, Stropliitus rugosus 93

perglobosa Baker, Lampsilis ventricosa 95

perplicata Conrad, C//ho 86

perpurpurea (Lea), Villosa 88

personata (Say), Plagiola 95

peruviana Lamarck, C/»io 86

phaseolus Hildreth, L'/n'o 138

phaseolus Scammon, Ptychobranchus 138

pilaris Lea, l/n/o 89

pilsliryi Marsh, C'»h'o 86

placitus Lea, C'?i/'o 97

plana Lea, Anodonta 92

platcolus (Rafinesque), Medionidus 88

plcasi Marsh, C'/n'o 94

plenum Lea, C'nio 91

plicata (Say), Amblcma 120

popenoi Call, L'/i/o 134

Potamilis Rafinesque 128

pouelli (Lea), Lampsilis 97, 121, 127

pi>wclli Lea, I'n/o 127

prasiinis Conrad, I'/i/'o 90

premorsa Rafinesciue, Plcurobona 91

Prolasmi(h)nta Ortmann 99

propinqua (Lea), Plagiola 95

148

Bulletin Museum oj Coniparatiic Zoi)loiJ.{i, Vol. 149, No. 2

Froptcra Rahncsciiit' 128

Ptychohnmchus Simpson 138

piiijuirdtd (Lainart'k), Projitcrci 98, 129

pusillii Kafiiiesciue, Ohlicpitirici 86

pusilla Simpson, Cyprogenia irrorata 136

))usill(i Siini^soii, Quddnthi ^raiiifcrii 90

ptistiilatus Lea, Uuio 90

pustulosa (Lea), Quudrula 90

)){/niini(l(ituni (Lea), Ptctirohciun 91

pyrainiddttis Lea, Lhiio 91

quadrula Rafinesque, Quudrula 90, 107

(piintardii Crayin. Quadrula perplicata 86

radiata (Ginelin), Lampsdis radiatii 99

rafinesqueana (Frierson), Actinonaias 97, 120

rafincsqueana Frierson, Lampsilis 120

rariplicata Deshayes, Unio 86

raveneliana (Lea), Alasmidonta 87, 88

recta (Lamarck), Ligumia 96, 107

reeviana (Lea), Villosa 98, 132

reevianus Lea, Unio 132

reflexa Rafinesque, Ohliquaria 96

refulgens Lea, Vnio 90

retusa (Lamarck), Ohovaria 96, 123

rimosus (Rafinesque), Lcmiox 87

rosacea DeKay, Unio 95

rotundata (Lamarck), Glebula 97

ndndus Lea, L^ruo 86

ruhigiiu>sus Lea, L/n/o 86

rubrum (Rafinesque), Pleurobema 91

rubrum (Rafinesque), Pleurobema obliquum 91

rugosus Barnes, C'»u'o 90

rugosus Swainson, Anodon 93

sageri Conrad, (7;uo 96

■sampsoni (Lea), Plagiola 95

sampsoniana Frierson, Quadrula flava 111

sapotalensis Lea, L^nu) 120

satur (Lea), Lampsilis 97

satur (Lea), Lampsilis ventricosa 97

satur Lea, C/fi/o 97

s«!// Ward, C'/n'o 92

schoolcraftensis Lea, (7nio 90

securis Lea, (7nto 95

selecta Wheeler, Fusconaia 89

siliquoidea (Barnes), Lampsilis radiata 95, 97, 107

simpsoni Ferriss, Lampsilis 95

simpsoniana Lea, Anodonta grandis 99, 107

sintoxia (Rafinesque), Pleurobema 91

solidus Lea, L/»uo 91

southalli Marshall, Cokeria 86

sowerbyana Frierson, Lampsilis 88

sparsa (Lea), Quadrula 88

sparsus Lea, t/nzo 88

stegaria (Rafinesque), Cyprogenia 96, 136

stegaria Rafinesque, Obovaria 136

sterkii Grier, Ellipfio dilatatus 90

stewardsoni (Lea), Plagiola HI

streckeri, Frierson, Lampsilis 98

streckeri Valentine and Stansberv, Actinonaias 98,

120
striata Rafinesque, Obovaria 90
strigillatus Wright, C/nio 89
steuartiana Lea, Anodonta 92

subeylindracea Lea, Anodonta 92

subelliptica Simpson, Pleurobema brevis 111

std)gibbosus Lea, t^>i((> 90

suborbiculata Say, Anodonta 92, 113

suborbiculata Say, Anodonta (Utterbackia) 113

suborbiculata Lamarck, (^n/'o 97

subrostrata Say, Ligumia 96

subrotunda (Lea), Fusconaia 89

subrotunda (Rafinescjue), Obovaria 62

subtentum (Say), Ptychobranchus HI

Subtentus Frierson 138

subtentus Say, t/nio 138

subviridis (Conrad), Lasmigona 88, 99

sulcatus Lea, (7n/o 95

superiorensis Marsh, t/mO 95

taeniata (Conrad), Villosa 88

teneltus (Rafinesque), Obliquaria 88

tennesseensis Frierson, Strophitus undulatus 93

^ere.v (Rafinesque), Lam))silis 95, 128

tetralasmus (Say), Uniomerus 92

texasensis Lea, Anodonta 92

texasensis Lea, C/nio 94

torulosa (Rafinesque), Plagiola 95

trabalis (Conrad), Villosa 88

trapezoides Lea, CZ/uo 96

trigonoides 'Frierson' Utterback, Fusconaia undata

89
trigonus Lea, t/nfo 89, 111
triquetra (Rafinesque), Plagiola 94
truncata Rafinesque, Truncilla 96
truncata Wright, Margaritana marginata 92
truncatus Spengler, C/n/o 96
tuberculata (Rafinesque), Cyclonaias 90
tuberculatus Barnes, t/n/o 90
tuberosa (Lea), Quadrula 87
tuberosus Lea, C/»n'o 87
tumescens (Lea), Fusconaia barnesiana 87
turgidula (Lea), Plagiola 87, 97
undata (Barnes), Fusconaia 89
undatus Barnes, C/nJo 86
undulata (Say), Alasmidonta 99
undulatus (Say), Strophitus 93, 99, 107
utterbacki Frierson, Pleurobema 111
Utterbackia F. C. Baker 113
Utterbackiana Frierson 113

utterbackiana Frierson, Quadrula tuberculata 90
u«;ju.vc'mt^«.sj.v (Lea), Villosa 87
vunuxemensis (Lea), Villosa vanuxemensis 113
variabilis Baker, Alasmidonta marginata 92
varicosa (Lamarck), Ahrsmidonta 99
ventricosa (Barnes), Lampsilis 95
ventricosus Barnes, C/ni'o 95
venusta Gregorio, Anodonta 113
Venustaconchu Frierson 120
venusttis Lea, (7n/o 94, 120
verrucosa (Rafinesque), Tritogonia 90
ti/K'.v (Conrad), \^i7/(>«; 98, 113
Villosa Frierson 130
vdlosus Wright, CZ/i/o 130
Li'rc'fjs Lea, Anodonta 92
virescens (Lea), Lampsilis 87, 88
t-/iT.vct'«.S' (Lea), t^/iu) 88, 128

Zoogeography of North American UNiONACEA'/o/m.son 149

viridis Rafinesque, Anodonta 92
wa^neri Baker, Fusconaia undata 89
wahhnnatensis Lea, Anodonta 99
wardii Lea, Unio 90

wheclcri Ortniann and Walker, Arkausia 97, 119
winncha^ocnsis Baker, Lcmipsilis venthcusa 95
winnebafioensis Baker, Strophitus rugosus 93

150 Bulletin Museum of ('.(iinjxirdtire Zoolo^ii. \\>l. 149, No. 2

Plate 1,

The drainages of North America relevant to the post-glacial distribution of Unionacea.

Maximum Nebraskan glaciation is regarded as extending, roughly, to the present Missouri and Ohio rivers.

Wisconsin glaciation is marked by a double dotted line (After Flmt, 1971, Fig. 21.1). No attention is paid to the several
driftless areas since they do not have any bearing on unionid distribution.

The regions, systems, and drainages are marked by v»^hite letters in black circles.

A. Ozarkian Region. Consists of the Osage, Gasconade, and Meramec river systems above the Ozark Crest, and below
It the St. Francis, White, Arkansas, Black and Red river systems. The latter two are not included on this map, but
they appear on the distributional maps.

B. Cumberlandian Region

C. Mississippian Region

D. Ohioan Region

E. Northern Atlantic Slope Region

F. St. Lawrence River System (partial)

CANADIAN INTERIOR BASIN Insert from Clarke 1973, inside of front cover in some copies.

G. Pacific Coastal Region
H. Arctic Drainage

I. Hudson Bay Drainage

J. St. Lawrence River system (continued)

Areas of former confluences are marked by letters in white circles.

A. Area of confluence between the Minnesota and Red rivers.

B. Confluence of the Fox and Wisconsin rivers in the vicinity of Portage, Columbia Co., Wisconsin. Dotted lines in Lake
Michigan indicate the now submerged Mackinac River system of which the Millecoqum River, in the eastern part of
the Upper Michigan Peninsula, and the Carp and Ocqueoc, in the northern part of the Lower Peninsula, were once
part.

C. Area of confluence of the Des Plaines River which, by way of the Chicago outlet, drained into the Illinois River.
Before the formation of Lake Michigan, the Muskegon, Grand, and St. Joseph rivers of western Michigan connected
with the Des Plaines River. The Muskegon River was tributary to the Grand, and these had a confluence with the
Saginaw River.

D. Confluence of the Wabash and Maumee rivers in the vicinity of Fort Wayne, Allen Co., Indiana. During the Trent
Stage, the Maumee flowed through the partially dry bed of Lake Erie into the Wabash. Many of the rivers presently
flowing into Lake Erie were once part of the Greater Maumee drainage. Among these are the Clinton River which
flows into Lake St. Clair; the Huron and Raisin rivers in eastern Michigan; the Sandusky, Vermilion, lower Cuyahoga,
and Grand Rivers in Ohio; as well as the Grand River in Ontario. The Rouge River between the Clinton and Huron
rivers was of later origin and was never connected to the Greater Maumee drainage.

E. Area of confluence of Lake Newberry, which inundated the Finger Lakes Basin and flowed south into the Susque-
hanna River.

F. Area of confluence between the Finger Lakes Basin and the Mohawk River.

Principal rivers and features mentioned in the tables and text which are not clearly visible on the map:

1. Ozark Crest

2. Black Water River

3. Meramec River

4. Monongahela River

5. Allegheny River

6. Tradewater River

7. Salt River

8. Ottawa River

Zoogeography of North American Unionacea- /o/uison 151

<'y

>x'„

f\

\^

>

it-

M:^^

.y

%i

trt-.

Rx-r

n.

,-' - f 1

.i^V

c;

Jt\

.SS\

)

^■^.

- ^ r —

/i^7:

152 Bulletin \hi.scun} of ('onipdrddic Zoology, Vol. 149, \(>. 2

Plate 2.

The distribution of Lampsilis abrupta (Say) (squares), and Lampsilis higginsi (Lea) (dots). Triangles represent localities
from which both species have been reported.

The distribution of these two closely related species suggests that L. abrupta had well established Pleistocene refugia
in the Tennessee and Cumberland rivers, and probably also in the Monongahela and Allegheny rivers. It is the only
Cumberlandian species to be found in the Mississippian region. While the Ozarkian species L. higginsi is found on both
sides of the Ozark Crest, it is probable that the Meramec River drainage population was the Pleistocene refugium of the
present Mississippian and Ohioan populations.

It appears that L. abrupta reached eastern Lake Erie during the Trent Stage, when the Maumee River flowed through
the partially dry bed of Lake Erie into the Wabash as did Proptera capax, Plate 3.

ZooGEor.RAPHY OF NoRTH AMERICAN Unionacea • Johnson 153

154 Bulletin Mtiscunt of Conifxirativc Zoology, Vol. lUJ, No. 2

Plate 3.

The distribution of Proptera capax (Green). This species appears to have had Pleistocene refugia in the Ozarkian Region,
but only below the Ozark Crest. It is presumed that capax reached eastern Lake Erie, and at least one tributary of
western Lake Ontario, dunng the Trent Stage, when the Maumee River flowed through the partially dry bed of Lake Erie
into the Wabash.

Zoogeography of North Amewcan U^io'sacea- Johnson 155

156 Bulletin Mtisi'uiu of (Uniipardtiic Zoolo^i/, Vol. 14!^), So. 2

Plate 4.

The distribution of Arcidens confragosus (Say). The unconfirmed record from the St. Lawrence River System: Great
Lakes Drainage (Lake Michigan). Illinois: Chicago, Cook Co. is represented by an open circle. Records from the James
River Drainage, South Dakota, of the Missouri River System, are beyond the limits of the map.

This species is w/idely distributed in the Gulf Coastal Region and in tributaries of the Mississippi south of the Ozark
Crest. It is possible that the Meramac River drainage was the Pleistocene refugium of the present Mississippian and
Ohioan populations. It is assumed that the one record from the Tennessee River is the result of the recent introduction
of fishes behind Kentucky Dam. If this is so, in all probability this species reached the Ohioan Region from the west in
post-glacial time, though it may have had a refugium in the Green River Drainage.

It is not impossible that A. confragosus reached Lake Michigan when the rivers of western Michigan, i.e. the Grand, the
Muskegon, and the Saginaw, connected with the Des Plaines River, through the Chicago outlet and Illinois River to the
Mississippi. Nevertheless, this record could not be confirmed and since it has never been reported elsewhere in Lake
Michigan, and since the open lake in the vicinity of Chicago does not appear to be a propitious environment for it, this
record is regarded as spurious.

ZOOCEOCRAPHY OF NORTH AMERICAN \JsiO\ACEA • Johnsoil 157

158 Bulletin Museum i)j Cumpdiaiiic Zoolv^ij, Vol- IM, Mo. 2

Plate 5.

The distribution of Anodonta suborbiculata Say. Records from tfie Big Blue and Elkhorn river drainages, Nebraska, of
the Missouri River System, are beyond the limits of the map.

The distribution of this species in the Gulf Coastal Region is discontinuous, as it is in the tributaries of the Mississippi,
south of the Ozark Crest. It is impossible to isolate a single refugium for the present Mississippian and Ohioan popu-
lations. It is assumed that the one record from the Tennessee River is the result of the recent introduction of fishes
behind Wheeler Dam. If this is so, then in all probability this species reached the Ohioan Region from the west in post-
glacial time, though it may have had a refugium in the Green River drainage.

Anodonta suborbiculata, like A. corpulenta, Vlllosa lienosa, Ligumia subrostrata, and Uniomerus tetralasmus, is a
species which does not usually inhabit large rivers, but is more often found in smaller streams, sloughs or lakes.

Z00GE0GR\PHY OF NORTH AMERICAN UnIONACEA • /(;/ni.so/l 159

^^

160 HulU'tin Museum of Coiitixirative Zoology, Vol. 149, No. 2

Plate 6.

The distribution of Cyprogenia aberti (Conrad) (squares) and Cyprogenia stegaria (Rafinesque) (dots).

Cyprogenia aberti (Conrad) is found only in the Ozarkian Region. Below the Ozark Crest it is widely distributed; above
the Crest, it is restricted to the Meramec River system.

Cyprogenia stegaria (Rafinesque). This species has persisted in the Cumberlandian Region, and may have had additional
refugia in the Green, Allegheny and Monongahela river drainages.

ZoOGEOGR\PHY OF \ORTH AMERICAN V'SIONACEA ' Jolliisoii 161

162 Bulletin Mu.scuni of Coinfxinitivc Zoology, Vol. 149, No. 2

Plate 7.

A. The distribution of Ptychobranchus occidentalis (Conrad).

Found only in the Ozarkian Region, below the Ozark Crest from the Red to the St. Francis river systems; above the
Crest, It is restricted to the Meramec River system.

Ptychobranchus fasciolare (Rafinesque) (Table 2 C, Table 3 A), a closely related species, that has persisted in the
Cumberlandian Region, and may have had additional refugia in the Green, Allegheny and Monongahela river drain-
ages, has a distribution in the Ohioan region, which is very similar to Cyprogenia stegaria (Rafinesque) (Plate 6,
dots). In addition, it spread into Lake Erie (Plate 1 D).

B. The distribution of Villosa reeviana (Lea).

Found only in the Ozarkian Region, below the Ozark Crest in the White River system; above the Crest, it occurs in
the Osage, Gasconade, and Meramec river systems.

Zoogeography of North American Unionacea -yo/i/iso/i 163

164 Bulletin Museum of (Uinipardtive Zoology, \'<)l. 149, No. 2

Plate 8.

A. The distribution of Fusconaia ozarkensis (Call) (dots), Arkansia wheeleri Ortmann and Walker (triangles), and Vlllosa
arkansasensis (Lea) (squares).

Fusconaia ozarkensis (Call) is found only in tfie Ozarkian Region, below tfie Ozark Crest, in the Ouachita River
drainage of the Black River system, the Neosho River drainage of the Arkansas River system, the White and St.
Francis river systems; above the Crest it is restricted to the Gasconade River system.

Arkansia wheeleri Ortmann and Walker is found only in the Ozarkian Region, belov»r the Ozark Crest, in the upper
Red and Black river systems.

Villosa arkansasensis (Lea) is found only in the Ozarkian Region, below the Ozark Crest, in the upper Black River
system.

B. The distribution of Actinonaias rafinesqueana (Frierson) (dots), and Lampsilis powelli (Lea) (triangles).

Actinonaias rafinesqueana (Frierson) is found only in the Ozarkian Region below the Ozark Crest, in the upper Red
and White river systems. The location of the Black River record is conjecture. This species is close to the Cumber-
landian species, Actinonaias pectorosa (Conrad). Lampsilis powelli (Lea) is found only in the Ozarkian Region, below
the Ozark Crest, in the Saline River drainage of the Black River system; the upper Arkansas, and upper White river
systems. The location of the White River record is conjecture. This species is close to the Cumberlandian species,
Lampsilis virescens (Lea).

Zoogeography of North American Vnionacea- Johnson 165

'j^ct

166 Bulletin Museum af (.oniixinitivc Zoology, Vol. 149, No. 2

Plate 9.

Fusconaia ozarkensis (Call)

Figure 1. Unio ozarkensis Call. Big Creek, |of Current River], Texas Co., Missouri. Paralectotype MCZ 5710 Length 55,
height 37, width 19 mm.

Figure 2. Pleurobema brevis subelliptica Simpson. Spring River, Hardy, Fulton Co., Arkansas. Paralectotype USNM
159952a. Length 43, height 28, vi/idth 16 mm.

Figure 3. Pleurobema utterbackll Frierson. White River, Hollister, Taney Co., Missouri. Holotype MCZ 271448, right valve.
Length 68, height 42, width 26 mm.

Figure 4. Quadrula flava sampsoniana Frierson. Elk River, McDonald Co., Missouri. Holotype MZUM 92226. Length 54,
height 37, width 19 mm.

Anodonta (Utterbackla) suborbiculata Say

Figure 5. Mercer Bayou, N Black Diamond, Miller Co., Arkansas. MCZ 293508. Length 71, height 58, width 26 mm.

Zoogeography of North American Unionacea -yo/i/ison 167

168 Bulletin Mii.scuin of C.oiiiixinilivc Zooloi^y, Vol. 14!^), So. 2

Plate 10.

Arcidens confragosus (Say)

Figure 1. Alasmidonta confragosa Say. Fox River of the [Little] Wabash [River, Richland Co., Illinois]. Topotype MCZ
53004. Length 88, height 68, width 45 mm.

Figure 2. Arcidens confragosa jacintoensis Strecker. San Jacinto River, Liberty Co., Texas. Topotype MCZ 23317. Length
82, height 57, width 32 mm.

Zoogeography of North American UNiONACEA-yo/mso/i 169

170 Bulletin Mtiscuin of Conipdratiic Zoolo^tj, Vol. 149, So. 2

Plate 11.

Arkansia wheeleri Ortmann and Walker

Figure 1. Arkansia wheeleri Ortmann and Walker. Old River |a bayou of the Ouachita River], Arkadelphia, Clark Co.,
Arkansas. Holotype MZUM 105514. Length 73.5, height 62, width 41 mm.

Actinonaias rafinesqueana (Frierson)

Figure 2. Lampsilis rafinesqueana Frierson. Illinois River, Moodys, Cherokee Co., Oklahoma. Holotype MZUM 87576.
Length 86, height 60, width 35 mm. Female.

Figure 3. Lampsilis rafinesqueana Frierson. Illinois River, Moodys, Cherokee Co., Oklahoma. Paratype MZUM 90665.
Length 69, height 43, width 26 mm. Female.

Figure 4. Lampsilis rafinesqueana Frierson. Illinois River, Moodys, Cherokee Co., Oklahoma. Paratype MZUM 90665.
Length 62, height 37, width 21 mm. Male.

Zoogeography of North American Unionacea -yohn.son 171

(

%

3

172 Bulletin Muscuin of Comparative Zoolo<iij, Vol. 149, No. 2

Plate 12.

Lampsilis (Lampsilis) abrupta (Say)

Figure 1. Ohio River, Cincinnati, Hamilton Co., Ohio. MCZ 37526. Length 82, height 63, width 40 mm. Male.

Figure 2. Unio abruptus Say. Wabash [River, New Harmony, Posey Co., Indiana). Figured holotype [lost] from Say.
Length 71, height 60, width 40 mm [based on original figure). Female.

Zoogeography of North American Union ace a 'Jo/ins on 173

174 Bullet in Museum nf Comparative Zoology, Vol. 149, No. 2

Plate 13.

Lampsilis (Lampsilis) higglnsi (Lea)

Figure 1. Unio higginsii Lea. (Mississippi River), Muscatine, [Muscatine Co.], Iowa. Holotype USNM 84823. Length 81,
height 59, width 48 mm. Male.

Figure 2. Mississippi River, near Aledo, Mercer Co., Illinois. MCZ 288383. Length 81, height 61, width 51 mm. Female.

Zoogeography of North American Unionacea • /o/i/ison 175

ITfi Rulli'tin Mitscttni of Comparative 7jOolo<s,y, Vol. 149, No. 2

Plate 14.

Lampsilis satur (Lea)

Figure 1. Unio satur Lea [Red River], Alexandria, [Rapides Parish], Louisiana. Holotype USNM 84505. Length 80, height
61, width 51 mm. Female.

Figure 2. Saline River, Benton, Saline Co., Arkansas. MZUM 33758. Length 74, height 55, width 35 mm. Male.

Zoogeography of North American Unionacea -yo/insoN 177

178 Bulletin Miiscudi of Comixinitivi' Zoolo^.!/, Vol. 149, \'(). 2

Plate 15.

Lampsilis excavata (Lea)

Figure 1. Saline River, Benton, Saline Co., Arkansas. MCZ 293585. Length 72, height 50, width 33. Male.

Figure 2. Saline River, Benton, Saline Co., Arkansas. MCZ 293585. Length 78, height 54, width 38 mm. Female.

Lampsilis powelli (Lea)

Figure 3. Unio powelli Lea. Saline River, Arkansas. Holotype USNM 85024. Length 80, height 56, width 29 mm. Male.

Figure 4. Unio powelli Lea. Saline River, Arkansas. Allotype USNM 85024a. Length 77, height 39, width 24 mm. Female.

Lampsilis hydiana (Lea)

Figure 5. Unio hydianus Lea. Teche River, [St. Landry Parish], Louisiana. Holotype USNM 85010. Length 64, height 37,
width 27 mm. Male.

Figure 6. Unio hydianus Lea. [Red River], Alexandria [Rapides Parish], Louisiana. Allotype USNM 85010a. Length 57,
height 37, width 28 mm. Female.

Figure 7. Unio placitus Lea. 1852. Alabama. Holotype USNM 85152. Length 34, height 22, width 15 mm. Male.

Zoogeography of North American Unionace a 'yo/inson 179

180 Bulletin Museum of Coiiipdnitivc Zoology, Vol. 149, No. 2

Plate 16.

Proptera capax (Green)

Figure 1. Mississippi River, Muscatine, Muscatine Co., Iowa. MCZ 5043. Length 79, height 59, width 54 mm.

Figure 2. Niagara River, Buffalo, Erie Co., New York. Collected by Elizabeth Letson, 1906, with the note, "Only one I
ever found." Buffalo Museum of Science (approx. nat. size).

Villosa arkansasensis (Lea)

Figure 3. Unio arkansasensis Lea. [Ouachita River] near Hot Springs, Garland Co., Arkansas. Holotype USNM 25710.
Length 40, height 28, width 18 mm. Male.

Figure 4. Saline River, Benton, Saline Co., Arkansas. MCZ 4989. Length 37, height 27, width 18 mm. Female.

Zoogeography of North American Unionacea • Johnson 181

182 Bttllctin Miiscuiii of C.oinixinitivc Zoology, Vol. 149, No. 2

Plate 17.

Villosa reeviana (Lea)

Figure 1. Unio reevianus Lea. (White River, Ell<ins, Washington Co., Arkansas). Holotype USNM 85025. Length 75, height
41, width 28 mm. Male.

Figure 2. Unio reevianus Lea. White River, Arkansas. Allotype USNM 85025a. Length 62, height 40, w^idth 24 mm. Female.

Figure 3. Unio breviculus Call. Jacks Fork |of Current River, Shannon Co.], Missouri. Lectotype MCZ 5020. Length 64,
height 40, width 24 mm. Male.

Figure 4. Unio breviculus Call. Big Creek [of Current River, Texas Co. [Missouri. Allotype MCZ 5023. Length 58, height
40, width 25 mm. Female.

Figure 5. Lampsilis brittsi Simpson. [Little Niangua River, Camden Co., Missouri. Topotype MCZ 293574, from J. P.
Britts. Length 47, height 30, width 18 mm. Female.

Villosa vibex (Conrad)

Figure 6. Lampsilis streckeri Frierson. Little Red River, Arkansas. Holotype MZUM 91075. Length 65, height 38, width
22 mm. Male.

Figure 7. Onion Creek [Colorado River system], Travis Co., Texas. MZUM 79921. Length 49, height 31, width 16 mm.
Male.

Figure 8. Onion Creek [Colorado River System], Travis Co., Texas. MZUM 79921. Length 53, height 32, width 19 mm.
Female.

Zoogeography of North American Unionacea 'yo/inson 183

184 Htillctiii Mttscinn of CoiDpanitirc 7a)()\o<^\i, Vol. 149, No. 2

Plate 18.

Cyprogenia stegaria (Rafinesque)

Figure 1. Obovaria stegaria Rafinesque. Ohio River. Lectotype ANSP 20215. Length 47, height 49, width 32 mm.

Figure 2. Unio irroratus Lea. Ohio River, Cincinnati. Hamilton Co., Ohio; T. G. Lea. Paratype USNM 84299; the figured
holotype is missing. Length 51, height 54, width 35 mm.

ZOOGEOCRAPHY OF NOHTH AviEmCAN VnIONACEA ' Johnson 185

186 Htillctin Muscitiu of Coinixirdtivc 'Zoology, Vol. 149, No. 2

Plate 19.

Cyprogenia stegaria (Rafinesque)

Figure 1. Cyprogenia irrorata pusllla Simpson. Green River, Mammoth Cave, Edmonson Co., Kentucky. Lectotype USNM
152004. Length 43, height 37, width 29 mm.

Figure 2. Tuscarawas River, New Philadelphia, Tuscarawas Co., Ohio. MCZ 14409. Length 50, height 45, width 27 mm.

Cyprogenia aberti (Conrad)

Figure 3. Saint Francis River, Wittsburg, Cross Co., Arkansas. MCZ 5492. Length 48, height 46, width 29 mm.

Zooceoc;r4phy of North American Unionacea • yo/in.son 187

188 Bulletin Muscuni of Coinparutivv Zoology, Vol. 149, No. 2

Plate 20.

Cyprogenia aberti (Conrad)

Figure 1 . Unio aberti Conrad. Rapids of Verdigris River, Chambers Ford [not located; Oklahomaj. Figured holotype [lost]
after Conrad. Length 40, height 34, width 15 mm.

Figure 2. Unio lamarckianus Lea. White River, Arkansas. Holotype USNM 84306. Length 37, height 30.5, w/idth 13 mm.

Figure 3. Unio popenoi Call. Fall River, Wilson Co., Kansas. Holotype MCZ 4943. Length 78, height 68, width 39 mm.

Figure 4. Meramec River, [near Franklin Co. line]. Crawford Co., Missouri. Specimen loaned by R. D. Oesch. Length 39,
height 37, width 24 mm.

Ptychobranchus occidentalis (Conrad)

Figure 5. Ptychobranchus clintonensis Simpson. Archies Fork of Little Red River, near Clinton, Van Buren Co., Arkansas.
Paratype MCZ 192371. Length 57, height 29, width 16 mm.

Figure 6. Unio occidentalis Conrad. Currant [Current] River, [Randolph "Co.], Arkansas. Figured holotype [lost] from
Conrad. Length 52, height 28, width 17 mm [based on original figure].

Zoogeography of North AMEmcAN UNiONACEA-yo/in.son 189

"^.

•*«-^

(US ISSN 0027-4100

SuUetin OF the

Museum of

Comparative

Zoology

/M c; 3 ^890

MUS. COMP. ZOOU
LIBRARY

FEB

A Phyletic Study of the Lake Tan^arlyllka-

Cichlid Genera Asprotilapia,

Ectodus, Lestradea, Cunningtonia,

Ophthalmochromis, and Ophthalmotilapia

KAREL F. LIEM

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBER 3
6 FEBRUARY 1981

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Breviora 1952-

bulletin 1863-

Memoirs 1864-1938

JOHNSONIA, Department of Mollusks, 1941-

OccASiONAL Papers on Mollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. I. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Tere-
dinidae (Mollusca: Bivalvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts, 02138, U.S.A.

© The President and Fellows of Harvard College 1980.

A PHYLETIC STUDY OF THE LAKE TANGANYIKA CICHLID
GENERA ASPROTILAPIA, ECTODUS, LESTRADEA,
CUNNINGTONIA, OPHTHALMOCHROMIS, AND
OPHTHALMOTILAPIA

KAREL F. LIEM'

CONTENTS

Abstract 191

Introduction 191

Materials and Methods 192

Osteological Aspects of Ectodus 193

Myological Aspects of Ectodus 195

Comparative Osteology 197

Comparative Myology 203

Phyletic Analysis 205

Acknowledgments 212

Abbreviations 212

Literature Cited 213

Abstract. On the basis of comparative osteology
and myology, Asprotilapia, Ectodus, Lestradea,
Cunningtonia, Ophthalmochromis, and Ophthal-
motilapia of Lake Tanganyika are hypothesized to
be members of a monophyletic lineage of cichlid
fishes. All members share the following suite of
characters: (1) the entopterygoid is widely separat-
ed from the palatine; (2) the posterior and dorsal
margins of the palatine form a 90° angle; (3) the
slender hyomandibula has a long symplectic pro-
cess and a very reduced hyomandibular flange; (4)
the anterior margin of the pterosphenoid is
notched; (5) the vertical depth of the metapterygoid
is shallow; (6) the operculum has a distinct auricular
process; (7) the transversus dorsalis muscle is re-
duced; and (8) the obliquus posterior muscle is en-
larged. These characters are considered specialized
when compared with the accepted generalized
mori^hology oi Astatotilapia. The phyletic relation-
ships of this lineage are documented by synapo-
morphies that distinguish subunits of decreasing
levels of universality within the assemblage. As-
protilapia represents a highly specialized branch
with six major skeletomuscular specializations. The
remaining five genera are pictured as a second lin-

' Museum of Comparative Zoology, Harvard Uni-
versity, Cambridge, Massachusetts 02138.

eage, of which Ectodus is the most generalized tax-
on. On the basis of recency of common descent,
Ophthalmochromis is synonymized with the genus
Ophthalmotilapia. Although there is no doubt that
the Ophthalmotilapia lineage has undergone ex-
tensive morphological radiation in both skull struc-
ture and dentition, the data on morphology, func-
tion, trophic ecology, and behavior of this and other
cichlid lineages have failed to establish unequivo-
cally that the morphological radiation is also adap-
tive. The morphological and functional pattern in
this lineage reinforces the paradox that morpholog-
ically and phylogenetically most specialized cichlid
taxa are not only remarkable specialists but also
jacks-of-all-trades.

INTRODUCTION

The precise phylogenetic interrela-
tionships of the endemic cichHds in Lake
Tanganyika are still unknown (e.g., Fryer
and lies, 1972: 507). Essentially the
problem is to distinguish groups that are
true, hierarchically evolved sister groups
showing increasing specialization or apo-
morphy, from groups that are gradal as-
semblages of polyphyletic ancestry
(Greenwood, 1974, 1979). Liem and
Stewart (1976) have shown that the lepi-
dophagous cichlids of Lake Tanganyika
represent a monophyletic lineage. Re-
cently Liem (1979) has postulated that
because of the basic homogeneity in spe-
cialized morphology and function, the in-
vertebrate picking cichlids of Lake Tan-
ganyika must have originated from a
common ancestor. Greenwood (1979) has
synonymized Limnotilapia with Simo-

Bull. Mus. Comp. Zool., 149(3): 191-214, Februar>-, 1981

192 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

chromis to show the close phyletic rela-
tionships of two genera previously pic-
tured as belonging to two different
lineages (Fryer and lies, 1972: 507). A
complete analysis of phyletic relation-
ships for the endemic cichlids is hin-
dered because of an apparent absence of
synapomorphic features which would en-
able one to build up the various levels of
relationship, and a phyletic analysis of
cichlids from other lakes and rivers. This
paper is part of a series dealing with the
establishment of monophyletic lineages.
Such an attempt represents a first step to-
ward the ultimate goal of interrelating
the different lineages on a sister group
basis.

This paper deals with six endemic gen-
era of Lake Tanganyika: Asprotilapia,
Boulenger, 1901; Ectodus, Boulenger,
1900; Lestradea, Poll, 1943; Cunningto-
nia, Boulenger, 1906; Ophthalmotilapia,
Boulenger, 1901; and Ophthalmochro-
mis. Poll, 1956. Except for Ectodus, this
assemblage is characterized as predomi-
nantly herbivorous, inhabiting rocky hab-
itats of Lake Tanganyika down to a depth
of 15 m. Ectodus is especially abundant
in shallower and sandy habitats. With the
exception of Ectodus descampsi, all
members of the group possess long intes-
tinal tracts (2.5-6 times the standard
length) and exploit filamentous and uni-
cellular algae, diatoms, crustaceans, and
insect larvae as food sources. Based on
comparative osteological and myological
data, I am presenting a hypothesis that
the assemblage is monophyletic, with
Asprotilapia as a very specialized form,
and Ectodus and Lestradea as the least
specialized representatives.

In this study only a limited number of
functional data have been gathered from
Lestradea and Ophthalmotilapia. How-
ever, much information from other cich-
lids (Liem, 1978, 1979, 1980) can be ap-
plied to this lineage in determining the
nature and possible evolutionary direc-
tion of the character complexes. A pro-
visional phylogenetic scheme is pre-
sented as a hypothesis, with each branch

defined by its unique morjihological spe-
cializations.

MATERIALS AND METHODS

Osteological studies are based on aliz-
arin preparations and some dried skele-
tons. Osteological nomenclature follows
Liem and Stewart (1976) and Barel et at.
(1976). The myology has been studied
from formalin fixed specimens, which are
stored in 70% ethyl alcohol. With very
few exceptions the nomenclature of Win-
terbottom (1974) and Anker (1978) is fol-
lowed. All drawings were made by
means of the Wild-M5 camera lucida.

The following specimens from the Brit-
ish Museum (BM) and the Museum of
Comparative Zoology (MCZ) have been
studied:

Asprotilapia leptura BM 1906.9.6.157
Asprotilapia leptura BM 671 (dried skeleton)
Aulonocranus dewindti BM 1960.9.30.4642-4656
Aulonocranus dewindti MCZ 49305
Callochromis macrops BM 1960.9.30.2821-2823
Callochromis melanostigma BM 1960.9.30.2845-

2859
Callochromis pleurospilus MCZ 49280
Cardiopharynx schoutedeni BM 1960.9.30.1574-

1615
Cunningtonia longiventralis MCZ 49243
Cunningtonia longiventralis BM 1960.9.30.1912-

1919
Ectodus descampsi BM 1961.11.22.113-119
Grammatotria lemairei MCZ 49277
Grammatotria lemairei BM 1960.9.30..3317-3330
Astatotilapia {"Haplochromis") burtoni BM

1960.9.30.2415-2433
Lamprologus elongatus BM 1960.9.30.6790-6793
Lamprologus elongatus MCZ 49251
Lamprologus moorei MCZ 49245
Leptochromis calliurus BM 1960.9.30.3351-3353
Lestradea perspicax perspicax BM 1960.9.30.1468-

1484
Limnochromis auritus BM 1960.9.30.1981-1985
Ophthalmochromis nasutus MCZ 49232
Ophthalmochromis ventralis BM 1960.9.30.1689-

1694
Ophthalmochromis ventralis MCZ 49232
Ophthalmotilapia hoops BM 1960.9.30.1720-1724
Ophthalmotilapia hoops BM 1960.9..30. 1716-1718
Sarotherodon nilotica MCZ 49312
Trematocara unimaculatum BM 1961.11.22.519-

528
Trematocara unimaculatum MCZ 49262
Tylochromis polylepis BM 1960.9.30.109-113
Xenotilapia ornatipinnis MCZ 49266
Xenotilapia sima BM 1961.11.22.208-211

CiCHLiD Phylogeny* Liem 193

Figure 1. Dorsal aspect of the neurocranium. A, Ectodus descampsi; B, Lestradea perspicax; C, Asprotilapia leptura;
D, Ophthalmotilapia ventralis; E, Ophthalmotilapia boops.

9,t^P^Tnn,^!i^^ ASPECTS ^ ^i^^^^^^ (Liem and Osse, 1975; Barel

OF ECTODUS ^^ ^i ^ 1^1^), such as the somewhat de-

The neurocranium of Ectodus retains curved dorsal profile to the preorbital

many primitive features found in Asia- face and the high cranial vault. However,

totilapia ("Haplochromis") burtoni and the ethmovomerine region is relatively

194 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

pa

pt

ps

eo
ic

V eoc
^boc

^^^-^^-^^'i^^'C::-^ ^

pa

sph soc i,..<t^

^

P§\^:^^^

Leo

^ St

j?r^

*^^boc

soc

eoc

boc

sph po

Figure 2. Lateral aspect of the neurocranium. A, Ectodus descampsi; B, Lestradea perspicax; C, Cunningtonia lon-
giventralis; D, Ophthalmotilapia boops; E, Ophthalmotilapia ventralis; F, Asprotilapia leptura.

CiCHLiD Phylogeny* Liem 195

long and is dorsally differentiated into a large, widely separated ascending pro-
well-developed rostral fossa, which is cesses of the dentary and anguloarticular.
made up by the ethmoid and the de- The ascending process of the anguloar-
pressed anteromedial wings of the fron- ticular is expanded posteriorly to form an
tals (Fig. lA:e,f). The ethmoid is sutur- adductor fossa, serving as the insertion
ally united with the ascending processes site of the A2 part of the adductor man-
of the vomer and with the lateral eth- dibulae muscle complex (Fig. 6A:am2).
moids (Fig. 2A:e,pf). The long axis of the The size, shape, and configuration of
basioccipital makes a 60-70° angle with the opercular apparatus (Fig. 5A) do not
that of the parasphenoid (Fig. 2). Both the differ from those of generalized cichlids.
basioccipital and parasphenoid bones In Ectodus the posterodorsal corner of
contribute to the formation of a promi- the operculum is raised to form an auric-
nent pharyngeal apophysis, which is ular process, which is not encountered in
therefore of the "Haplochromis" type, generalized cichlids.

Another specialization can be found in Of the well-developed five or six cir-
the pterosphenoid in the form of a deep cumorbitals, the large lacrimal is distinct-
notch in the anterior margin of the ptero- ly shaped and possesses six sensory pores
sphenoid (Fig. 2A:pts). The saccular bul- (Fig. 5A:la). The ventral margin is con-
la is greatly enlarged (Figs. 2A, 3A). vex, while the posterodorsal margin is

Morphologically the suspensory appa- concave. The anterodorsal margin is
ratus of Ectodus deviates from the more straight. Anteriorly the lacrimal ends in
generalized configuration of that in As- an anteriorly directed angular process.
tatotilapia hurtoni (Fig. 4A; Liem and The branchial skeleton is not de-
Osse, 1975). The hyomandibula is slen- scribed here because the elements retain
der with an elongate symplectic process a relatively generalized condition and do
and a much reduced flange area below not contribute to the solution of phylo-
the anterior condyle. The two hyoman- genetic relationships in this group of
dibular condyles are distinctly separated, cichlids.
while the symplectic process is connect-
ed to the metapterygoid, which is a char- %As/r\i n^r-ir^Ai Aooirr^-ro r\c
acteristically shallow bone. Ihe reduced p_.^_^.._
entopterygoid is separated from both the

ectopterygoid and the palatine. Half of The lateral head muscles of Ectodus
the entopterygoid's lateral surface is (Fig. 6A) retain their unspecialized con-
overlapped by the quadrate. The palatine dition in respect to topography, configu-
bone is unusual in having its posterior ration, and structure. Thus the descrip-
and dorsal border meet at a 90° angle, tion of the head muscles of As^afoff/apia
Posteriorly the suspensory apparatus is elegans by Anker (1978) also applies to
delineated by a large preoperculum of those of Ectodus. The adductor man-
which the outer rims of the horizontal dibulae complex (Ai, A2, A3, A^) of Ec-
and vertical limbs make a distinct 90° an- todus is identical to that of A. burtoni,
gle. The anterior border of the suspen- except that the muscle fibers in the for-
sory apparatus formed by the ectoptery- mer are considerably more elongate, a
gold and the palatine is oblique, making feature which is probably correlated with
a 60° angle with the horizontal plane. the much larger orbit and the longer and

The jaw apparatus in Ectodus (Fig. 4A) more shallow suspensory apparatus. The
possesses a generalized premaxilla re- levator arcus palatini in Ectodus (Fig.
sembling that of Astatotilapia burtoni, 6A:lap) is less voluminous than in A. bur-
hut the maxilla has a prominent earlike toni (Liem and Osse, 1975). Because the
postmaxillary process. The elongate, hyomandibular flange zone is much more
slender mandible is characterized by restricted, most of the fibers insert on the

196 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

muscular process of the metapterygoid. tion of pharyngol^ranchials 2 and 3,

The other lateral head muscles have not inserting on both bones. Levator internus

undergone any specialization and will not 2 (or posterior) passes ventrocaudally

be described here (Fig. 6A:aap,do,lo,ep). along the outer margin of the second

The ventral muscles of the head of Ec- pharyngobranchial to insert on the third

todus (Fig. 7A:gh,sh) do exhibit several pharyngobranchial near the junction of

specializations if compared with those of the third and fourth epibranchials. The

A. hurtoni and A. elegans. The left and levatores interni are equally developed,

right geniohyoideus anterior (Fig. cylindrical muscles. The levatores exter-

7A:gha) are clearly separated, cylindrical ni 1-3 are relatively slender, straplike

muscles, each attaching to the dentary muscles running caudoventrally to insert

near the mandibular symphysis ventral to on the dorsal aspects of the epibranchials

the small parallel-fibered intermandibu- 1-3 respectively. As in all cichlids (Liem,

laris. A transverse myosept interrupts the 1974), the fourth levator externus is the

fiber course of the geniohyoideus just an- dominant component of the externi com-

terior to the hyoid symphysis. From the plex, lying in a parasagittal plane. Its fi-

transverse myosept, the geniohyoideus bers pass ventrally to converge on the

posterior runs laterally to attach to the lat- muscular process of the lower pharyngeal

eral aspect of the hyoid above the first 2 jaw. The insertion appears as a complex

branchiostegals. The hyohyoideus com- mixture of muscular and tendinous ele-

plex (Fig. 7A:hhs,hhi) lies between the ments. Only a few lateral fibers attach to

hyoids, the branchiostegals and the me- the dorsal side of the fourth epibranchial

dial aspect of the opercular apparatus, by means of an aponeurotic plate.

We may distinguish three parts: (1) the The obliquus posterior muscle (Fig.

hyohyoideus inferior (Fig. 7A:hhi) aris- 8A:obp) is characteristically a highly de-

ing from the anteroventral corner of the veloped muscle in Ectodus originating

hyoid ramus runs caudally and slightly from the dorsal surface of the fourth epi-

lateral to insert on the first to fourth bran- branchial. From this expanded origin, the

chiostegal rays by splitting into four fibers run laterally and caudoventrally

heads; (2) the hyohyoideus superior (Fig. converging on the muscular process of

7A:hhs) runs between the rays and the the lower pharyngeal jaw. The insertion

medial aspect of the gill cover; and (3) site is medial and ventral to that of the

the hyohyoideus transversus (Fig. 7A:hht) fourth levator externus.

runs between the first left and first right Joining the insertion sites of the fourth

branchiostegals. The hyohyoideus trans- levator externus and obliquus posterior

versus is much better developed than in muscles on the muscular process of the

A. hurtoni. lower pharyngeal jaw is the fifth adductor

In the branchial musculature, the ven- muscle (Fig. 8A:ad). This spindle-shaped

tral muscles correspond with those of muscle attaches to the medial surface of

more generalized cichlids (e.g., Astatott- the dorsal end of the shank of the fourth

lapia elegans. Anker, 1978) with respect epibranchial adjacent to its articulation

to topography, structure, attachment sites ^i^h ^^e fourth ceratobranchial. From

and shape. Therefore, the description ^j^^^ P^i^t, the muscle runs ventrally to

here will be focussed mainly on the dor- 1^^ muscular process, to which it attaches

1 1 /!?• 7 A QA\ laterally and posteriorly to the insertion

sal muscles (rigs. <A, oA). •, c .\ r .^.v i • .

'' ' ' site of the fourth levator externus.

Both the levatores externi and interni j^ye two transversi dorsales muscles in
(Figs. 7A, 8A:le,li) originate from the Ectodus (Fig. 8A:tda,tdp) have under-
muscular process of the prootic. The le- gone a reduction as compared with those
vator internus 1 (or anterior) runs medi- of Astatotilapia elegans (Anker, 1978).
ally and ventrocaudally towards the junc- Basically, the transversi dorsalis anterior

CiCHLiD Phylogeny • Liem 197

muscles run between the second pha- trally to insert musculously on the pos-
ryngobranchials and the second epibran- terodorsal surfaces of both the third and
chials. The drastically reduced transver- fourth pharyngobranchials.
sus dorsalis posterior is a small, straplike Three ventral muscles, the pharyngo-
muscle which runs between the left and hyoideus and the pharyngocleithralis ex-
right fourth epibranchials. The fibers run ternus and internus, play an important
without interruption between the two at- role in the operation of the lower pharyn-
tachments. As in more generalized cich- geal jaw (Liem, 1974, 1978). Although
lids, the transversus dorsalis anterior, al- none of the three muscles (Fig.
though reduced in Ectodus, is 7A:ph,pci,pce) exhibits morphological
differentiated into four heads (Anker, specializations which can be used in a
1978). The posterior head (M. transversus phylogenetic analysis, they will be brief-
epibranchialis 2 pars dorsalis) arises mus- ly described as a basis for future, more
culously from the dorsal surface of the functionally focussed studies. As in most
anterior fork of the second epibranchial cichlids, the fibers of the pharyngohyoi-
and runs medially over the second pha- deus are interrupted by an aponeurotic
ryngobranchial towards a large aponeu- sheet halfway along the length of the
rosis. Very closely associated with the muscle. Anteriorly, the muscle attaches
posterior head is a more ventrally located to the urohyal. Near this insertion site,
component (M. transversus epibranchia- the left and right pharyngohyoideus
lis 2 pars ventralis) with its attachment to merge inseparably into one unit. Poste-
the junction of the second pharyngobran- riorly, the muscle fibers blend into a
chial and second epibranchial and con- strong tendon attached to the lateral sur-
necting with the median aponeurosis, face of the anteroventral portion of the
The third head (M. transversus pharyn- lower pharyngeal jaw. The pharyngo-
gobranchialis 2) is the anterior one, run- cleithralis externus arises from the an-
ning between the rostromedial surfaces terolateral surface of the cleithrum from
of the left and right pharyngobranchials. nearly the ventral tip and dorsally until
Finally, the fourth head (M. cranio-pha- about one-third of its vertical depth,
ryngobranchialis 2) arises from the neu- From this origin, the fibers run in a par-
rocranium just anterior to the articulation allel fashion dorsally and slightly ante-
with the first pharyngobranchial, and riorly converging on an aponeurotic sys-
runs posterolaterally and ventrally to in- tern which merges with the insertion
sert on the lateral edge of the anterior as- tendon of the pharyngohyoideus. How-
pect of the second pharyngobranchial. ever, a subdivision of this aponeurotic

The obliquus dorsalis muscle (Fig. system takes up a more lateral course to

8A:od) originates from the area lateral to attach on the fourth epibranchial. The

the pharyngeal apophysis of the third pharyngocleithralis internus originates

pharyngobranchial and runs caudolater- from a fossa to a point about midway on

ally to insert on both the dorsal aspect of the vertical depth of the cleithrum; its fi-

the third epibranchial just medial to the bers run anteroventrally to converge on

insertion site of the third levator externus a tendon which inserts on the lower pha-

and the anterodorsal aspect of the fourth ryngeal jaw just medial and posterior to

epibranchial. the insertion site of the pharyngohyoi-

The large retractor dorsalis muscle deus muscle.
(Fig. 8A:rd) runs parasagittally close to

the medial plane. Its extensive origin is COMPARATIVE OSTEOLOGY
irom the exoccipital, the vertebral bodies

of the first three vertebrae, and the Having outlined the main morpholog-

apophysis of the third vertebra. From ical features of the most generalized rep-

these points, the fibers run anteroven- resentative, i.e., Ectodus, we can consid-

198 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

Figure 3. Ventral aspect of the left half of the neurocranlum. A, Ectodus descampsi; B, Lestradea perspicax; C, As-
protilapia leptura; D, Ophthalmotilapia ventralis; E, Ophthalmotilapia boops.

CiCHLiD PhylogenY' Liem

199

er the anatomy of the other members of
the assemblage. The comparative de-
scription will not be a comprehensive
one. Instead, emphasis will be placed on:
(1) features that deviate from those found
in Ectodiis; (2) structural specializations
characterizing each genus; and (3) mor-
phological elements which can be used
to interrelate the different lineages on a
sister group basis.

Neurocranium (Figs. 1, 2, 3). The neu-
rocranium of Lestradea is identical to
that of Ectodiis except for a somewhat
shorter ethmovomerine complex (Fig.
2B). In Cunningtonia the ethmovomer-
ine complex is distinctly shorter, while
the basioccipital is excluded from the
pharyngeal apophysis (Fig. 2C). The eth-
movomerine region in Ophthalmotilapia
is greatly abbreviated so that the prefron-
tal is vertically oriented. The neurocrania
of Ophthalmotdapia and Ophthalmo-
chromis resemble each other very closely,
both having enlarged sensory canals and
foramina. As in Cunningtonia, the basi-
occipital is excluded from the pharyngeal
apophysis (Fig. 3C,D,E). However, in
Ophthalmochromis the abbreviation of
the ethmovomerine region is much less
pronounced, while the enlargement of
sensory canals and foramina is much
greater. Ophthalmochromis and
Ophthalmotilapia differ in the orienta-
tion of the ventral process of the basi-
sphenoid, being vertical in the latter and
oblique in the former (Fig. 2:bs). The
most specialized neurocranium is repre-
sented in Asprotilapia: the ethmovomer-
ine region is elongate with the ethmoid
in an almost horizontal position; the lat-
eral ethmoids are enlarged, forming pro-
nounced preorbital processes; the inter-
orbital width is much reduced; the orbit
is greatly enlarged; and the sensory ca-
nals and foramina are reduced in size.

Opercular apparatus (Fig. 5). The
opercular apparatus of all members of
this lineage conforms with that of Ecto-
diis.

Suspensory apparatus (Fig. 4). The de-

scription of the suspensory apparatus for
Ectodiis also applies to the other five
genera of this lineage. However, in Cun-
ningtonia the palatine is further special-
ized by the development of a posteriorly
directed slender process at the postero-
dorsal corner. Although the vertical
depth of the suspensorium at the level of
the metapterygoid is characteristically
shallow for all members of this lineage,
the feature is least developed in
Ophthalmotilapia and most pronounced
in Cunningtonia. Cunningtonia also pos-
sesses a very elongate and nearly hori-
zontal symplectic, and the horizontal
limb of the preoperculum is longer than
the vertical one. In Ectodiis the horizon-
tal and vertical limbs of the preopercu-
lum are of equal length, while in Aspro-
tilapia, Lestradea, Ophthalmochromis,
and Ophthalmotilapia the horizontal
limb of the preoperculum is the shorter
one. The sensory canals and pores of the
preoperculum are enlarged in Ophthal-
mochromis and Ophthalmotilapia.

Jaw apparatus (Fig. 4). As discussed
above, the mandible of Ectodiis is spe-
cialized in having an extensive shelflike
adductor fossa in the anguloarticular and
a laterally expanded ascending process of
the dentary. The mandible of Lestradea,
Cunningtonia, Ophthalmochromis, and
Ophthalmotilapia exhibits a more gen-
eralized configuration with very restrict-
ed adductor fossae. The most specialized
features are found in the mandible oi As-
protilapia: the lower jaw is elongate and
quite slender, with a drastic reduction in
the depth of the unit; the large ascending
processes of the dentary and articular are
widely separated and form the bulk of the
mandible. Morphologically, the premax-
illae and maxillae in this lineage are very
uniform, resembling those of Ectodus.
But in Lestradea the body of the maxilla
is stout, being only as long as the neck
and associated condyles, and having a
pronounced posteriorly directed post-
maxillary process. In general, the ascend-
ing processes are shorter than in Ecto-

200

Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

appm

ppm

mx

Figure 4. Lateral aspects of the suspensorium, mandible, and upper jaw. A, Ectodus descampsi; B, Lestradea perspi-
cax; C, Cunningtonia longiventralis; D, Asprotilapia leptura, in which only parts of the suspensorium are depicted, since
the elements have been disarticulated; E, Ophthalmotilapia ventralis; F, Ophttialmotilapia boops.

dus. However, both the premaxilla and diiced and shifted into a more forward

maxilla in Asprotilapia (Fig. 4D) exhibit position. The shape of the shortened

several unique specializations. In the maxilla deviates greatly from that of other

premaxilla, the articular process is re- cichlids. The greatly enlarged cranial

CiCHLiD Phylogeny* Liem 201

^ pts bs

sph^pa

B

pts

soc

sph ^.^-^

bs

pf

eo

appm^

mx

"^ms.

'^ ' "^V;;;^rbp^P^'"

soc

d ^cf^/

ra

M-

/ -■' ^' / \

-^fo

I •"*

"^V V \: \ /^

-/^sop

■ :V eA^-'^y;:

T K*-'P

PS mpt ^y— ^^^--^

sph Ppr ^

\:

mpt sy^

Figure 5. Lateral aspect of the skull. A, Ecfodus descampsi; B, Lestradea persp/cax; C, Cunningtonia longiventralis;
D, Ophthalmotilapia ventralis; E, Ophthalmotilapia boops.

202 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

do ep

do ep

artTr7+— ^r' P°P
lap lop

do lo

pop

Figure 6. Lateral aspect of the cephalic musculature after removal of the lacrimal, circumorbitals, and eyes. A, Ectodus
descampsi; B, Lestradea perspicax; C, Cunningtonia longiventralis; D, Ophthalmotilapia ventralis; E, Ophttialmotilapia
boops; F, Asprotilapia leptura; G, deeper dissection of the adductor mandibulae complex after removal of part A. in
Asprotilapia.

CiCHLiD Phylogeny • Liem

203

condyle and premaxillary process consti-
tute the bulk of the bone, with its dorsal
extremity directed anteriorly as a flat-
tened process. The angle between the
body of the maxilla and the part bearing
the cranial condyle and premaxillary pro-
cess is more acute than that of other
members of the lineage.

Dentition. Much of cichlid classifica-
tion and phylogeny has been based sole-
ly on dental characters. In this lineage
dental characters are of little value in es-
tablishing phyletic relationships at the
generic level, because the polarity of
character transformation is not known,
and intraspecific and ontogenetic plastic-
ity further complicates the use of dental
characters. No new information is added
here to our already extensive knowledge
on the teeth of members of this lineage
(Poll, 1943; Poll and Matthes, 1962). In
Ectodiis descampsi the teeth are conical
and arranged in two to three rows in such
a fashion that the outer rows on the lower
jaw are directed outward. Lestradea per-
spicax perspicax, a subspecies from the
northern half of Lake Tanganyika, pos-
sesses conical teeth arranged in three to
four rows. The teeth of the outer row are
much larger and are directed outward. In
juveniles of L. perspicax stappersi the
teeth are invariably tricuspid, while in
the adult all teeth can become conical.
Cunningtonia, Ophthalmochromis, and
Ophthalmotilapia have teeth with long
flexible stalks, giving them a mobile na-
ture, which is regarded as an adaptation
to raking and scraping algae off irregular
rock surfaces. In Cunningtonia the tri-
cuspid teeth are numerous, forming a
broad band in each jaw. Characteristical-
ly, the distal end of each tooth is bent
backwards rather sharply. In Ophthal-
motilapia most teeth are tricuspid
throughout the life history of the fish, ex-
cept for the conical ones toward the outer
corners of the jaws. It is remarkable that
the dental patterns in the subspecies of
Lestradea closely parallel those in the
subspecies of Ophthalmochromis.
Ophthalmochromis ventralis ventralis is

a conical-toothed form, geographically
restricted to the northern half of Lake
Tanganyika. In the southern subspecies,
Ophthalmochromis ventralis heterodon-
tus, the juveniles have tricuspid teeth,
while later in development a mixed den-
tition of tricuspid and conical teeth ap-
pears. Asprotilapia has small teeth with
very long slender stalks and expanded
tricuspid crowns arranged in two rows.

COMPARATIVE MYOLOGY

The lateral head musculature in mem-
bers of this lineage is relatively uniform
(Figs. 6, 7, 8). The description given for
Ectodus also applies to other members of
the lineage except for Asprotilapia,
which exhibits many unique specializa-
tions in the lateral head musculature. In
Asprotilapia (Fig. 6F,G) the elongate ad-
ductor mandibulae part Ao is the domi-
nant muscle occupying the greater part of
the cheek region. From their origin on
the preoperculum, the fibers run parallel
to insert on the extensive adductor fossa
of the mandible. Part A, of the adductor
mandibulae complex in Asprotilapia is
a small muscle originating from the pre-
operculum and symplectic processes of
the hyomandibula; its fibers insert on a
tendon at a point about halfway between
origin and insertion. Thus the tendon of
part A] is exceptionally long; just beyond
the ascending process of the articular it
gives off a ventral branch which merges
with the tendon of the intramandibular
(A^v) part of the adductor mandibulae
complex. The A, tendon inserts on the
medial aspect of the maxilla in the area
between the cranial condyle and premax-
illary process. The adductor mandibulae
complex of Lestradea, Cunningtonia,
Ophthalmochromis, and Ophthalmoti-
lapia deviates from that of Ectodus in the
dominance of the A, part. Both in cross-
sectional area and in volume, the A, is
the largest component of the complex
(Fig. 6).

The configuration of the remaining lat-
eral head muscles and the ventral head

204 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

bsr

Figure 7. Ventral aspect of the ventral cephalic musculature with the hyohyoideus transversus partially or In B, entirely
removed. A, Ectodus descampsi; B, Asprotilapia leptura; C, lateral view of the branchial musculature after removal of
the gills and mucous membranes In Ophthalmotilapia nasutus.

CiCHLiD Phylogeny • Liem

205

muscles in all members of this lineage
conforms with that of Ectodus (Fig. 6).

In general, the branchial musculature
of members of this lineage resembles that
of Ectodus (Figs. 7, 8). In Asprotilapia
the transversus dorsalis anterior muscles
have been reduced to only one head, run-
ning between the 2nd pharyngobranchi-
als (Fig. 8C:tda). Ophthalmotilapia ex-
hibits two specializations in the dorsal
branchial musculature: both the obliquus
dorsalis and retractor dorsalis are greatly
enlarged and are subdivided into two dis-
tinct heads (Fig. 8D:od,rd).

PHYLETIC ANALYSIS

Although an in depth study on phyletic
relationships is not yet available, it is
generally accepted that the endemic la-
custrine cichlids of Lake Tanganyika
have been derived from several unrelat-
ed fluviatile lineages. The fluviatile gen-
era ''Tilapia,' Tylochromis, "Haplo-
chromis," and Lamprologus are
considered the primitive sister lineages
of the many derived lineages of endemic
cichlids in Lake Tanganyika (e.g., Regan,
1920; Fryer and lies, 1972). Unfortu-
nately, no attempts have been made to
interrelate the different lineages witliin
Lake Tanganyika on a sister group basis.
Since the phyletic relationships of the la-
custrine lineages are still poorly known,
the determination of their sister group in-
terrelationships with the more general-
ized fluviatile lineages will be postponed
until a much broader data base is avail-
able. Without the information on the re-
lationships of the lacustrine species and
their respective fluviatile sister lineages,
it is impossible to make precise outgroup
comparisons. Despite these limitations,
it is possible to construct a classification
from which a theory of relationships is
recoverable. Clearly its precision and ef-
ficiency can be improved, but that must
await a phyletic analysis of fluviatile taxa.

In this phyletic analysis comparisons
have been made between the lacustrine
assemblage Asprotilapia, Ectodus, Les-

tradea, Cunningtonia, OpJithalmo-
chromis, and Ophthalmotilapia and the
fluviatile taxa Astatotilapia burtoni and
Sarotherodon nilotica. Tylochromis,
Sarotherodon, Tilapia, and Lamprologus
are specialized along very different lines
(Liem, in prep.) and are not closely re-
lated to the Ophthalmotilapia lineage.
Although an outgroup comparison is ten-
tative, I have considered the morphology
of A. burtoni to be generalized (Liem, in
prep.) and therefore representative of a
plesiomorphous group with which the
apomorphous Ophthalmotilapia assem-
blage can be compared. In the succeed-
ing phyletic analysis I have made the ba-
sic assumption that the morphological
condition of the various structures in As-
tatotilapia burtoni and A. elegans rep-
resents the primitive state. Any devia-
tions from this basic morphological
configuration are considered specializa-
tions. Monophyletic lineages are then
formulated on the basis of shared spe-
cialized characters.

The Ophthalmotilapia Assemblage
AS A Monophyletic Lineage

A comparison of the osteology and
myology of all cichlid genera of Lake
Tanganyika and the four fluviatile species
Sarotherodon nilotica, Astatotilapia
burtoni, Tylochromis microlepis, and
Lamprologus congolensis has resulted in
the identification of a suite of eight de-
rived characters shared by Asprotilapia,
Ectodus, Lestradea, Cunningtonia,
Ophthalmochromis, and Ophthalmoti-
lapia. This group of genera will be re-
ferred to as OA (Ophthalmotilapia As-
semblage) in the following discussion.

The entopterygoid is separated from
the palatine in the OA (Fig. 4:ent). In
generalized cichlids the entopterygoid is
attached to the posterior margin of the
body of the palatine (Goedel, 1974; Van-
dewalle, 1972; Liem and Osse, 1975;
Barel et al., 1976) or the two elements
abut against each other. A similar con-
nection is also present in several special-

206 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

ebi.4

Figure 8. Dorsal aspect of the branchial musculature. A, Ectodus descampsi, in which the dissection has been carried
out in greater detail to show the precise insertions and origins of the muscles; B, Lestradea perspicax; C, Asprotilapia
leptura; 0, Ophthalmotilapla boops.

ized lineages of Lake Tanganyika, such as
the herbivores (e.g., Simochromis and
Petrochromis), piscivores (Liem, 1978),
scale eaters (Liem and Stewart, 1976) and
invertebrate pickers (Liem, 1979). Sepa-
ration of the palatine and the entoptery-
goid (Fig. 4:ent,p) is considered as a de-
rived character state. This specialization
also occurs in Xenotilapia (Liem and
Osse, 1975), Aiilonocraniis, Callochro-
mis, and Cardiopharynx indicating the
possibility that these genera are related
to the OA. However, on the basis of the
present data it is impossible to ascertain
that Xenotilapia, Aulonocranus, Calloch-
romis, and Cardiopharynx are members
of a sister lineage of the OA.

The functional significance of the sep-
aration of the entopterygoid from the pal-
atine remains unclear, because no corre-
lation can be found between this

specialized feature and specializations in
either the mobility and shape of the sus-
pensorium and jaws, or the morphologi-
cal characteristics of the adductor man-
dibulae complex.

A unique specialization of the OA can
be found in the palatine bone (Fig.
4E,F:p). The dorsal margin and the ver-
tically directed posterior border of the
palatine meet at a 90° angle. Because the
palatine is expanded in the region of this
90° angle, it has a characteristic shape
when viewed laterally (Fig. 4:p). The 90°
posterodorsal angle surmounting a pos-
terodorsal expansion of the palatine is not
found in any of the other Lake Tangan-
yika cichlids and deviates from the con-
dition in generalized cichlids (e.g., Asta-
totilapia burtoni, Liem and Osse, 1975;
A. elegans, Barel et al., 1976). This
uniquely specialized shape of the pala-

CiCHLiD Phylogeny • Liem

207

tine is found in all members of the OA,
indicating that the group may represent
a monophyletic lineage.

The hyomandibula of cichlids exhibits
remarkable diversity in shape and size.
However, among the diverse forms one
can recognize distinct themes, each of
which may reflect a particular ancestry.
The hyomandibula of Astatotilapia bur-
toni (Liem and Osse, 1975) and A. ele-
gans (Barel et al., 1976) is characteristi-
cally stout, with an expanded flange
associated with the relatively short sym-
plectic process. In Tilapia and Saro-
therodon (Vandewalle, 1972; Goedel,
1974) the symplectic process is elongate
and the hyomandibular flange below the
anterior head is only moderately devel-
oped. Both Lamprologus and Tylo-
chromis possess a strongly modified hyo-
mandibula. The slender hyomandibula of
all members of the OA (Fig. 4:hm) resem-
bles that of Tilapia, but it has a relatively
longer symplectic process, and has lost
its flange. However, a reduced flange is
present in Asprotilapia. Because the spe-
cialization of the hyomandibula in mem-
bers of the OA has progressed further
than in Tilapia, the feature is considered
derived and indicative of the monophy-
letic nature of the group. A similarly spe-
cialized hyomandibula is found in Xen-
otilapia.

The anterior margin of the pterosphe-
noid in members of the OA is peculiarly
notched (Fig. 2:pts). A liplike process at
the base of the notch serves as the origin
of a strong ligament which attaches to the
eye ball. Although the notched anterior
margin of the pterosphenoid and the as-
sociated ligament occurs in some other
specialized cichlids, it is absent in the
generalized forms. The presence of the
pterosphenoid notch in all members of
the OA is considered as supporting evi-
dence for the hypothesis of monophyly of
the OA.

All members of the OA possess a very
shallow suspensory apparatus at the level
of the metapterygoid (Fig. 4). This spe-
cialization is the result of a sharp de-

crease in the vertical depth of the meta-
pterygoid, which in all generalized
cichlids has a much greater vertical
depth. Actually among Lake Tanganyika
cichlids this specialization and the pe-
culiar position of the ectopterygoid is
characteristic of the OA assemblage.

Although the presence of a well-devel-
oped auricular process (Fig. 5:op) on the
operculum occurs in several specialized
cichlid groups, the character is here in-
cluded in the suite of derived characters,
upon which the hypothesis of monophyly
of the OA is based.

Two myological specializations in the
branchial musculature of all members of
the OA offer further evidence for the
probable monophyletic nature of the
group. In sharp contrast to the condition
in generalized cichlids, there is a clear
reduction in the development of the
transverse dorsalis muscle accompanied
by a hypertrophy of the obliquus poste-
rior muscle (Fig. 8:obp). Reduction of the
transversus dorsalis and hypertrophy of
the obliquus posterior have not been
found in Astatotilapia elegans (Anker,
1978) and other groups of specialized
cichlids in Lake Tanganyika. In view of
this admittedly tentative evidence, the
reduction of the transversus dorsalis and
hypertrophied obliquus posterior are
considered derived characters support-
ing the notion that the OA is actually
monophyletic.

The OA is viewed as a monophyletic
lineage (Fig. 9: cladogram of synapomor-
phies [1]-[301) because its members share
the following suite of characters:

(1) The entopterygoid is widely sepa-
rated from the palatine.

(2) The posterior and dorsal margins of
the palatine form a 90° angle.

(3) The slender hyomandibula has a
long symplectic process and no or a
very reduced hyomandibular flange.

(4) The anterior margin of the ptero-
sphenoid is notched.

(5) The vertical depth of the metapter-
ygoid is shallow.

208 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

(6) The operculum has a distinct aiiric-
uhir process.

(7) The transversus dorsaHs is reduced.

(8) The obHquus posterior is enlarged.

No other cichlid in Lake Tanganyika
possesses this suite of derived characters,
although none of the above mentioned
characters, with the possible exception of
(2), is autapomorphic. The sister group of
the Ophthalmotilapia lineage is still un-
known. Possibly, Xenotilapia, Callo-
chromis, and Aulonocranus may repre-
sent members of a sister group of the
Ophthalmotilapia lineage since they do
share two of the above mentioned de-
rived characters (i.e., [1] and [3]). How-
ever, the precise relationships of this lin-
eage to others must await more data and
further analysis now in progress.

This phyletic hypothesis for the exis-
tence of an Ophthalmotilapia lineage dif-
fers from the previous hypotheses (Fryer
and lies, 1972, and Regan, 1920) by the
inclusion of Ectodus. In the previous
schemes, Ectodus has been regarded as
a derivative of a ''Haplochromis" or
"Haplochromis-\ike" ancestor, while the
other members of the Ophthalmotilapia
lineage were seen as derivatives of "one
or more unknown ancestors," which are
not related to "Haplochromis J' Unfortu-
nately the inferred relationships of the
relevant genera in the previous hypoth-
eses have been presented with virtually
no documentation, and were not based
on Hennigian methodology.

Phyletic Relationships of

Members of the

Ophthalmotilapia Lineage

The proposed phyletic relationships
are depicted in the cladogram in Fig. 9.
Two major lineages are recognized: one
contains the monotypic genus Asproti-
lapia, and the second lineage contains
the five remaining genera. The following
discussion deals first with the synapo-
morphies for the entire complex, and
then with those derived characters that

distinguish subunits of decreasing levels
of universality within the assemblage.

Asprotilapia represents a very special-
ized lineage, with six autapomorphic
characters:

(9) The elongate, slender mandible has
an expanded adductor fossa (Fig.
9:9) serving as the insertion site for
the adductor mandibulae part A2
(Fig. 6F,G).

(10) The posterior head of the transver-
sus dorsalis anterior is absent (Figs.
9:10, 8C:tda).

(11) The lateral ethmoids are greatly en-
larged (Figs. IC, 3C:pf).

(12) The interorbital width of the neu-
rocranium is drastically reduced
(Fig. IC).

(13) The reduced articular process of the
premaxilla is in a more forward po-
sition (Fig. 4D:ap).

(14) The greatly enlarged cranial con-
dyle and the premaxillary process
constitute the bulk of the maxilla
(Fig. 4D:cc,pmp).

The rare species Asprotilapia leptura
represents a highly specialized lineage,
with a very elongate body and a much
attenuated caudal region. Because of the
ventral mouth and the tricuspid teeth ar-
ranged in two series on both jaws, As-
protilapia is considered herbivorous,
feeding primarily on epilithic algae.

The remaining five genera are regard-
ed as a monophyletic group on the basis
of their sharing two derived characters:

(15) Anteriorly the enlarged lachrimal is
differentiated into a distinct process
(Figs. 9:15, 5:1a). The blunt process
is directed anteroventrally, except
in Cunningtonia in which it is di-
rected anterodorsally.

(16) The adductor mandibulae part A,
has become the dominant compo-
nent of the adductor mandibulae
complex (Fig. 6:ami). Its cross-sec-
tional area surpasses that of the oth-
er parts. The origin from the pre-
operculum has expanded ventrally

CiCHLiD Phylogeny • Liem

209

ASPROTILAPIA ECTODUS LESTRADEA OPHTHALMOTILAPIA CUNNINGTONIA

VENTRALIS NASUTUS BOOPS

Figure 9. Proposed cladogram depicting the phylogenetic relationships of the Ophthalmotilapia lineage on the basis
of the following synapomorphies: (1) entopterygoid separated from palatine; (2) posterior and dorsal margins of palatine
form a 90 angle; (3) slender hyomandibula with long symplectic process and reduced hyomandibular flange; (4) ante-
rior margin of pterosphenoid notched; (5) vertical depth of metapterygoid shallow; (6) operculum with auricular
process; (7) transversus dorsalis muscle reduced; (8) obliquus posterior muscle enlarged; (9) elongate, slender
mandible with expanded adductor fossa; (10) transversus dorsalis anterior absent; (11) prefrontals greatly enlarged;
(12) interorbital width greatly reduced; (13) reduced articular process of premaxilla in forward position; (14) cranial
condyle and premaxillary process of maxilla greatly enlarged; (15) lacrimal enlarged with a distinct anterior process
and elaborate sensory system; (16) adductor mandibulae part A, is dominant; (17) saccular bulla enlarged; (18)
horizontal and vertical limbs of preoperculum of equal length; (19) elongate gut which is at least 2.5 times the
fish's standard length; (20) edentulous anterior process of the lower pharyngeal jaw is only half as long as the
toothed part; (21) stout body of maxilla with prominent postmaxillary process; (22) jaw teeth mobile with long stalks;
(23) first pelvic fin ray is greatly elongate; (24) straight vertical limb of preoperculum forms a 90' angle with horizontal
limb; (25) symplectic elongate; (26) a spinelike process at junction of dorsal and posterior margins of palatine;
(27) long-stalked tricuspid jaw teeth with posteriorly recurved crowns; (28) distal end of elongate first ray uniquely
bifid and widened into spatulae; (29) trend toward enlarged sensory canals and pores; (30) hypertrophied re-
tractor dorsalis muscle subdivided into two distinct heads.

at the expense of the adductor man-
dibulae part Ag.

Ectodiis inhabits shallow waters of
sandy areas, feeding on a mixed diet of
insect larvae, crustaceans, and filamen-
tous algae. According to the present phy-
logenetic hypothesis, Ectodus represents
the most primitive of the Ophthcdmotila-
pia lineage. Two autapomorphies distin-
guish Ectodus from the other genera:

(17) The saccular bulla is greatly en-
larged (Figs. 2, 3).

(18) The horizontal and vertical limbs of
the preoperculum are of equal
length (Figs. 4, 5:pop).

The next monophyletic subunit is com-

posed of Lestradea, Ophthalmotilapia,
Ophthalmochromis, and Cunningtonia,
which all share one notable feature:

(19) The gut is elongate and is at least
2.5 times the fish's standard length.

Lestradea possesses a gut which is 2.5-4
times the standard length. In Ophthal-
motilapia and Ophthalmochromis the
gut measures 3-3.5 times the standard
length. The longest gastrointestinal tract
is that of Cunningtonia, measuring 5-6
times the fish's standard length. The rel-
ative increase in the length of the diges-
tive tract is correlated with an increased
emphasis on a herbivorous diet.

Within this subunit, Lestradea can be

210 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

distinguished by two autapomorphic
characters:

(20) The edentulous anterior process of
the lower pharyngeal jaw is only
half as long as the toothed part.

(21) The body of the maxilla is stout and
has a prominent postmaxillary pro-
cess (Fig. 4).

Cunningtonia, Ophthalmochromis , and
Ophthalmotilapia form a distinct sub-
unit sharing three specialized characters:

(22) The jaw teeth have long stalks and
are mobile.

(23) The first pelvic fin ray is greatly
elongate.

(24) The posterior margin of the vertical
limb of the preoperculum is straight
and forms a 90° angle with that of
the horizontal limb (Fig. 4).

Within this monophyletic subunit,
Cunningtonia possesses three autapo-
morphic features:

(25) The symplectic is very elongate,
lengthening the suspensorium (Fig.
4:sy).

(26) At the 90° junction of the dorsal and
posterior margins of the palatine is
a spinelike process directed dorso-
posteriorly (Fig. 4:p).

(27) The long-stalked tricuspid jaw teeth
have posteriorly recurved crowns
(Fig. 6).

Finally, the close relationship of
Ophthalmochromis and Ophthalmotila-
pia can be established on the basis of
three derived characters:

(28) The distal end of the very elongate
first ray of each pelvic fin in the
male is uniquely bifid and widened
into spatulae.

(29) All members of this subunit show a
trend toward enlargement of the
sensory canals and pores of the head
(Figs. 1, 2, 5).

(30) The hypertrophied retractor dorsalis
is subdivided into two distinct heads
(Fig. 8:rd).

The genus Ophthalmochromis was es-
tablished by Poll (1956) on the basis of
O. ventralis, its type species possessing
conical teeth in adults as well as juve-
niles. Ophthalmotilapia hoops, on the
other hand, has mainly tricuspid teeth,
although conical teeth are present on the
posterolateral corner of the jaws. The
neurocranium of O. hoops has a shorter
ethmovomerine region and a steeper eth-
movomerine slope than that of Ophthal-
mochromis ventralis. However, the con-
dition with respect to the ethmovomerine
region in the skull of Ophthalmochromis
nasutus is intermediate between that of
hoops and ventralis. The validity of
Ophthalmochromis and Ophthalmotila-
pia as distinct genera is questionable.

Originally the distinction between
Ophthalmochromis and Ophthalmotila-
pia was based on morphological gaps in
the general and skull morphology. With
the discovery of Ophthalmochromis na-
sutus, the few morphological gaps of
characters in the skulls of ventralis and
hoops are bridged. Ophthalmochromis
nasutus has an ethmovomerine slope in-
termediate between those of ventralis
and hoops. The subtle differences in the
decurvature of the skull anteriorly be-
tween Ophthahnochromis and Ophthal-
motilapia are well within the range
found intragenerically in Perissodus
(Liem and Stewart, 1976). In the related
genus Lestradea, the contrast between
tricuspid and conical teeth has been in-
terpreted to indicate a difference at the
subspecies level for, respectively, L. per-
spicax stappersi and L. perspicax perspi-
cax. In Lestradea tooth shape changes
ontogenetically from tricuspid to bicus-
pid and conical. In ventralis, nasutus,
and hoops geographical variation has
been demonstrated very convincingly
(Poll and Matthes, 1962). Conical teeth
do occur in all species, even though in
hoops the tricuspid form is dominant. In
view of the fact that tooth form in some
members of this lineage can change dras-
tically from tricuspid to conical during
ontogeny, and that conical teeth actually

CiCHLiD Phylogeny • Liem 211

occur in all forms varying only in relative
abundance, the presumed "morphologi-
cal gap" between ventralis and nasutus
on the one hand, and hoops on the other
hand, is nonexistent and does not justify
the recognition of two genera.
^ More importantly, the phylogenetic re-
lationships as depicted in Fig. 9 indicate
that ventralis, nasutus, and hoops share
a recent common ancestor not shared by
any of the other taxa. I propose, therefore,
that Ophthalmotilapia and Ophthalmo-
chromis be considered synonyms. The
name Ophthalmotilapia established by
Boulenger in 1901 has priority over
Ophthalmochromis (Poll, 1956).

"Adaptive Radiation" of the
Ophthalmotilapia Lineage

Once the phyletic nature and relation-
ships of the Ophthalmotilapia lineage
have been established, a more meaning-
ful statement can be made concerning
their adaptive radiation.

Unlike most cichlid lineages, the radia-
tion in the Ophthalmotilapia lineage es-
pecially has involved behavioral patterns
in courtship and specific recognition sig-
nals, although the feeding apparatus does
exhibit some morphological diversity. All
are oral brooders, but in Cunningtonia
and Ophthalmotilapia the pelvic fins of
the males are drawn out into long slender
filaments whose tips reach posteriorly to
the margin of the anal fin and terminate
in brightly colored spots which function
not only for species recognition, but
probably also as egg dummies (Fryer and
lies, 1972: 110, 121). The strikingly
colored tips of the elongated pelvic fins
play an important role in specific, and,
thus inevitably, in sex recognition in the
complex communities where Cunning-
tonia and Ophthalmotilapia occur. Both
Cunningtonia and Ophthalmotilapia
produce only a few eggs, measuring 3.5-
4 mm in diameter. Egg collecting pre-
cedes fertilization. Ectodus, the most
primitive member of the lineage, and
Lestradea produce larger clutches (20-

30) of smaller eggs (2 mm) and lack the
specialized brilliantly colored tips on the
elongate pelvic fins. The males of Les-
tradea and Ectodus possess different
specific recognition signals, and their
courtship behaviors differ from those of
Cunningtonia and Ophthalmochromis.
In Lestradea and Ectodus egg laying,
fertilization, and collection by the female
take place rapidly in that sequence. The
courship behavior of Asprotilapia is un-
known.

Members of the Ophthalmotilapia lin-
eage occupy three types of habitat. Gen-
erally, Asprotilapia, Cunningtonia, and
Ophthalmotilapia inhabit the offshore
margins of the shallow, rocky littoral ad-
jacent to the sandy bottom. O. ventralis
often penetrates to greater depths of up
to 5 m. Lestradea inhabits shallow areas
with both sandy and rocky bottoms. L.
perspicax stappersi is common at greater
depths down to 50 m. The littoral habitat
of Ectodus is much narrower, being con-
fined to the shallow waters with sandy
bottoms. Except for Ectodus, all mem-
bers of the Ophthalmotilapia lineage are
not restricted to a particular habitat but
move freely between habitats with hard
and soft bottoms. The absence of strong
habitat restriction means the crossing of
"alien" habitats whether or not condi-
tions are adverse. As a result, a spectrum
of food resources is exploited. Judging
from the anatomy, stomach contents, and
behavior in the laboratory, members of
the Ophthalmotilapia lineage possess an
extensive feeding repertoire. Thus the
feeding pattern exhibited by the
Ophthalmotilapia lineage resembles that
of other cichlids in which increased func-
tional and morphological differences in
the trophic apparatus (dentition, jaw
structure, skull morphology, and pharyn-
geal myology) do not necessarily lead to
a greater separation on the food axis
(Liem, 1980).

The comparative anatomical and phy-
logenetic data presented in this study
reinforces the paradox (Liem, 1980;
Greenwood, in prep.) that the morpho-

212 Bulletin Museum of Comparative Zoology, Vol. 149, No. 3

logically and phylogenetically most spe-
cialized cichlid taxa are not only remark-
able specialists but also jacks-of-all-trades.
How the morphological radiation of the
feeding apparatus has evolved remains a
major problem in understanding the
causal factors underlying the spectacular
diversifications in the cichlid trophic
structures. If specialists are simulta-
neously jacks-of-all-trades, they defy the
commonly accepted ecological notion
that broadening the range of usable re-
sources prevents species from specializ-
ing on individual types. So far, most stud-
ies have attempted to correlate the
morphological diversity in the feeding
apparatus of cichlids with adaptation and
therefore fitness. However, Carson (1970)
and Fryer (1977) have suggested that spe-
ciation and adaptation are not necessarily
synchronous processes. Carson has of-
fered evidence that in Hawaiian dro-
sophilids speciation may have preceded
adaptation.

Although there is no doubt that the
Ophthabnotilapia lineage has under-
gone extensive morphological radiation
in both skull structure and dentition, the
data on morphology, function, trophic
ecology, and behavior of this and other
cichlid lineages (Liem, 1980) have failed
to establish unequivocally that the mor-
phological radiation is also adaptive.

ACKNOWLEDGMENTS

I am greatly indebted to Dr. P. H.
Greenwood of the British Museum of
Natural History for making available lab-
oratory facilities and collections and for
his thorough and critical review resulting
in a much improved manuscript. Karsten
Hartel of the Museum of Comparative
Zoology, and Gordon Howes and James
Chambers of the British Museum have
offered their skills in solving numerous
problems during the course of this study.
Sara Fink has produced several of the
drawings and Christine Fox has typed
and edited the manuscript and legends.
William Fink and George V. Lauder, Jr.

have made important contributions and
suggestions throughout the progress of
this study. To all of the above individuals
I am much indebted. This research was
supported by National Science Founda-
tion Grants DEB 79-00955 and DEB 76-
04386.

ABBREVIATIONS

a: anguloarticular

aap: adductor arcus palatini

ad: adductor branchialis

ad,: fifth adductor

af: adductor fossa

am,: adductor mandibulae pars A,

amj: adductor mandibulae pars Aj

amg: adductor mandibulae pars A3

ao: adductor operculi

ap: articular process

apa: ascending process of articular

apd: ascending process of dentary (coronoid)

appm: ascending process of premaxilla

apu: apophysis of upper pharyngeal jaw

bh: basihyal

boc: basioccipital

bs: basisphenoid

bsr: branchiostegal ray

ca: rostral cartilage

car: rostral cartilage

cbj: first ceratobranchial

cb,: fifth ceratobranchial (lower pharyngeal jaw)

cl: cleithnim

con: cranial condyle

d: dentary

do: dilatator operculi

e: ethmoid

eb: epibranchial

eb,: first epibranchial

eb3_4: third and fourth epibranchial

ect: ectopterygoid

ent: entopterygoid

eo: epiotic

eoc: exoccipital

ep: epaxial muscles

f: frontal

gha: geniohyoideus anterior

ghp: geniohyoideus posterior

hh: hyohyoideus

hhi: hyohyoideus inferior

hhs: hyohyoideus superior

hht: hyohyoideus transversus

hm: hyomandibula

by: hyoid

ic: intercalar

ini: interniandibularis

iop: interoperculum

la: lacrimal

1,: palatomaxillary ligament

CiCHLiD Phylogeny • Ltem 213

I2: palatopalatine ligament BoULENGER, G. A. 1900. Materiaux pour la faune

li: craniopalatine ligament du Congo. Poi.sson.s nouveaux du Congo. Six-

I4: ethmopalatine ligament ieme Partie: Mormyres, Characins, Cyprins,

lap: levator arcus palatini Silures, Acanthopterygiens, Dipneu.stes. Ann.

164: fourth levator externus Mus. Congo, Zool. 129-164.

li,_2: first and second levator internus . 1901. Diagnoses of new fishes discovered

lim: interoperculomandibular ligament by Mr. J. E. S. Moore in Lakes Tanganyika and

lo: levator operculi Kivu. II. Ann. Nat. Hist. 7: 1-6.

Ip: levator posterior . 1906. Fourth contribution to the ichthyol-

md: mandible ogy of Lake Tanganyika. Report on the collec-

mml: medial mandibular ligament tion of fishes made by Dr. W. A. Cunnington

mpt: metapterygoid during the third Tanganyika Expedition 1904-

mx: maxilla 1905. Trans. Zool. Soc. London 17: 537-576.

n: nasal Carson, H. L. 1970. Chromosome traces of the

obp: obliquus posterior origin of species. Science, N.Y. 168: 1414-

od: obliquus dorsalis 1418.

op: operculum DuLLEMEljER, P., AND C. D. N. Barel. 1977.

p: palatine Functional morphology and evolution. Pages

pa: parietal 83-117 in Hecht, Goody, and Hecht, eds. Major

pbj: second pharyngobranchial patterns in vertebrate evolution,

pb.-,: third pharyngobranchial FRYER, G. 1977. Evolution of species flocks of

pbj: fourth pharyngobranchial cichlid fishes in African lakes. Z. Zool. Syst.

pee: pharyngocleithralis externus Evolut.-forsch. 15: 141-165.

pci: pharyngocleithralis internus FRYER, G., AND T. D. Iles. 1972. The cichlid fish-

pf: lateral ethmoid complex es of the great lakes of Africa: Their biology

ph: pharyngohyoideus and evolution. Oliver and Boyd, Edinburgh,

pm: premaxilla GOEDEL, W. 1974. Beitrage zur vergleichenden

pop: preoperculum und funktionellen Anatomie des Kopfes von

pp: postmaxillary process of maxilla Tilapia (Cichlidae, Teleostei). Teil 1. Zool. Jb.

ppm: protractor pectoralis (Anat.) 92: 220-274.

ps: parasphenoid GREENWOOD, P. H. 1974. The cichlid fishes of

pts: pterosphenoid Lake Victoria, East Africa: The biology and

pv: vomer evolution of a species flock. Bull. Br. Mus. Nat.

q: quadrate Hist. (Zool.) Suppl. 6: 1-134.

ra: retroarticular . 1979. Towards a phyletic classification of

rd: retractor dorsalis the "genus" Haplochromis (Pisces, Cichlidae)

rf: rostral fossa and related taxa. Bull. Br. Mus. Nat. Hist,

sh: sternohyoideus (Zool.) 35: 265-322.

so: circumorbital LlEM, K. F. 1974. Evolutionary strategies and mor-

soc: supraoccipital phological innovations: Cichlid pharyngeal

sop: suboperculum jaws. Syst. Zool. 22: 425-441.

sph: sphenotic . 1978. Modulatory multiplicity in the func-

st: pterotic (supratemporal) tional repertoire of the feeding mechanism in

sy: symplectic cichlid fishes. I. Piscivores. J. Morph. 158:

tda: transversus dorsalis anterior 323-360.

tdp: transversus dorsalis posterior . 1979. Modulatory multiplicity in the feed-

tf: trigeminofacialis chamber ing mechanism in cichlid fishes, as exemplified

uh: urohyal by the invertebrate pickers of Lake Tanganyi-

x: foramen of vagus nerve ka. J. Zool., Lond. 189: 93-125.

. 1980. Adaptive significance of intra- and

interspecific differences in the feeding reper-
toires of cichlid fishes. Am. Zool. 20: 295-314.

LITERATURE CITED Liem, K. F., and G. v. Lauder. 1980. The evo-
lution and interrelationships of the actinopte-

Anker, G. C. 1978. The morphology of the head rygian fishes. In R. Davis and R. G. Northcutt,

muscles of a generalized Haplochromis species: eds. Neurobiology of Fishes. The University of

H. elegans Trewavas 1933 (Pisces, Cichlidae). Michigan Press, Ann Arbor (in press).

Neth. J. Zool. 28: 2.34-271. Liem, K. F., and J. W. M. Osse. 1975. Biological

Barel, C. D. N., F. WiTTE, AND M. J. P. VAN Oijen. versatility, evolution, and food resource ex-

1976. The shape of a generalized Haplo- ploitation in African cichlid fishes. Am. Zool.

chromis species: H. elegans Trewavas 1933 15: 427-454.

(Pisces, Cichlidae). Neth. J. Zool. 26: 163-265. LlEM, K. F., AND D. J. Stewart. 1976. Evolution

214 Bulletin Museuin of Comparative Zoology, Vol. 149, No. 3

of the scale-eating cichlid fishes of Lake Tan-
ganyika. A generic revision with a description
of a new species. Bull. Mus. Comp. Zool. 147:
319-350.

Poll, M. 1943. Descriptions de poissons nou-
veaux du Lac Tanganyika appartenant aux fa-
milies des Characidae et Cichlidae. Rev. Zool.
Bot. Afr. 37: 305-318.

. 1956. Poissons Cichlidae. Resultats scien-

tifiques explorations hydrobiologiques du Lac
Tanganika (1946-1947), 3, fasc. 5b: 1-619.

Poll, M., and H. Matthes. 1962. Trois poissons

remarquables du Lac Tanganika. Annls. Mus.

R. Afr. Cent. Ser. 8vo (Sci. Zool.) Ill: 1-26.
Regan, C.T. 1920. The classification of the fishes

of the family Cichlidae. I. The Tanganikan

genera. Ann. Mag. Nat. Hist. 5: 33-53.
Vandewalle, p. 1972. Osteologie et myologie de

la tete de Tilapia gutneensis Bleeker (Pisces,

Cichlidae) Annls. Mus. R. Afr. Cent. (Sci. Zool.)

196: 1-50.
WiNTERBOTTOM, R. 1974. A descriptive synonymy

of the striated muscles of the Teleostei. Proc.

Acad. Nat. Sci. Philadelphia 125: 225-317.

(US ISSN 0027-41(

jSulletin OF THE

Museum of

Comparative

Zoology

MUS. COMP. ZCOi

AUG 4 lyni

i--- A, •":;'■/

A Case History in Retrograde Evolutfom

The Onca Lineage in Anoline Lizards.

II. Subdigital Fine Structure

<«n-rY

J. A. PETERSON AND E. E. WILLIAMS

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBER 4
21 JULY 1981

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Breviora 1952-

BULLETIN 1863-

Memoirs 1864-1938

JOHNSONIA, Department of Mollusks, 1941-

OccASiONAL Papers on Mollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. I. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Tere-
dinidae (Mollusca: Bivalvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts, 02138, U.S.A.

© The President and Fellows of Harvard College 1981.

A CASE HISTORY IN RETROGRADE EVOLUTION;
THE ONCA LINEAGE IN ANOLINE LIZARDS.
II. SUBDIGITAL FINE STRUCTURE

J. A. PETERSON"
E. E. WILLIAMS-

CONTENTS

Introduction 216

Materials and Methods 216

Anolis annectens Moq^hology 217

Comparison with Anolis (Tropidodactylus)

onca 224

Adult Morphology 224

Juvenile Morphology 226

Comparison with Anolis chrysolepis 229

Comparison with Anolis auratus 232

Discussion 238

Summary of Comparative Data 238

The Mon^hological Series 240

Gross Scale Shape 243

Scale Contouring 243

The Fine Structural Series: Spine to

Seta 243

Functional and Ecological Implications — 245
Possible Functional Significance of

Lamellae 246

The Relationship between Lamellae

and Setae 246

Implications for the "Retrograde"

Series 248

Possible Functional Significance of Scale

Contouring 249

Contours as "Traction" Devices 249

The Evolution of Multicarinate Scales

in Anolis 250

The Evidence from Non-Anole Iquanids

and Geckos 253

Possible Functional Significance of the

Spine to Seta Series 255

The Spine to Seta Series: Adhesion or

Prehension 255

Implications for the "Retrograde"

Series 259

A. auratus and Grass Surfaces — 259

' Department of Biology, U.C.L.A., Los Angeles,
California 90024.

- Museum of Comparative Zoology, Hai-vard Uni-
versity, Cambridge, Massachusetts 02138.

A. onca and Diminished Selection for

Setae 260

A. annectens and Juvenile onca as Alterna-
tive Models for the "Retrograde"

Series 261

Conclusions 265

Acknowledgments 267

Literature Cited 267

Abstract. Anolis annectens is a presumed inter-
mediate between the padless anole A. onca and
more typical Anolis with adhesive toe pads. This
study uses scanning electron microscopy to com-
pare adult A. annectens with adult and juvenile A.
onca and two species, A. chrysolepis and A. aura-
tus, that were considered (Williams, 1974) to show
stages in "retrograde" evolution of the anoline sub-
digital pad. Differences are found in scale shape, in
scale contouring, and in the fine structural series.
The fine structural series begins with short hook-
like spines and terminates with setae which have a
long stalk and a triangular tip. In generalized anoles,
setae cover the surface of the expanded lamellar
scales which form the pad. The spines appear to
function as prehensile devices, while the setae may
combine prehensile and adhesive mechanisms. A.
auratus parallels A. annectens in showing dediffer-
entiation of the pad (reduction in pad width, in
lamellar number, and in the distinctness of the pad
from the phalanx i region), but the distribution of
setae is expanded in A. auratus, not restricted as in
A. annectens. The subdigital specializations of
A. auratus may reflect adaptation to the complex
surfoce architecture of grass blades and stems. A.
annectens and A. onca appear to be on a very dif-
ferent adaptive pathway. Short spines partly or com-
pletely replace setae, and the lamellar scales are
partially or completely replaced by multicarinate
scales. Comparative data from a variety of anoles,
non-anole iguanids, and geckos suggest that the
subdigital specializations of A. annectens and A.
onca may be correlated with radiation into sandy
substrates and thorn scrub vegetation. The absence
of definite keels in juvenile onca, in contrast to

Bull. Mus. Comp. Zool., 149(4): 215-268, July 1981

2irs BuIIcliit Miiscuni of C(>)ii])(ir(iliic' Zoology, Vol. 149, No. 4

their extreme development in adult A. onca, is be-
lieved to he an ontoj^enetic adaptation whieh eo-
incidentally parallels the jjhyletic seqnence. A.
chnjsolepis, with somewhat narrowed |iads and a
few snhdigital keels, is a suitable struetural aneestor
for A. auratus as well as A. annectens and A. onca.
The alternative adaptixe pathways leading to A.
(lurntus and A. onca emphasize that "retrograde"
is purely a descriptive term calling attention to the
dedifferentiation of a typical Anolis digital pad, not
a statement of loss of adaptive value for the end
points in the "retrograde" series.

INTRODUCTION

In an earlier paper one of us (Williams,
1974) described a new species, Anolis
annectens, whose subdigital morphology
represented an intermediate between the
adhesive toe pad of generalized anoles
and the keeled subdigital scales of A.
(Tropidodactylus) onca. On this basis
and other comparative data, the earlier
paper: 1) proposed a series of morpholog-
ical stages in the loss of the subdigital
adhesive pad, 2) suggested that A. an-
nectens was derived from an A. chryso-
lepis "stage" of the mainland beta anoles
and represented an intermediate stage in
the evolution of the very specialized con-
dition in onca, and 3) discussed probable

correlations of habitat and subdigital
morphology.

In this paper we amplify and examine
these ideas, drawing upon new and more
detailed morphological data obtained by
scanning electron microscopy (SEM)
from A. annectens, A. onca, their proba-
ble relatives, and still other species that
appear to provide insight into the mor-
phological sequence of pad loss.

MATERIALS AND METHODS

The size, sex, and source for speci-
mens used in the study are given in Ta-
ble 1.

In the previous paper the fourth toe of
the hindfoot formed the basis for inter-
specific comparisons (Williams, 1974).
Because Anolis annectens is known only
from the unique type, we have shifted
our attention to a toe not described or fig-
ured in the original paper. All descrip-
tions of fine structure in annectens and
other species are based on the third toe
of the hindfoot. However, except in an-
nectens, we have also examined the
fourth toe of each species.

Phalangeal dimensions were taken

Table 1. Specimens examined.

Snout-vent length (S\'L)

Anolis annectens
Anolis onca

Anolis chnjsolepis

Anolis auratus

*holotype FMNH 5679

*MCZ 158248

MCZ 158249

MCZ 158250
*MCZ 158251
*MCZ 154068

MCZ 154067

*LACM 74485
MCZ 43856

*LACM 72781
MCZ 77428

68.5 mm male

67 mm male
83 mm male
77 mm female
69 mm male
26 mm (juvenile)
25 mm (juvenile)

62 mm male
62.5 mm male

54 mm female
47.5 mm female

* Measurements and illustrations are largely based on these animals.
Abbreviations:

FMNH Field Museum of Natural History, Chicago, Illinois

LACM Los Angeles County Museum of Natural History, Los Angeles, California

MCZ Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts

The o.vcA Lineage: Subdigital Fine Structure • Peterson dud Willimm 217

witli dial calipers or an ocular microme-
ter. In discussing the subdigital scales we
nimiher them from claw to palm. Phalan-
geal number is counted in the same fash-
ion from distal to proximal.'

Specimens were prepared for micros-
copy by cutting the third and fourth toes
from the foot at the metatarsal-phalangeal
joint. The toes were then critical-point
dried and coated with carbon followed by
gold/palladium. Coatings were applied in
15-20 sec. bursts to avoid the possibility
of heat damage.

Measurements of fine structure were
made from photographs corrected for the
magnification given by the instrument.
Manufacturer's specifications suggest
that the error in reported magnification
may be as high as 20%, although with
some care it is unlikely to be more than
10%. The annectens specimen was pho-
tographed twice, with eight weeks sepa-
rating the sessions. Measurements of seta
and scale dimensions taken from the two
sets of photographs differed by 3-8%.
Where photographs were taken at an un-
determined angle to the plane of the
structure, an angular correction is esti-
mated, and the data are reported as ap-
proximate. More precise measures are re-
ported as the mean of 6-20 individual
measures along with the 95% confidence
limit of the mean.

ANOLIS ANNECTENS MORPHOLOGY

There are 20 subdigital scales on the
third toe: one over the claw base, seven
under phalanx i (2-8), six under phalanx
ii (9-14), and six under phalanx iii (15-

' Numbering the phalanges from distal to proxi-
mal and ignoring the claw has been the convention
in taxonomic descriptions of Anolis since at least
Boulenger (1885). It has therefore been used here,
as it was in Williams (1974). Morphologists, since
they emphasize the number of phalanges in each
digit, have preferred to count from proximal to dis-
tal (e.g., Russell, 1975). We have preferred consis-
tency within this series and within taxonomic lit-
erature to congruence with the morphological
literature. The discrepancy is radical, but in each
case context should make the usage clear.

20) (Fig. 1). The subdigital scales fall into
three morphological groups: a) multicar-
inate scales, b) lamellar or lamellar-like
scales, and c) unique scales of an inter-
mediate type.

a) Scales 1-8 (i.e., all those distal to
phalanx ii) and 16-19 are multicarinate
(scale 20 was cut at the palm, but presum-
ably belongs in this group). These are rel-
atively narrow, have robust, unfrayed dis-
tal margins studded with "mechano-
receptors" (see Schmidt, 1920; Miller
and Kasahara, 1967; Hiller, 1968), and
closely resemble the scales on the lateral
margins of the toe.

A typical multicarinate scale has three
to five keels. The keels originate near the
proximal base of the scale and expand
distally into large rounded bosses. The
crests of the keels are bare and devoid of
surface structure apart from a few stria-
tions (Fig. 3). The striations are prob-
ably produced by wear. Mechanorecep-
tors are usually positioned just distal and
in the lee of the keels.

The keels throw the surface of the scale
into sharp relief. The slopes of the keels
are covered with spines (for terminology
of spine to seta series see Fig. 21). Al-
though there are intermediate sizes and
variation in spine size is actually contin-
uous, two major size classes of spines can
be distinguished. The larger, more nu-
merous spines are 0.9-1.2 /jl tall and 0.5-
0.75 fjL. in diameter near their base (Fig. 3).
They occur in a density of 1.75 spines/sq /x.
These are interspersed with nubbin-like
small spines (0.1-0.5 /x tall; 0.2-0.3 fi in
diameter near the base). The shape of the
spine tip varies from sharply recur\'ed to
almost straight. Recurved spines with the
tip pointing distally and toward the crest
of the keel are the most common. The
spine bases are formed by buttresses or
struts (Ruibal, 1968), so that the scale sur-
face appears to be pinched or pidled up
into spines. The centers of the spines are
0.6-0.8 fx apart, and regions like that
shown in Fig. 3c have a density of 2.0
spines/sq /jl. The spines are separated

218 Biillcliii MiisciiDi ofCoinixirdlirc Zoolofitj, Vol. 149, No. 4

The onca Lixeage: Subdigital Fine Structure • Peterson (ind Williams

219

Figure 1. The third toe of A. annectens. a) Shows an overall view. Phalanx i{b), ii(c), and iii(d) scales are shown as
montages; (b) and (c) match at " + " and scale 9; (c) and (d) match at the "•" and scale 15. Scales are numbered from
the claw.

into irregularly shaped fields by narrow
(0.6 ijl) "trackways." The trackways pre-
sumably represent the borders of the epi-
dermal cells (Ruibal, 1968).

The general pattern described above is
modified slightly in the proximal portion
of the scales of phalanx iii. In these re-
gions the scale surface is slightly con-
toured into low "hillocks" (Figs. 1, 3)

which appear to correspond to a single
epidermal cell as each is completely
bounded by a trackway (see also Figs. 6
and 7 of adult A. onca). The density of
spines increases slightly between the
trackway boundary and the crest of the
hillock. This variation also occurs on the
lateral digital scales.

b) Subdigital scales 10-14 resemble la-

220 Bulletin Muscuiii of Cinuparative Zoology, Vol. 149, No. 4

Figure 2. Oblique view of the phalanx ii scales of the third toe in A. annectens. The keels and bosses on the lamellae and
intermediate scales are indicated by small arrows. Note the relief of the keels and bosses on scales 9, 14, and 15. The
arrow adjacent to the dimensional scale indicates the direction of the claw (distal). This convention is used in subsequent
figures.

mellae in having a distinct, thin, frayed
free margin and in being laterally ex-
panded to form a pad.' They differ from
typical Anolis lamellae in having small
bosses or keels along their lateral mar-
gins (Fig. 2). The surface architecture of
the bosses is similar to that of the keels
on the multicarinate scales.

The central portion of each of the la-
mellar scales is covered with setae (Fig.
4). The setae are about 20 /x tall (shorter
near the base of the scale and margins),
have a terminal "spatula," and occur with
a spacing of about 1 fx (density of 1-2
setae/sq fx). Midway between the spatula
and the scale surface, the shaft of the
seta is round in cross section and 0.48 ±
0.08 jx in diameter. The shaft tapers slight-
ly (to 0.35-0.45 fx in diameter) toward
the spatula. At about 1.4 /u. proximal to the
spatula the setae are usually inflected
slightly toward the claw and the shaft be-
comes oval in cross section (dimensions
are about 0.25-0.35 /jl by 0.2 ix). Just proxi-

' The outer keratin layer wa.s ah.sent on scale 10,
but there are some probable "bos.s" sites and a rel-
atively thin margin. We have included it with the
lamellar-like scales on the basis of these characters.

mal to the spatula the shaft is constricted
and bent toward the metatarsal-phalangeal
joint. The two inflections in the shaft are
responsible for orienting the spatula so
that its free distal margin faces the meta-
tarsal-phalangeal joint, and its expanse is
30°-90° from the vertical plane of the shaft
(orientation varies significantly). The dis-
tal margin of the spatrda is usually frayed,
and it may be concave toward the lamel-
lar surface. The spatula itself is shaped
like an isosceles triangle with a base of
0.73 ± 0.09 Amanda height of 0.58 ± 0.02 fx.
The area of the spatula is 0.211 ± 0.022
sq fx. The morphology of the setae in an-
nectens is very similar to that in gener-
alized anoles (e.g., A. carolinerisis and A.
sagrei as described by Ruibal and Ernst,
1965).

c) Subdigital scales 9 and 15 (Figs. 2,
5) exhibit a blend of lamellar and multi-
carinate scale features.

Scale 9 is laterally expanded, but less
so than the more proximal lamellar
scales. Unlike the lamellae, the distal
margin is quite robust, and there are
eight small bosses or keels along the dis-
tal margin. One of these shelters a me-

The o\ca Lineage: Subdigital Fine Structure • Peterson and Williams 221

Figure 3. The fine structure of multicarinate scale ^7 \r\ A. annectens. a) Shows the central keel and adjacent fields of
spines, b), c), and d) Are progressively higher magnifications of the spinate surface. Note the low hillocks in (b) and (c)
and the presence of many small spines and knobs in (d) (compare with Figs. 21 and 22).

chanoreceptor. The surface of the scale
appears "velvety," but less so than the
lamellar scales. Based on gross characters
it is neither a lamellar nor a multicarinate
scale.

The fine structure also suggests the in-
termediate character of the scale (Fig. 5;

Fig. 21 for definition of types). The sur-
face architecture of the bosses is similar
to that of the keels on the multicarinate
scales. In a few regions, particularly ad-
jacent to the central keels, there are fields
of small spines (0.2-0.3 /u, diameter at
the base; a density of 2.2 spines/sq fx).

222 Bulletin Museum of Coinixirdfive Zoology, Vol. 149, No. 4

Figure 4. The setae on scale 12 (lamella 3) In A. annectens. a) Shows setae along the distal margin of the lamella, b)
Is a view of the spatulae as seen from the claw.

These are taller, but otherwise similar to spines (Fig. 5b and lower left and center

the small spines on the multicarinate top of Fig. 5c). These spines have the di-

scales. The zones of small spines are ameter (0.55 fx at the base) and spacing

bounded by regions of somewhat larger characteristics of the larger spines on the

The onca Lineage: Subdigital Fine Structure • Peterson atid Willicnns

223

Figure 5. The morphology of scale 9 in A. annectens. a) Shows an overall view, b) A montage of half the scale
surface illustrates the heterogenous fine structure and relationship of the keels and spine/spike/seta series. The region
indicated with an "x" is shown in higher magnification in (c). Note the spines to the left of (c), the abrupt transition from
spines to prongs along the bottom of (c), and the gradual transition from spines to spikes to prongs along the top of
(c). A few seta shapes occur to the lower right (black arrow).

224 Bulletin Museum of Coniptintlivc Zooloj^ij, Vol. 149, No. 4

miilticarinate scales, hut i^rade up to 5 /x Scale 15 is quite similar to scale 9, al-

in height (hy definition; see below). Most though there are more definite keels

of the scale, however, is covered by even along its lateral margins (Figs. 1, 2).

taller, more robust "spine-like" forms These are almost certainly serially ho-

which grade through a variety of shapes mologous with those on multicarinate

into a few setae. The much taller (above scale 16 and lamellar scale 14. The adja-

5 ^t) and thicker (0.7-0.8 /jl at the base; cent multicarinate scales have a domi-

0.5 IX midpoint diameter) structures with nant central keel which is missing on

strongly tapered, needle-like tips we re- scale 15. The fine structure of scale 15

fer to as spikes (Fig. 5 center top; Figs, was difficult to examine thoroughly be-

21, 22). The gradation from spines to cause of the debris on its surface. There

spikes may be abrupt, as along the center are spines grading into spikes, as on scale

of Fig. 5c, or very gradual, as along the 9, but we were unable to confirm that se-

top of Fig. 5c. The spikes in the lower tae are present. Scales 9 and probably 15

and right portions of the figure are inter- combine the features of lamellar and

spersed with structures, here called multicarinate scales and suggest a model

"prongs," which have a straight or re- for evolutionary intermediate moipholo-

curved tip which is less tapered than the gy.
spike (see also Figs. 21, 22). There are

seta shapes (defined by the expanded tri- COMPARISON WITH ANOLIS
angular tip) m the lower left of the figure ^tROPIDODACTYLUS) ONCA
(note black arrow; see also l^ig. 21). Se-
tae are rare compared to spikes and Adult Morphology
prongs. In just a few areas adjacent to the

"bosses" and over the central portion of There are more subdigital scales in A.

the scale, spikes and prongs grade into onc« than in A. flnnecfen.s (24 rather than

setae. The dimensions of the spatula on 20, Figs. 1, 6). This reflects a) the longer

these setae are comparable to those of phalanges in an onca of approximately

spatulae on the lamellar scales, and the the same snout-vent length as the annec-

midpoint diameter of the stalk is also sim- tens type (phalanges ii and iii are 2.3 mm

ilar. Setae on scale 9 differ only in their and 2.8 mm compared to 1.4 mm and 2.4

shorter stalk (6-7 ix compared to about 20 mm, respectively, in annectens), and b)

fx for the lamellar setae). Although we differences in the size of the scales in some

have not encountered setae this short regions. The distribution of scales rela-

elsewhere, height is probably not a rea- tive to the phalanges in onca versus an-

sonable basis for differentiating them nectens is: 1/1 claw base; 9/7 phalanx i;

from the setae on the lamellar scales be- 6/6 phalanx ii; and 8/6 phalanx iii. In

cause 1) a few of the setae from the mar- onca the scales under phalanx ii are long-

ginal and proximal areas of the lamellar er, and those under phalanx iii are shorter

scales are almost this short (about 10 ix), than in annectens.

and 2) stalk height is a variable character While there is some regional variation

within and between anole individuals in the morphology of the subdigital

and species (Peterson and Williams, un- scales, all of them are multicarinate (Fig.

published data). 6). Four or five robust keels are present,

Figure 6. The third toe of A. onca. a) Only the distal 22 scales are shown. The "x" indicates the scale shown in (b)
center, b) Shows multicarinate scales 8, 9, 10, 11, and 12 at the i/ii interphalangeal joint. Note the more prominent
median keel on scale 1 1 , the asymmetry of the keels, and the great relief of the keels on the lateral digital scales, c) Shows
the structure of an individual scale. Hillocks occur toward the base of the scale, and the two pairs of lateral keels are
most prominent, d) Illustrates the bare surface of the keel and junction with adjacent spinate areas. Note the "wear
striations" along the crest and proximal slopes of the keel.

The osca Lineage: Subdigital Fixe Structure • Peterson and Williants 225

226

Bullet iit Musctmi of Cnrnpanitivc Zoo/o^'f/, Vol. 149, No. 4

and the distal margins of the scales have
additional bosses. The distal margin is
definitely not thin or frayed, and numer-
ous mechanoreceptors occur in the lee of
the keels and bosses. The regions be-
tween the keels and toward the base of
the scale are somewhat velvety in ap-
pearance and heavily contoured into hil-
locks (Figs. 6, 7). The subdigital and lat-
eral scales have very similar gross
morphology.

All the subdigital scales in onca resem-
ble the multicarinate series (scales 1-8,
16-20) in annectens. There is a differ-
ence, however, in the prominence and
placement of the keels. The pattern in
both animals is based on five possible
keels: a median keel, a pair of lateral
keels, and a pair of far lateral keels. In
onca the four lateral keels are consis-
tently well developed with the lateral
pair being the most robust and promi-
nent, while the fifth or median keel is
conspicuous only adjacent to the inter-
phalangeal and metatarsal-phalangeal
joints (Fig. 6). This differs from the con-
dition in annectens in which the median
keel is always well developed and the
most prominent of the five keels.

The fine structure of the keels in onca
appears similar in every respect to that of
the keels in annectens (Fig. 6).

The areas between the keels are cov-
ered with spines. Distally on the scale
where hillock contouring is not well de-
veloped, the spines are relatively uni-
form (Fig. 7c). They are 0.8-1.2 /x tall and
0.3-0.45 ^t in diameter near the base (0.1-
0.2 IX in diameter at their midpoint) and
occur in a density of about 1.2 spines/sq
IX. The tips are sharp and recurved, usu-
ally toward the crest of the keel.

Most of the scale surface, however, is
heavily contoured, and in these areas the
spines are more varied in morphology
and spacing (Fig. 7d). The range of spine
height becomes 1.0-1.8 ^l, and there is
marked variation in robustness. While
some spines resemble those near the dis-

tal margin, others have extremely large
buttresses or struts, so that their diameter
at the base (0.8 /x) is as much as half their
height. Spacing between the center of
the spines ranges from 0.7 /x near the
crests of the hillocks where the buttress-
es of adjacent spines interdigitate to more
than 3 /u, in the depths of the valleys be-
tween the crests.

The spinose regions of the scale in
onca differ from those in annectens only
in a) the scarcity of small spines (these
are present, but rare), and b) the more
extreme development of contouring and
associated spine variation.

Juvenile Morphology

Juvenile onca have a different subdig-
ital morphology from that of the adults
(Maderson and Williams, 1978). The
number (24) and distribution of the sub-
digital scales relative to the phalanges do
agree with the condition in the adults.
But, in contrast to the adults, none of the
scales are strictly multicarinate (Fig. 8a).
There are two poorly differentiated se-
ries: a) scales in the phalanx ii and iii re-
gions (scales 10-24) and the distal and
proximal scales in the phalanx i region
(scales 1, 8, 9), which have less robust
ridges near the distal margin and a more
velvety appearance (Fig. 8b), and b)
scales in the mid-phalanx i region (scales
2-7), which have more robust ridges and
a less velvety texture (Fig. 9a).

a) Scales 8-24 have a narrow but la-
mellar-like shape particularly because
they flare distally, but they lack the
frayed distal margin which is character-
istic of lamellae (Fig. 8b, 8c). In most
cases the scales are proportionally wider
than in the adult, but the difference is
often very small. Instead, the major dif-
ferences between hatchling and adult
morphology are in the more velvety ap-
pearance and the absence of robust keels
with sharp crests and "bare areas" (com-
pare scale 13 shown in Fig. 8b with scale
11 shown in the lower portion of Figs.

The onca Lineage: Subdigital Fine Structure • Peterson and Williams Til

Figure 7. The third toe of A. onca. a) Shows detail of the surface of scale 7. A small keel occurs in the lower right, b)
Shows detail of the hillocks, c) Is taken on scale 7; d) is from scale 11. c) and d) Illustrate the differences in spine
morphology between countoured and uncountoured regions.

6b and 6c). In the hatchling, the surface
of the scales is gently contoured into ridges
which occur in positions corresponding
to the robust lateral and far lateral keels
of the adult. There is, however, no ridge
that anticipates the median keel. Mech-
anoreceptors often occur in the lee of the
lateral contours, and the scale bases are
also contoured into low hillocks.

Setae are not present on any of the
third or fourth toe scales. The distal half
or two-thirds of the scale surface, includ-
ing the contoured ridges, is covered with
spikes and prongs (for definitions see
Figs. 21, 22). Shorter spikes at the margin
itself grade into taller spikes proximal to
the margin (7.2 ± 0.51 /x tall; 0.56 ± 0.08

fx in diameter). In some areas the spikes
grade into forms with a more robust
prong-like tip (in Fig. 8d spikes in the
upper left are shown grading into prongs
in the lower right). Between the con-
toured ridges as well as proximally, the
scale surface is covered with spines (Fig.
9b). There is continuous variation in
spine height from about 0.75 /u. up to the
spike morphology at 5 /x.

Although its shape is quite different,
scale 1 is included in this series because
its contouring and fine structure are com-
parable to those of scales 8-24. The sur-
face of the scale is covered with spikes
and prongs 5-7 jx tall.

b) Scales 2-7 in the juvenile have a

228 Bulletin Museum of Comparative Zoologij, Vol. 149, No. 4

Figure 8. The third and fourth toes of juvenile A. onca. a) Shows an overall view, b) Shows scale 13; the lamellar-
like scales in the phalanx ii region of the third toe have contours in the position of the lateral keels, c) Shows a robust
distal scale margin, d) The surface of the scale is covered with spikes (upper left portion) and prongs (lower right
portion).

gross shape similar to their homologues The scale surface adjacent to the distal

in the adult, and the position of the ridges margin and that over the distal portion of

in most cases anticipates the position of the ridges is covered with spines (Fig.

keels in the adult. 9a). On scale 6, spines in the distal region

The onca Lineage: Subdigital Fine Structure • Peterson and Williams

229

Figure 9. The second, third, fourth, and ftfth scales of the third toe (phalanx i region) of juvenile A. onca. a) Shows 3-
4 fx tall spines along the distal margin and over the ridges. The region between the contours and toward the base of the
scale is covered with shorter spines, b) Illustrates the border between the distal and proximal spines on scale 5.

are 3.37 ± 0.75 /x tall and 0.47 ± 0.08 fx
in diameter (see the top portion of Fig.
9b). These grade more or less abniptly
into spines which are more comparable
to those of the adult (1-2 /x tall; 0.5 /jl
diameter at the base; and a range of den-
sities of 0.7-1.2 spines/sq jx; see the low-
er portion of Fig. 9b). In some proximal
areas fields of still smaller spines occur
(0.5 /x tall; 0.3 fx diameter at the base; and
2 spines/sq fx; see the lower right corner
of Fig. 9b).

Scales 7 and 8 at the boundary between
the series are not very different in shape,
contouring, or fine structure. The "spines"
on scale 7 are 4-5 /x tall, while the spikes
on scale 8 are 6-7 fx tall. The subdigital
scales are regionally differentiated more
than in the adult, but much less than in
annectens or chrysolepis.

Some of the lateral digital scales have
robust keels with bare areas, but most of
the scales immediately adjacent to the
subdigital series have ridges (Fig. 9a)
rather than keels as in the adult (Fig. 7b).

COMPARISON WITH
ANOLtS CHRYSOLEPIS

A. chrysolepis has 29 subdigital scales
(a second specimen has 27; Fig. 10). The
distribution of these scales relative to the
phalanges in chrysolepis versus annec-
tens is: 1/1 claw base; 10/7 phalanx i; 11/
6 phalanx ii (9/6 in the second specimen);
and 7/6 phalanx iii. The differences in
distribution reflect smaller scales over
phalanges ii and iii in chrysolepis and a
longer second phalanx in chrysolepis (2.1
mm compared to 1.4 mm in annectens;
phalanx iii is 2.3 mm in chrysolepis and
2.4 mm in annectens).

The subdigital scales of chrysolepis in-
clude a) lamellae (scales 13-22), b) gen-
eralized subdigital scales which are de-
fined as non-lamellar and unkeeled
(scales 10-12, 22-29), and c) scales with
small keels or boss-like areas (scales 1-9;
Fig. 12).

a) The lamellar scales lie under pha-
lanx ii, form a pad, and resemble the la-

230 Btillctiii MttsciiDi of CoiDparative Zoology, Vol. 149, No. 4

Figure 10. Composite of the third toe in A. chrysolepis. The distal scale series (a) fits with the pad series (b) at the
"•"; the proximal series (c) matches (b) at "x."

The o\ca Lineage: Subdigital Fine Structure • Peterson and Williams 231

Figure 11. The setae in A. chrysolepis. a) Is taken on the eighth lamellae (scale 20) adjacent to the distal margin (the
material below the spatulae is debris), b) Is taken on scale 28 just distal to the metatarsal-phalangeal joint. The setae on
scales 28 and 30 are slightly narrower than on the lamellar scales (0.84 ± 0.05 /x compared to 1.01 ± 0.07 /x; the height
of the spatula is not different), and the stalk is less tapered and thicker just below the spatula. The difference in density
of setae is an artefact of the sites chosen.

mellae of generalized anoles. There is no
evidence of small keels or bosses like
those of annectens. The surface of these
scales is covered with setae (Fig. 11). The
orientation and height of the setae are
comparable to those of annectens, but
the stalk diameter (0.79 ± 0.09 ^t at the
midpoint) and spatular dimensions (base
[width] 1.01 ± 0.07 m; height [length] 0.71
± 0.05 ^x■ area 0.358 ± 0.037 sq ^x) are larg-
er. The chrysolepis measurements are un-
usually large for Anolis. The dimensions
in annectens are similar to those of West
Indian anoles investigated by us and by
Ruibal and Ernst (1965).

b) Proximal and distal to the lamellae
there are unspecialized subdigital scales
(Figs. 10, 12). These have a fairly robust
rounded distal margin and a velvety sur-
face texture. Mechanoreceptors occasion-
ally occur along their distal margins. The
shape of these scales is highly variable.
The series of scales proximal to the pad
continues beyond the metatarsal-phalan-
geal joint and into the sole of the foot.
Scales with these gross characters occur

proximal and distal to the pad in most
anoles we have examined.

There is a poor correlation between
this gross morphology and the fine struc-
ture of the scale. In chrysolepis the scale
series proximal to the pad (scales 23-29
and beyond into the sole of the foot) bears
setae with well-defined spatulate tips
(Fig. 11). Distal to the lamellae only one
of the scales with this gross morphology
bears setae. This is scale 12, which is ad-
jacent to the lamellae and part of the
raised pad. Scale 11, distal to the pad,
bears tall spikes with tips which com-
bine the shape characters of the seta
and the prong (Fig. 13). The tip is flat-
tened or compressed like the spatula,
but not laterally expanded to form a
triangular surface and broad distal mar-
gin (in generalized anoles this unexpand-
ed tip shape is present in the very prox-
imal portion of the lamella as part of the
morphological grade from spines to se-
tae; Fig. 23). The seta/prong intermediate
morphology occurs over the central por-
tion of scale 11 (the distribution is evi-

232 Btillctiii Miisciitu of Coinpanitivc Z()ol()<s,y, Vol. 149, No. 4

Figure 12. Scales distal to the pad in A. chrysolepis. a) Shows the scales just distal to the pad. b) Is an oblique view
of scales 1 and 2. Ridges (scale 1), small keels (e.g., scale 8), or bare areas (e.g., scale 2) are present on the series 1-9.

dent from Fig. 13a). Toward the lateral
and proximal portions of the scale the
shape grades into a blunt prong and fi-
nally a short spine with a recurved tip.
The spinose regions of the scale exhibit
contouring similar to that seen in annec-
tens.

c) The still more distal groups of scales
(scales 1-9) grossly resemble scales 10 and
1 1 with the addition of small keels or keel-
like bare areas (Figs. 12, 14). The fine
structure of the keels and bare areas is
similar to that of the keels on the lateral

scales and the keels in annectens and
onca. The scale surface adjacent to these
regions is covered by spines with needle-
like tips. The spines over most regions
are 1.4-1.7 /x tall, but on scales 8 and 7
they are as tall as 3.2 /x and 2.4 fx, re-
spectively.

COMPARISON WITH ANOLIS AURATUS

A. auratus has 22 subdigital scales
(Fig. 15). Their distribution relative to
the phalanges is quite similar to that in

The onca Lineage: Subdigital Fine Structure • Peterson and Williams 233

Figure 13. The morphology of scale 11 \r\ A. chrysolepis. a) Shows the distal portion of the scale (Fig. 10) and the
grade from spikes Into the taller seta-prong intermediate morphology, b) Shows the flattened tip and frayed margin of
the intermediate shape. Note that some of the tips appear to be bifurcated and not simply frayed.

A. annectens; the distribution in auratus
versus annectens is: 1/1 claw base; 7/7
phalanx i; 8/6 phalanx ii; and 6/6 phalanx
iii. Overall toe length, individual phalan-
geal lengths (phalanx ii is 1.3 mm in au-
ratus and 1.4 mm in annectens), and
scale sizes are very similar in spite of the
smaller body size of auratus (Table 1).

The auratus scales include a) lamel-
lae (scales 9-15), b) generalized subdig-
ital scales (scales 7, 8, 16-22), and c) in-
termediate scales (scales 1-6).

a) The lamellae are grossly similar to
those in annectens (apart from the ab-
sence of small lateral keels and bosses)
and chrysolepis. The series of lamellae
particularly resembles annectens in
forming a narrower pad which is not
clearly set off from the distal scale series
(the "Norops" condition;^ Figs. 1, 15, 18).

' The "Norops type condition" was defined by
Williams (1974; 17) as "loss of distinctness of the
anterior margin of the pad (that under phalanx ii) as
against the scales under phalanx i" and was regard-
ed as the third morphological stage in a sequence
of "retrograde evolution." The present study has

The lamellae are covered with setae
(Fig. 16). The seta stalk is shorter than
that in annectens (mean height over the
front half of the lamella is 13.2 ± 0.5 ^;
maximum height is 15.7 /x; height at the
distal margin is 12 /a), but the orientation,
stalk diameter (0.55 ± 0.06 ^t), and spa-
tidar dimensions (base [widthl 0.60 ± 0.04
ix; height [lengthl 0.57 ± 0.04 fx; area
0.173 ± 0.020 sq jx) are similar to those
in annectens. The spatula is slightly less
broad than that in most West Indian An-
olis.

shown that the condition depends upon the absence
of one or more small scales deep to the pad at the
ii/i interphalangeal joint. These scales, when pres-
ent, "raise" the first phalanx relative to the pad
{Anolis condition fide Boulenger, 1885: 11); in
their absence the pad is "not raised" (Norops con-
dition fide Boulenger, p. 95). The two conditions
are less sharply distinct than Boulenger believed,
although the end points of the series are indeed
conspicuously different. Also, as we shall show be-
low, the "Norops type condition" is not, as Williams
(1974) supposed, truly retrograde in the sense of
any loss of adaptive or even adhesive value; it is
instead an alternative adaptation.

234 BuUefiit Museum of Conijxinifive Zoology, Vol. 149, No. 4

Figure 14. Scale 8 in A. chrysolepls. a) Shows a keel "k. " b) Shows spines. This fine structure is typical of scales 1-8.

b) Scales 7, 8, and 17-22 resemble the
generalized subdigital scales (non-lamel-
lar and unkeeled) in chrysolepis and are
unlike any of the scales in annectens.
Scales 8 and 17 (Fig. 15) which flank the
series of lamellae are more laterally ex-
panded and could be considered lamel-
lae except that the distal border is not
thin and frayed.

All of the scales in this group bear setae
(Fig. 16). The stalk height, particularly
maximum height, is less than that of setae
on the lamellar scales, but the other di-
mensions are not different. The setae are
distributed over the central and distal re-
gions of the scale. Proximally and later-
ally the morphology grades into prongs,
spikes, and finally spines. The fine struc-
ture of the scales is generally similar to
that of the proximal scale series (scales
22-29) in chrysolepis.

c) A. (luratus, like A. annectens, has
individual scales with an unusually het-

erogeneous fine structure which com-
bines the keels or bosses of multicarinate
scales with the setae typical of lamellae.
In auratus these intermediate scales ex-
tend from the claw almost to the pad.

Grossly, the intermediate scales (scales
1-6; Figs. 15, 18) have a fairly robust dis-
tal margin and well-developed keels. A
few bosses occur over the median portion
of the distal margin on scales 2, 3, and 4,
but keels occur on all the scales. From
one to four pairs of keels are present in
median-lateral (on scale 3 only), lateral,
far lateral, and extreme lateral positions
(Fig. 18). The number of keels along the
side of the toe is relatively constant, but
their distribution on the subdigital versus
the lateral digital scales differs. Exclud-
ing the median-lateral keel, the number
of keels on the subdigital scales increases
from proximal to distal as the subdigital
scale expands around the phalanx. Proxi-
mally (e.g., scale 6) only the lateral keel

Figure 15. The third toe in A. auratus. The phalanx i(a) and ii(b) regions match at scale 9; the phalanx ii(b) and iii(c)
regions match at scale 16. Note the difference in morphology of the distal borders of scales 15 and 16.

The o\'ca Llneage: Subdigital Fine Structure • Peterson and Williams 235

236 Bulletin Museum of Comparcitivc Zoology, Vol. 149, No. 4

I't

if*.

Figure'16. The setae in A. auratus. a) Illustrates setae on the fourth lamellae (scale 12). b) Illustrates setae on the claw
base (scale 1). c) Shows setae on scale 21 just distal to the metatarsal-phalangeal joint.

is present on the subdigital scale, while
the lateral digital scale is multicarinate.
Distally (e.g., scale 3) where the subdi-
gital scale wraps 180° around the pha-
lanx, all three pairs of lateral keels (plus
an additional median-lateral keel) are
present on the subdigital scale, while the
lateral digital scale is small and unicari-
nate.

The fine structure of the subdigital
keels and bosses is comparable to that in
annectens, adult onca, and chrysolepis.
But there is a major difference in keel
placement. In those species, a median
keel is present and often the dominant
scale feature. In auratus the homologous
central region of the scale is covered with
setae (Fig. 17). The dimensions, density,
and orientation of these setae are com-
parable to those of the lamellar setae
(with the possible exception of stalk

height, which may be reduced). The se-
tae grade into prongs and tall spikes ad-
jacent to the lateral keel. Over the lateral
region of the scale, including the zones
between the keels (e.g., the lateral and
far lateral keels), the scale surface is cov-
ered with spikes and prongs.

While the combination of features, par-
ticularly the grade from setae to spines to
keel, is very similar to that on scale 9 in
annectens (compare Figs. 5 and 17),
there are some interesting differences. 1)
In annectens the grade from setae to
spines occurs on the medial side of nu-
merous bosses scattered along the distal
margin of the scale. The field of setae is
much less coherent than in auratus. 2)
The spikes and prongs are numerically
dominant on the transitional scales in an-
nectens, while in auratus the setae are
much more common.

Figure 17. The morphological grade from spikes to setae adjacent to the keel on scale 1 in A. auratus. a) Illustrates an
oblique view of the scale, claw "cl" and keel "k" position. The "[ ]'s" indicate the position of the transect which begins
at the left of (b) and continues to the right through (c). The (b) and (c) sections fit together at the point indicated by the
"x." Spikes and prongs are mixed through (b). At the left in (c) are prongs grading into the seta/prong intermediate
shape and finally setae at the far right. The grade continues to the right of (c), toward the center of the scale, as the
spatulae become broader (see Fig. 14b). The difference between morphotypes can be most readily seen by comparing
regions which are far apart (e.g., the right of [b] and [c]) and using Fig. 21 as a guide. The gradation of one morpho-
type into another is clear in regions such as (c) center. The tip shapes in these regions are not readily classified.

The onca Lineage: Subdigital Fine Structure • Peterson and WiUiain.s 237

J

238 Bulletin Museum of Comparative Zoolofiy, Vol. 149, No. 4

Figure 18. Distal portion of the third toe in A. auratus.
Note the variation in keel number and placement in scales
1-7.

The lateral scales in auratus are multi-

DISCUSSION

Summary of the Comparative Data

The comparisons among annectens,
adult and juvenile onca, chrijsolepis, and
auratus suggest that there are significant
interspecific differences but relatively
little novelty in suhdigital morphology
per se. The fine structure of setae, keels,
spines, etc. is relatively similar among
the species, and, in three of the species,
the regional variation in structure ap-
proximates the range of interspecific vari-
ation. The dominant evolutionary pattern
is change in the distribution of keels, se-
tae, and spines relative to the phalanges.

If the species are arranged in the phy-
logenetic series proposed in the earlier
paper (Williams, 1974), the major inter-
specific differences appear as trends for
a) expanded setae distribution in auratus
relative to chrysolepis, and b) the pro-
gressive substitution of multicarinate.

or unicarinate, studded with mechano- spinate morphology for lamellae and se-
receptors, and covered with spines. The tae in the chrysolepis-annectens-onca se-
spines occur in several sizes (Fig. 19). ries (Fig. 20).

Figure 19. The diversity of spine shapes and sizes in A. auratus. The figures are taken on the infradigital scale adjacent
to scale 5. a) Shows three spine sizes, b) Is a higher magnification view of the junction between two fields of spines.

The onca Lineage: Subdigital Fine Structure • Peterson and Willianis

239

/I. on CO

phalanx I -narrow, multiccrinate, spmate

A

II -narrow, multicarinate, spmate

III -narrow, multicarinate, spmate

SETAE OVER PHALANX ii "lOST"
REPLACED BY MULTICARINATE FEATURES

A. onnecfens

phalanx i - narrow, multicarinate, spmate

phalanx ii - scale 9: lamellarshape, multicarinate, spines and setae
scales 10-14; lamellae, keels or bosses, setae
scale 15: lamellarshape, multicarinate, spines and setae (?)

phalanx in - narrow, multicarinate, spmate

A

SETAE RESTRICTED TO PHALANX ii
INCREASED KEELING ON PHALANGES i & iii
INTRODUCTION OF KEELS ON PHALANX ii

A. ourotus

phalanx i - shapes various, multicarinate,
setae and spines

phalanx ii - lamellae, acarmate, setae

phalanx m - narrow, acarmate, setae

SETAE DISTRIBUTION EXPANDED TO ALL
SUBDIGITAL SCALES

INCREASED KEELING ON PHALANX i SCALES

'/I. chrysolepis

phalanx i - scales I - 9 : narrow, ridges and bare areas on robust margin and spmate
- scales 10, ll ; narrow, acarmate and spinate
-scale 12: narrow, acarmate and setae

phalanx II -lamellae, acarmate, setae

phalanx iii -narrow, acarmate, setae

Figure 20. Summary of the comparative data.

A.nnectens differs from chrysolepis in
three respects: a) the substitution of mul-
ticarinate, spinate scales on phalanx iii
(scale shape is similar), b) the encroach-
ment of multicarinate, spinate characters
on the phalanx ii scales (the "lamellae"
are slightly narrower, keels are present
on the lateral portion of the lamellae, and
the first and last "lamellae" are multicar-

inate with a surface dominated by spines
and spikes), and c) the expansion of the
multicarinate, spinate moiphology to all
the phalanx i scales.

Adult onca goes beyond annectens
only in the morphology of the phalanx ii
scales. The lamellar scales and setae have
been replaced by narrower, multicarinate,
spinate scales.

240 Bulletin Mitseuni of Couiparative Zoologu, Vol. 149, No. 4

The presence of an almost pad-like tens. For example, the differentiated re-
scale series in juvenile oncii alon^ with gion in juvenile onca is not synonymous
the existence of a unique annectens with any of the regional series in annec-
specimen places particular importance tens.

on the similarities and differences among It is highly implausible that annectens

annectens, adult onca, and juvenile is an "immetamorphosed" onca hatch-

onca. The subdigital mori:)hology of ju- ling. Juvenile onca are more similar to

venile onca resembles the adult mor- the adult onca than they are to annec-

phology in several respects: a) in the tens, chrysolepis, anratiis, or generaWzed

number and distribution of the scales, b) "Norops" species (e.g., A. tropidonotus

to some extent in the shape of the scales, and A. notopholis).

c) in the correspondence between ridges Clearly juvenile onca displays still
in the juvenile and keels in the adult, d) another morphology intermediate be-
in the absence of lamellae and setae, and tween more generalized anoles, such as
e) in the poor differentiation among pha- chrysolepis, and adult onca. Indeed, the
langeal regions. The morphology of the implications of annectens and juvenile
juvenile onca differs from that of the onca for a "retrograde" sequence are
adult in a) the absence of definite keels, contradictory. A. annectens morphology
although scale contouring in some re- implies a sequence in which robust keels
gions is pronounced, and b) the pres- invade and progressively replace setae
ence of tall spines, spikes, and prongs. In and lamellae. In contrast, juvenile onca
the first of these differences, the juvenile morphology would appear to imply that
morphology is more conservative in a substantial "retrograde" shift in the
terms of the proposed phylogenetic se- spine to setae series (see below) and de-
ries than that of annectens or even chrys- differentiation of the pad occur prior to
olepis, which have keeled subdigital the appearance of definite keels,
scales. The absence of well-developed The juvenile onca morphology is an
keels in juveniles may not be unique to ontogenetic adaptation. It may or may not
onca. A. cristatellus adults have marked also represent a phylogenetically inter-
keels and bare areas in the phalanx i re- mediate morphology in the "retrograde"
gion while hatchling-sized (28.7 mm sequence. We discuss these alternatives
SVL) cristatellus have spine-covered in light of additional functional and com-
ridges like those of onca juveniles and a parative considerations in the final para-
few bare areas which are not associated graphs of the discussion.

with the ridges. Most of the contours on _. ,, ,. , ■ , « ■

.,,.1^1 , J i^u The Morphological Series

the lateral scales are also covered with ^ ^

spines as in juvenile onca. In the second The extraordinary diversity in subdig-
of the differences relative to adult onca, ital scale morphology may be related to
the juvenile morphology is less conser- the wealth of structural intermediates
vative than that of annectens which has and to the poor correlation among the
the complete spine to seta series and defi- characters. These factors result in a myr-
nite lamellae (see below and Fig. 21). iad of possible character combinations for
There is little resemblance between a given scale. Setae occur not only on la-
juvenile onca and annectens. In both mellae but also on narrow multicarinate
forms the spine-seta series is more com- scales. Keels occur on lamellae as well as
plete than in adult onca, and there is on scales with a robust distal margin,
some regional differentiation in the sub- Some "generalized subdigital scales"
digital scale series. But neither of these bear setae while others which are not
parallels translates into a convincing sim- otherwise different do not. This particu-
ilarity between juvenile o?ica and annec- lar group of species may exhibit more

The osca Lineage: Subdigital Fine Structure • Peterson and Williams 241

ir

A= -

c o
\ E

c
o

II

r

/i\

in (D
-c r:

-F

A 5-

o

(/)

3

■o

^_

O)

- Cl

3

CP

^ <3^.=

a>

CL

A

(U

c

CL

Q

UJ

CO

LU

_J

(T

!^uj

<

t—

<J)

UJ

I

-J _i

o

Q

en

UJ

< o

UJ <

q:

UJ
(O

CO

2

OJ

^

_j

o

c

UJ

U)

UJ

Q.

o

OQ

Spinulate small s
Oberhautchen

K

E "
o °

A

^<-^

^

A

9-

C
CO

O

rO

I

O

a.

O
00

I
in

3,
o

in

3,
in

I
in

O)

E
o

13

o

Q.

-o
E

— o

3

c
o

in

en

c

CD

■o

3,

in

T3

2 3

Q)

o

CD

E

n.

o

in

■o

o

Zl,

Q.

in

3

;

CP

c
o
o

Q.

3,

T

cr
to
\

lO
lO

O

C\J

<D

c
'q.

CO

CO

0>
CO

«
o

O)

o
o

Q.

k_

o

E
«

I-

3

—

T)

u

O)

*

>

+

\-

n.

3

O

I

a>

t-

\

<rt

\-

O)

QJ

CT

3

QJ

C

o

C

Ti

E

Q.

u

n

CO

3,

-o

Q.

in

o

3k

CO

3,

rO

— 3.
1

1

in

o^

rO

242 Bulletin Musctiiti of Coinparative Zoology, Vol. 149, No. 4

The o\ca Ldjeage: Subdigital Fine Structure • Peterson and Williams 243

"creativity" in character combinations
than most anoles, but their morphology
is a cogent argument for the existence of
more than one morphological series or
character suite in the evolution of scale
morphology.

The diversity in scale moiphology ap-
pears to be interpretable in terms of three
semi-independent morphological series,
each one operating within a different di-
mensional range. These are: gross scale
shape, contouring of the scale surface by
keels, hillocks, etc., and fine structural
specialization of the Oberhautchen (i.e.,
the spine to seta series; see Figs. 21, 22,
23).

Gross Scale Shape

Gross scale shape is determined by the
length/width dimensions of the scale, the
character of the borders, the cross-sec-
tional shape, and the junctions with
adjacent scales. In the present context
our concern is the differentiation be-
tween lamellar and non-lamellar scales
in anoles (A. Russell informs us that la-
mellar morphology in geckos is differ-
ent). Lamellae are highly specialized
scales, but they can be defined rather
simply by the presence of a frayed, thin
distal border and a length/width ratio <1.
Both features are required to characterize
lamellae. Wide scales with a robust mar-
gin occur fairly frequently, particularly
adjacent to the claw and interphalangeal
joints (e.g., auratus scale 3), but with rare
exceptions (e.g., the most proximal la-
mella in piilchellus) all scales with a free,
thin distal border are wider than they are
long.

Scale Contouring

Variation in the surface contouring of
scales can be arranged in a series based
on increasing elevation and extent of the
raised contours. The series begins with
a relatively flat surface and extends to
hillocks, which involve a single epider-
mal cell (e.g., Fig. 7), to ridges and boss-
es, and finally to keels. The hillock mor-
phology apparently occurs only over the
proximal and central portion of the scale.
It is not clear that the hillocks are ho-
mologous with the rest of the series.
Ridges, which extend into the central
portion of the scale, and bosses, which
occur adjacent to the distal margin, ap-
pear to be small keels. While both these
features usually have a bare, smooth sur-
face, the ridges, at least, may be covered
with spines as in juvenile onca.

The Fine Structural Series:
Spine to Seta

The Anolis Oberhautchen has spinules
(Fig. 22; Ruibal and Ernst, 1965; Ruibal,
1968; Maderson, 1970). Such spinules are
the beginning term in a morphological
series which can be described as follows
(Fig. 21): a) the spinulate Oberhautchen
(Fig. 22a) consists of spinules and knobs
which are 0.3-1.0 fx tall, 0.15 jx in diame-
ter and about 0.3 /Lt apart (or 10-15 spi-
nules/sq /x), b) the small spine (Fig. 22b
center), which differs from the spinules
by greater stalk height, a curved, more
tapered tip, and reduced density (only
some spinules are longer; see discussion
following), c) the subdigital spine (Fig.
22b, c), which is marked by the absence
of an understory of spinules, increased

Figure 22. a) The spinulate Oberhautchen from a dorsal digital scale in A. evermanni. b) Is taken on a lateral digital scale
of A. eugenegrahami at one-quarter the magnification of (a). The central patch of small spines with a spinule understory
has the same diameter and overall density as the spinulate Oberhautchen. The adjacent larger spines illustrate the
diameter and spacing features which are typical of the subdigital spine/seta series, c) Illustrates fairly tall spines in A.
pulchellus. d) Shows the even taller spikes in A. evermanni. Prongs are shown in the lower right of (e), and a few seta-
prong intermediates are visible in the upper center of (e). The specimen is Anolis sp. n. near eulaemus. f) Shows a
heterogeneous field of spikes, prongs, and seta-prong intermediates in A. cuvieri.

244

Bulletin Muscudi of Coinixirative Zoolo^ij, Vol. 149, No. 4

Figure 23. The subdigital spine to seta morphological series in A. evermanni. The figures are taken in the proximai
portion of lamellae 8. The toe has been sectioned proximodistally. a) Is a ventral view at a slightly larger magnification
than the lateral view (b). In this particular specimen the hook tips are not strongly recurved or highly tapered (cf. Fig.
20) so that the spike morphology is not well shown. The seta-prong intermediate "s/pi" shape is shown along the section
margin and lateral to the margin (see also Figs. 13 and 20). A prong, "p," is also indicated. Setae significantly increase
in height distal to the figure margin.

The onca Lineage: Subdigital Fine Structure • Petersoi aiul WiUiam.s

245

diameter of the stalk (about 0.5 /jl at
the midpoint or base, if the stalk is
very short), decreased density (2 or less/
sq fx; spacing about 1 /x), and often in-
creased height (up to 5 /x), d) the spike
(Fig. 22d), which differs from the spine
by increased stalk height, e) the prong
(Fig. 22e, f), which has a blunt or curved,
but almost untapered tip, 1) the seta-
prong intermediate (Figs. 22e, f, 13b), in
which the prong tip is dorso-ventrally
compressed and stalk height is often
greater, and g) the seta (e.g.. Figs. 4, 11, 16,
etc.), in which the flattened tip is ex-
panded into a spatulate, triangular shape
(the spatula is about 0.75 fx wide by 0.6 /u,
long). Several of the morphologies in this
series may occur within the borders of
one epidermal cell, and generalized ano-
les usually have all members of the se-
ries. Subdigital scales usually exhibit
only the terminal (right) portion of the
series (c-g). The structural series be-
tween spines (c) and setae (g) typically
occurs in the proximal portional of la-
mellae (Ruibal and Ernst, 1965; Fig. 23).
The spinule/spine portion of the series
(a-c) occurs on the lateral digital and
body scales, although spines with the di-
mensions of subdigital spines are rare
except on the lateral digital and plantar/
palmar scales.

The spinule (a) to spine (c) series ap-
pears to model the transition from the
spinulate Oberhautchen to the subdigital
series. It involves alteration in the height
and diameter of the spinules and a major
change in their spacing or density. The
spine (c) to seta (g) series is based on vari-
ation in stalk height and tip shape. Stalk
diameter and spacing or density are rel-
atively constant between the spine or
spike and the setae. Russell (1976) pro-
posed a morphological series for the ori-
gin of gecko setae based on interspecific
variation in Cyrtodactylus which ap-
pears to show spine-spike-seta compo-
nents. Dimensions apart from height are
not indicated, so it is difficult to compare
the two series in detail.

The aspinulate, bare areas are typical
of major scale contours, but also occur in
regions with no relief (Figs. 8, 9). The
relationship of this Oberhautchen spe-
cialization to the spinulate series is un-
clear. In fact, the surface morphology of
these bare areas is unlike any of the three
generalized Oberhautchen types (spinu-
late, lamellate, or dentate) described bv
Ruibal (1968) and Maderson (1970). A
number of observations suggest that the
morphology is not an artefact of the shed-
ding cycle nor produced from a spinulate
Oberhautchen by wear, a) The bare areas
are often slightly raised or elevated rel-
ative to the plane of the adjacent spinate
areas — they are not worn down, b) The
immediately adjacent spines do not show
excessive wear, c) The bare surface is ac-
tually striated as if by use, and these stria-
tions do not occur in specimens where
the Oberhautchen has separated post-
mortem, d) Limited comparisons (for
West Indian species) suggest that there
is consistency in the placement of these
areas between individuals. The bare,
smooth Oberhautchen morphology char-
acterizes all of the subdigital scales in the
primitive iguanid, Morunasaurus groi
(Peterson and Williams, unpublished data).
It is not clear whether this surface mor-
phology is a specialization of and within
the spinulate Oberhautchen series or
whether it represents an even simpler
and distinct Oberhautchen type. The
problem is compounded by the differ-
ence and implied increase in bare areas
between juvenile and adult specimens of
cristatellus and onca (see Maderson and
Williams, 1978, and in preparation, for dis-
cussion of onca).

Functional and Ecological
Implications

Without actual behavioral and experi-
mental data, discussion of the functional
significance of variation within and
among the three morphological series has
to be limited. But we can construct and

246 Bullet in Museum of Coiiiparafive Zoo/o^'//, Vol. 149, No. 4

evaluate a number of hypotheses on the ments more than compensate for the re-
basis of comparative data. duction in himelhir width. A. auratus has

at least as many setae as an A. evermanni

Possible Functional Significance of (a Puerto Rican trunk-crown ecomorph

Lamellae with a wide pad and many lamellae) of

The Relationship between ^^^ jrl^^ ^^"dy ''^''- The situation is rath-

Lamellae and Setae ^' differen for annectens. While ,t is

very possible that the ratio oi setae/body

The shape of a lamella appears to re- weight might be altered without com-
flect its function in bearing adhesive prising essential function, the "safety fac-
hairs or setae. Given that the density of tor" reflected in generalized anole ratios
setae is constant over the exposed portion would have to be very great for the re-
of the scale surface, the number of setae duction in the annectens ratio not to have
is determined by the total surface area behavioral consequence. The restriction
(more extensive data regarding this are in pad width coupled with the short sec-
given in Peterson and Williams, in prepar- ond phalanx (1.4 mm compared to 2.1 mm
ation). Lateral expansion of the scales in a comparably sized specimen of chry-
which bear setae produces the character- solepis) and the restriction in seta distri-
istic lamellar shape and, at least in the bution to the second phalanx could mean
West Indian ecomorphs we have exam- that the setae would have to "cooperate"
ined, is a major vehicle for adjusting the with other devices to support the animal,
number of adhesive hairs to the body Apart from the relationship with body
weight (loading) (Peterson and Williams, size, the functional advantage of narrow
in preparation). There are, however, two lamellae is unclear. As suggested in the
additional and alternative means to vary earlier paper, narrow lamellae may con-
seta number: a) change in the length of stitute a stage in pad reduction per se,
the phalanges which have setae-bearing but clearly they do not always imply a
scales (varying the length of the "pad" in- reduction in the number of setae (cf.
dependently of lamellar number or pad again auratus) nor presumably de-
width), and b) restricting or expanding the creased competence in adhesion,
distribution of setae relative to the pha- Lamellar scale morphology probably
langes. It is important to note that packing also reflects adaptation for applying se-
more or less lamellae into a pad of a given tae. Lamellar scales, particularly the thin
length and width does not significantly distal margins, are unusually flexible and
alter the number of setae. compliant. Flexibility of the scale along

The relatively narrow pad of many an- with the length of the seta stalk permit
oles — chry solepis, auratus, and annec- the spatulae to "find" the environmental
tens among them — would itself provide surface and conform to a curved or oth-
for fewer setae in proportion to the ani- erwise grossly irregular substrate (Hecht,
mal's body weight than in the West In- 1952, discusses lamellar shape in terms
dian ecomorphs unless the alternative of pad flexibility; see also Hiller, 1968,
adjustments are made. In auratus, where and the discussion of seta function fol-
the pad is particularly narrow, the pha- lowing). For a pad of a given size, an in-
langes are long in relation to snout-vent crease in lamellar number might permit
length, and setae occur on all the subdig- the seta-bearing area to better conform to
ital scales (i.e., two compensating adjust- surfaces with a smaller radius of curva-
ments are made). Crude estimates of seta ture. This aspect of lamellar function may
number (based on seta densit\' and seta- partially explain the interspecific in-
bearing area) indicate that these adjust- crease in lamellar number with increas-

The onca Lineage: Subdigital Fine Structure • Peterson and WiUiams 247

ing perch elevation (and decreasing lamellae are significant adaptations for

perch diameter) (Collette, 1955; Wil- applying setae and transmitting tensile

Hams, 1974). forces.

The lamellae of geckos (Dellit, 1934; The function of lamellar tendons in

Russell, 1975, 1976, 1979) receive the in- transmitting tensile forces may also con-

sertion of numerous tendons originating tribute to the correlation between lamel-

from the metatarsal-phalangeal joint cap- lar number and body size (Hecht, 1952;

sules and plantar tendon system. Ulti- Collette, 1955; Williams, 1974). Scaling

mately the tendons operate in series with for increased tensile loads could occur by

the digital and ankle flexors. The spe- an increase in tendon diameter or by an

cialized tendon arrangement coupled increase in the number of tendons (by in-

with the flexibility of lamellae have sug- creasing the number of lamellae). Ten-

gested to Hiller (1968) and Russell (1975, don robustness does vary, but, at least in

1976) that lamellae have a dynamic role the West Indian anoles, the dominant

in applying setae. Russell (1975) argues pattern is to multiply the number of ten-

that the subdivision of the pad into scan- dons (and therefore the number of la-

sorial plates (i.e., lamellae) increases the mellae). This pattern maintains a rela-

effectiveness of the setae because each tively size-independent lamellar area/

lamella can be individually controlled by tendon relationship while scaling total

the tendons. Once the setae tips "find" tendon cross-sectional area to total seta

the surface, flexion of the lamellae via the stalk cross-sectional area. A size-correlat-

tendons places the setae under tension ed increase in tendon diameter by itself

and transmits force between the adherent would scale the cross-sectional areas of

epidermis and the skeleton (Russell, the load-bearing tissues, but drastically

1975). Hiller's (1968) hypothesis for seta increase the lamellar area served by each

function in geckos involves an initial ap- tendon. Perhaps the role of lamellae and

plication of the spatulae via an active lamellar tendons in applying setae has

flexion of the toes and lamellae. (Russell required criteria of scaling similarity

[1975] and Hiller [19681 also implicate which include lamellar area,

inflation of digital venous sinuses, which An interesting caveat to the argument

are present in some geckos, in the adhe- that setae and lamellae are functionally

sion process.) associated is the existence of setae on

Anole lamellae have a similar set of ten- non-lamellar scales (or the converse — la-
dons (Peterson and Williams, unpublished mellae which do not bear setae). This oc-
data; Russell, personal communication), curs not only in chrysolepis and auratus.
Given the morphology of the setae in an- but in several West Indian and Central/
oles, tension exerted on the lamellae. South American species (Peterson and
which is produced as the foot approaches Williams, in preparation). The non-lamel-
contact with the substrate and the digital lar setae are slightly different from typical
flexor musculature fires to meet weight- lamellar setae. In aurflfw.S' the stalk is short-
bearing conditions, would wipe the distal er. In chrysolepis the stalk is shorter and
edges or flattened expanse of the setae into less tapered, and the spatula is somewhat
the surface and establish the "grip" (see narrower. But, in each case the range of
discussion of seta function). Once the variation in lamellar setae includes these
"grip" is established, tensile stress can be morphologies. Unless very similar seta
transmitted from the setae/epidermis to morphology can function in a totally dif-
the skeleton via the lamellar tendons as in ferent fashion on lamellar and non-la-
geckos. If these hypotheses are correct, mellar scales, the flexibility, compliant
the internal morphology and flexibility of margin, and internal tendons of lamellae

248 Bulletin Museum of Comparative Zoologij, Vol. 149, No. 4

are simply not necessary for seta func- species. There is an enormous diversity
tion. in the architecture of vegetational sur-
There are several possible, comple- faces ranging from the waxy rubble, often
mentary explanations for the variation in trichome-studded surface of deciduous
distribution of setae compared to lamel- leaves to the grooved and "rough" sur-
lae. a) Pad morphology (i.e., lamellae) ftice of grass blades, to the very gross sur-
might be less effective in positions adja- face relief of cork and older bark (Martin
cent to the claw or sole of the foot. In and Juniper, 1970). While the claws prob-
some anoles the pads extend to the sole, ably contribute most of the grip on rough
but lamellar morphology does not seem bark, setae appear to be responsible for
to occur in the phalanx i region of any purchase on very dissimilar cuticular sur-
anole (cf. geckos; Russell, 1976, 1979). b) faces, e.g., on grass blades and herba-
If there is a necessary relationship be- ceous leaves. The grass anole, A. aura-
tween external lamellar morphology and tus, has at least as many setae for its size,
internal digital tendons, the distribution fewer lamellae, and a high proportion of
of lamellae might be limited to certain non-lamellar setae compared to eco-
phalanges by constraints on the distri- morphs which use leaf, stem, and twig
bution of the tendons. There are a num- vegetational surfaces. The functional as-
ber of potential mechanical constraints sociation between lamellae and setae
on extending tendons into the phalanx i presumably is reduced in auratus, sug-
region. For example, long tendons which gesting that lamellae may not be as ad-
are not controlled by a retinaculum or vantageous on grass as on most herba-
placed very deeply in the toe could act ceous leaves. Considerably more data on
not only on their lamella of insertion, but plant cuticles, setae, and setae/lamellae
"bowstring" or alter the more proximal distribution are required to approach this
pad contour and disengage setae. The ar- problem, but the functional advantage
rangement in geckos (Russell, 1975, and significance of lamellar morphology
1976) appears to avoid this problem with may vary with the properties and fine
a deeper placement of the tendons and a structure of vegetational surfaces,
more conservative (distal) branching of

Implications for the
"Retrograde" Series

the tendons serving individual lamellae.
But it is also interesting that the terminal
and basal pads in several gecko genera

are associated with separate internal ten- This general discussion suggests a

don systems (Russell, 1976). Any of these framework for evaluating the morpholog-

"potential constraints" beg the question ical trends in the auratus and annectens-

of an altered tendon origin and arrange- onca lineages. In both lineages the num-

ment in anoles, but nonetheless, mechan- ber of lamellae and pad area are reduced

ical constraints may be operating to limit relative to chrysolepis, and the "Norops"

the distribution of lamellae to certain condition appears. These morphological

phalanges, c) Lamellar morphology may parallels are probably not indicative of

increase the effectiveness of setae by en- precise parallels in behavioral and eco-

gaging more of them, but the advantage logical specialization.

of lamellar setae may be significant only The functional significance of the

on certain environmental surfaces. Both "Norops" condition is unclear at this

the number of lamellae (Collette, 1955; point. The distribution of the character

Williams, 1974) and seta distribution rel- within Anolis and its association with

ative to the lamellae (Peterson and Wil- trends in fine structure are not readily

Hams, in preparation) appear to be related interpretable. In the annectens-onca line

to the environmental surfaces used bv the the "Norops" condition is associated

The ONCA Lineage: Subdigital Fine Structure • Peterson diul Williams 249

with the invasion of multicarinate mor- these more terrestrial, but still general-

phology into the phalanx ii or pad region, ized anoles, there are no parallels for an-

while in auratus it is associated with the nectens. The mori:)hology suggests that

expansion of pad or adhesive function selection for "adhesion" based on setae

into the phalanx i region. In both cases is diminished in annectens and that the

morphological differentiation of the pha- range of vegetation and behavioral rep-

lanx i and ii regions is decreasing, but the ertoire of the animal are altered relative

effect appears to be functional diver- to generalized anoles. Annectens proba-

gence. The "Norops" condition, we em- bly is much more dependent on its claws,

phasize again, is not invariably a "retro- and surfaces which will not accept the

grade" change and an indication of claw, like grass blades and leaves, may

decline in adhesive function. not be as readily used.

While the number of lamellae and pad

area in auratus are reduced relative to Possible Functional Significance of

chrysolepis, they are comparable to those Scale Contouring

of West Indian grass anoles (e.g., pul- CONTOURS AS "Traction" Devices
chellus, where, however, the fane struc-
ture of the pad is different). Even if there Scale contours which are relatively rig-
were no data on the behavior and habitat id and permanent are probably "traction"
preferences of auratus (Kiester et al., devices which work with moderately
1975), there would be little basis to ques- large (on the order of 0.5 mm-1 mm) sur-
tion the functional competence of the face irregularities (e.g., on bark) or create
pad. The trends in auratus appear to rep- such irregularities in loose sand or dirt,
resent body size and habitat specializa- Keels, bosses, ridges, and even the distal
tion of the "adhesive" structures rather margins of carinate scales will interdigi-
than selection for an alternative "non-ad- tate with the gross surface irregularities
hesive" function. of bark or terrestrial substrates, and the
The situation is different in annectens. relative rigidity of multicarinate scales
If, as we have argued above, pad area is permits these contacts to transmit pro-
crudely indicative of the number of setae pulsive and reaction forces. Given irreg-
available to support the animal (in this ularities of the proper size or loose sand
species there are few non-lamellar setae) or dirt, carinate scales will increase the
and lamellae variously function in apply- static frictional coefficient of the contact
ing setae, and if interspecific compari- between the foot and environment (Gans,
sons can be used to infer what the abso- 1974) so that more force can be transmit-
lute values mean in terms of limits on ted across the contact without slippage,
adhesive behavior and vegetational hab- The keel can transmit forces applied a)
itat, then there are rather serious ques- parallel to the toe if the distal end of the
tions whether the annectens pad is func- keel is engaged by an environmental con-
tionally competent to support the animal, tour, and/or b) perpendicular to the long
For its size annectens has fewer lamellae axis of the toe if the crest of the keel is
and less seta area than 1) grass anoles like braced against an environmental contour.
auratus and pulchellus, 2) chrysolepis, Both probably occur, and the ostensibly
which uses a wide range of vegetational parallel keels on adjacent digits are able to
surfaces close to and including the forest "find" and exploit environmental irregu-
floor, and 3) the West Indian trunk- larities with a variety of spatial orienta-
ground ecomoiphs, such as cmtote//u.s and tions because the hind limb digits (1-4)
cybotes, which often sleep on leaves, but are placed at angles of 30°-110° to the di-
climb and forage from the lower trunk rection of motion (see also Robinson,
(see also Kiester et al., 1975). Among 1975). The long third and fourth toes are

250 Bulletin Muscttm oj CoDiparative Zoolo^ij, Vol. 149, No. 4

nearly perpendicular to the direction of" Keels may also represent adaptation for
motion (there is considerably greater di- behaviors which rely on the rapid appli-
versit\' in arboreal postural positions). The cation of large forces, e.g., running. From
interdigitating contacts created by the these hypotheses one would predict that
keels on these toes can very effectively keels and related scale contours would
transmit propulsive forces applied peipen- characterize climbing, trunk-dwelling,
dicular and parallel to the direction of mo- and terrestrial species, but not those spe-
tion. The proximo-distal orientation of the cializing in twig or smooth leaf surfaces,
keels may also be advantageous in provid- Among the West Indian anole eco-
ing traction as these toes rotate about their morphs, subdigital keels occur on some
long axis during the propulsive phase scales in the phalanx i region in the
(Laerm, 1973 describes this rotation for trunk-ground ecomorphs (i.e., A. cri.sta-
Basiliscus, but it occurs to some degree telliis, A. lineatopus, and A. cybotes);
in many lizards). Keels are rotated into in these species the other phalanx i
contact with the substrate during limb re- scales have striated, bare areas along
traction, and like caterpillar or tank tread, the robust curved margins of the scales,
could "find" or create contacts which The keels and bare areas have some-
maintain a high coefficient of static fric- what more relief, but are generally sim-
tion. In many terrestrial cursors (e.g., ilar to those in chry.solepis (Fig. 10).
Dipsosaurus) and some forms that com- Both morphologies are absent in twig
bine scansorial and terrestrial habits (e.g., anoles, e.g., A. valencienni and A. occul-
Enyalioides) keel shape, placement, and tii.s. The anole ecomorphs which use the
relief often display a preaxial/postaxial trunk and large diameter branches, e.g.,
asymmetry that is probably related to the A. distichus and A. stratulus, lack sub-
long-axis rotation of the digits during the digital keels, but have prominent bare
propulsive phase. areas on the phalanx i scales similar to

^ ^ those in chrysolepis and cristatellus.

The Evolution of Multicarinate ^hese contoured margins may be able to
Scales in Anolis function like keels in increasing traction
An evolutionary trend to substitute relative to forces applied parallel to the
contoured, multicarinate scales for un- toe (e.g., digital flexion in a leap), but
contoured scales may represent adapta- scale contouring appears to be less prom-
tion for using environmental surfaces inent in the more strictly trunk-dwelling
which would have low coefficients of species than in the trunk-ground eco-
static friction to a smooth scale. Low coef- morphs. Keels are not present in the West
ficients of static friction might arise for a Indian grass anole species we have ex-
smooth scale if the environmental surface amined (A. pulchellus, A. poncensis, A.
is formed a) by particles with low cohe- olssoni, and A. semilineatus) although
sion (e.g., sand) or b) by rigid irregulari- "bare areas" occur on some of the pha-
ties which offer too little contact area for lanx i scales.

the scale. The latter situation might occur Subdigital keels or marked scale con-
on tree bark or branches. Whether it does touring into ridges appear in several Cen-
or not depends on the relationship be- tral and South American anole species
tween the spacing of the surface irregu- (e.g., A. notopholis, A. tropidonotus, A.
larities and the dimensions of the digits, nebuloides, A. nebulosus, A. qiiercorum.
Keels and other contour features would A. liogaster, and A. doUfusianus) which
appear to be advantageous for locomotion combine terrestrial substrates with
on bark surfaces where the distance be- perches close to the ground. The keels
tween irregularities is somewhat less are confined to the phalanx i region and
than the width of the digit or on sand, are comparable to those in cristatellus in

The o.vcA Lineage: Subdigital Fixe Structure • Peterson and Williams 251

relief (the position of the keels on the than most of these iguanids. Leaves,

scale and the distribution of multicari- twigs, and other grossly smooth vegeta-

nate scales within the phalanx i area are tional surfaces are often used as sleeping

quite variable). sites or rarely as auxiliary perches. This

The available intergeneric compari- greater repertoire of surfaces may make
sons largely corroborate the pattern in it advantageous to maintain the pad and
Anolis. Multicarinate subdigital scales introduce keels in only a limited area ad-
occur in many (e.g., Agama agama, A. jacent to the claw. The trunk-ground
stellio, Dipsosaurus, Phrynosoma, Cro- structural niche might begin to approxi-
taphytus, Morunasaurus, Sceloporus, mate the iguanid scansorial/terrestrial
Enyalioides, Tropiduriis, and Iguana), habitat with an increase in body size or
but not all (e.g., Diplolaemus and Lacer- in extreme xeric areas where a) the vege-
ta viridis), the terrestrial and terrestrial/ tation may be deciduous and the leaves
scansorial lizards. The subdigital scales are often too narrow and small to support
are acarinate in Chamaeleo and Phena- the animals, and b) the presence of loose,
cosaurus, which are slow moving and sandy soil might increase selection for
characteristically use small diameter, scale contours as "traction" devices,
smoother surfaces similar to those used Body size is generally conservative
by A. valencienni. However, Polychrus among trunk-ground forms, but the West
shares these behaviors and habitat pref- Indian ecomorphs offer two cases to ex-
erences and has subdigital scales which amine whether scale contour features are
are crowded with low bosses and keels exaggerated and the pad characters re-
(Peterson and Williams, unpublished). duced when trunk-ground forms radiate

Although there are exceptions and the into arid regions. These are: the radiation

character of bark and terrestrial sub- of Puerto Rican cristateUus stock into

strates is more complex than we can con- scrub forest as A. cooki and the radiation

sider, the distribution of multicarinate of Hispaniolan cybotes stock into thorn

scales is generally consistent with their scrub forest as A. whitemani (Williams,

role as a "traction" device and with ad- 1963).

aptation for locomotion on surfaces Comparison of A. cristateUus and A.
which have a low coefficient of friction cooki indicates no marked difference in
because of gross surface irregularity or the contour features of the phalanx i re-
the low cohesion of surface particles, gion and no contour features present in
Within Anolis, keels occur in species the phalanx ii and iii regions, although
which use terrestrial substrates, but not lamellar number and pad width appear to
in the more strictly scansorial or grass- be reduced in cooki. Comparison of A.
dwelling forms (A. auratus appears to be cybotes and A. whitemani reveals that
an exception in possessing keels). The some A. whitemani individuals have
distribution of keels within the genus three or more robust keels on all the pha-
suggests that the trunk-ground structural lanx i scales while cybotes has laterally
habitat (broadly defined to include forms compressed scales with robust distal mar-
such as chrysolepis) might be the starting gins distally and two or three keeled
point for the elaboration of multicarinate scales proximally. A. whitemani shows
leatures in the annectens-onca lineage. no expansion of keels into the phalanx ii

It is interesting that trunk-ground ano- or iii regions, and there is no indication

les hav^ limited keels and an adhesive of reduction in lamellar number or pad

pad while most other scansorial/terrestri- size.

al iguanids have multicarinate subdigital The multicarinate scales in A. white-
scales. The smaller body size of trunk- mani combine the features of annectens
ground anoles permits them to use a and onca. Scale shape resembles the
broader range of vegetational surfaces onca phalanx i scales rather than the long

252

BitHetin Museum of Coinparative Zoology, Vol. 149, No. 4

cylindrical scales of annectens (Fig. lb).
The relief of the keels is less than that in
onca and comparable to that of most an-
nectens keels, but the lateral digital
scales have many sharp, protruding keels
as in onca. Hillock contouring between
the keels and toward the base of the scale
is comparable to that in annectens. The
positions and relative development of the
keels show no consistent resemblance to
annectens or onca. The more proximal
phalanx i scales have a dominant median
keel and a pair of lateral keels. This is
similar to the annectens phalanx iii pat-
tern or onca scales at the interphalangeal
joints. Midway between the claw and the
pad the pair of lateral keels are dominant,
and a second pair of lateral keels may be
added. This is similar to onca scales in
the mid-phalangeal regions. The most
distal scales are similar in all three
species.

A, whitemani occurs over a range of
arid habitats. It can occur in the same
general localities as A. cybotes. In these
areas cybotes is found in small oases of
trees adjacent to springs, while A. white-
mani occurs in surrounding open dry
scrub and out into sparsely vegetated
areas (Williams, 1963). Extreme white-
mani habitat is similar to that of onca.
Unfortunately, our series of whiteTuani
specimens is too small to examine the in-
dividual variation in subdigital keels in
terms of habitat variation, but the ap-
pearance of robust keels in at least some
specimens of a species which has radiat-
ed into sandy, extreme xeric habitat com-
parable to that of onca is striking.

The results of the two comparisons are
inconsistent: each case is unique, and
there are too few behavioral and ecolog-
ical data to interpret their differences.
But each appears to exhibit nascent
trends which parallel the annectens-onca
lineage: a) reduction in lamellar number
and pad width in cooki, and b) an in-
crease in the number and robustness of
the keels in some specimens of white-
mani.

It is, however, the coincidence of these
trends which characterizes the annec-
tens-onca lineage, and, to our knowl-
edge, the annectens-onca lineage is
unique among anoles in the combination
of trends for pad reduction and marked
expansion of multicarinate scales. If we
view this combination of trends as an ex-
treme in the spectrum of adaptation, rather
than as a truly unique evolutionary pat-
tern, the comparisons among anoles sug-
gest the following interpretation of the
evolutionary sequence in keel elaboration.
a) A few low subdigital keels occur in
the phalanx i region of a substantial
number of anole species which use ter-
restrial substrates. These less "arboreal"
or trunk-ground species also evidence
narrower pads and/or reduction in la-
mellar number compared to trunk and
canopy species of the same body size.
Grass anoles may parallel the trend in
pad reduction, but keels are infrequent.
If, as is suggested by juvenile onca
morphology, the "retrograde" sequence
begins with pad dedifferentiation, then
either grass or trunk-ground species
might be the starting point for the ra-
diation. But if, as is suggested by the
annectens morphology, the "retrograde"
sequence involves the progressive ex-
pansion of multicarinate features into
seta-bearing areas, then the most proba-
ble starting point for the annectens-onca
lineage is a trunk-ground habitat (broadly
defined), b) To our present knowledge,
multicarinate features are elaborated in
only three anole species: whitemani
(some individuals), annectens, and onca.
In the cases where the general habitat
associated with this shift is known, it is a
xeric region. Although our survey has
not been exhaustive, we have seen no
evidence of keel elaboration in trunk-
ground species with other habitat asso-
ciations. A. notopholis, a mesic trunk-
ground species, comes the closest to
being an exception. In notopholis all but
one of the phalanx i scales have long low
keels. Keel position and relief is compa-

The osca Lineage: Subdigital Fine Structure • Peterson and Williams 253

rable to that in auratus, and relief is
much less than that in the whitemani
specimens described above. Other eco-
morphs, particularly grass (e.g., auratus,
poncensis, and olssoni) and trunk (e.g.,
brevirostris) anoles, also radiate into xe-
ric areas. With the exception of auratus,
multicarinate scales do not seem to ap-
pear in xeric grass or trunk forms. Elab-
oration of multicarinate features in anoles
is not always (e.g., cooki) or perhaps ex-
clusively (e.g., notopholis) associated
with the invasion of xeric regions by
trunk-ground species, but there does ap-
pear to be some relationship among the
habitat, ecomorph, and subdigital keels.
c) Presumably the invasion of xeric areas
translates into diminished selection for
an "adhesive pad" (see the section on
behavioral ecology of oncfl) and selection
for scale contours which are more effec-
tive on lower friction surfaces like sand
and rough bark. Continued or extreme
selection for locomotion on these sub-
strates might be correlated with the en-
croachment or substitution of multicari-
nate scales for seta-bearing lamellae in
annectens and onca.

The Evidence from Non-Anole
Iguanids and Geckos

Subdigital keels are relatively uncom-
mon in anoles, but quite common in non-
anole iguanids. Some insight into the
habitat and behavioral correlates of keels
can be gathered from the non-anole igua-
nids if we examine not the origin or ap-
pearance of keels, but the interspecific
variation in keel shape and position. The
comparisons among non-anole iguanids
can be constructed to evaluate whether
a) variation in keel shape and position is
correlated with habitat differences simi-
lar to those implicated by the examina-
tion of anole species, and b) whether the
multicarnate scales of onca and annec-
tens are convergent on those of desert-
dwelling iguanids. A complete survey is
beyond the scope of this paper, but the

comparisons which we have constructed
as "test cases" proved quite interesting.

Enijalioides microlepis and £. hetero-
lepis are mesic forest species which dif-
fer in the proportion to which they use
bark and forest floor substrates (K. Mi-
yata, personal communication). Keel shape
and position are very similar in the two
species.

The pattern of keel placement is also
similar in two scansorial Tropidurus
species: T. torquatus, from mesic open
areas (Rand and Rand, 1966) and T. oc-
cipitalis from xeric coastal Peru (Dixon
and Wright, 1975). Both species have a
dominant median keel flanked by subsid-
iary lateral keels similar to the pattern in
the Enyalioides species and on the pha-
lanx iii scales of annectens. The xeric
species, occipitalis, however, does differ
from torquatus in several other features:

a) the keel tips and crests are sharper, and

b) keels on the lateral digital scales have
more relief. These features and the mor-
phology of the phalanx i region are con-
vergent on annectens.

Scale shape and keel development are
different in Tropidurus peruvianus, a xe-
ric Peruvian coastal species which is pre-
dominantly terrestrial and lives on
beaches, sand dunes, and boulders (Dix-
on and Wright, 1975). Keel asymmetr>' is
marked in peruvianus, but there is a pair
of dominant keels (probably the median
and postaxial lateral) except at the inter-
phalangeal joints where three robust
keels (the median and the pair of lateral
keels) are present. While the asymmetry
and probable homologies of the keels are
different, the emphasis on two robust
keels rather than a single dominant me-
dian keel is similar in onca and peruvi-
anas.

The more terrestrial, sand-dwelling
iguanids which we have examined {Cro-
taphytus wizlizeni, C. coUaris, Dipsosau-
rus dorsalis, Phrynosoma sp. and
Callisaurus) resemble T. peruvianus and
onca in having two, or usually more, rel-
atively equal keels in lateral positions

254 BuUetin Muscutu of Couipardtiic y.oolo'^ii. Vol. 149, No. 4

rather than a dominant median keel (tlie on a rigid irregular surface would prob-
homologies of the lateral and far lateral ably decrease the coefficient of friction
keels are not particularly clear because of by decreasing the area of contact or pre-
torsion and asymmetry in the scales, e.g., venting keels from projecting into crev-
C. wizlizeni has four or more keels all on ices). The xeric scansorial forms (Scelop-
the postaxial side of the scale).' In con- orus olivaceus and Tropidurus
trast, Basiliscus vittatus, a mesic cursor- occipitalis) utilize terrestrial substrates
ial/scansorial form, has a single keel that for foraging (Blair, 1960; Dixon and
is asymmetrical in orientation. Scelopo- Wright, 1975), and both display an in-
rus olivaceus, a xeric climber, has a ro- crease in keel number or robustness com-
bust median keel, although on some pared to mesic forms. This could be re-
scales the pair of lateral keels are as large lated to foraging on sand, but there may
as the median. In Iguana, there are two be other habitat differences, and the re-
relatively equal keels placed so close to- lationship is likely to be more complex,
gether in the median plane that they The difference in keel placement be-
might well operate as a single keel. tween annectens and onca is mirrored in

The xeric terrestrial iguanids appear to the difference between Tropidurus oc-
have a larger number of keels and more cipitalis and T. peruvianus, and the dif-
robust keels in the lateral portion of the ferences in subdigital scales between oc-
scale than the terrestrial forms in mesic cipitalis and torquatus are generally
habitats or than the more scansorial forms, convergences between occipitalis and
which as a group emphasize a medially annectens. The resemblance and com-
placed keel. This provides some corrob- parative data are not complete enough to
oration of the relationship between elab- suggest a probable habitat and behavior
oration of keeled scales and xeric habitats for annectens. But the difference in keel
in Anolis. Presumably the aspect of xeric placement and shape between annectens
habitats which is functionally significant and onca appears to be significant in
for subdigital scales is sandy or loose soil light of data from the non-anole iguanids.
substrates. It is possible that the in- Quite apart from the difference in pad re-
creased number and lateral placement of tention, the morphology of the mluticar-
the keels are more effective in stabilizing inate scales suggests that annectens may
sand to apply propulsive force while a be adapted for a somewhat different
single median keel is more effective on structural habitat than onca. Both pad re-
a curved surface or in working into bark tention and the morphology of the keels
crevices (a large number of keels acting in annectens appear to indicate a more
traditional anole structural habitat.

' Fringe-toed lizards such as Uma scoparia. The most striking parallel for the sub-

Acanthodactylus scutellatus, and Aporosaurus an- stitution of a keeled surface for the ad-

chietae are apparent exceptions to the generaUza- i • j r ^ d

.. r „ r 1 1 \ J J n- hesive pad may come irom geckos, nus-

tion oi greater keel numbers in sand-dwelling ,, .,^1,^ ^nrrA\ i -i i.u i

species. The subdigital scales in these forms are Sell (1976, 1979) describes the reduction

laterally compressed with a single median or med- or loss of subdigital pads in twO grOUps

ian preaxial keel and subsidiary keels are absent. of geckos which have radiated into xeric,

Tropidurus thoracicus, a relative of T. peruvianus ^^^^ habitats. In these cases the pad is

which even sand swims has similar subdigital i i i n i i

morphology (Dixon and Wright, 1975). Presumably replaced by small, erect granular scales

this morphology is adapted to trap sand between covered with spiues similar tO those in

the "fringe" on the lateral digital scales and the primitively padless, sand-dwellillg forms

subdigital keel , , , , (Russell, 1979). It is possible that the

Apparently there are a number OI alternative ad- n . i i i i-i • j- -j i

. J .. u 11 • *u CQ ,. c ij small, erect scales behave like individual

aptations which will increase the coerhcient of slid- i i i i i

ing friction between the animal and a loose sand keels and stabilize sand under the toot as

substrate (see also sidewinding [Gans, 1974]). propulsive forces are applied. Although

The onca Lineage: Subdigital Fine Structure • Peterson und William.'

255

little is known of the behavior of these
animals (they appear to be more terres-
trial than onca, see Werner and Broza,
1969), the parallel reduction in adhesive
structures in sand-dwelling geckos and
anoles is striking. It is possible that sand
injures or clogs the setae (Underwood,
1954, makes a comparable argument for
forest floor geckos), but anoles in captiv-
ity do not appear to have any difficulty in
climbing after walking or running over
sand, and pad retention in the xeric
trunk-ground species would suggest that
sandy substrates do not strongly select
against a pad and setae in anoles. We
would interpret the trend in Anolis in
terms of a) diminished selection for pad
morphology in arboreal/climbing behav-
iors (also discussed in the subsequent
section), and b) enhanced selection for
keels and "traction" devices which are
effective in loose sand or dirt.

Possible Functional Significance of
the Spine to Seta Series

The Spine to Seta Series:
Adhesion or Prehension

The moiphological series in fine struc-
ture is remarkably continuous and al-
most certainly reflects the interface of de-
velopmental processes and function. In
order to develop functional hypotheses
for the differences reflected in the series,
we will assume, for the present, that the
morphological series is also a phyloge-
netic one and that the spinule morphol-
ogy is a "starting point" for the subdigital
series.

There appear to be two distinct por-
tions of the series: the first involves the
transition from the generalized Ober-
hautchen spinules to the subdigital
spine-seta series (Figs. 21, 22), and the
second comprises variation within the
subdigital series (Figs. 21, 23).

The functional significance of the spi-
nulate Oberhautchen is not well under-
stood. The spinules may represent ad-
aptation for protection from physical

abrasion (Ruibal, 1968), the diffusion of
some wave lengths of radiant energy
(Porter, 1967), facilitation of shedding
(Maderson, 1970), and/or some yet un-
appreciated factor.

The transition from the spinules to the
subdigital spine-seta series might have
occurred in response to selection for
greater abrasion resistance, but this
would not seem to explain the acquisi-
tion of a recurved tip. It is more likely
that selection for purchase or grip is re-
sponsible. Initially, the spinulate surface
itself might in some fashion interdigitate
with the environmental surface and in-
crease the frictional coefficient of the
contact. The longer spinules (see Fig.
22a) might even be able to engage envi-
ronmental irregularities like small hooks
(the "velcro" principle). Selection to aug-
ment the microscopic prehension or in-
terdigitation of some of the spinules
could result in exaggeration of the stalk
height and acquisition of a recurved,
more hook-like tip (e.g., the small spine
morphology shown in the central portion
of Fig. 22b; note in the figure that not all
the spinules have been exaggerated into
small spines: there is an understory of
very short spines and knobs similar to
those shown in Fig. 22a). Presumably the
height increase is selectively advanta-
geous because a consistent increase in
stalk height would produce too dense a
"pile" for the tips to function. The selec-
tive increase in height establishes or ag-
gravates a functional differentiation be-
tween spinules in the understory which
do not function in prehension and small
spines which do, and it alters the func-
tional density. The spinules in Fig. 22a
occur in a density greater than 10/sq /x;
the small spines (excluding the under-
story in Fig. 22b) are about half as dense.
Continued selection for prehensile, mi-
croscopic hooks could then result in in-
creased stalk height to augment the ef-
fective range of the hook and in increased
diameter to tolerate larger tensile load-
ings. The increase in stalk diameter is

25f) Bulletin Museum of Comparative TjOoIo^ij, Vol. 149, No. 4

necessarily linked with reduction in den- snbdigital surface were studded with

sity and occurs at the expense of the spi- spines.

nules in the understory. The generalized Within the anole snbdigital series

snbdigital spine morphology is reached there are two trends operating to create

with a height of at least 1 ^t (usually great- the different morphologies: one in stalk

er, see Figs. 22, 23), a diameter close to height and one in tip shape.

0.5 IX near the base, and a density of 2 or 1) The variation in stalk height is con-

less/sq fx. tinuous and not closely correlated with tip

The collective alterations involved in shape (e.g., the sharp, recurved tip and
the shift from spinules to snbdigital spatula occur over a significant range of
spines could create a "prehensile" scale stalk heights). Stalk height affects the in-
surface capable of augmenting the claw, dependence of the tip from the scale sur-
It might augment the claw in such a way face. A longer stalk permits the tip a
as to permit acrobatic and perch-walking greater range of positions and increases
maneuvers over a wider range of sub- the odds that the tip will contact the en-
strates. The effect of a "prehensile" scale vironmental surface. Because the stalks
surface would presumably be least on the are stressed in tension, variation in their
tree trunk and greatest in the bush/can- length does not alter their strength nor
opy branch and twig zones where the sur- require a compensatory change in diam-
faces accept the claw less well and the eter (Peterson ef a/., in preparation). While
radius of curvature of the perch is small, there is no purely mechanical disadvan-
Some of the foot positions which anoles tage or cost associated with increasing
use in acrobatic maneuvers on small stalk length, presumably there is a de-
perches would not be possible unless some velopmental and material cost. Excep-
grip independent of the claw were per- tionally long stalks would also be more
mitted. If spine-covered subdigital scales, likely to tangle and trap debris, and the
like the series of phalanx i and iii scales in properties of the keratin would have to
many anoles, could contribute to grip, the be such as to maintain the tip orientation
shift from spinules to the subdigital spine relative to the scale surface regardless of
series would have significant impact on stalk length.

the range of vegetational surfaces and the The longest stalks generally occur on la-
repertoire of behaviors. mellar setae (there is significant interspe-

Hiller (1968) argues that, while the cific and regional variation in setae stalk
spine tip might behave like a microscopic height; Peterson and Williams, in prepar-
hook, the behavioral significance of ation). The length of the stalk coupled
spines is negligible. His argument is with the flexibility of the lamellae permit
largely based on experiments with gee- the spatulate tips to reach surfaces which
kos adhering to a series of artificial sur- have a small radius of curvature or are
faces which are much smoother than irregular at the ultrastructural level (see
twigs or branches, and probably too discussion of tip function following),
smooth for the small number of spines This is an extremely important aspect of
present to develop much gripping force seta function in anoles because the total
(these experiments are discussed in more number of spatulate tips is not particu-
detail below). Sanding of the surface to larly high (cf. the branched setae of gee-
roughen it resulted in a slight increase kos) and the total area of the spatulae is
(value not reported) in the gripping force actually less than that of the underlying
(Hiller, 1968). It is simply not clear that lamellae (Peterson and Williams, in prep-
Hiller's (1968) objection would hold for aration). The advantage of setae regardless
natural surfaces or if major portions of the of their mechanism is not exclusively in

The onca Lineage: Subdigital Fine Structure • Peterson and Williams 257

the area or number of tips which they of-
fer to the environment, but in the abiHty
of those tips to "find, " mold to, and en-
gage the substrate, and this substantially
depends on the dimensions and proper-
ties of the stalk.

Short stalks occur in narrow zones on
the periphery of seta, prong, or spike-cov-
ered areas (Fig. 23), often where the scale
surface is shadowed by another scale,
and occur widely over the surface of
more rigid, often contoured scales. The
tips on stalks along contour slopes and
valleys will contact the environmental
surface if a) the contour interdigitates
with a permanent irregularity of the sub-
strate, or b) the substrate molds to the
scale contour. But given scale relief an
order of magnitude greater than the max-
imum observed stalk length, "finding"
the substrate does not depend on stalk
length. Possibly the fine structure of con-
toured scales is not selected to "find" the
substrate as setae do, but to interact
with particular portions of the substrate
which "find" the stalks by offering resis-
tance or purchase to the scale contours.
Short stalks, regardless of the associated
tip shape, may be functionally advanta-
geous on multicarinate scales it, by
"roughening" the keel slopes, the stalks
increase the frictional coefficient of the
scale surface.

2) Although the variation in tip shape is
continuous, there are two basic and nu-
merically dominant shapes, the hook and
spatula (Fig. 21). The hook tip on spines
and spikes (and even prongs) functions
like a microscopic hook to engage irreg-
ularities of the environmental surface
(see above and Hiller, 1968). There is
a variety of hypotheses on spatular mech-
anisms (see Russell, 1975 and Hiller,
1968 for review), but two seem most like-
ly for anoles. Compression and lateral ex-
pansion of the hook tip into a spatula
could be an adaptation a) to shift from a
prehensile mechanism to an adhesive
one (Hiller, 1968 for geckos), or b) to add

the option of an adhesive mechanism to
the existing prehension. In either case,
the difference between the hook and the
spatula seems to be adhesion. An adhe-
sive gripping force depends a) on the sur-
face energy or tension of the environ-
mental surface and of the seta keratin,
and b) on the area of contact between
them. Given that the seta keratin itself is
preadapted for adhesion, selection might
increase the area of contact of the tip in
the following sequence (Fig. 21): a) re-
duction in the stalk taper to yield a thick-
er, larger tip — the prong, b) compression
of the round or oval prong to form a pla-
nar surface and increase the effective
area — the seta-prong, and c) expansion of
the flattened surface — the seta. The
prong and seta-prong shapes are very rare
in generalized anoles (Fig. 23) and ap-
pear to represent developmental and
evolutionary transitions linking the hook
and the spatula.

There is no direct evidence for an ad-
hesive nor for a prehensile mechanism in
anoles. Hiller (1968) argues strongly for
an adhesion mechanism based on the
morphology of the spatula and on exper-
iments measuring the clinging force of
geckos on a series of artificial surfaces.
The force required to dislodge a gecko
from a horizontal surface was directly
proportional to the surface tension of the
material (Hiller, 1968). Considerable
force was required to dislodge the animal
from glass which has high surface tension
(contact angle of water droplet with glass
is 0°), while the animal had little or no
adhesive strength on certain plastics of
low surface tension (water-drop contact
angles of 80° or more). Hiller (1968) ar-
gues that "velcro"-like prehensile mech-
anisms can be excluded because a)
roughening the surface of a low surface
tension material results in only a slight
increase in gripping force, and b) pulling
the animal perpendicular rather than par-
allel to low ridges on the unroughened
surface of a low surface tension material

258 Btillctiii Mtisctiiii of Coinixinitirc y.oology. Vol. 149, No. 4

resulted in only a slight increase in grip- many surfaces to better than 110°; in Hel-
ping force (values were not reported in loway's (1969) study of the leaf surfaces
either case). In Hiller's adhesion hypoth- of 40 connnon plant species, 24 have con-
esis whole animal adherence occurs as tact angles greater than 110°, and an ad-
the spatula is brought into intimate con- ditional 6 have contact angles between 90°
tact with a flat, high surface tension sub- and 110°. Presumably the small size and
strate. Surface roughness or ultramicros- active application of setae will permit the
copic irregvdarities will interfere with spatulae to "find" the smooth portions of
this process so that the animal is unable the surface better than a water droplet
to cling to such surfaces without its which rides on the peaks of the irregu-
claws. Russell (1976) augments Miller's larities, so that the contact angles in ex-
argument with the hypothesis that the cess of 110° may not themselves be rel-
waxy secretions on leaves and smooth evant. But, major portions of common
bark provide the initial advantage to seta- vegetational surfaces (leaves, young bark,
bearing geckos. stems, and particularly grass blades) are
There are two major difficulties in ac- unsuitable for adhesion. They have been
cepting or extending the adhesion hy- selected for surface irregularity and low
potheses to anoles. 1) The surface prop- surface tension to inhibit the adhesion of
erties of vegetation suggest that, if an water droplets (Martin and Juniper,
adhesive mechanism alone is possible, 1970). 2) There is little indication that
anoles could not use some of the sub- anoles in natural habitats are selecting
strates they clearly thrive on. The major- particular plant species or surfaces which
ity of plant surfaces appear to have lower might have high surface tension and low
surface tension than Hiller (1968) found surface irregularity (but see Kiester e^ a/.,
was required in geckos (contact angles of 1975 for auratus). In the laboratory, ano-
80° or less were required for even modest les will adhere to smooth, high surface
adhesive force). The common chemical tension glass and waxy, low surface ten-
constituents of plant waxes have (water sion leaves (neither surface will accept
drop) contact angles from 94°-108°; the the claw). The discrepancy among Hill-
isolated smooth surfaces of the native er's findings, the properties of natural
waxes yield angles between 80° and 108° surfaces, and the behavior of anoles need
(Martin and Juniper, 1970). The accu- to be addressed directly with adhesive
mulated SEM data on leaf and stem sur- force experiments on natural surfaces,
faces (see review in Martin and Juniper, but provisionally we would suggest that
1970) indicate that the surfaces are also spatular morphology permits coexisting
highly irregular and rough in the dimen- adhesion and prehension mechanisms,
sional scale of the spatulae: waxes com- The eccentric attachment of the spatula
monly exist in projecting rods or semi- to the stalk and the broad leading margin
crystalline excrescences; the epidermis could permit the setae to function as a
may be grooved or the individual cells hook. It seems to us possible that the foot
convex or granular; trichomes often pro- and lamellar movements wipe the spat-
ject from the surface of leaves; stems and ulae over the environmental surface until
twigs have cracks, ridges, rough scars, the leading edge of the spatula catches
and considerable debris. Only a portion, on a wax rod/crystal or cuticular irregu-
and often a small portion (e.g., 28% on larity or until the spatular surface finds a
corn leaves, Martin and Juniper, 1970), of smooth, high surface tension area of cu-
the leaf or bark area is flat and smooth tide. Either would establish the grip and,
enough to make close and complete con- since most leaf and twig surfaces are
tact with the spatula. Fine structural ir- rough/smooth mosaics, even adjacent se-
regularity raises the contact angle of tae might exhibit different gripping

The onca Lineage: Subdigital Fine Structure • Peterson and Williams 259

mechanisms. Anole setae probably rep-
resent an adaptation for this mosaic sur-
face structure and a variety of surface ten-
sions.

Implications for the
"Retrograde" Series

This general discussion of the spine-
seta series provides a context within
which to evaluate the trends in the an-
nectens-onca and auratus lineages. The
two lines offer an extraordinary contrast,
not in the morphology of members of the
series, for all but onca have quite similar
representatives of all stages, but in spe-
cialization for long or short stalks and for
hook or spatulate tips. In generalized
anoles the subdigital surface is covered
with spatulate and hook tips in about the
proportions 60/40 or 70/30, depending on
body size and ecomorph. In auratus the
series is shifted even further toward the
spatula and relatively long stalks. All the
subdigital scales bear setae, and spines
are quite rare. In the annectens-onca
line, the series is shifted in the opposite
direction, until in adult onca only spines
are present. The functional significance
of this divergence is at least in part relat-
ed to the substrate features.

A. auratus and Grass Surfaces

The characteristic substrates of aura-
tus, grass blades, sheaths, and stems
(culms), will not accept the claw, and
most of the surface irregularities are too
small for scale contouring to be effective.
Purchase largely, if not entirely, depends
on adaptations of the spine-seta series.
Substrate selection should favor long
stalks to permit the tips to "find" the sur-
face, since grass is too rigid to comply
and "find" the tips, and either a hook
(spike) or spatulate (seta) tip, depending
on surface tension and irregularity. The
surface characteristics of the Panamanian
grasses are not known, but the available
data suggest that most Graminaceae have
a) ribs or keels on stems, sheaths, and

blades, b) prominent surface corrugations
or groove/ridge contours between the ribs
on the blades, c) trichomes, and d) a coat-
ing of surface wax (Martin and Juniper,
1970; Challen, 1960; personal observa-
tion). Challen (1960) found contact angles
of 135° and 1 10° on two temperate grasses
{Agropijron repens and Festuca praten-
sis). Removing the surface wax reduced
the angles to about 67° and 77°, respec-
tively (Challen, 1960). The common Pan-
amanian grasses {Paniculum pilosum,
Paspalum conjugatum, and P. panicula-
tum; Croat, 1978) probably used by au-
ratus have surface ribs and corrugations
at least as prominent as those of the
species investigated by Challen (1960).
The surfaces auratus uses are clearly
"rough" in the dimensional scale of the
tips and probably have a relatively low
surface tension. Systematic descriptions,
however, indicate regional variation in
some of the sources of surface roughness
(e.g., the stem may lack trichomes while
the sheath or blades are pubescent;
Croat, 1978). Although a non-adhesive
mechanism appears to be required and
dominant, there is sufficient variation
among the regions of the grass plant and
on an ultrastructural level (e.g., smooth,
moderate surface tension cuticle exposed
between wax crystals and adjacent to the
ribs) to permit the possibility of supple-
mental or occasional adhesive contacts.
If these adhesive contacts are behavior-
ally significant and if anole setae can
function in either prehension or adhe-
sion, selection related to grass substrates
would favor setae or a combination of se-
tae and tall spikes. The West Indian grass
anole, A. puIcheUus, exhibits the second
condition: setae and large numbers of
spikes occur on a well-developed pad. In
auratus, the seta spatula is somewhat
narrower, but setae cover almost all of the
subdigital surface. The distribution of se-
tae to all phalanges is apparently corre-
lated with the absence of an expanded
phalanx ii pad and with dimensions and
density parameters within the spike to

2fS() Bullcliii Mu.sciiiti of CoDifxmitivc V.oolo^ii, Vol. I4h), No. 4

seta series (see discussion of lamellar
scale shape). The auratus morphology
represents adaptation for the use of grass
or grossly smooth vegetational surfaces
and for a narrow pad region. The differ-
ences between auratus and pulchellus
in pad width and in the specialization for
setae versus the combination of spikes
and setae are major, but we cannot yet
interpret their functional significance. It
does seem reasonably clear, however,
that if the environmental surface is gross-
ly smooth, and fine structure is respon-
sible for whole animal adherence, setae
or the spike-seta portion of the morpho-
typic series is selectively advantageous.

A. onca and Diminished Selection
for Setae

While onca (and presumably annec-
tens) climb, setae would not necessarily
be appropriate or advantageous, given
the climbing behavior and vegetation
(see Williams, 1974; Collins, 1971). The
vegetation in onca habitat is unusual for
an anole; there are no broad leaf plants
and few trees. Fence posts and the dense
maze of thorn scrub bushes are readily
negotiated with claws. The thorn scrub
branches are used in combination (two or
more perches support the animal) so that
foot positions can be chosen to take ad-
vantage of the claws and even opposing
combinations of digits. Dipsosaurus, a
predominantly terrestrial iguanid that
also clambers into sage and other scrub
bushes, exhibits very similar "climbing"
behavior. In captivity onca climb bark
and apparently "rough" surfaces and
"walk" through a closely spaced mesh of
small diameter perches as well as any
other anole. If onca are placed in a gen-
eralized anole habitat — the more open, ar-
boreal network of perches or a trunk-
ground interface — they do not seem to dis-
play the repertoire of fast acrobatic and
single perch maneuvers which most oth-
er anoles have. When the distance be-
tween perches is greater than a comiort-
able step or a short hop, they usually drop

from the perch rather than leaping. Al-
though the scrub bush is superficially
like the arboreal habitat where anoles ob-
viously depend on seta/pad function, the
mesh of perches is dense relative to adult
onca body size, and there is a meter or
more distance between bushes. These
factors effectively preclude the single
perch and fast acrobatic behavioral strat-
egies of generalized trunk-ground anoles.
The vegetational structure permits onca
(and Dipsosaurus) to use climbing be-
haviors that require precise foot place-
ment but are not otherwise very different
from those used in terrestrial locomotion.
Given the character of the vegetational/
structural habitat and the behavioral
strategies of anoles, it is likely that hab-
itat selection for seta/pad function is di-
minished in adult onca.

Juvenile onca are found in thorn bush-
es and the dense dry grass at the base of
the bush (K. Miyata, personal communica-
tion). Sandy areas surrounding the thorn
bush islands seem to be rarely encoun-
tered, and this could reduce the selec-
tion for multicarinate features in hatch-
lings. Also, the body size range over
which the juvenile morphology occurs
is about half that of the adult (Mader-
son and Williams, in preparation), so
that the relationship between body size
and perch/perch matrix dimensions
is more like that of a grass anole than
that of adult onca. It is not known
whether juvenile onca utilize the behav-
ioral strategies of generalized grass or
trunk-ground anoles, but it seems feasi-
ble for juveniles to perch on a single
twig, grass blade, or thorn bush leaf
while it does not seem feasible for the
adults. The surface properties of these
substrates might permit adhesion, so that
it is not clear that the advantage of setae
would be diminished for the juvenile to
the same degree as it appears to be for
the adult. But, grass, twigs, and probably
the available leaf surfaces would almost
certainly provide microscopically rough
substrates which would be engaged by

The oxcA Lineage: Subdigital Fine Structure • Peterson and Williams 261

the prehensile elements of the fine struc-
tural series (prong/spike/spine). Robust
keels like those of the adult would prob-
ably be less effective in dealing with this
range of substrates than setae or the pre-
hensile forms in the fine structural series.

Juvenile onco morj^hology underlines
the argument that diminished selection
for setae is not by itself an adequate ar-
gument for the substitution of multicari-
nate scales and the convergence toward
sand-dwelling iguanids demonstrated by
the adult onca morphology. If lamellae
and setae are developmentally expensive
(Williams, 1974) and not particularly ad-
xantageous in locomotion, it is naively
reasonable to expect the substitution of
spine/spike-covered generalized subdi-
gital scales like those in the phalanx iii
region of many anoles or in juvenile onca
or in some of the para-anoline genera
(Peterson and Williams, in preparation). A
number of semiaquatic, streamside anole
species retain the pad and gross charac-
ters of generalized anoles, but spikes and
prongs have completely replaced setae
on the subdigital scales (Peterson and
Williams, in preparation).

Adult onca morphology reflects a par-
allel but more extreine "retrograde" shift
in the spine-seta series coupled with a
trend to contour the scales. It appears
that this is related to the combination of
sand and bark substrates. The surface ir-
regularity which dominates bark or must
be created in sand is at least an order of
magnitude larger than that on grass
blades. Purchase and traction on bark and
sand depend on adaptations — claws,
scale shape, and keels — at a larger di-
mensional scale than spikes and setae.
The subdigital surface may be selected
for short stalks and hook tips to increase
the frictional coefficient of the keel
slopes or even to contribute some grip by
prehension, but much of the propulsive
force is probably transmitted indepen-
dent of the spines.

The comparatively reduced signifi-
cance of the spine-seta series in actual

grip may be indicated by the onca spine
morphology itself. Except where the
spines are associated with a contour, as
on a hillock or adjacent to a keel, their
height and diameter are less than those
of generalized subdigital spines. With a
diameter of 0.3-0.45 /x and a density of
1.2 spines/sq fi, most of the scale surface
is bare. By itself, the reduced size of the
spines suggests a simple retrograde shift
toward Oberhautchen spinules, but the
combination of low spine density and
small individual spines appears to indi-
cate a general trend toward repression of
the spine-seta series and a completely
bare contoured surface similar to that of
multicarinate scales in some iguanids,
e.g., Dipsosaurus.

A. annectens and Juvenile onca as
Alternative Models for the
"Retrograde" Series

The earlier paper outlined a five-stage
sequence in the "retrograde" evolution
from the typical anoline pad morphology
to the keeled subdigital surface in an-
nectens and onca (Williams, 1974).
Through the first three steps — narrowing
of the lamellae, reduction in the number
of lamellae, and loss of the raised char-
acter of the pad — the pad dedifferentiates
from the adjacent phalanx i and iii scales.
If phalangeal length is constant, narrow-
ing of the pad and reduction in the num-
ber of scales convert the lamellar scale
shape into that of generalized subdigital
scales. Pad dedifferentiation characteriz-
es a variety of "Norops" species, includ-
ing auratiis, and is approached by A.
chnjsolepis. The fourth and fifth steps in
the sequence comprise the shift from
generalized subdigital scales to multicar-
inate scales. Keels are introduced into
the phalanx iii series in annectens (step
four), and in onca the entire series of sub-
digital scales is multicarinate (step five).

The additional comparisons and fine
structural data alter our view of the se-
quence in several ways. 1) It is no longer

262 Bulh'lin Museum of Compaaitivc Zoolo^ij, Vol. 14'^), .Vo. 4

clear that auratus, as an example of a he considered an ontogenetic adaptation

"Norops" species, is an appropriate mod- with no relationship to the phylogeny of

el for the early stages in the secjuence. adult morphology.

The expanded seta distribution in aura- The critical difference between these

tus demonstrates that dedifferentiation of sequences is whether the adhesive pad

the pad is not necessarily linked to a de- dedifferentiates prior to the appearance

crease in the seta-bearing area nor to the of multicarinate scales or whether mul-

substitution of multicarinate scales. 2) ticarinate scales progressively encroach

There appear to be two alternative se- on and replace lamellae. These different

quences. modes of origin for the multicarinate

If juvenile onca morphology is used to scales correspond to two possible models

model the intermediate stages, the se- for transitional scales. Onca juvenile

quence might a) begin with a trunk- scales (e.g., scale 13) have the gross

ground or even more probably a grass an- shape of the adult multicarinate scales

ole with gross morphology similar to that and rounded contours without the bare

of auratus, b) progress by more extreme areas which characterize keels. A "retro-

dedifferentiation of the pad region (in- grade" shift in the spine to seta series has

eluding loss of the frayed lamellar border occurred over most of the scale. The cen-

and extreme narrowing of the lamellae) tral portion has prongs, a "more conser-

combined with the replacement of setae vative" morphology than the spikes. The

by spikes, prongs, and spines to yield the latter occur distally and laterally, sug-

gross morphology of the onca juvenile, gesting that the trend progresses inward

and c) culminate with keels and marked from the scale margins. But, on the

contour features throughout the series of whole, the scale surface is reasonably ho-

subdigital scales (the onca condition). In mogeneous, and adjacent scales have

this scheme A. annectens would repre- very similar achitecture.

sent a parallel lineage independent of A. annectens scale 9 suggests a very

that of onca after the first step in the se- different mode of evolution in scale ar-

quence. chitecture. Here the transitional scale

If, on the other hand, A. annectens shape is close to that of the lamellae, but

morphology is considered representative the distal margin is robust and set with

of the intermediate stages, the "retro- bosses or keels. The bosses are set into

grade" sequence could a) begin with a fields of spines, spikes, prongs, and a few

trunk-ground anole (broadly defined) not rare setae. The fine structure is extra-

unlike A. cJirysolepis, i.e., perhaps with ordinarily heterogeneous, apart from the

narrow lamellae, but a defined pad re- distal border which is covered witli spines,

gion and a few phalanx i keels, b) contin- Compared to the morphology of scales in

ue with elaboration of the keel features the onca hatchlings, the fine structure of

in the phalanx i region (the A. its/nfemrtn/ scale 9 appears chaotic and disrupted,

variant morphology), c) follow this by in- Scale 9 is also markedly different from the

vasion of keels into the phalanx iii region lamellae and multicarinate scales on either

of generalized subdigital scales (the A. side of it. The small keels which are pres-

annectens morphology), and d) finally ent on the lamellae appear to become

substitute multicarinate scales for lamel- progressively' more robust toward the

lae and setae in the phalanx ii region (the interphalangeal joints (at scales 9 and 15).

onca condition). In this scheme the "Nor- The morphology itself suggests that the

ops" condition may arise at any point transition begins as keels invade the sur-

before the substitution of keels for la- face of lamellae.

mellae and setae in the phalanx ii region, It is not clear which of these two se-

and the juvenile oncfl morphology would quences or models of the transitional

The onca Lineage: Subdigital Fixe Structure • Peterson and WilliaDis

263

morphology are correct. We cannot pro-
vide incisive arguments for or against
either of them, but the second or A. an-
nectens model seems somewhat more
plausible.

1) If onca and annectens represent par-
allel lineages which have been indepen-
dent from the chrysolepis-\\\.e stage of
the sequence, it is difficult to explain the
close resemblance in non-digital charac-
ters (Williams, 1974). There are marked
differences in digital characters (e.g.,
phalangeal length, scale shape, keel
placement, and the retention of lamellae
and setae), but these could represent ad-
aptation to somewhat different structural
habitats.

2) Each of the model intermediates has
equivocal status or relevance to the evo-
lution of onca. A. annectens is a unique
specimen which closely resembles onca
in non-digital characters. There is the
possibility it is an individual variant of
onca. Elements of the morphology might
be relevant to the sequence in this event,
but the particular combination of char-
acters on which the sequence is based
could as readily reflect developmental
patterns as the phylogenetic sequence.
The juvenile onca morphology may, on
the other hand, imply the phylogenetic
sequence or reflect ontogenetic adapta-
tion for a smaller body size and different
habitat and behavior. There are two crit-
ical features on which the juvenile mor-
phology makes any claim as an interme-
diate between generalized anoles and
onca: the absence of definite keels and
the presence of spikes and prongs. The
absence of definite keels on the subdig-
ital scales in the hatchling of another an-
ole species (A. cristatelliis) suggests that
this character is not restricted to onca ju-
veniles. It may be a feature of juvenile
anoles and an ontogenetic adaptation
which coincidentally parallels the phy-
logenetic sequence. The presence of
well-developed contours usually in the
position of the adult keels and on all the
subdigital scales also suggests a very de-

rived status in terms of keel elaboration.
Juvenile cristatelliis have setae and the
same fine structural series as the adults,
but with decreased stalk height. The oppo-
site difference — increased stalk height —
in the juvenile onca appears to be a fea-
ture that is unique to onca juveniles. If
the onca juvenile morphology is con-
verted into a hypothetical adult by the
addition of bare areas to the contours to
form keels, the result is a form which is
very similar to onca and an apparently
later member of the transitional series
than annectens. If the hypothetical adult/
juvenile onca is used as the basis of a
revised "retrograde" sequence, it sug-
gests very little about the early stages of
the lineage, particularly about keel origin
and expansion, so that a number of se-
quences are possible, including the early
part of the annectens sequence. Any of
the juvenile onca sequences and the an-
nectens sequence differ markedly on the
issue of a sequential versus a simultane-
ous shift to multicarinate scales in the
phalanx ii and iii regions. Even the hy-
pothetical adult version of the juvenile
onca morphology still leaves the two
model intermediates as mutually exclu-
sive forms in the onca lineage.

3) The alternative sequences and
models for transitional scale types can
also be compared in terms of circumstan-
tial support from the comparative data.

a) The pattern of transitional scales
based on the onca juvenile morj^hology
has some comparative support although
it is unique in most features. In the liono-
tits species group there is a general trend
to replace setae with spikes, and on sev-
eral lamellae in A. lionotus only the cen-
tral portion of the scale has setae while
the surrounding areas are covered with
spikes. The "retrograde" shift in fine
structure appears to occur from the lateral
and distal portions of the scale inward,
but, in this case, the shift occurs on a la-
mellar scale and there are no contours
present. The annectens scale 9 fine struc-
ture is to our knowledge unique, but

264 Bulletin Museum of Comparative Zoology, Vol. 149, No. 4

apart from annectens, there are no ex- diffieult to find a model ancestor for the
amples of keels invading lamellae. Given juvenile onca morphology. All of the pro-
that keels are being added to the scale spective models with which we are fa-
surface, the disrupted, heterogeneous miliar (eg., A. chry. sole pis, A. barkeri, A.
fine structure is no less plausible than the notopholis, A. tropidonotus, A. auratus,
more homogeneous fine structure of the A. nebulosus) have keels or bare areas.
onca juvenile. Based on the morphology of these more

b) The transitional scale model^id se- generalized related species, either keels
quence based on the onca juvenile are and bare areas are lost in the early por-
unusual for anoles because they imply so tion of the onca juvenile sequence and
little regional differentiation and sequen- then "re-acquired" at the onca stage or
tial change. The "retrograde" shift from the absence of bare areas is not descrip-
setae to spikes in A. lionotus appears to tive of the phylogenetic intermediate. If
have occurred in a scale-by-scale se- one accepts, as is argued above, that the
quence beginning at the i/ii and ii/iii in- absence of bare areas on the contours is
terphalangeal joints and progressing to- an ontogenetic adaptation and that the
ward the center of the pad. This is similar adult intermediate morphology includes
to the suggested pattern of keel encroach- keels, then any of the forms cited above
ment in annecten.s and is markedly dif- could model the starting point of the
ferent from the simultaneous change in onca lineage. The gap between any of
two or more phalangeal regions suggest- the prospective initial morphologies and
ed by juvenile onca. The expansion or that of juvenile onca is so great that it
retraction of setae in the phalanx ii region simply is not clear how the sequence
(the phalanx iii and iv regions of the might start. The "Norops" condition is
fourth toe) of West Indian anoles also ap- combined with narrow but definite la-
pears to occur with a sequential scale-by- mellae in A. notopholis and A.
scale pattern. Marked differentiation in tropidonotus, but there is no evidence of
fine structure of the phalanx i and ii re- a "retrograde" shift in fine structure or
gions occurs even in some extreme "Nor- conversion of phalanx ii lamellae into
ops" species like A. aequatorialis, A. no- generalized subdigital scales. A. auratus
topholis, and A. meridionalis. The exhibits the "Norops" condition, narrow
dominant morphological and evolution- lamellae and poor regional differentia-
ary patterns within the genus are regional tion in fine structure, but the latter is as-
differentiation and sequential change sociated with an expansion of setae into
(scale-by-scale and phalangeal region-by- the phalanx i region rather than a "retro-
phalangeal region). Clearly the onca lin- grade" shift. A. barkeri exhibits a "retro-
eage could be exceptional in this regard, grade" shift in the fine structure of the
but the sequence and transitional mor- phalanx ii and iii regions (prongs are
phology implied by annectens is quite present), but there is little indication of
consistent with the patterns we have en- pad dedifferentiation or replacement of
countered elsewhere in the genus while lamellar scales. The differences between
those suggested by the onca juvenile are generalized anoles and juvenile onca
not. suggest that three trends occurred in the

c) There is no great difficulty in deriv- intervening stages: dedifferentiation of
ing A. annectens from generalized trunk- the pad and lamellae, a "retrograde" shift
ground forms such as A. chrysolepis, and in the fine structure, and expansion of
parallels for the early stages in the se- contours and keels into the phalanx ii and
quence can be found among other lin- iii regions. But there is no real indication
cages (e.g., A. whitemani). But it is very of the order or relative emphasis of these

The osca Lineage: Subdigital Fixe Structui^ • Peterson and Williaiits 265

trends during the early portion of the ra- CONCLUSIONS
diation.

Summary: The relative efficacy of the The significant differences in the sub-

annectens and juvenile onca specimens digital morphology of annectens, onca,

as phylogenetic models for the onca lin- chrysolepis, and auratus can be de-

eage cannot be evaluated with any de- scribed in terms of changes in three mor-

gree of security from the available data, phological series: a) scale shape, b) scale

The possibility that annectens is an aber- surface contouring, and c) fine structure

rant version of onca or represents an in- or the spine-seta series. The series are

dependent radiation has to be weighed quite similar in the different species, and

against the possibility that the differ- the dominant evolutionary pattern is

ences between onca juveniles and adults change in the distribution or relative

coincidentally parallel the phylogenetic abundance of members of the series,
sequence. But the odds appear to be There appear to be two lineages: the

shifted slightly in favor of annectens as chrysolepis-auratus lineage and the

a model. 1) The similarity between an- chrysolepis-annectens-onca lineage. An

nectens and onca in non-digital charac- alternative hypothesis is that the latter

ters is so great that it seems unlikely that species are divergent and that the mor-

they represent as distantly related forms phology of onca juveniles rather than

as the onca juvenile morphology would that of annectens is the relevant model

imply. 2) The individual variation we for the evolution of adult onca morphol-

usually encounter (e.g., whitemani) in- ogy.

volves a single morphological series. In the auratus lineage there are trends
while that in annectens involves all three for: a) pad dedifferentiation, b) limited
series plus scale number and phalangeal keel elaboration in the lateral portion of
lengths. A. annectens is thus less likely the phalanx i scales, and c) expansion of
to be an individual variant than if pad re- seta distribution to cover not only the pad
tention were the only character distin- scales, but also the entire phalanx i re-
guishing it from onca. 3) Comparison gion. In the annectens-onca lineage
with the juvenile of another anole there are trends for: a) pad dedifferentia-
species and consideration of the more tion, b) expansion in the distribution and
generalized anoles related to onca sug- increased size and number of keels, and
gest that the absence of definite keels in c) a "retrograde" shift from setae to
juvenile onca is an ontogenetic adapta- spines in the pad region. The combina-
tion which coincidentally parallels the tion of these trends results in the pro-
phylogenetic sequence. If this is the gressive substitution of multicarinate
case, it is only the presence of longer scales for lamellae and for smooth gen-
stalks in the fine structural series which eralized subdigital scales. In onca this
makes the case for juvenile onca as a substitution has been complete. In an-
model intermediate; the rest of the fea- nectens it has occurred in the phalanx i
tures are highly derived and resemble and iii regions and begun in the phalanx
adult onca. 4) The models for transitional ii area, but setae and lamellae are still
scales based on annectens scale 9 and present.

onca juvenile scales seem equally plau- Dedifferentiation of the pad (reduction

sible. The annectens sequence is more in lamellar number and width and the

consistent with the evolutionary patterns absence of a raised distal border [i.e., the

for regional differentiation and sequen- "Norops" condition]) occurs in both lin-

tial change that we have encountered cages, but the significance of the trend is

elsewhere in the genus. probably different in each case. In au-

266 Build ill Museum of Comparative Zoology, Vol. 149, No. 4

ratus pad dedifferentiation is associated
widi the expansion of setae into the pha-
lanx i region. The seta-bearing area
which is "lost" through decreased lamel-
lar width is more than compensated for
by the increased phalangeal length and
expansion of setae onto the non-lamellar
phalanx i scales. There is no indication
that pad dedifferentiation in auratus is
associated with decreased competence of
the setae for gripping or with selection
for another mechanism. In contrast, pad
dedifferentiation in the annectens-onca
lineage is associated with reduction in
the seta-bearing area (relative to body
size). Almost certainly, the competence
of the fine structural series in maintain-
ing grip is compromised relative to gen-
eralized anoles.

In the annectens-onca lineage the sub-
stitution of keels and a highly contoured
scale surface for lamellae and setae ap-
pears to reflect selection for an alterna-
tive "gripping" or "traction" device.
Keels probably function as "traction" de-
vices and presumably are more effective
than the fine structural series on surfaces
with gross irregularity, like most mature
bark, or on surfaces with low cohesion of
particles, like loose sand or dirt. Within
Anolis subdigital keels occur in the pha-
lanx i region of species which use the
trunk-ground structural niche (broadly
defined) and are rare in other ecomorphs.
In at least one case keels are elaborated
in a trunk-ground species which has in-
vaded habitat similar to that of onca.
Limited comparisons among non-anole
iguanids suggest that locomotion on
loose sand or dirt selects for additional
and larger keels. The comparative data
suggest that the annectens-onca lineage
represents a radiation of trunk-ground an-
oles into extreme xeric habitats where
there is enhanced selection for keels and
"traction" devices which are effective on
the combination of bark and loose sand
(see also below).

The fine structural series includes a
number of morphotypes which act like

microscopic hooks in addition to setae
which may alternate this prehensile
mechanism with adhesion. The fine
structural series is adapted to utilize a)
microscopic surface irregularity, or b)
high surface tension, smooth regions of
a grossly smooth environmental surface
like that of leaves, stems, and twigs. The
trend to increase the distribution and to
emphasize the fine structural series in
auratus correlates with behavioral spe-
cialization for perching and acrobatics on
single grass blades, stems, and twigs on
which robust keels like those of onca
would provide too little contact between
the subdigital and environmental sur-
faces.

The extreme xeric habitats of onca of-
fer relatively little broad leaf and grossly
smooth vegetation which can be used by
an animal with adult onca body size. Giv-
en the body size of adult onca, the dense
maze of small diameter thorn bush twigs
precludes most of the single perch and
fast acrobatic behavioral strategies which
select for pad function in auratus and
generalized anoles. The character of the
vegetation in onca habitat is such that
selection for "grip" based on the fine
structural series is probably diminished.

Several alternative sequences of evo-
lutionary morphological paths toward an-
nectens and onca are discussed. The
most likely sequence, in our judgment,
begins with a generalized trunk-ground
form, which, like chrysolepis, has a nar-
row pad and a few keels in the phalanx
i region. Radiation into a more xeric hab-
itat in which bare soil and sand becomes
the common terrestrial component of the
trunk-ground structural habitat, but in
which trees and large open shrubs are
still present, would select for the elabo-
ration of the keels in the phalanx i region
(the variant A. whitemani morphology)
and perhaps expansion of keeled scales
into the phalanx iii region (the annectens
morphology). Finally, multicarinate scales
replace lamellae and setae in the phalanx
ii region (the onca morphology). The

The os'ca Lineage: Subdigital Fine Structure • Peterson and Williams 267

complete substitution of multicarinate
scales seems to be associated with the in-
vasion of extreme xeric habitats a) where
vegetation is scarce and selection for lo-
comotion on loose sand is even more en-
hanced, and b) where the residual vege-
tation does not offer enough smooth
surfaces nor select for behaviors which
would maintain the setae and spike por-
tions of the fine structural series.

ACKNOWLEDGMENTS

We gratefully acknowledge the exper-
tise of Mr. Edward Seling of the Museum
of Comparative Zoology SEM facility and
Mrs. Sarah Beydler of the UCLA Depart-
ment of Pathology SEM facility.

We also thank Dr. John Wright, Curator
of Herpetology, Los Angeles County Mu-
seum of Natural History and Hymen Marx
of the Field Museum of Natural History
for permission to use specimens under
their care; Alfonso Coleman and Laszlo
Meszoly for preparation of the plates;
Mr. Eric Wilson and Dr. Franz Engel-
mann for assistance in the translation of
Hiller (1968); and Carl Lieb for discus-
sion of habitat correlations in Central
American anoles. Dr. A. P. Russell of the
University of Calgary made many valu-
able suggestions on the manuscript.

The study was supported in part by the
UCLA Biomedical Research Fund and
University Research Grants to Jane Pe-
terson and by National Science Founda-
tion Grant DEB 77-03302 to E. E. Wil-
liams.

LITERATURE CITED

Blair, F. 1960. The Rusty Lizard. University of

Texas Press, Austin.
BOULENGER, G. 1885. Catalogue of the lizards in

the British Museum (Natural History), 2nd

edition. 2. Taylor and Francis, London.
Challen, S. B. 1960. The contribution of surface

characters to the wetability of leaves. J. Pharm.

and Pharmacol. XII: 307-311.
CoLLETTE, B. 1955. Correlations between ecology

and moq)hology in anoline lizards from Ha-

vana, Cuba and southern Florida. Bull. Mus.
Comp. Zool. 125: 137-162.

Collins, J. 1971. Ecological observations on a lit-
tle known South American anole: Tropidodac-
tyliis onca. Breviora Mus. Comp. Zool. No.
370: 1-6.

Croat, F. B. 1978. Flora of Barro Colorado Island.
Stanford Univ. Press, Stanford.

Dellit, W. F. 1934. Zur Anatomie und Physiolo-
gic der Geckozehe. Jena Z. Naturwiss. 68:
613-656.

Dlxon, J. R., AND J. W. Wright. 1975. A review of
the lizards of the iguanid genus Tropidurus in
Peru. Contrib. Sci. Mus. Nat. Hist. Los Angeles
Co. No. 271: 1-39.

Cans, C. 1974. Biomechanics: An Approach to
Vertebrate Zoology. J. B. Lippincott, Philadel-
phia, pp. 72-100.

Hecht, M. K. 1952. Natural selection in the lizard
genus Aristelliger. Evolution 6: 112-124.

Hiller, U. 1968. Untersuchungen zum Feinbau
und zur Funcktion der Haftborsten von Repti-
lien. Z. MoHjh. Tiere 62: 307-362.

Holloway, p. J. 1969. The effects of superficial
wax on leaf wettability. Ann. Appl. Biol. 63:
14,5-153.

Kiester, a. R., G. C. Gorman, and D. C. Arroyo.
1975. Habitat selection behavior of three
species of Anolis lizards. Ecology 56: 220-225.

Laerm, J. 1973. Aquatic bipedalism in the basilisk
lizard: the analysis of an adaptive strategy. Am.
Midi. Nat. 89: 314-333.

Maderson, p. F. a. 1970. Lizard glands and lizard
hands; Models for evolutionary study. Forma
et Functio 3: 179-204.

Maderson, P. F. A., and E. E. Williams. 1978. A
"metamorphic event" in an iguanid lizard.
Am. Zool. 18: 648.

Martin, J. T. and B. E. Juniper. 1970. The Cu-
ticles of Plants. Edward Arnold (Publishers)
Ltd., Great Britain, pp. 1-296.

Miller, M., and M. Kasahara. 1967. Studies on
the cutaneous innervation of lizards. Proc. Cal-
if Acad. Sciences 4th Series, XXXIV, No. 16:
549-568.

Porter, W. 1967. Solar radiation through the liv-
ing body walls of vertebrates, with emphasis on
desert reptiles. Ecol. Monogr. 37: 27.3-296.

Rand, A. S., and P. J. Rand. 1966. Aspects of the
ecology of the iguanid lizard Tropidurus tor-
qiiatus at Belem, Para. Smithson. Misc. Coll.
151: 1-16.

Robinson, P. L. 1975. The functions of the hooked
fifth metatarasal in lepidosaurian reptiles.
Problemes Actuel de Paleontologie — Evolution
des Vertebres. Colloq. Internat. CNRS No. 218:
461-483.

RuiBAL, R. 1968. The ultrastructure of the surface

of lizard scales. Copeia 1968: 698-704.
RuiBAL, R., and V. Ernst. 1965. The structure of

268

BuUciin Musctiin of Coiiipdralirc y.nolof^y. Vol. 149, No. 4

the digital setae of lizards. J. Mon^h. 117: 271-
294.

Russell, A. 1975. A contribution to the functional
analysis of the foot of the tokay, Gckko ^ecko
(Reptilia: Gekkonidae). J. Zool. Lond. 176:
437-476.

. 1976. Some comments concernint^ inter-
relationships among gekkonine geckos. Pages
217-244 in A. d'A. Bellairs and c! B. Cox (eds.),
Morpholog) and Biology of Reptiles. Linnean
Society Symposium, Series 3, Academic Press,
London.
-. 1979. Parallelism and integrated design in

the foot structiue of gekkonine and diplodac-

tyline geckos. Copeia 1979: 1-22.
Schmidt, K., axd R. Inger. 1957. Living Reptiles

of the World. Douhleday, New York.
Schmidt, W. J. 1920. Einiges uber die Hautsin-

nesorgane der Agamiden, insjjesondere von
Calotes nehst Bermerkungen uber diese Or-
gane bei Cieckoniden and Iguaniden. Anat.
Anz. 53: 113-139.

Undehvvood, G. 1954. (categories of adaptation.
Evolution 8: 365-377.

Werner, Y. L., AND M. Broza. 1969. Hypothetical
function of elevated locomotory postures in
geckos (Reptilia: Gekkonidae). Israel J. Zool.
18: 349-355.

Williams, E. E. 1963. Anolis whitemani, new
species from Hispaniola (Sauria, Iguanidae).
Breviora Mus. Comp. Zool. No. 197: 1-8.

. 1974. A case history in retrograde evolu-
tion: the onca lineage in anoline lizards. I. An-
olis annectens new species, intermediate be-
tween the genera Anolis and Tropidodactyhts.
Breviora Mus. Comp. Zool. No. 421: 1-21.

(US ISSN 0027-41 0(

bulletin OF the

Museum of

Comparative

Zoology

MUS. COM P. zoc

AUG 4 1981

The American Orb-Weaver Genera waro
Dolichognatha and Tetragnatha
North of Mexico
(Araneae: Araneidae, Tetragnathinae)

HERBERT W. LEVI

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBERS
21 JULY 1981

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Breviora 1952-

BULLETIN 1863-

Memoirs 1864-1938

JOHNSONIA, Department of VIollusks, 1941-

OccASiONAL Papers on Mollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. I. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Tere-
dinidae (Mollusca: Bivalvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts, 02138, U.S.A.

© The President and Fellows of Harvard College 1981.

THE AMERICAN ORB-WEAVER GENERA
DOLICHOGNATHA AND TETRAGNATHA NORTH OF MEXICO
(ARANEAE: ARANEIDAE, TETRAGNATHINAE)

HERBERT W. LEVI*

Abstract. One species of Dolichognatha and fif-
teen species of Tetragnatha are found in North
America north of Mexico. Of these, three are new:
T. earmra from the Everglades; T. shoshone, wide-
spread in south central Canadian provinces and
north central states; and T. branda, found from
Connecticut to Mississippi.

Tetragnatha harrodi is a synonym of T. dear-
mata, described from northern Eurasia. Tetragna-
tha antillana is a synonym of the cosmotropical T.
nitens.

Males of Tetragnatha are readily separated by
the shape and structure of the palpal conductor and
paracxnibium; they cannot be separated by the
structure of the chelicerae. Females can be sepa-
rated by the configuration and placement of seminal
receptacles, as revealed by simple dissection; they
cannot be separated by the structure of the chelic-
erae.

Several Tetragnatha species are distributed from
Canada to the tropics. Several uncommon species
probably have specialized habitats: T. viridis in
pines; T. vermiformis on reeds; and T. branda in
salt marsh grass.

INTRODUCTION

One of the aims of a generic revision
is to find diagnostic characters that sep-
arate the species. Reliable identification
is essential for any work by ecologists
and physiologists. Another aim, of course,
is to reveal the relationships among the
species, and, more important, the rela-
tionship of the genus to other genera.

Tetragnatha species are among the
most abundant spiders worldwide. North
American species can be determined
only with difficulty using Chickering's
papers (1957a, b, c, 1959). Chickering
(with one exception) detemiined species

' Museum of Comparative Zoology, Harvard Uni-
versity, Cambridge, Massachusetts 02138.

correctly, but could determine only
males; his drawings are awkward, often
failing to tell what position is illustrated
or whether the structure illustrated is
from the left or right palpus. Also, his
Tetragnatha reports are limited geo-
graphically to Michigan and to some trop-
ical American regions. Tetragnatha is
the second largest araneid genus north of
Mexico, after Araneus.

Despite a conscious effort not to
change names, two synonymies had to be
resolved. To avoid further name changes,
neotypes were designated for several old
names, the use of which has been based
mainly on tradition rather than on the
type method. The neotypes were desig-
nated in accordance with Article 75
(1961) of the International Code of Zoo-
logical Nomenclature.

ACKNOWLEDGMENTS

I would like to thank the following per-
sons for the loan of specimens: J. Beatty;
D. Boe; G. Byers and M. Rothschild
(Snow Entomological Museum, Univer-
sity of Kansas); J. Carico; C. Dondale
(Biosystematics Research Institute, Agri-
culture Canada, CNC); M. Grasshoff
(Senckenberg Museum); M. Hubert (Mu-
seum National d'Histoire Naturelle,
Paris); J. Kaspar; B. J. Kaston; C. Okuma;
B. Opell; W. Peck (Exline-Peck collec-
tion); N. I. Platnick (American Museum
of Natural History, AMNH, and Cornell
University Collection, CUC); J. Proszyri-
ski; W. F. Rapp; A. Riedel (Polish Acad-
emy of Science); W. T. Sedgwick; W.
Shear; W. Star^ga; F. Wanless (British

Bull. Mus. Comp. Zool., 149(5): 271^18, July, 1981

271

272 Bulletin Museum of Coinparafive Zoology, Vol. 149, No. 5

Museum, Natural History): H. K. Wal-
lace; H. V. Weeins and G. B. Edwards
(Florida State Collection of Arthropods,
FSCA). Most of the research was done
with the Museum of Comparative Zool-
og>- (MCZ) collections.

The American Museum of Natural His-
tory provided about 4000 vials of com-
mon Tetragnatha species; the Museum
of Comparative Zoology between 2000
and 2500 vials; the Canadian National
Collections 600 vials; the Florida State
Collection of Arthropods and the Exline-
Peck collection about 300 vials each.

Mapping, rewriting, and typing were
done by Cecile Villars; some rewriting
was done by L. R. Levi. John Hunter and
Edward Seling made the scanning elec-
tron micrographs.

National Science Foundation Grant
DEB 76-15568 made the research possi-
ble. Publication costs of this study were
covered in part by a grant from the Wet-
more Colles Fund.

METHODS OF STUDY

1. Background. Washed sand is the
ideal surface on which to examine the
long-legged Tetragnatha in various
views. The palpus does not usually have
to be removed to reveal its diagnostic fea-
tures; however, magnification of at least
50 to 80 times is required to examine the
shape of the tip of the conductor.

2. Seminal Receptacles. To detennine
females it is necessary to examine the
seminal receptacles. At first this was
done by removing the genital area and
clearing it in clove oil or mounting it on
a slide: the cleared genital area has been
illustrated for all species. Because of
their large size, it is more difficult and
time-consuming to clear the whole spi-
der.

An easier method is to lift the exoskel-
eton of the genital operculum, carefully
separating it from clinging tissues, and
search for the lightly sclerotized seminal
receptacles. Most species can readily be

detemiined by the position and shape of
the seminal receptacles (together with
eye placement).

To lift the genital cover, cut with a
mounted and sharpened minutenadel,
starting from the anterior corner between
the book-lung spiracle and the genital
cover, cutting in a posterior direction to-
ward the genital opening. After cutting,
carefully lift the exoskeleton and remove
tissue clinging to it. Fold the cover back
and probe for the seminal receptacles in
the tissues adjacent to the cut. Some prac-
tice is necessary. Many Tetragnatha
species are sufficiently common so that
a few specimens can be sacrificed for the
purpose. While penultimate female spec-
imens may have an indistinct genital cov-
er, the seminal receptacles are not pres-
ent.

The final method used was to peel the
genital cover to the side, place the whole
spider in clove oil, and watch the genital
area clearing: the area between the sem-
inal receptacles clears first; later the tis-
sue around the more dorsal median re-
ceptacle. (The spider should then be
removed and the oil washed out in alco-
hol.)

3. Eye Distance. The distance between
anterior and posterior lateral eyes is vari-
able. Nevertheless, the relative distance
between the two lateral and the two me-
dian eyes is consistent and useful. It is
best seen from the side with the speci-
men lodged in sand. The distances were
measured from the center of the eye.

4. Chelicerae. I first tried to study the
genus by illustrating the chelicerae of
males and females of various species. But
I could not consistently separate species
using the shape and number of cheliceral
teeth in the fang groove, the shape of the
fang, presence of fang cusps (a hump at
point of greatest curvature), or length of
the chelicerae. Such characters vary
enonnously within species (Figs. 83-87).
Chickering also found the chelicerae un-
reliable, and depended on palpi to sep-
arate the males. Only one mistake was

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 273

found in his many determinations of
males: he labeled the rare T. viridis as
T. versicolor in collections. However,
Chickering resorted to characters ol the
chelicerae to determine females, and
many of these determinations were
wrong. The teeth and chelicerae are use-
ful nevertheless in limited ways: male T.
deannata and T. pallescens always lack
the first, distal, long tooth; T. elongata
generally has longer chelicerae than T.
versicolor.

Some authors have detennined imma-
ture North American specimens on the
basis of the cheliceral teeth, burdening
students of the genus with much ques-
tionable reporting. To elevate chelicerae
to a reliable character, it would be nec-
essary to raise the inhabitants of egg-sacs
of different species and study the allo-
metric growth and development of their
chelicerae and teeth. Tetragnatha
species, like other Araneidae and many
other spiders, mature after a variable
number of molts; some mature speci-
mens are twice the size of others. In ad-
dition, perhaps due to homional influ-
ences, the chelicerae of some individuals
are more developed than those of others
(e.g., T. elongata. Fig. 83 and T. palles-
cens from Florida, Figs. 134, 135). Other
species have larger chelicerae in south-
em localities than in northern ones (e.g.,
r. elongata. Figs. 84-87). I judged the
previous determinations of immatures
doubtful because the identifiers often
could not place adult specimens correct-

ly.

5. Measurements. Measurements of to-
tal length are from the clypeus to the pos-
terior tip of the abdomen. Initially ten to
twenty specimens of each species were
measured to establish variation; subse-
quently only extremely large and small
individuals were measured. If there was
a geographical difference in size, this was
noted. Total length is not a good mea-
surement, as it is difficult to repeat, and
the measurements are intended only to
give a general idea of size range, not as

a statistical tool to distinguish between
species. Furthennore, the smallest size of
mature females is only an estimate; dis-
section was not always made to demon-
strate maturity.

6. Determination. In this study, for
which perhaps 30,000 to 40,000 speci-
mens were available, 1 had to take some
shortcuts to get data for the revision. I
did not detemiine the contents of all
vials, nor could I sort some mixed collec-
tions. Usually only a male was selected
from the vial and carefully examined and
recorded. In collections that consisted
only of females, I did not examine a spec-
imen unless it was from a poorly collect-
ed area or a less common species was sus-
pected. It is possible to detennine adult
females, but it is less time-consuming to
detemiine males. Gathering data for this
revision had a higher priority than label-
ing collections.

7. Dubious Species. In the past when
there was doubt as to the distinctness of
species (or genera), I tended to lump
them, considering it more parsimonious
to lump than to split. However, I am
aware that some mistakes have been
made. Araniella (Levi, 1974) of Europe,
for instance, contains a number of
species, all with similar genitalia but dif-
ferent habits. Unfortunately, I had ex-
amined only a few specimens and failed
to consult colleagues who knew the
species in the field. Also, the genus La-
trodectus in the family Theridiidae
(Levi, 1959) is now known to contain
many species difficult to identify by gen-
italia. Renewed studies are needed for
both.

Similarly, what is considered Tetra-
gnatha versicolor here may turn out to be
several related species. It is particularh
disconcerting that the genitalia of T. vir-
idis, certainly a distinct species, are \e\y
similar to the genitalia of southeastem T.
versicolor; T. viridis differs by being
green or reddish in coloration, in having
the lateral eyes as far apart as the medi-
ans, and in having unusually long mac-

274 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

rosetae on the le^s. While the genitalia lowing characters, some shared with

of T. viridi.s strikingly resemble those of species of other genera, are evidence of

southeastern T. versicolor, there is great- specialization: the palpus has a separate

er genitalic difference between south- paracymbium (also in Pachygnatha and

eastern and western T. versicolor (Figs. Linyphiidae); there is no tapetum in the

104-109), suggesting that there may be secondary eyes {Dolichognatha and

other sibling species. Azilia share this feature; Pachygnatha

8. Variation. Tetragnatha species vaiy lacks tapetum only in the posterior me-
as widely in size as do other Araneidae, dian eyes); the lateral eyes are often sep-
some individuals being almost twice as arated (also in Azilia); the chelicerae are
large as others of the same species. Is the modified as holdfasts for mating (also in
variability in size a preadaptation to Pachygnatha, Dolichognatha, and per-
dwarf males found in various genera of haps other genera); the endites are
the family? Large and small individuals elongate and widened distally (also in
of the same species have different den- Pachygnatha); and when resting, the
tition on the chelicerae; smaller individ- femoral-patellar joint is straight (Plates 3,
uals have fewer teeth, and the teeth may 4) (also in most relatives and in Metinae,
be relatively larger than those of larger- but not in Dolichognatha [Plate Ij).
sized individuals (Figs. 83-87). Primitive features include: the rela-

9. Illustrations. The conventions of il- tively long-lived males; the lack of an
lustrations were followed. The light embolus cap on the male palpus, allow-
comes from the upper left. In dorsal ing males to mate several times; and, per-
view, anterior is placed on top; in side haps, the use of both palpi simultaneous-
view, anterior is on the left. These con- ly in sperm induction.

ventions were not followed for palpi; The female genitalia are referred to re-

they are considered separate structures peatedly in the literature as haplogyne

for ease of visualization. and primitive. The absence of an epigy-

10. Scanning Electron Micrographs, num (separate copulating pore) is almost
Scanning electron micrographs (SEM) certainly adaptive and specialized. The
were made of conductors and their tips strong coupling of male and female che-
after the illustrations had been complet- licerae when mating obviates the need
ed. It was hoped the SEM would add de- for another coupling guide. When mat-
tails for the diagnostic features. While the ing, the male's venter faces that of the
photographs added these details (Plates female, and the female's genital opening
5-7), drying the palpi prior to SEM twist- is directly underneath the male's palpus,
ed the soft pleats of tissue and gave the The female can curl her abdomen to fa-
structures a greater cui-vature; some tips cilitate the insertion of the embolus
were turned. The drawings were only (Plate 4).

touched up after examining the photo- The median "seminal receptacle " of

graphs; they were not changed. It ap- the female genitalia is almost certainly

pears that scanning electron micrographs not that. Wiehle (1967) could not find

alone are not a substitute for illustrations sperm in this structure, nor does the

made under a dissecting microscope, but shape of this structure suggest a spemi

are a valuable additional tool. storage function. I think it was Stefan

Heimer (in correspondence) who sug-

PHYLOGENETIC CONSIDERATIONS gested that it might be a holdfast for the

tip of the conductor. Evidence tor this is

1. Tetragnatha. Tetragnathids are not seen in the shape of the conductor. In

primitive spiders, various contrary pub- those species in which the conductor tip

fished opinions notwithstanding. The fol- is a large hook, there is little or no me-

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 275

dian structure {viridis, laboriosa, palles- median eyes and the canoe-shaped ta-

cens, caudata, vermiformis). Species in petum has been lost. The canoe-shaped

which the conductor tip is a knob have a tapetum is still present in the lateral eyes

spherical median structure {guatemalen- of both Glenognafha and Pachijgnatha,

sis, elongata, straminea, vermiformis, but has been lost in the lateral eyes of

nitens). In one species the conductor Tetragnatha, Dolichognatha, and Azilia.

tip is filifonn; here the structure is elon- 3. Uloborids. Uloborids share a num-

gate, apparently matching the filiform tip ber of characters with Tetragnatha: sep-

(dearmata. Fig. 83). Perhaps the primi- aration of the lateral eyes; loss of tapetum

tive, haplogyne median seminal recep- in secondary eyes; presence of tricho-

tacle is homologous and was a preadap- bothria proximally on the femora; fre-

tation for this structure. To verify these quent absence of an epigynum (Opell,

ideas it will be necessary to preserve 1979); and similarities of palpal structure,

mating individuals and examine the in- Most important perhaps is that many

serted palpi. uloborids make an orb-web, as do tetra-

2. Glenognatha and Pachijgnatha. The gnathids and all Araneidae. Are these
presence (or absence) of an elaborate tra- shared characters synapomorphies?
cheal system is of minor phylogenetic im- The separate lateral eyes may be a syn-
portance, and may reflect small size and apomorphy; more likely not. This feature
large surface, or may be an adaptation is no doubt related to the resting position
against water loss in a more arid environ- of the spiders: unlike most Araneidae,
ment. With a more elaborate and func- Tetragnatha (and uloborids) extend the
tional tracheal system, the posterior tra- legs when resting (Plates 3, 4) rather than
cheal spiracle moves to an anterior pulling the anterior legs back so the fem-
position in Glenognatha (as well as in oral-patellar joint is flexed. With the an-
other groups, e.g., some Dipoena species terior legs flexed, there is little space to
and Anyphaenidae). Tracheae are well pemiit lateral vision (Plate 1). In most
known to be plastic and adaptive char- Araneoidea the lateral eyes touch; their
acters, and developed as a result of need separation in Tetragnatha is no doubt a
(Levi, 1967). The distribution of tracheae specialization.

is part of a built-in pattern of body Loss of the tapetum, like loss of any
growth. Although this general pattern of structure, is a poor character for combin-
growth is fixed, variation in detail can be ing groups. The loss in Tetragnatha in-
influenced by local need. For instance, dicates a specialization from the Ara-
by implanting glands that have a high neus-like eye: the loops of rhabdoms to
oxygen requirement, it has been dem- the median side of the canoe-shaped ta-
onstrated in insects that the demand for petum fill the back of the eye (Figs. 21,
oxygen stimulates an exaggerated pro- 22) and the tapetum is lost. There is no
duction and outgrowth of new tracheae indication, as there is in Tetragnatha,
and tracheoles which invest and pene- that the ancestor of Ulohorus had a ca-
trate the implanted organ. The growing noe-shaped tapetum. Perhaps rhabdom
tracheal cells seem to be attracted by rows got scrambled and rows will still be
sites of reduced oxygen tension (Wig- found in some members of the family,
gles worth, 1972). The trichobothria on the femora (Fig.

The palpal structures and female gen- 17) may be a second synapomorphy. But

ital structures of G/enognaf/ia and Fflc/iy- perhaps these trichobothria have a spe-

gnatha share several derived characters, cific sensory function, evolved in re-

synapomorphies, (Levi, 1980a). Most im- spouse to a similar need in uloborids and

portant, in both, as in Tetragnatha, the Tetragnatha. I do not know how com-

median rhabdom loops fill the posterior mon femoral trichobothria are in other

276 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

spiders, but they are certainly absent in even within a given species. Individuals

most Araneidae and probably in most Ar- that undergo fewer molts are smaller in

aneoidea. matiire size than those that matured after

The absence of the epigynum among a greater number of molts. The size range

Tctrap,natha is probably related to the is greater among adults of araneid species

coupling of chelicerae when mating, than in species of other families. Usually

There is no such correspondence in ulob- in any araneid population some mature

orids. Again, loss of a structure is a poor individuals are twice the size of others,

character to relate groups. and there may be large geographical size

The complex orb construction behavior variation as well. Perhaps such size vari-
can perhaps be considered a synapomor- ations preadapt for the occurrence of
phy. Several striking points set uloborids dwarf males in various araneid genera,
apart from Araneoidea, including tetra- The variability in number of instars of
gnathids, and make a close association spiders differs from the steady number in
unlikely: 1) the cribellum, generally con- most mites and pseudoscorpions.
sidered a primitive structure, has been If individuals mature at various sizes,
lost a number of times in spiders that are the small and large adults proportion-
gave up webs in favor of hunting, but it al or are the adults the result of allometric
is present in uloborids, as is a calamis- growth? This knowledge is important not
trum; 2) the loss of poison glands in the only for those who want to use random
chelicerae (also reported in the unrelated numerical methods to differentiate
liphistiomoiph Heptathela) is a very un- species, but also for those who want to
usual specialization, possible only in spi- differentiate Tetragnatha species by pro-
ders that overpower their prey by wrap- portions and cheliceral teeth,
ping in silk; 3) the rhabdom arrangement Illustrations and measurements were
within the eye is not like that found in Ar- made of a large male T. elongata from
aneoidea (Homann, 1950, 1971); and 4) Texas (Fig. 87), and a small male from
the important structure of the male palpi Wisconsin (Fig. 85). The total length of
is not close to that of Araneoidea gener- the large male was 8.6 mm; the small one,
ally, although closest to that of Tetrag- 6.2 mm (72% the total length of the large
natha. Other unique attributes of ulobor- one). The length of the smaller carapace,
ids are the hair covering, the individual as measured on the illustration, was 72%
hairs often being constricted at the base; of that of the larger one. The smaller che-
the elongate labium; and the laterally licerae measured 60% of the cheliceral
bulging sternum. length of the larger spider. The length of

Tetragnatha is close to Meta, and spi- the palpal tibia-tarsus was 60%, but the

ders related to Meta are probably the smaller palpal bulb was 80% of the larger

most primitive Araneoidea, having most one, and the length of the conductor was

plesiomorph characters of the superfam- 88%. The seminal receptacles of a large

ily. The orb-web is present in Meta but Texas female are slightly smaller than

has been lost in some specialized Ara- those of a smaller female from Massachu-

neidae (Archemorus, Mastophora) and setts. But more important is that the che-

Uloboridae (Miagrammopes). The orb- licerae and their teeth, commonly used

web must be a primitive structure, and to differentiate Tetragnatha species, dif-

the webs of theridiids and linyphiids de- fer widely. (The left chelicera pictured in

rived (see also Brignoli, 1979; Levi, Figure 85 had two teeth next to each oth-

1980b). er, but this condition was not present on

Growth and Size. All spider individu- the right chelicera of the same individu-
als pass through a succession of molts be- al.) While the large individuals of T.
fore maturing; the number of molts varies elongata had many more cheliceral teeth

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 277

than did the small individuals, the most
characteristic teeth as well as the spur
were present in both sizes. Of great in-
terest are relative changes in proportions
of the palpal cymbium. Does the cym-
bium correspond more with leg length
than with that of the palpal bulb? The
cymbium of the large Texas specimen
extends beyond the tip of the conductor
(Fig. 87); that of the small Wisconsin in-
dividual is shorter than the conductor
(Fig. 85).

To summarize, growth is allometric,
and the genitalia of larger and smaller in-
dividuals are more alike than other body
parts. Random measurements might gen-
erally separate size classes and are not so
useful as the study of genitalia for the
separation of species. Also, the chelic-
erae are not as diagnostic as the genitalia.
Seeley's study (1928), which revised
North American Tetragnatha using
mainly the amiature of the chelicerae, is
only of limited use. Chickering's publi-
cation on the Tetragnatha of Michigan
(1959), and those of Mexico, Central
America, and Jamaica ( 1957a, b,c and
1962), used mainly the palpi and are still
very usable although limited by awkward
illustrations.

Dolichognatha O. P. -Cambridge

Dolichognatha O. P. -Cambridge, 1869, Jour. Lin-
nean Soc. London, 10: 387. Type species by
monotvpv D. nietneri O. P. -Cambridge, 1869,
ibid., p. 388, pi. 12, figs. 39^5, 6 , from Ceylon.
Male holotype in the British Museum, Natural
History, examined.

Landana Simon, 1883, Ann. Mus. Genova, 20; 184,
erroneously placed in Archaeidae. Type species
by monotypy L. petiti Simon, 1883, ibid., p. 185,
figs. 1-5, 6, from the Congo. Male holot>'pe in
the Museum National d'Histoire Naturelle, Paris,
examined. NEW SYNONYMY.

Paraebius Thorell, 1894, Bihang Kongl. Svenska
Vet.-Akad. Handl., 20: 43. Type species by mono-
typy Paraebius mandibularis Thorell, 1894,
ibid., p. 44. Male holotype from Sumatra in the
Natural History Museum, Stockholm, examined.
NEW SYNONYMY.

Prolochus Thorell, 1895, Descriptive catalog of the
spiders oi Burma, London, p. 122. Type species
by monotypy Prolochus longiceps Thorell, 1895,

Plate 1. Unnamed Dolichognatha sp. from New Guinea
hanging on debris; spider facing left.

ibid. Male holotype in the Natural History Mu-
seum, Stockholm, examined. NEW SYNONYMY.

Nicholasia Bryant and Archer, 1940, Psyche, 47: 60.
Type species bv monotypy Epeira pentagona
Hentz, 1850. NEW SYNONYMY.

Afiamalu Marples, 1955, Jour. Linnean Soc. Lon-
don, Zoology, 42: 495. Erroneously placed in Ar-
chaeidae. Type species by monotvpy Afiamalu
richardi Maiples, 1955, ibid., figs. 5, 7, 10, 9 , 6 ,
from western Samoa. NEW SYNONYMY.

Synonymies. Landana contains those
species of Dolichognatha in which the
males have longer chelicerae than fe-
males, and have the distal end of their
chelicerae modified. As there are inter-
mediates, it appears advisable to synon-
ymize Landana with Dolichognatha.
The Paraebius synonymy is uncertain.
The male type examined lacked the char-
acteristic four posterior abdominal humps.
The cheliceral teeth were not all distal
but spread out. The palpus is like that of
Dolichognatha.

Prolochus belongs here, judging by the
structure of the chelicerae, the palpus,
and the poorly preserved abdomen. Pro-
lochus longiceps differs in lacking pos-

278 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

D. pentagona

Map 1. Distribution of Dolict)ognattia pentagona (Hentz).

teiior median eyes, but in some species
of Dolichognatha the posterior median
eyes are reduced in size. Nicholasia was
erected by Bryant and Archer (1940: 60)
for D. heptagona, which "differs from the
Ceylonese genotype in the fomi of web,
as well as in several important structural
characters. " Bryant and Archer repeated
E. Simon's report (1894: 743) that Doli-
chognatha makes a horizontal sheet web.
This is Simon's error. According to
Bryant and Archer, Nicholasia differs
from Dolichognatha by the "truncate tip
of the labium, fewer teeth on the interior
margin of the mandible and by the male
palpus which has a dorsal apophysis on
the basal third of the cymbium and a
much simpler palpus." Afiamalu resem-
bles Landana in "the shape of the cara-
pace and in having spines on the legs,
but differs in the annature of the chelic-
erae and the presence of a serrula [on the
enditesl."

Diagnosis. Dolichognatha, like other
tetragnathids, lacks a tapetum in the sec-
ondary eyes, and the rhabdoms are ar-
ranged in looping rows (Figs. 7, 8). It dif-

fers from both Tetragnatha and Azilia in
the shape of the abdomen: higher than
long, with two pairs of posterodorsal
humps (Plate 1; Figs. 1, 2, 10); it differs
from Tetragnatha in lacking trichoboth-
ria on the femora. Like Azilia, but unlike
Tetragnatha and Pachygnatha, Dolicho-
gnatha has a simple epigynum in the fe-
male (Fig. 6), and the male palpi do not
have the cymbium modified (Fig. 13).
Unlike Azilia, Dolichognatha exhibits a
palpal paracymbium which is a compli-
cated sclerite, different in different
species (Fig. 13); it is not a free sclerite
as in Tetragnatha. Dolichognatha' s most
distinctive character is the shape and col-
oration of the carapace: head, often dark,
with parallel sides, and carapace evenly
rounded, always light-colored. The short-
fanged chelicerae (Figs. 1-3, 10, 11) may
be so long as to resemble those of Ar-
chaeidae, and if they are greatly elongat-
ed, the head may lose the characteristic
parallel sides. The posterior median eyes
touch (Figs. 2, 3, 11), or are only slightly
separated. They may be minute in size
or, among species previously placed in
Prolochus, are lost completely.

Description. Carapace: head brown
with darker marks; thorax yellow-white.
Chelicerae usually brown. Sternum yel-
low-white with dark marks. Coxae yel-
low-white with dark spots. Legs contrast-
ingly banded, dark bands nanow. Dorsum
of abdomen with white and black pig-
ments; abdominal humps are light-col-
ored posteriorly, dark-colored anteriorly
(Figs. 1, 2). Sides of abdomen dark pos-
teriorly. Ventrally, area between genital
groove and spinnerets with paired white
patches and black pigment. Sides and
head parallel if chelicerae small. Tho-
racic depression barely visible. Lateral
eyes separated from each other by about
their diameter (Figs. 1, 8, 10). Height of

Plate 2. Webs of Dolichognatha pentagona. Upper photograpn, line of debris to the left of web's center (arrow), out
of focus; web 12 cm diameter. Bottom photograpfi, witfi line of debris behind upper right outer spirals (arrow); web 7.5
cm diameter. Webs dusted with corn starch.

DoLicHOGNATHA AND Tethagnatha North OF MEXICO • Levi 279

* .4

ig^^_^ ^- A^* '^-

K

ik

>^«^-

•%'

280 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

clypeus equal to diameter of the anterior
median eyes. Anterior median eyes usu-
ally the largest (Figs. 3, 11). All lack ta-
petum, and secondary eyes have rhah-
doms in rows as in Tetragnatha (Figs. 7,
8). Legs with strong macrosetae and
sometimes denticles. Abdomen always
higher than long and with four humps
(Figs. 1, 2, 10). Femora without tricho-
bothria, tibiae with a row or two, and
metatarsi with one trichobothrium dor-
sally on the proximal end.

Epigynum lightly sclerotized and
transverse (Fig. 6). Palpus simple, lightly
sclerotized, except for the embolus (Figs.
13-15). Cymbium entire, with large at-
tached paracymbium (P in Fig. 13).

Natural History. I have seen the webs
and habitats of four species of Dolicho-
gnatha: D. pentagona in the southeast-
ern United States, two species in north-
em South America and one species in
New Guinea. All constructed diagonal
orb-webs between the buttress roots at
the bases of large trees in relatively
moist, dark forests. The web has a line of
debris and egg-sacs across it or hanging
vertically behind it (arrows, Plate 2). The
spider rests among the debris. It is not
surprising that collectors have at times
marked specimens as an unknown species
of Cyclosa. But the species is easily over-
looked by collectors due to the special-
ized habitat, and individuals are not col-
lected by sweeping. The web, often
messy, does not stand out from other de-
bris at the base of the tree, and the spider
may not be on the web. Unlike other tet-
ragnathids, Dolichognatha pulls the legs
in at rest (Plate 1).

Species. Species of Dolichognatha are
common worldwide in wann or tropical
climates. The type species comes from
Ceylon; there are specimens of different
species from Africa in the Museum of
Comparative Zoology collections. The
genus is found in southeast Asia, and I
have collected a species in New Guinea
(Plate 1). Species differ only slightly in
the shape of the abdomen and epigynum;

there are greater differences in length
and structure of chelicerae, in the inter-
nal female reproductive ducts, and in the
structure of the paracymbium of the male
palpus. Males of species with greatly
elongated chelicerae also may have the
cephalic carapace of diverse shapes and
not have sides parallel as in D. penta-
gona. Males of some species have the
fang groove at the distal end of the che-
licerae modified and expanded. Meta
quadrituherculata Keyserling, 1883, from
Peru, belongs to the genus, as does Cy-
closa minuscula Mello-Leitao.

Note. This genus does not belong to
the family Archaeidae and is not related
to Archaea, despite certain superficial re-
semblances. Archaea, unlike Dolicho-
gnatha, has scale-like body hair cover-
ing; has a more or less long, constricted
neck between head and thorax; and has
reduced lateral eyes. The endites of Ar-
chaea curve around the labium, and there
is an elongate sternum. The unique spin-
nerets are the most important difference,
and one that excludes the Archaeidae
from the Araneoidea; the median spin-
nerets are minute, and the posterior spin-
nerets are much smaller than the anterior
ones. Zearchea has only two spinnerets.
The anal tubercle oi Archaea is relatively
large. The legs are thin and spindly, and
there are only one or two trichobothria
on the tibiae, one distally (unlike Ara-
neoidea) on the metatarsus, none on the
tarsus. The reduction of the spinnerets
suggests association with the Zodarioi-
dea; the loss of spinnerets is otherwise
rare in Labidognatha.

Forster (1955) has illustrated an epig-
ynum and a pair of seminal receptacles
of archaeid spiders. His illustrations of
archaeid palpi show few sclerites and
much diversity among related species;
the paracymbium is always absent. The
sclerites are difficult to homologize with
those of Araneoidea.

A key and illustrations to the Archaei-
dae of Madagascar can be found in Le-
gendre (1970); a key to the genera was

DOLICHOGNATHA AND TeTRAGNATHA NORTH OF MEXICO • Lcvi 281

M'Sb%.

Figures 1-15. Dolichognatha pentagons (Hentz). 1-9. Female. 1. Lateral. 2. Carapace and abdomen. 3. Eye region and
chelicerae. 4. Left chelicera, posterior. 5. Epigynum, cleared, dorsal. 6. Epigynum. 7. Posterior median eyes. 8. Lateral
eyes. 9. Labium and endites. 10-15. Male. 10. Left lateral. 11. Eye region and chelicerae. 12. Left chelicera, posterior.
13-15. Left palpus. 13. Lateral, expanded. 14. Ventral. 15. Lateral.

Scale lines. 0.1 mm; except Figs. 1, 2, 10, 1,0 mm.

Abbreviations. C, conductor; E, embolus; P, paracymbium; T, tegulum.

282 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

published by Forster (1955); and a list
with distributions of all known species is
in Legendre (1977).

Dolichognatha pentagona (Hentz),
new combination
Plates 1, 2; Figures 1-15; Map 1

Epeira ? pentagona Hentz, 1850, J. Boston Soc.
Natur. Hist., 6: 18, pi. 3, fig. 1, 9 . Specimens from
Alabama, destroyed.

Ci/rtophora tubercuhita: — McCook, 1893, Ameri-
can Spiders, 3: 236, pi. 17, fig. 11, 9. Probably
not D. tuberciilata Keyserling.

Dolichognatha tubcrculata: — F. P. -Cambridge,
1903, Biologia Centrali-Americana, Araneidea, 2:
447. Probably not D. tiiberculata Keyserling.

Nichohisia pentagona: — Bryant and Archer, 1940,
Psyche, 47: 61, fig. 1, 9,(5. Archer, 1940, Paper
Alabama Mus. Natnr. Hist., 14: 27, pi. 2, fig. 1,
web, pi. 5, figs. 1, 2, 9, d.

Description. Female. Posterior median
eyes 0.5 diameters of anterior medians;
anterior laterals 0.8; posterior laterals 0.7
diameters of anterior medians. Posterior
median eyes touching, 2.5 to 3 diameters
from laterals (Figs. 1-3). The chelicerae
are elongate with an anterior boss and a
posterior hump (Figs. 1, 3, 10). There are
three teeth on the anterior margin, two
small ones on the posterior margins, and
a longer one at the base of the fang (Fig.
4). Total length, 3.6 mm. Carapace, 1.7
mm long, 1.2 mm wide. First femur, 2.0
mm; patella and tibia, 2.4 mm; metatar-
sus, 1.6 mm; tarsus, 0.7 mm. Second pa-
tella and tibia, 1.9 mm; third, 1.0 mm;
fourth, 1.2 mm.

Male, from Georgia. Eye arrangement
as in female. Chelicerae much longer
than those of female (Fig. 11). The palpus
is rather small (Figs. 10, 14, 15). Total
length, 3.2 mm. Carapace, 1.7 mm long,
1.1 mm wide. First femur, 2.2 mm; pa-
tella and tibia, 2.4 mm; metatarsus, 1.7
mm; tarsus, 0.7 mm. Second patella and
tibia, 1.9 mm; third, 0.9 mm; fourth, 1.2
mm.

Variation. Total length of females, 2.6
to 4.0 mm; carapace, 1.6 to 1.8 mm long;
first patella and tibia, 1.8 to 2.3 mm.
Males vary from 2.6 to 3.2 mm in total
length; carapace, 1.6 to 1.9 mm long; first
patella and tibia, 2.2 to 2.7 mm.

Diagnosis. This species differs from
others by the shape of the chelicerae
(Figs. 3, 11), by the embolus and para-
cymbium in the small palpus (Figs. 14,
15), and by the shape of the ducts and
seminal receptacles in the female (Fig.
5).

Natural History. This forest species
makes its web between tree roots, just
above the ground and at an angle nearly
horizontal. Behind the web a vertical line
of debris hangs from an upper frame
thread, held together by silk and includ-
ing egg-sacs (arrows, Plate 2). According
to W. T. Sedgwick (personal communi-
cation), the spider rests at night in the
hub of the web. During daytime the spi-
der is among the line of debris. The
species is probably much more common
than records indicate; unless one espe-
cially searches for it, it is overlooked. The
web is about 6 to 12 cm in diameter (Plate
2). Mature males and females can be
found all summer.

Distribution. Southeastern states (Map
1). The northernmost record is from Lit-
tle Grand Canyon, Jackson Co., Illinois,
from several collections (J. Beatty).

Tetragnattia Latreille

Tetragnatha Latreille, 1804, Tableau methodique
des insectes, in Nouveau Dictionnaire d'Histoire
Naturelle, 24: 135 (unavailable). Type species by
monotypy is T. extensa. The name is feminine.

Eugnatha Audouin in Savigny, 1825, Explications
sommaires des Planches d'Arachnides de I'Egypte
et de la Syrie, p. 119. New name for Tetragnatha
(an objective synonym).

Eucta Simon, 1881, Arachnides de France, 5: 5.
Type species by monotypy is E. gallica.

Arundognatha Wiehle, 1963, Tetragnathidae in

Plate 3. Upper left, male Tetragnattia versicolor, Michigan; upper rigtit, female T. viridis, Cape Cod; bottom, web of T.
versicolor with closed hub, Friday Harbor, Washington (web photo B. Opell).

DoLicHocNATHA AND Tetragnatha North OF MEXICO • Levi 283

284 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Tierwelt Deutschlands, 49: 47. Type species T.
.striata L. Koch designated by Wiehle.

Note. The characters that separate Euc-
ta and Arundo^natha from Tetrafi,uatha
are not those of distinct groups. The two
North American species with tails {=Euc-
ta), T. caudata (Figs. 140-148) and T.
branda (Figs. 167-175), are not closely
related. Judging by similarity of genita-
lia, T. caudata is close to T. pallescen.s;
both have the conductor tipped by a hook
(Figs. 139, 148). Tetragnatha branda is
close to T. vermiformis; both lack the
pleat next to the conductor (Figs. 175,
184); females of both have the seminal
receptacles on each side dorsoventral to
each other (Figs. 170, 179).

Some rare specimens of T. laboriosa
have the lateral eyes farther apart than
medians, and could be included in Arun-
dognatha. But, most have the lateral eyes
slightly closer together than the medians.
Tetragnatha laboriosa, judging by the
genitalia, is close to T. versicolor, T. vir-
idis, and T. pallescens. Tetragnatha ver-
sicolor always has the lateral eyes closer
than the medians; T. viridis, the same
distance apart; T. pallescens, always far-
ther apart. All belong to the same species
group.

Diagnosis. Tetragnatha differs from
Pachygnatha and Glenognatha by lack-
ing a tapetum in the lateral eyes (Figs.
21, 22); it differs from Dolichognatha as
well as from the two previous genera by
having an elongate to tubular-shaped ab-
domen (Fig. 23), usually with silver pig-
ment. It further differs from Dolichogna-
tha in lacking an epigynum and by hav-
ing trichobothria on the femora (Fig. 17).
It differs from Leucauge by lacking ta-
petum in the secondary eyes.

Description. Carapace slightly elon-
gate, weakly sclerotized, lacking setae,
and often with an indistinct median lon-
gitudinal gray band. Eyes on black spots.

Sternum weakly sclerotized, surrounding
coxae (Fig. 16). Chelicerae and legs yel-
lowish white to brown. Abdomen silver,
consisting of tiny pigment spots least
dense on venter, most dense on dorsum.
Dorsum either has gray or black folium,
or lacks pigment and folium altogether;
the venter may have a median longitu-
dinal dark band.

Eyes closely grouped, lateral eyes less
than two diameters from medians (Figs.
24, 31). The lateral eyes may touch, but
more commonly are slightly separated or
far apart, with the anterior lateral eyes
having moved away in an anterior lateral
direction (Figs. 177, 182). The placement
of the eyes is somewhat variable within
species. Clypeus height equals one to
t^vo diameters of anterior median eyes
(Figs. 24, 130). Chelicerae more or less
enlarged, often with complex teeth (Figs.
24, 25, 31). Labium rebordered (distally
swollen); endites elongate and distally
widened (Fig. 16). Legs very long, first
longest, third shortest. Abdomen cylin-
drical, often swollen anteriorly (Fig. 23)
or with sides almost parallel (Fig. 149);
in some species a tail overhangs spinner-
ets (Figs. 140, 167).

Males are only slightly smaller than fe-
males. Male chelicerae are larger and
more projecting than those of female,
with a spur (apophysis) on anterior face
to hold the female fang when mating
(Figs. 31, 52). The chelicerae have a very
long tooth (the main tooth) distally from
a row of smaller teeth (Figs. 62, 80) (ab-
sent in T. dearmata and T. pallescens).
These smaller teeth become progressive-
ly smaller proximally. Distal from the
large tooth is a slanting tooth (Figs. 62,
80), and at the base of the fang a smaller
knob called a fang guide (Locket and
Millidge, 1953; Wiehle, 1963). The den-
tition is much more complex in many
species (Figs. 24, 31).

Plate 4. Above, female Tetragnattia pallescens, Florida; below, mating T. straminea, Massachusetts, male on left with
left palpus inserted, right palpus held forward (photo J. Coddington).

DoLicHOGNATHA AND Tetragsatha North OF MEXICO • Levi 285

286 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

There are tricholjothria on each femur
(Fig. 17). Tetrag,natha lahoriosa females
have 13 trichobothria on the first femur,
11 on the second, 11 on the third, and 13
on the fourth. There are the usual trich-
obothria on the tibiae, and there is one
trichobothrium on the third metatarsus.

Genitalia. The male palpus is slightly
more complex than that of Pachygnatha.
A spherical tegulum (T in Figs. 19, 20)
encloses the duct. The duct makes only
a single coil. The filifomi embolus (E) is
distally wrapped within a complex con-
ductor (Figs. 19, 20). The cymbium (Y)
appears more complex than that of other
araneids; it is small and may be more or
less lobed (Figs. 18-20). The paracym-
bium (P) is a separate sclerite (Fig. 18) as
in Pachygnatha and the family Linyphi-
idae.

The paracymbium (P) is straight, but
usually has a small lobe on the lateral
cymbium side (Fig. 18). Tetragnatha
guatemalensis and T. eannra have a dis-
tal finger projecting from the paracym-
bium (Figs. 44, 54); that of T. nitens has
a large ventral transparent lobe and a dis-
tal notch (Fig. 33). In other species the
paracymbium is not diagnostic.

The most complex part of the palpus is
the conductor (C in Figs. 19, 20 and
Plates 5-7), which often has pleats lat-
erally (L in Fig. 45). These pleats are ab-
sent in T. branda (Fig. 175), T. vermi-
formis (Fig. 184), the European T.
striata, and the cosmotropical T. nitens
(Fig. 34). When expanded, it can be seen
that the pleats enlarge, since they are he-
matodocha (C in Figs. 19, 20). The distal
tip of the conductor is sclerotized and
seems the most useful character for iden-
tifying species (Plates 5-7).

The female genitalia are quite similar
to those of Pachygnatha (Levi, 1980a).
There is no epigynum; the genital open-
ings are at the posterior end of a median
ventral lobe from the epigastric area (Fig.
26). The seminal receptacles are anterior,
near the level of the lung spiracles; there
are two on each side (Fig. 48), only one

in T. nitens (Fig. 27). There is a median,
lightly sclerotized structure, comprising
the so-called median seminal receptacles
(Figs. 27, 48), which probably hold the
conductor when mating. The shape and
position of the seminal receptacles are
the most useful species diagnostic char-
acters of females.

Species. There are Tetragnatha species
on all continents and islands, and in arc-
tic, temperate, and tropical climates.

Judging by the similarity of the geni-
talia, most of the North American species
are closely related to each other: elon-
gata, versicolor, viridis, lahoriosa, pal-
lescens, caudata, straminea, and proba-
bly shoshone. The cosmotropical T.
nitens is distinct from any other species
in the area: it has only two seminal re-
ceptacles (Fig. 27) and, like T. vermifor-
mis, lacks the pleats on the palpal con-
ductor (Fig. 34). Tetragnatha venniformis
(Figs. 176-184) and T. branda (Figs.
167-175) I believe are closely related to
the European T. striata. Tetragnatha
guatemalensis and T. earmra belong to
a fourth group, judging by the more spe-
cialized paracymbium (Figs. 44, 54) and
tip of the conductor (Figs. 45, 55). Tet-
ragnatha extensa and T. dearmata are
closest to the first group mentioned and
perhaps to Old World relatives of the T.
versicolor group.

Species that have the lateral eyes rel-
atively close together also have the larg-
est amount of black pigment, and have
the abdomen swollen anteriorly. Species
that have the lateral eyes far apart tend
to have less dark pigment; the abdomen
tends to be tubular and silvery.

Diagnostic Species Characters. The
best characters to separate species are: in
males, the shape of the conductor tip
(Plates 5-7); in females, the shape and
position of the seminal receptacles,
which can be seen only after lifting the
exoskeleton of the genital lobe on one
side (see Methods of Study). Coloration,
amount of black pigment, and the dis-
tance between the lateral eyes are help-

DOLICHOGNATHA AND TeTRAGNATHA NORTH OF MEXICO • Levi 287

ful. The chelicerae, faithfully illustrated only when temperatures are warm and it

by earlier authors, have unique charac- is not windy. It seeks a new location each

ters in only a few species. night; the web size is controlled by the

However, the separation of Tetragna- physical dimensions of the space be-
tha species is still reasonably difficult, tween vegetation. The spiders feed most
While the distance between lateral eyes actively from sundown until just after
compared with the distance between me- dark (Le Sar and Unzicker, 1978). Tet-
dians is an excellent superficial charac- ragnatha feed nonselectively on dipter-
ter, there is considerable variation within ans, but Le Sar reports that their T. la-
species, perhaps most noticeable in T. boriosa also feeds on Miridae and
laboriosa. Most T. lahoriosa have the lat- Cicadellidae. The importance of Tetra-
eral eyes slightly closer together than the gnatha as a mosquito predator has been
medians, but some have them the same described in papers by Luczak and her
distance apart, and in some specimens colleagues (1966, 1968, 1970). Adults of
they are slightly farther apart. North American species are found almost

Many of the old North American pub- all summer, and live for the season; per-
lished records of Tetragnatha species haps there are two generations in sum-
were based on determinations of juvenile mer (Le Sar and Unzicker, 1979). The
specimens or adults, using chelicerae, males are almost as common as females,
and are in error. and therefore must have a similar life

Natural History. Most Tetragnatha span,

species are found near water. Some may There is no obvious courtship. The

seem rare only because spider collectors male grasps the female's chelicerae with

are generally terrestrial (e.g., T. branda his fangs (Plate 4); the coupling is appar-

and T. vermiformis). The most abundant ently so secure that no holdfast organ is

species in collections, T. laboriosa and necessary adjoining the copulatory open-

T. versicolor, are the two species living ing (epigynum). Perhaps this is the cir-

farthest away from water. Several species cumstance that has pemiitted the loss of

have very specialized habitats: T. viridis the epigynum. The female may coil her

is found in conifers, especially pines; the abdomen while mating. Palpi are alter-

European T. pinicola is also found in nated, as in most entelegyne spiders

trees. The rarity of T. viridis may reflect (Plate 4). Unlike most other Araneidae,

its reluctance to drop within reach. Tet- the male does not court between the use

ragnatha branda has been collected only of left and right palpi. Like haplogyne

from salt marshes. spiders and Palpimanidae, males may

The orb-web ranges from horizontal to use both palpi together in spenn induc-

vertical. The hub, usually but not always tion. A week after mating, the female may

open (Plate 3), is cut out after the web is produce several egg-sacs, fluffy struc-

completed (Wiehle, 1963). The spider tures broadly attached to vegetation. The

rests on a plant or in the middle of the young hatch soon thereafter and overwin-

web, legs extended in a paraxial direction ter in immature stages. Adults of the spe-

(Plates 3, 4). The web of T. laboriosa is cialized T. viridis are only found in late

made in the evening, before nightfall (Le spring. In one instance a late August

Sar and Unzicker, 1978), as is the web of young was raised. It readily fed on the

T. nigrita (Wiehle, 1963), and perhaps of drosophila flies which were caught with-

other species as well. According to out web (in a vial) and matured in De-

Crome (1954), Tetragnatha kaestneri (= cember in the laboratoiy after two molts.

Eucta kaestneri) does not make a scaf- As an adult male, it continued to feed on

folding before the viscid spiral. drosophila flies. Luczak and Dabrowska-

Tetragnatha laboriosa makes its web Prot (1966) also describe the ability of

288 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Map 2. Distributions of Tetragnatha nitens (Audouin), T. earmra n. sp., 7". guatemalensis O. P.-Cambridge, T. extensa
(Linnaeus), and T. dearmata Thorell in North America.

Tetragnatha to catch flies without the
use of webs.

Key to Species. The distance of the lat-
eral eyes from each other compared to
the distance of the median eyes is an easy
character to see. The distance is mea-
sured from the center of the eyes. Unfor-
tunately there is considerable variation,
especially in the commonest species, T.
laboriosa. While in most specimens of
this species the lateral eyes are slightly
closer together than the medians or an
equal distance apart, in a very rare spec-
imen the eyes are very slightly farther
apart.

Key to Female Tetragnatha North
OF Mexico

1. Lateral eyes farther apart than medians

(Figs. 130, 141, 150, 159, 168, 177) 2

Lateral eyes as far apart as medians or
closer together (Figs. 24, 36, 75, 111) -— 7

2(1). Abdomen with a tail extending beyond

spinnerets (Figs. 140, 167) 3

Abdomen without a tail (Figs. 129,
176) 4

3(2). Anterior median seminal receptacles
closer to each other than to other mem-
ber of pair (Figs. 142, 143); Canada,
Northwest, eastern states (Map 4)

caudata

Anterior and posterior seminal recep-
tacles on each side next to each other
(Figs. 169, 171); New England to Gulf
Coast (Map 4) branda

4(2). Abdomen only silver on dorsal half,
edge of silvery area forming a lateral
line with area below lacking silver pig-
ment (Fig. 149); seminal receptacles
on each side about 1 to 2 times their
diameter apart and connected by a
stalk (Fig. 153); Alberta, Utah, and
East (Map 4) straminea

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 289

Tetragnotho viridis

Map 3. Distributions of Tetragnatha elongata Walckenaer, T. versicolor Walckenaer, T. viridis Walckenaer, and 7.
laboriosa Hentz.

Abdomen all silvery, except for mid-
venter (Fig. 158); seminal receptacles
otherwise (Fig. 162) 5 -

5(4). Seminal receptacles elongate, paraxi-
al, and members of a pair on top of one
another (Figs. 178, 180); abdomen at 8(7).
most 3 times longer than wide or high
(Fig. 176); Ontario, northwest to Flor-
ida (Map 4) vermiformis -

Seminal receptacles pear-shaped and

side by side (Figs. 131, 133, 160, 162); 9(8).

length of abdomen about 4 or 5 times

width or height (Figs. 129, 158) 6

6(5). Members of pair of seminal recepta-
cles on each side almost touching (Fig. -
162); Great Slave Lake, Ontario, In-
diana to Pacific (Map 4) shoshone

Members of pair of seminal recepta- 10(9).

cles on each side separated by an in-
termediate lobe (Fig. 133); widespread
(Map 4) pallescens

7(1). Genital cover twice as long as wide,

narrowed anteriorly, posterior rim -

notched (Fig. 26); only two seminal re-

ceptacles (Fig. 27); Florida, Texas,

California (Map 2) nitens

Genital cover square or wider than
long (Figs. 49, 77); four seminal recep-
tacles (Figs. 50, 69, 78) 8

Lateral eyes about as far apart as me-
dians or slightly closer together (Figs.

57, 111, 121) _'_ 9

Lateral eyes clearly closer together
than medians (Figs. 47, 66, 75, 91) ._ 11
Green; first and second tibial macro-
setae length 3 to 5 times diameter of
article (Fig. 115); New Brunswick to

eastern Texas (Map 3) viridis

Silvery and yellowish white; macro-
setae length at most 3 times diameter

of article 10

Prominent pouch between pair of sem-
inal receptacles hiding anterior one in
ventral view; posterior seminal recep-
tacle pear-shaped (Fig. 124); wide-
spread (Map 3) laboriosa

No pouch between members of pair of
seminal receptacles; posterior seminal

290

Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Map 4. Distributions of Tetragnatha pallescens F. P.-Cambridge, T. caudata Emerton, T. straminea Emerton, 7. Sho-
shone n. sp., T. branda n. sp., and T. vermiformis Emfirton.

receptacle elongate and transverse
(Fig. 60); Alaska, Canada, northeastern
states. Rocky Mountains (Map 2) .. extensa
11(8). Anterior seminal receptacles of each
pair elongate, paraxial, parallel to re-
ceptacle of other pair (Figs. 76, 78);
widespread (Map 2) elongata

- Anterior seminal receptacles oval or
pear-shaped; if elongate, not placed
parallel (Figs. 39, 69) 12

12(11). Seminal receptacles of each pair on top
of one another, paraxial (Figs. 39,
50) 13

- Seminal receptacles of each pair side

by side (Figs. 69, 94) 14

13(12). Seminal receptacles elongate-oval (Fig.

50); widespread (Map 2) guatetnalensis

- Seminal receptacles subspherical (Fig.
39); Everglades and Florida Keys
(Map 2) earmra

14(12). Posterior lateral seminal receptacle of
each pair U-shaped (Fig. 69); Alaska to
northern States (Map 2) dearmcita

- Posterior lateral seminal receptacle of
each pair oval (Fig. 94); widespread
(Map 3) versicolor

Key to Male TETRAGNATHA North of Mexico

1. Lateral eves farther apart than medi-
ans (Figs. 137, 146, 155, 164, 173,
182) 2

- Lateral eyes as far apart as medians or
closer together (Figs. 31, 52, 71, 80,
126) 7

2(1). Abdomen with a tail extending beyond

spinnerets (Figs. 145, 172) '_ 3

- Abdomen without tail (Figs. 136,
181) 4

3(2). Conductor with lateral pleats (Figs.
147, 148); tip bird head-shaped with a
lateral spur (Fig. 148; Plate 7c); Can-
ada, northwest to eastern states (Map
4) caudata

- Conductor without lateral pleats (Figs.
174, 175), tip with lateral notch (Fig.
175; Plate 7g, h); New England to Gulf
Coast (Map 4) branda

4(2). Conductor without pleats (Figs. 183,
184); tip tapering to a point, small,
bent over (Fig. 184; Plate 7i); On-
tario, northwest to Florida (Map
4) vermiformis

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 291

Conductor with pleats (Figs. 138, 139);
tip large with a neck (Figs. 138,
166) 5

5(4). Tip of conductor a large recurved hook
(Fig. 139; Plate 7a); widespread (Map
4) pallescens

- Tip of conductor not a large recurved
hook (Figs. 157, 166) 6

6(5). Tip of conductor with a narrow neck
(Figs. 156, 157; Plate 7d); only dorsal
half of abdomen silver (Fig. 154); Al-
berta, Utah, and East (Map 4) straminea

- Tip of conductor with a wide neck
(Figs. 165, 166; Plate 7e, f); abdomen
except for midventer all silver (Fig.
163); Great Slave Lake, Ontario, In-
diana to Pacific (Map 4) shoshone

7(1). Paracvmbium with distal notch or with

distal'finger (Figs. 33, 44, 54) 8

- Paracvmbium not modified, distally
rounded (Figs. 18, 20, 81) 10

8(7). Paracymbium wide with distal notch
(Fig. 33); conductor without pleats
(Fig. 34; Plate 5a, b); Florida, Texas to
California (Map 2) nitens

- Paracymbium narrow with distal fin-
ger (Figs. 44, 54) 9

9(8). Chelicerae about half length or longer
than carapace (Figs. 51, 52); conductor
tip with lateral finger (Fig. 55; Plate

5d); widespread (Map 2)

guatemalensis

- Chelicerae about half carapace length
(Figs. 40^2); tip of conductor cres-
cent-shaped in ventral view, concave
laterally (Fig. 45; Plate 5c); Everglades
and Keys (Map 2) earmra

10(7). Conductor tip with neck and sclero-
tized on lateral side, pointing laterally
to side of paracymbium (Figs. 63, 64;
Plate 5e); Alaska, Canada, northeast-
ern states. Rocky Mountains (Map
2) extensa

- Conductor tip pointed distally, re-
curved or pointing mesallv (Figs. 73,

82, 128; Plates 5f, g, 6b, h)' 11

11(10). Conductor tip bird head-shaped (Fig.
128; Plate 6h, i); pleat closest to con-
ductor lifted up (Fig. 128); widespread
(Map 3) lahoriosa

- Conductor tip otherwise (Figs. 73, 82,
98, 119); pleats' curvature about par-
allel to conductor (Figs. 73, 82) 12

12(11). Lateral eyes about as far apart as me-
dians (Fig. 117); green; length of first
and second tibial macrosetae 3 to 5
times diameter of article (Fig. 115);
New Brunswick to eastern Texas (Map
2) viridis

- Lateral eyes closer together than me-
dians (Figs. 80, 96); silver-brown to
gray; length of first and second tibial

macrosetae at most 3 times diameter of
article 13

13(11). Conductor tip tapering to a minute
point, conductor concave laterally near
tip (Fig. 82; Plate 5g, h); chelicerae
usually longer than carapace (Figs. 79,
80); widespread (Map 3) elongata

- Conductor tip otherwise (Figs. 73, 98);
chelicerae usually shorter than cara-
pace (Figs. 70, 7l', 95, 96) 14

14(13). Conductor tip a pennant, pointing
mesally (Figs. 72, 73; Plate 51); Alaska
to northern states (Map 2) dearmata

- Conductor tip a recurved hook (Fig.
98, 105-109; Plate 6b-f); widespread
(Map 2) versicolor

Tetragnatha nitens
(Audouin in Savigny)
Plate 5a, b; Figures 23-34; Map 2

Eugnatha nitens Audouin in Savigny, 1825, Expli-
cation sommaire des Planches d'Arachnides de
I'Egypte et de la Syrie, p. 118, pi. 2, fig. 2, 9.
Specimens from near Rosetta [Rashid], Egypt,
lost.

Eugnatha pelusia Audouin in Savigny, 1825, Ex-
plication sommaire des Planches d'Arachnides de
I'Egypte et de la Syrie, p. 119, pi. 2, fig. 3, 6.
Specimen from island of Rosetta [Rashid], Egvpt,
lost. NEW SYNONYMY.

Tetragnatha andina Taczanowski, 1878, Horae Soc.
Entom. Rossicae, 14: 144, pi. 1, fig. 2, 9. Numer-
ous female and male synt>'pes from Amable Ma-
ria, Prov. Tarma, Peru in the Polish Academy of
Sciences, examined. NEW SYNONYMY.

Tetragnatha antiUana Simon, 1897, Proc. Zool. Soc.
London, p. 868. 13 male and 17 female and 5 ju-
venile syntypes from St. Vincent Island, Lesser
Antilles, in the British Museum, Natural Histor\',
examined. Seelev, 1928, Bull. New York State
Mus, 278: 105, figs. 1-i, 9, d. Roewer, 1942, Kat-
alog der Araneae, 1: 988. Chickering, 1957, Bull.
Mus. Comp. Zool., 116:306, figs. 1-6, 9,6; Bre-
viora, 68: 2. Bonnet, 1959, Bibliographia Araneo-
rum, 2: 4318. Chickering, 1962, Bull. Mus. Comp.
Zool., 127: 428, figs. 1-6, 9, 6. NEW SYNONY-
MY.

Tetragnatha vicina Simon, 1897, Proc. Zool. Soc.
London, p. 869. Three male syntypes from St.
Vincent, Lesser Antilles in the British Museum,
Natural History, examined. NEW SYNONYMY.

Tetragnatha peninsulana Banks, 1898, Proc. Cali-
fornia Acad. Sci., ser. 3, 1: 246, pi. 15, fig. 12, 9,
6 . Four female syntypes from San Jose del Cabo,
Baja California in the Museum of Comparative
Zoo'ogy, examined. Male synt>'pe in California
Academy of Sciences, destroyed. NEW SYNON-
YMY.

Tetragnatha galapagoensis Banks, 1902, Proc.
Washington Acad. Sci., 4: 61, pi. 1, fig. 10, 9,6.
One male, one female syntype from Galapagos

292 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Islands in the Museum of Comparative Zoology,
examined NEW SYNONYMY.

Tctra^nutha aptatts Chamberlin, 1920, Bull. Brook-
lyn Mus. Sci., 3: 41, figs. 7, 8, 9. Female para-
types from Chincha Island, Peru in the Museum
of Comparative Zoology, examined.

Tctra^natha eremita Chamberlin, 1924, Proc. Cal-
ifornia Acad. Sci., ser. 4, 12: 645, figs. 89, 90, 6 .
Right male palpus of holotype from Puerto Es-
condido, Baja California in the Museum of Com-
parative Zoology, examined. NEW SYNONYMY.

Tctragnatha seminola Gertsch, 1936, Amer. Mus.
Novitates, no. 8.52: 10, figs. 22, 23, 6. Male Ho-
lot\pe from the northeast shore of Lake Okee-
chobee, Florida in the American Museum of Nat
ural History, examined. NEW SYNONYMY.

Tetragnatha steckleri Gertsch and Ivie, 1936
Amer. Mus. Novitates, no. 858; 19, figs. 31-3^,
9, (5. Male holotype from Tucson, Arizona ir, the
American Museum of Natural History, examined.
NEW SYNONYMY.

Tetragnatha nitens: — Roewer, 1942, Katalog der
Araneae, 2: 978. Bonnet, 1959, Bibliographia Ar-
aneorum, 2: 4345. Okuma, 1968, Acta Arachno-
logica, 21: 40, figs. 9-16, 9,6.

Tetragnatha elmora Chamberlin and Ivie, 1942,
Bull. Univ. Utah, biol. ser., 7(1): 62, fig. 160, 9.
Female holotype from Bluevvater Lake, New
Mexico in the American Museum of Natural His-
tory', examined. NEW SYNONYMY.

Tetragnatha festina Bn,ant, 1945, Bull. Mus. Comp.
Zool., 95: 407, figs. 38, 39, 41, 6 (not 9). Male
holotype from foothills of Cordillera Central,
south of Santiago, Dominican Republic in the
Museum of Comparative Zoologv, examined.
NEW SYNONYMY.

Tetragnatha haitensis Bryant, 1945, Bull. Mus.
Comp. Zool., 95: 408, fig.' 37, 9 . Female holotype
from Ennery, Haiti in the Museum of Compara-
tive Zoology, examined.

Note. Roewer (1942) cites the page of
Savigny as 323 and the date as 1827. This
is a later printing of the work. Plate 2c of
Audouin in Savigny shows the diagnostic
posterior cusp on the fang of the female.
The posterior distal cheliceral teeth of
the female are shorter in Audouin's illus-
tration than those in our specimens. Te-
tragnatha pelusia has never before been
synonymized with T. niten.s, although

the palpus and chelicerae are illustrated
and the illustrations are more diagnostic
than those of the female.

Variation. The length of the palpal tib-
ia may be greater or less than the cym-
bium. The lateral eyes are sometimes as
far apart as the medians, usually slightly
closer together. On Panamanian speci-
mens the diagnostic tooth at the posterior
base of the fang is as long as the chelic-
erae are wide, and is sometimes smaller
than illustrated on the most northern
specimens. Total length of females, 7.4 to
11.0 mm; total length of males, 5.2 to 10.1
mm.

Diagnosis. Both sexes of this species
are easily separated from other North
American species. The lateral eyes are as
far apart as the medians, sometimes clos-
er together or, rarely, farther apart. There
are gray pigment marks on the body (Fig.
23).

The paracymbium of the male palpus
has a distal notch and a ventral transpar-
ent lobe (Fig. 33); the conductor lacks
pleats along the distal edge (Plate 5a;
Fig. 34).

The female chelicerae have a large
posterior lateral tooth at the insertion of
the fang (Figs. 24, 25) and a cusp poste-
riorly on the base of the fang (sometimes
absent) (Fig. 25); the genital area is long-
er than wide, constricted anteriorly, with
a notch above the opening (Fig. 26). Un-
like all other North American species,
there are only two seminal receptacles
(Fig. 27). The female chelicerae are as
long as the carapace or longer (Figs. 23,
24).

Natural History. The species has been
collected above wet ground and near a
creek in Mexico, and in woods in Puerto
Rico.

Plate 5. Tetragnatha male, left palpi; scanning electron micrographs of conductors and their tips, a, b, T. nitens,
compare to Fig. 34. c, T. earmra (tip broken on right), compare with Fig. 45. d, 7. guatemalensis, compare to Fig. 55. e,
T. extensa, compare to Fig. 64. f, T. dearmata, compare to Fig. 73. g, i, T. elongata, compare to Fig. 82. (g, h, locality
unknown; i, Florida). Scale lines. 0.05 mm.

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 293

294 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Distribution. Cosmotropical, in North
America north to northern Florida, north-
vxn Texas, New Mexico to northern Cal-
ifornia (Map 2). Since there are several
similar species in tropical America, in-
cluding T. peruviana Taczanowski and
T. piscaforia Simon, T. nitens may have
originated in South America.

Records. Florida. Alachua Co.: Caines-
ville; Newnan's Lake. Clades Co.: NE
shore of Lake Okeechobee. Highlands
Co.: Sebring. Monroe Co.: Johnston Key.
Texas. Baylor Co.: Mabelle. Hidalgo Co.:
Hargill. New Mexico. Valencia Co.: Blue-
water Lake. Arizona. Maricopa Co.: Salt
River near Blue Point. Pima Co.: Tucson.
California. San Diego Co.: Escondido
Reservoir; San Diego Mission. San Louis
Obispo Co.: Atascadero Lake. Shasta Co.:
McArthur.

The Museum of Comparative Zoology
has specimens from several areas not
shown on Map 2. Galapagos Islands.
Peru. Chincha Isl. Brazil. Rio dos Ma-
cacos; Mendez; Sao Paulo. Paraguay. Rio
Parana; 75 km NE Concepcion; Hernan-
drias. Madagascar. New Zealand.

Tetragnatha earmra new species
Plate 5c; Figures 35-45; Map 2

Holotype. Male from Eannra South East Island,
Everglades National Park, Florida, 28 January
1973 (A. Sheldon), in the Museum of Compara-
tive Zoology. The name is an arbitrary combina-
tion of letters.-

Descriptio7i. Female. Carapace brown,
head region darker. Chelicerae brown;
labium dark, except for swollen area.
Sternum dark brown, legs brown. Eyes

^ The exact location of Earmra South East Island
could not be located in Everglades National Park.
To avoid a name change, if Eannra turns out to be
a misspelling, the name is considered an arbitrary
combination of letters.

with distinct black rings (Fig. 36). Dor-
sum and abdomen with several pairs of
dark patches on sides (Fig. 35). Venter
with black band having parallel sides
(Fig. 35). The lateral eyes are closer to
each other than the medians, and situated
together on a tubercle (Fig. 36). Posterior
eyes 0.8 diameters of anterior medians;
anterior laterals 0.5 diameters of anterior
medians. Anterior median eyes their di-
ameter apart, slightly less than two di-
ameters from laterals. Posterior median
eyes about 4 diameters apart, slightly less
than four diameters from laterals. Total
length, 6.0 mm. Carapace, L9 mm long,
L3 mm wide. First femur, 4.2 mm; pa-
tella and tibia, 5.1 mm; metatarsus, 4.3
mm; tarsus, LO mm. Second patella and
tibia, 2.7 mm; third, LI mm; fourth, 2.4
mm.

Male. Male lighter colored than fe-
male. Diameter of posterior eyes 0.8 di-
ameters of anterior medians. Anterior lat-
eral eyes 0.6 diameters of anterior
medians. Anterior median eyes their di-
ameter apart, about L5 from laterals. Pos-
terior median eyes about two diameters
apart, about L8 from laterals (Figs. 41,
42). The chelicerae are not long, but the
guide tooth has shifted anteriorly; the
main tooth is the biggest (Figs. 41, 42).
The base of the fang is attached to an area
that seems narrow and projecting (Fig.
42). Total length, 3.8 mm. Carapace, 1.4
mm long, 0.9 mm wide. First femur, 3.1
mm; patella and tibia, 3.4 mm; metatar-
sus, 3.2 mm; tarsus, 0.9 mm. Second pa-
tella and tibia, 1.9 mm; third, 0.8 mm;
fourth, 1.7 mm.

Diagnosis. Tetragnatha earmra has the
lateral eyes closer together than the me-
dians, and on a tubercle; there are large
black rings around the eyes (Figs. 36, 41,
42).

Plate 6. Tetragnatha male, left palpi; scanning electron micrographs of conductors and their tips; a-f, T. versicolor,
compare to Figs. 98, 104-109. a, b, (Georgia); c, (Virginia), d, (Rhode Island); e, (Michigan); f, (California), g, T. viridis,
compare to Fig. 119. h, i, 7. laboriosa, compare to Fig. 128. Scale lines. 0.05 mm.

DoLicHOGNATHA AND Tetragnatha North OF MEXICO * Levi 295

296 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Like the larger T. guatemalensis, the
paracyiiil)ium of the male has a distal
bent finger (Fig. 44), but the tip of the
conductor (Plate 5c; Fig. 45) appears
twisted quite differently from that of T.
guatemalensis.

The female has smaller chelicerae
(Figs. 35, 36) than those of T. guatema-
lensis. There are two seminal recepta-
cles, one behind the other on each side:
one is lighter, ventral, pear-shaped and
the other is darker, dorsal, more angular
(Fig. 39). The seminal receptacles of T.
guatemalensis are more elongate.

Paratijpes. Florida. Everglades Nation-
al Park: Paradise Key, 29 March 1951, 6
(J. Vockeroth, CNC); Pineland Trail, 21
June 1964, 9 (K. J. Stone, FSCA) (Map
2).

Tetragnatha guatemalensis O. P. -Cam-
bridge
Plate 5d; Figures 46-55; IVlap 2

Tetragnatha guatemalensis O. P. -Cambridge, 1889,
Biologia Centrali-Americana, Araneidea, 1: 8, pi.
2, figs. 6, 7, 9,6. Male lectotype here designated
from Guatemala [Cahabon or Laguna de los Co-
heteros near Coban] in the British Museum, Nat-
ural History, examined. F. P. -Cambridge, 1903,
Biologia Centrali-Americana, Araneidea, 2: 431,
pi. 40, figs 12, 13, 9 , d. Chickering, 1959, Bull.
Mus. Comp. Zool., 119: 482, figs. 13-21, 9,6.
Bonnet, 1959, Bibliographia Araneorum, 2: 4334.

Tetragnatha hanksi McCook, 1893, American Spi-
ders, 3: 262, pi. 25, fig. 6, pi. 28, fig. 4, 6 . Types
from Florida and Wisconsin in the Academy of
Natural Sciences, Philadelphia, lost.

Tetragnatha intermedia Banks, 1898, Proc. Califor-
nia Acad. Sci., ser. 3, 1: 247, pi. 15, fig. 14, 6 . One
female, two male syntypes from Tepic, Mexico in
the Museum of Comparative Zoology, examined,
NEW SYNONYMY.

Tetragnatha seneca Seeley, 1928, Bull. New York
State Mus., 178: 134, pi. 4, figs. 44^8, 9,6. Male,
female syntypes from Lodi Landing, Seneca
Lake, New York in the American Museum of Nat-
ural History, examined. Kaston, 1948, Bull. Con-
necticut Geol. Natur. Hist. Surv., 70: 271, fig. 872,

9. Bonnet, 1959, Bibliographia Araneorum, 2:
4355.
Tetragnatha laudativa Gertsch and Mulaik, 1936,
Amer. Mus. Novitates, no. 863: 15, fig. 33, 6
(only). Male holotype from Kingsville, Texas in
the American Museum of Natural History, ex-
amined. NEW SYNONYMY.

Variation. The male chelicerae are at
times as long as the carapace, sometimes
shorter. Total length of females, 5.4 to
11.5 mm; males, 5.2 to 10.2 mm.

Diagnosis. This species has the lateral
eyes always closer together than the me-
dians and has a variable amount of dusky
gray to black markings.

The male can readily be separated
from that of other species except T. earm-
ra by the presence of a finger extending
from the tip of the palpal paracymbium
(Figs. 53, 54); from T. earinra by the nar-
row lateral (in ventral view) extension on
the tip of the conductor, straight or con-
vex laterally (Plate 5d; Fig. 55). In con-
trast, that of T. earmra is concave later-
ally.

The female can only be separated from
Tetragnatha versicolor and T. dearmata
by examining the seminal receptacles:
each member of the pair is oval, paraxial
in position; one is dorsal and slightly to
the side of the other (Fig. 50). The sem-
inal receptacles are longer (Fig. 50) than
the similar but more spherical ones of T.
earmra. Tetragnatha guatemalensis has
longer chelicerae (Figs. 46, 47) than T.
earmra.

Natural History. Specimens have been
collected from herbs; vegetation along
marsh trail; sumac (Rhus) foliage and red
pine {Pinus resinosa) in Ontario; sweep-
ing honeysuckle (Lonicera) in West Vir-
ginia; and pine trees and web along
grassy lakeshore in Wisconsin. In Florida
specimens have been collected from
weeds bordering streams; Australian

Plate 7. Tetragnatha male, left palpi: scanning electron micrographs of conductors and their tips; a, 7. pallescens,
compare to Fig. 139. b, c, T. caudata, compare to Fig. 148. d, 7. straminea, compare to Fig. 157. e, f, T. stiostione,
compare to Fig. 166. g, h, T. brands, compare to Fig. 175. i, T. vermiformis, compare to Fig. 184. Scale lines. 0.05 mm.

DoLicHOGNATHA AND Tetragnatha North OF MEXICO * Levi 297

298 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

pine (Casuarina); garden vegetation;
tangerine tree; orange grove; and citrus
grove. All these habitats were probably
near the edge of rivers, lakes, and other
water.

Distribution. Nova Scotia, Minnesota,
Kansas, New Mexico, Southern Califor-
nia, south to Panama, Cuba, Jamaica. Bor-
der records are: Annapolis Roval, Nova
Scotia, 1 Aug. 1971, 6 (C. D. Dondale);
Rocky Point, Lake of the Woods Co.,
Minnesota, 6 (B. Cutler); Meade County
State Park, Kansas, 9, S (W. Ivie); Blue-
water Lake, 4 Sept. 1941, 9 (W. Ivie);
Sabino Canyon, Pima Co., Arizona, Sept.
1952, 6 ; San Diego Co., California (sev-
eral records) (Map 2).

Tetragnatha extensa (Linnaeus)
Plate 5e; Figures 56-64; Map 2

Aranea extensa Linnaeus, 1758, Systema Naturae,
10th ed., p. 621. Specimens from Sweden.

Tetragnatha extensa: — Tullgren, 1947, Entomol.
Tidskr., 68: 130, figs. 1, 2, 7, 8, 16, 17, 19, 20, 9,
6 . Locket and Millidge, 1953, British Spiders, 2:
100, figs. 64a, b, 66a, 67a, 68a, 69b, 9, d. Wiehle,
1963, Tierwelt Deutschlands, 49: 12, fig. 5, web,
figs. 13-19, 9,6, egg-sac. Bonnet, 1959, Biblio-
graphia Araneorum, 2: 4323 (in part only).

Tetragnatha manitoha Chamberlin and Ivie, 1942,
Bull. Univ. Utah, biol. ser., 7(1): 61, figs. 153-158,
9,3. Male holotype from Churchill, Manitoba in
the American Museum of Natural History, ex-
amined.

Tetragnatha rusticana Chickering, 1959, Bull. Mus.
Comp. Zool., 119: 489, figs. .36-40, 9,6. Male
holotype from Ba\ County, Michigan in the Mu-
seum of Comparative Zoologv, examined. NEW
SYNONYMY.

Note. Most American citations to the

name T. extensa are misidentifications
and actually refer to T. versicolor.

Variation. The female fang may have
a cusp on its base (Fig. 57), or this may
be absent. The distal tooth on the poste-
rior margin of the chelicerae of females
may be asymmetrical or may have a cusp.
There may be a diastema behind this
tooth, or this may be absent. One female
examined from Colorado had one on one
side, not on the other. Total length of fe-
males, 5.8 to 8.7 mm; males, 4.3 to 7.6
mm.

Diagnosis. The lateral eyes are almost
as far apart as the medians (Figs. 57, 62).
The male has the chelicerae slightly
shorter than the length of the carapace
(Fig. 61), the female considerablv shorter
(Fig. 56).

The male can readily be separated
from T. versicolor, T. shoshone, and oth-
er North American species by the tip of
the conductor, which is sclerotized lat-
erally with the tiny tip pointed laterally
toward the paracymbium (Plate 5e; Figs.
63, 64).

The female generally has a black ster-
num. The dorsum of the abdomen is sil-
very; the venter of the abdomen has a
black band bordered on each side by a
narrower, bright silver band. The poste-
rior seminal receptacles of each pair are
longer than the anterior and in a trans-
verse position (Figs. 58, 60).

Natural History. This species has been
collected sweeping the understory of as-
pen woods in British Columbia and Col-
orado, and also in wet meadows in Mon-

Figures 16-22. Morphology of Tetragnatha laboriosa. 16. Sternum, labium, coxae, and endites of male. 17. Left femora
of female, legs one and four, showing trichobothria, prolateral. 18-22. Male. 18-20. Left palpus, expanded. 18. Cymbium
and paracymbium, lateral. 19. Subventral. 20. Mesal. 21. Posterior median eyes. 22. Left lateral eyes.

Figures 23-34. Tetragnattia nitens (Audouin). 23-29. Female. 23. Lateral, without legs. 24. Chelicerae and eye region.
25. Left chelicera from below. 26-29. Genital area. 26. Ventral. 27. Dorsal. 28. Ventral, cleared. 29. Left seminal recepta-
cle, ventral. 30-34. Male. 30. Lateral. 31. Chelicerae and eye region. 32-34. Left palpus. 32. Ventral. 33. Lateral. 34. Con-
ductor and embolus, ventral, and conductor tip, mesal.

Scale lines. 0.1 mm; except Figs. 16, 17, 23-25, 30, 31, 1.0 mm.

Abbreviations. C, conductor; E, embolus; h, hematodocha; P, paracymbium; T, tegulum; Y, cymbium.

DoLicHOGNATHA AND Tetbagnatha North OF MEXICO • Levi 299

300 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

tana, Wyoming, and Colorado. It occurs
up to timber! ine in the Colorado Rocky
Mountains, about 3000 m elevation.

Distribution. Circumboreal, south to
Maine, New Hampshire, New York,
Michigan; in Rocky Mountains south to
Chiricahua Mts., in the West to Idaho and
Washington (Map 2). The southernmost
record is from Herb Martyr Dam, Chiri-
cahua Mts., Cochise Co., Arizona, 9 , 1
Sept. 1956 (A. F. Archer).

Tetragnatha dearmata Thorell
Plate 5f, Figures 65-73; Map 2

Tetragnatha dearmata Thorell, 1873, Remarks on
Synonyms of European Spiders, p. 462. Speci-
mens from Sweden. Tullgren, 1947, Entomol.
Tidskr., 68: 139, figs. 32-44, 9, S. Wiehle, 1963,
Tierwelt Deutschlands, 49: 41, figs. 65-74, 9,
S . Bonnet, 1959, Bibliographia Araneorum, 2:
4321. Holm, 1973, Zool. Scripta, 2: 98.

Tetragnatha harrodi Levi, 1951, Amer. Mus. Nov-
itates, no. 1501: 17, figs. 32-37, 6. Male holotype
from Harrowell, Prince Edward County, Ontario
in the American Museum of Natural History.
Chickering, 1959, Bull. Mus. Comp. Zool., 119:
484, figs. 22-25, 9,6. NEW SYNONYMY.

Variation. Total length of females, 5.2
to 10.0 mm; total length of males, 5.9 to
6.8 mm.

Diagnosis. The lateral eyes are always
much closer together than the medians
(Figs. 66, 71).

Male Tetragnatha dearmata can be
confused with T. elongata. However, the
chelicerae of T. dearmata are shorter
than the carapace (Fig. 70); T. dearmata
lacks the first large (main) tooth on the
anterior margin, and the tooth closest to
the base of the fang is larger (Fig. 71).
The similar European T. montana male
has a large main tooth. The tibia of the
palpus is shorter than the cymbium (Fig.
72). Unlike that of T. elongata, the mid-
dle of the conductor appears to have a
thin fold; there is no basal, lateral, trans-
parent shield on the conductor; and the
tip of the conductor is a pennant, bent
about 45° (Plate 5f; Figs. 72, 73).

The female is separated from T. elon-
gata by the smaller chelicerae with few-
er teeth (Figs. 65, 66), and from all Tet-

ragnatha species by the doubled-up,
U-shaped posterior lateral seminal recep-
tacles (Fig. 69). Both margins of the silver
streak on sides of the abdomen are wavy
(Fig. 65).

Natural History. Specimens have been
collected from white spruce in Northwest
Territories; trees in New Brunswick;
shrubs, tall grasses in a swamp, sweeping
grass and forest, and a cottage in Ontario.

Distribution. Northern Eurasia, Alaska
to northern states (Map 2). The southern-
most localities are Trenton Falls, New
York; Point Comfort, Winnebago Co.,
Wisconsin; Saint Peters, Nicollet Co.,
Minnesota; and near Forked Lake, Ea-
gletown, McCurtain Co., Oklahoma, 6
without date (Map 2).

Tetragnatha elongata Walckenaer
Plates 5g-i; Figures 74-89; Map 3

Tetragnatha elongata Walckenaer, 1805, Tableau
des Araneides, p. 69. Name for Bosc manuscript
fig. 5, 9 , "Aranea gibba" from the Carolinas. Pho-
tocopy in the Museum of Comparative Zoology,
examined. Male neotype from Raleigh, North
Carolina in garden, 21 to 31 August 1944 (C. S.
Brimley) in the Museum of Comparative Zoology,
here designated. Seeley, 1928, Bull. New York
State Mus., 278: 109, pk 1, figs. 11-13, pi. 2, figs.
14-16, 9,6. Comstock, 1940, Spider Book, rev.
ed., p. 425, figs. 424, 425, 9,6. Roewer, 1942,
Katalog der Araneae, 1: 992. Kaston, 1948, Bull.
Connecticut Geol. Nattir. Hist. Surv., 70: 270,
figs. 853, 854, 865, 867, 9,6. Bonnet, 1959, Bib-
liographia Araneorum, 2: 4322. Chickering, 1959,
Bull. Mus. Comp. Zool., 119: 480, figs. 9-12, 9,
6.

Tetragnatha grallator Hentz, 1850, J. Boston Soc.
Natur. Hist., 6: 26, pi. 4, figs. 1,2,9,6. Syntypes
from Pennsylvania, North and South Carolina,
Alabama, destroyed.

Variation. Total length of females, 8.2
to 13.2 mm; of males, 4.8 to 10.5 mm. The
smallest individuals are from the north-
ern states, the largest from Texas. The
tibia of the palpus is usually longer than
the cymbium, occasionally shorter. Che-
licerae from large and small specimens
are illustrated by Figures 83-89.

Diagnosis. This species has the lateral
eyes closer together than the medians,
and both sexes have long chelicerae
(Figs. 74, 75, 79, 80).

DoLiCHOGNATHA AND Tetragnatha North OF MEXICO • Levi 301

Figures 35-45. Tetragnatha earmra n. sp. 35-39. Female. 35. Lateral. 36. Chelicerae and eye region. 37-39. Genital
area. 37. Dorsal. 38. Ventral, cleared. 39. Left seminal receptacles, ventral. 40-45. Male. 40. Lateral. 41. Carapace and
chelicerae. 42. Chelicerae and eye region. 43^5. Left palpus. 43. Ventral. 44. Lateral. 45. Conductor, ventral, and
conductor tip, mesal.

Figures 46-55. Tetragnatha guatemalensis O. P.-Cambridge. 46-50. Female. 46. Lateral. 47. Chelicerae and eye region.
48-50. Genital area. 48. Dorsal view. 49. Ventral, cleared. 50. Left seminal receptacles, ventral. 51-55. Male. 51. Lateral.
52. Chelicerae and eye region. 53-55. Palpus. 53. Ventral. 54. Lateral. 55. Conductor, ventral, and tip of conductor,
mesal.

Scale lines. 0.1 mm; except Figs. 35, 40, 46, 47, 51, 52, 1.0 mm.

Abbreviations. L, pleats of conductor (Fig. 45).

302 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

The male is easily recognized by the
conductor, which tapers evenly to its tip
(Fig. 82) and tenninates in a minute hook
(Plate 5i). The distal third of" the conduc-
tor is convex laterally, especially notice-
able when viewed from slightly ventro-
laterally. The tip bends toward the
cymbium. There are usually three, some-
times two, pleats, and the base of the con-
ductor has a transparent shield (Figs. 82,
88, 89). The conductor tip (Plate 5g) is
much smaller than that of T. straminea.
The palpus in ventral view can be con-
fused with that of T. dearmata, but it
lacks the dorsal extension of the tip. Un-
like Tetragnatha versicolor, the male
chelicerae are almost always longer than
the carapace (Figs. 79, 84-87), and the
recui"ved hook at the tip of the conductor
is minute (Plate 5g-i; Fig. 82).

The female can readily be separated
from other species: the chelicerae usu-
ally are at least equal to or greater than
carapace length, usually with a lateral
cusp outside near base of fang (Figs. 74,
75, 83). Unlike all other North American
species, the median of each pair of sem-
inal receptacles is longer than twice the
width, and placed paraxial in the abdo-
men, parallel to each other (Figs. 76-78).

Natural History. The large, nearly hor-
izontal (sometimes vertical) webs are
placed over small streams or other run-
ning water, often in woods or shaded
branches. The web has 4 to 5 turns out-
side of the hub, a free space, and 30 to 40
spirals.

Distribution. Eastern North American
from New Brunswick, Quebec, Manitoba
south to Cuba, Jamaica, central Mexico.
Uncommon in Western states. Western
records are: Salmon Ann and Wellington,
British Columbia; northeast of Fruitland
and Notus, Idaho; in Washington, Silver
Lake, Cowlitz Co., Chase Lake, King Co.,
and Seattle; Eugene, Philomath, and
Corvallis, Oregon; Monterey, California;
and Sierra Laguna, Baja California (Map
3).

Tetragnatha versicolor Walckenaer
Plates 3, 6a-f; Figures 90-109; Map 3

Tetragnatha versicolor Walckenaer, 1841, Histoire
Naturelle des Insectes Apteres, 2:215. Name giv-
en to Abbott's illustration of Georgian spiders, p.
20, fig. 231, ? and p. 37, fig. 466, 9. Photocopy
in the Museum of Comparative Zoology, exam-
ined. Male neotype from Bar-M Ranch, south of
Boston, Thomas Co., Georgia, 4 April 1973 (W. T.
Sedgwick) in the Museum of Comparative Zool-
ogy here designated. Kaston, 1948, Bull. Con-
necticut Geol. Natur. Hist. Surv., 70: 270, figs.
852, 862-864, 9,6. Chickering, 1959, Bull. Mus.
Comp. Zool., 119: 497, figs. 53-57, 9,6.

Tetragnatha convexa Banks, 1898, Proc. California
Acad. Sci., ser. 3, 1: 247. "Several specimens from
San Jose del Cabo " Baja California in the Cali-
fornia Academy of Sciences, destroyed in the Cal-
ifornia earthquake of 1906.

Tetragnatha extensa: — Seelev, 1928, Bull. New
York State Mus., 278: 113, figs. 17-20, 9, d. Corn-
stock, 1940, Spider Book, rev. ed., p. 425. Bonnet,
1959, Bibliographia Araneonmi, 2: 4323 (in part
only). Not T. extensa Linnaeus.

Tetragnatha limnocharis Seelev, 1928, Bull. New
York State Mus., 278: 129, figs'. .32-35, 9,6. Male
lectotype here designated from Nixon's Ham-
mock, Okefenokee Swamp, Georgia in the Amer-
ican Museum of Natural Historv, examined. NEW
SYNONYMY.

Tetragnatha munda Chamberlin and Gertsch, 1929,
J. Ent. Zool., 21: 103, pi. 4, figs. 41-45, 9,6. Male
holotype from Clear Lake, California, in the
American Museum of Natural Historv, examined.
NEW SYNONYMY.

Tetragnatha marianna Archer, 1940, Paper Ala-
bama Mus. Natur. Hist, 14: 20, pi. 1, fig. 1, 6.
Male holotype from Randon's Creek, Monroe Co.,
Alabama in the American Museum of Natural
History, examined. NEW SYNONYMY.

Note. The Museum of Comparative Zo-
ology collection has one male of T. con-
vexa from Sierra Laguna, Baja California,
N. Banks collection; the labels are in
A.M. Chickering's handwriting. I am un-
certain if this is a syntype of T. convexa.

Variation. This is a variable species.
The abdominal patterns vary from com-
pletely silver to contrasting paired dorsal
markings (Fig. 90). The base of the con-
ductor of the palpus has at times a dis-
tinct transverse fold in the middle (Figs.
98, 104), but often this break is not visi-
ble. The tip of the conductor in western
and northern specimens is sclerotized and

DOLICHOGNATHA AND TeTRAGNATHA NORTH OF MEXICO • Levi

303

Figures 56-64. Tetragnatha extensa (Linnaeus). 56-60. Female. 56. Lateral. 57. Chelicerae and eye region. 58-60. Genital
area. 58. Dorsal. 59. Ventral, cleared. 60. Left seminal receptacle, ventral. 61-64. Male. 61. Lateral. 62. Chelicerae and
eye region. 63, 64. Left palpus. 63. Ventral. 64. Conductor, ventral, and tip, mesal.

Figures 65-73. Tetragnatha dearmata Thorell. 65-69. Female. 65. Lateral. 66. Chelicerae and eye region. 67-69. Genital
area. 67. Dorsal. 68. Ventral, cleared. 69. Left seminal receptacles, ventrolateral and lateral. 70-73. Male. 70. Lateral. 71.
Chelicerae and eye region. 72, 73. Palpus. 72. Ventral. 73. Conductor, ventral, and tip, mesal.

Scale lines. 0.1 mm; except Figs. 56, 57, 61, 62, 65, 66, 70, 71, 1.0 mm.

304 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

dark (Fig. 98); in southeastern specimens hers of each pair are less than their di-

it is usually white (Figs. 104-109). Those ameter apart (Figs. 92-94). The placement

with light tips for a while were consid- and shape are different from those of oth-

ered a separate species since there were er Tetragnatha species with lateral eyes

also differences in shape. But in exam- closer together than medians,

ining a light-tipped specimen and a dark- Natural History. Specimens have been

tipped one collected together from On- collected from mixed coniferous forests;

tario, and a few dark-tipped ones from the spruce forests; flood plain forests, decid-

South, I fovuid that the differences were uous trees; grasses along edges of fields,

in degree of sclerotization; there are no always near water; and various species of

consistent differences in shape (Plates trees and shrubs, including aspens in

6a-f). Wyoming, oranges in Florida, and

The tibia of the palpus is often longer sweeping sagebrush {Artemia sp.) in

than the cymbium in southeastern spec- Colorado.

imens; others from the same area have Distribution. Alaska to Nicaragua and

the tibia shorter than the cymbium. Cuba (Map 3).

One collection from Centreville, Wil-

l^'f'lo.^.T'^^ Mississippi, January to jetragnatha viridis Walckenaer

July 1944 (A. F. Archer) has the lateral pj^^^g 3 g ^. iio-119; Map 3

eyes separated by the same distance as => ^ < f

the medians. Tetragnatha viridis Walckenaer, 1841, Histoire Na-

Total length of females, 5.4 to 13.3 mm; turelle des Insectes Apteres, 2:216. Name given

total length of males, 4.3 to 9.2 mm. The to fig. 236 9 p. 20 and fig. 471, 9 p 37 of Abott's

, " . ^ r^ T- r • manuscript 01 Georgian spiders, rhotocopies m the

largest specmiens are from Calllornia, Museum of Comparative Zoology, examined,

some of the smallest from the Everglades Male neotype from Brier Creek, 7 mi. north of

region in Florida. Sylvania, Screven County, Georgia, 12 April 1943

The neotvpe and the original speci- l^" I^^^) J" the American Museum of Nahiral

■11 i. i. J u A 1 1 1. r History, here designated. Kaston, 1948, Bull.

mens illustrated by Abbot come from Connecticut Geol. Natur. Hist. Surv., 70: 272.

Georgia, where males have a lightly Tetragnatha pinicola Emerton, 1915, Trans. Con-

Sclerotized conductor tip. necticut Acad. Sci., 20: 139, figs. 7a, b, 9, (?. One

Diagnosis. This species has the lateral ™^lf ' ^Z' ^"^'"tlf T^^'^""' from Nantucket, Mas-

1 , 1 .1 .1 .1 1- sachusetts, in the Museum 01 Comparative Zo-

eyes closer to each other than the medi- ^i^g,,^ e.xamined. Not T. pinicola L. Koch, 1870.

ans (Figs. 91, 96), and some dusky mark- Tetragnatha pinea Seeley, 1928, Bull. New York

ings on cephalothorax and abdomen (Fig. State Mus., 278: 133, figs. 21-24, 9,6. New name

QQ^ for T. pinicola Emerton preoccupied by L. Koch.

rr-i 1 1 J. J r .1 Bonnet, 1959, Bibliographia Araneorum, 2; 4351.
Ihe male can be separated from other

species by the recurved tip of the con- Note. Chickering detennined T. viridis

ductor, pointing proximally; there is a specimens in collections as T. versicolor.

slight swelling at the base of the tip Variation. Some specimens are all

(Plate 6a-f; Fig. 98). The male can be green, some afe partly reddish. A green

separated from T. elongata by the chelic- juvenile from Massachusetts turned into

erae, usually shorter than the carapace an adult male after two molts (in a vial),

length (Figs. 95, 96). Males have been with carapace greenish on sides; sternum

confused with T. viridis. Tetragnatha reddish; legs light brown with red spots;

versicolor is never green, as is T. viridis; dorsum of abdomen with a wide longi-

the lateral eyes in T. versicolor are closer tudinal red band having parallel white

together than the medians, in T. viridis sides; sides of abdomen dark green; ven-

as far apart as the medians. ter silvery.

The female has two seminal recepta- The green and red color fade rapidly in

cles on each side, separated; the mem- alcohol preserving fluid. Total length of

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 305

Figures 74-89. Tetragnatha elongata Walckenaer. 74-78. Female. 74. Lateral. 75. Chelicerae and eye region. 76-78.
Genital area. 76. Dorsal. 77. Ventral, cleared. 78. Left seminal receptacles, ventral. 79-82, Male. 79. Lateral. 80. Chelicerae
and eye region. 81 , 82. Left palpus. 82. Conductor and embolus, ventral, and conductor tip, mesal. 83-89. Variation. 83.
Carapace and left chelicera of two adult females collected together (Massachusetts). 84-87. Male carapace, left chelicera
and left palpus. 84. (Massachusetts). 85. (Wisconsin). 86. (Washington state). 87. (Texas). 88, 89. Conductor, ventral. 88.
(Wisconsin). 89. (Texas).

Scale lines. 0.1 mm; except Figs. 74, 75, 79, 80, 83-87, 1.0 mm.

306 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

females, 5.7 to 7.4 nun; total length of
males, 4.4 to 6.7 mm.

Diagnosis. Faded, preserved speci-
mens look like T. lahoriosa but have gen-
italia like those of T. versicolor. The lat-
eral eyes are about as far apart as the
medians (Figs. Ill, 117).

Specimens can be distinguished from
T. lahoriosa by the green or reddish col-
or; in males by the similarity of the palpal
conductor tip to that of T. versicolor
(Plate 6g; Fig. 119); and in females by the
seminal receptacles (Fig. 114).

This species can easily be distin-
guished from T. versicolor, which has
similar genitalia, by the separated lateral
eyes (Figs. Ill, 117). These are farther
apart than the medians in the male, as far
apart as the medians in the female. Other
distinguishing features are the green to
red coloration, and the abdominal mark-
ings: bands with straight margins (Fig.
110) rather than the lobed gray markings
on the dorsum of T. versicolor. T. viridis
also differs from T. lahoriosa and T. ver-
sicolor by the long macrosetae on the
legs; those of the first tibia have a length
5 to 8 times the diameter of the tibia (Fig.
115).

The green coloration and long leg mac-
rosetae make it possible to detennine ju-
veniles.

The only differences that could be
found in the conductors of T. viridis and
T. versicolor are in the pleats near the
base at the lateral end: in T. viridis they
look as though lifted up (Fig. 119).

Natural History. This species lives in
pines, occasionally in other conifers, and
is difficult to find since it rests appressed

to the needles. It is also difficult to dis-
lodge by sweeping or beating. In Mas-
sachusetts the species is common in pitch
pine {Pinus rigida) in sandy areas of
Cape Cod. In Massachusetts adults can
be found from late May to early July. Ju-
veniles are found later in the season, and
they overwinter. This short season is un-
usual for Tetragnatha.

Distrihution. New Bnmswick to North
Carolina along coastal pine forest, south-
eastern states to northern Florida, eastern
Texas (Map 3).

Records. New Brunswick. Fredericton,
6-9 July 1970, 9 in balsam fir (T. R. Re-
nault). Massachusetts. Barnstable Co.:
Chatham, South Chatham, Harwich, be-
tween Falmouth and Bourne, many spec-
imens. Essex Co.: Essex 24 June 1878,
pine trees, 9, 6 (J. H. Emerton). Con-
necticut. New Haven Co.: Mt. Camiel, 12
Sept. 1936, juv. (B. J. Kaston). Neiv Jer-
sey. Ocean Co.: Lakehurst, 13 May 1909,
9,6. North Carolina. Pasquotank Co.:
Elizabeth City, May 1900, 6 (J. H. Emer-
ton). Georgia. Thomas Co.: Bar-M Ranch,
28 April 1973, 9 (G. B. Edwards). Flori-
da. Jackson Co.: 30 mi. southeast of Mar-
ianna, 1 May 1933, 9 (H. K. Wallace). Jef-
ferson Co.: Monticello, 11 April 1968, 9
(A. M. Chickering). Libertv Co.: Rock
Bluff, 4 April 1927, 6 (C. R. Crosby). Al-
ahama. Lee Co.: Chewacla Creek State
Park, April 1947, S (A. F. Archer). Tus-
caloosa Co.: Tuscaloosa, 13 April 1941,
6 (A. F. Archer). Arkansas. Calhoun Co.:
5 Nov. 1964, juv. on pine (L. O. Warren).
Texas. Harris Co.: 9 mi. west of Tomball,
5 Dec. 1968, juv. in freshlv felled pine
(E. Sabath).

Figures 90-109. Tetragnatha versicolor Walckenaer. 90-94. Female (California except 94, Michigan). 90. Lateral. 91.
Chelicerae and eye region. 92-94. Genital area. 92. Dorsal. 93, Ventral, cleared. 94. Left seminal receptacles, ventral.
95-98. Male (California). 95. Lateral. 96. Chelicerae and eye region. 97, 98. Left palpus. 97. Ventral. 98. Conductor,
ventral, and tip, mesal. 99-109. Variation. 99, 100. Female (Florida). 99. Lateral. 100. Chelicerae and eye region. 101,
102. Male (Georgia). 101. Lateral. 102. Chelicerae and eye region. 103, 104. Palpus (Georgia). 103. Ventral. 104. Con-
ductor, ventral, and tip, mesal. 105, 106. Conductor tips, ventral and mesal (both Oxford Mills, Ontario). 107-109. Con-
ductor tips, mesal. 107. (New Brunswick). 108. (Neel Gap, Georgia). 109. (California).

Scale lines. 0.1 mm; except Figs. 90, 91, 95, 96, 99-102, 1.0 mm.

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 307

308

Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

Tetragnatha laboriosa Hentz

Plate 6h, i; Figures 16-22, 120-128;
Map 3

Tetragnatha ]a])oriosa Hentz, 1850, J. Boston Soc.
Natiir. Hist., 6: 27, pi. 4, fig. 3, 6 . Specimens from
United States, destroyed. Male neotype from
Holliston, Massachusetts, 23 June 1929 (N.
Banks) in the Museum of Comparative Zoology
here designated. Seeley, 1928, Bull. New York
State Mus., 278: 123, figs. 25-31, 9, d. Comstock,
1940, Spider Book, rev. ed., p. 426. Roewer, 1942,
Katalog der Araneae, 1: 992. Kaston, 1948, Bull.
Connecticut Ceol. Natur. Hist. Surv., 70: 269,
figs. 850-851, 859-861, 9, d. Chickering, 1959,
Bull. Mus. Comp. Zool., 119: 486, figs. 26-30,
9,6. Bonnet, 1959, Bibliographia Araneorum, 2:
4335.

Tetragnatha illinoiensis Keyserling, 1879, Ver-
handl. Zool. Bot. Gesell. Wien, 29: 318, pi. 4, fig.
18, 9 . Female from Illinois in the British Mu-
seum, Natural History, examined.

Tetragnatha numa Levi, 1955, Canadian Field Nat-
ur., 69: 37, figs. 19-23, 9 , d . Male holotype from
Bowman Lake, Glacier National Park, Montana
in the American Museum of Natural History, ex-
amined. NEW SYNONYMY.

Variation. The lateral eyes are usually
as far apart as the medians, sometimes
slightly closer together or very slightly
farther apart. Occasionally there is a dark
folium on the dorsum of the abdomen;
the folium is usually more distinct pos-
teriorly than anteriorly. Total length of
females, 5.2 to 9.0 mm; total length of
males, 3.8 to 7.4 mm.

Diagnosis. The lateral eyes are almost
as far apart or as separate as the medians
(Figs. 121, 126). Similar eye distances are
also found in T. extensa, T. shoshone, and
the green T. viridis. There is very little
gray pigment on the body.

The male differs from other species by
the bird head-shaped tip of the conductor
(Plate 6i; Fig. 128); the tip is often slight-
ly turned to point dorsally to the cym-
bium. The conductor is twisted near mid-
way, and there are about two pleats, the
proximal pleat lifted up away from the
conductor (Plate 6h; Figs. 127, 128).

The female can only be separated from
similar species by the seminal recepta-
cles. Each pair has a large pouch be-
tween the anterior and posterior recep-

tacles which partly hides the anterior
receptacle in ventral view (Fig. 124).
This is unlike that of any other species.

Natural History. Specimens have been
found in grass, often dry fields some dis-
tance from water; in soybean fields in
North Carolina; in cotton, sweet potato,
and alfalfa fields in Arkansas; in tomatoes
in California; in pasture in Texas; in high
meadows to 3000 m elevation in Colo-
rado; and in bog in New Brunswick. It
has rarely been found in trees.

Distribution. Alaska to Panama (Map
3).

Tetragnatha pallescens F. P.-Cam-
bridge
Plates 4, 7a; Figures 129-139; Map 4

Eugnatha pallida Banks, 1892, Proc. Acad. Natur.
Sci. Philadelphia, p. 51, fig. 88, 6 . Two male and
one female syntypes from Ithaca, New York in
the Museum of Comparative Zoology, examined.
Not T. pallida O. P.-Cambridge, 1889.

Tetragnatha pallescens F. P.-Cambridge, 1903,
Biologia Centrali-Americana, Araneidea, 2: 436.
New name for T. pallida Banks preoccupied by
T. pallida O. P.-Cambridge. Seeley, 1928, Bull.
New York State Mus., 278: 131, pi. 3, figs. 40-43,
9 , d . Comstock, 1940, Spider Book, rev. ed., p.
429. Roewer, 1942, Katalog der Araneae, 1:992.
Kaston, 1948, Bull. Connecticut Geol. Natur.
Hist. Surv., 70: 272, figs. 875-876, 9, d. Chick-
ering, 1959, Bull. Mus. Comp. Zool., 119:487,
figs. 31-35, 9, d. Bonnet, 1959, Bibliographia
Araneorum, 2: 4350.

Eugnatha pallidula Banks, 1910, Bull. U.S. Natl.
Mus., 72, p. 36. New name for T. pallida Banks,
preoccupied by T. pallida O. P.-Cambridge.

Variation. Females may have either
large or small chelicerae (Figs. 134, 135).
Total length of females, 7.3 to 12.2 mm;
males, 6.5 to 9.6 mm.

Diagnosis. The lateral eyes are farther
apart than the medians (Figs. 130, 137).
The abdomen is entirely covered with
minute silvery spots, the spots least
dense on the venter.

The male can be separated from other
species having similar eye spacing by the
conductor of the palpus: the lower sur-
face of the proximal end is concave (Figs.
138, 139), and the tip has a bulge above

DOLICHOGNATHA AND TeTBAGNATHA NORTH OF MEXICO • Levi

309

.^Oe>o»

Figures 110-119. Tetragnatha viridis Emerton. 110-115. Female. 110. Lateral. 111. Chellcerae and eye region. 112-114.
Genital area. 112. Dorsal. 113. Ventral, cleared. 114. Left seminal receptacles, ventral. 115. First left patella and tibia with
macrosetae. 116-119. Male. 116. Lateral. 117. Chelicerae and eye region. 118, 119. Left palpus. 118. Ventral, 119, Con-
ductor, ventral, and tip, mesal.

Figures 120-128. Tetragnatha laboriosa Hentz. 120-124. Female. 120. Lateral. 121. Chelicerae and eye region. 122-124.
Genital area. 122. Dorsal. 123. Ventral, cleared. 124. Left seminal receptacles, ventral. 125-128. Male. 125. Lateral. 126.
Chelicerae and eye region. 127, 128. Palpus. 127. Ventral. 128. Conductor with embolus, ventral, and conductor tip,
mesal.

Scale lines. 0.1 mm; except Figs. 110, 111, 116, 117, 120, 121, 125, 126, 1.0 mm.

310 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

the neck (in ventral view) under the ter-
minal hook (Plate 7a; Fig. 139). There is
one pleat (Fig. 139).

The female is separated from T. cau-
data, T. branda, and T. vermifonnis by
the long, slender shape of the abdomen,
lacking a tail (Fig. 129). Females are sep-
arated from T. straminea by the abdomen
having silver sides (Fig. 129); there is no
distinct line where the silver ends and
the pigmentless area starts, but some-
times the dorsal silver grades into a less
dense ventral silver area. Females can be
confused with those of T. laboriosa, but
T. laboriosa has some gray pigment on
the sternum and venter of the abdomen,
absent in T. pallescens. In the West In-
dies and Central America, where T. .stra-
minea does not occur, Tetragnatha pal-
lescens may have a line on each side,
above which there is silver and below
which there are only scattered silver
spots.

Females from the northern part of the
range can be confused with those of T.
shoshone. The seminal receptacles of
each pair are more than their distance
apart in T. pallescens, with the interme-
diate connection being a lobe (Fig. 133);
in T. shoshone they are almost touching.

Natural History. Specimens have been
collected from pine trees on Nantucket,
Massachusetts, and sandflats and lake
shore in Pennsylvania. In Florida the
species has been found in parlor palm
(Chamaedorea erumpens), mangroves,
tall grasses, roadside weeds, vegetation
along lakeshore and pond in waste fields.
It has also been found in rice in Arkansas,
and grasses in water in Texas.

Distribution. Southern Canada to Pan-
ama, West Indies (Map 4).

Tetragnatha caudata Emerton
Plate 7b, c; Figures 140-148; Map 4

Tetragnatha caudata Emerton, 1884, Trans. Con-
necticut Acad. Sci., 6: 335, pi. 39, figs. 16, 22, ? .
Penultimate instar female holotype in poor phys-
ical condition from Maiden, Massachusetts in the

Museum of Comparative Zoology, examined.
Seeley, 1928, Bull. New York State Mus., 278:
107, figs. 5-10, 9,6. Kaston, 1948, Bull. Con-
necticut Geol. Natur. Hist. Surv., 70: 273, figs.
873, 874, 9 . Chickering, 1959, Bull. Mus. Comp.
Zool., 119: 479, figs. 1-8, 9, d.

Tetragnatha lacerta: — Comstock, 1940, Spider
Book, rev. ed., p. 429, fig. 428, 9 . Roewer, 1942,
Katalog der Araneae, 1 : 993.

Eucta lacerta: — Bonnet, 1956, Bibliographia Ara-
neorum, 2: 1807.

Note. This species has often been re-
ferred to as T. lacerta Walckenaer. Tetra-
gnatha lacerta is a doubtful name and
has also been used for Argyrodes fictil-
ium (Hentz), Theridiidae.

Variation. Tail length is quite variable,
often very much longer than illustrated.
Total length of females, 7.6 to 11.5 mm;
of males, 5.7 to 8.9 mm. The lateral eyes
are farther apart than the medians (Figs.
141-146).

Diagnosis. Both sexes are readily sep-
arated from other Tetragnatha species
(except T. branda) by the presence of a
tail behind the spinnerets (Figs. 140-
145).

Tetragnatha caudata males have pleats
on their conductor, and a hook on the tip
of the conductor (Plate 7b, c; Fig. 148),
both lacking in T. branda. In males the
proximal end of the conductor is twisted
below the pleats (Fig. 148); the distal
hook faces mesally and points dorsally
(Plate 7b; Fig. 148). At the base of the
hook is a small spur seen in ventral view,
absent in other species (Plate 7c; Fig.
147).

In females the seminal receptacles are
relatively small, the anterior medians of
each pair closer together than to the other
member of the pair (Figs. 142-144).
Those of T. branda are close to each oth-
er on each side. The abdomen is entirely
silver; on the dorsum the silver spots are
fused, whereas they are less dense in a
longitudinal line on the venter.

Natural History. Specimens have been
collected from bogs; marshes; weeds
along canal; swamp grass; lakeshores;

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 311

Figures 129-139. Tetragnatha pallescens F. P.-Cambridge. 129-135. Female. 129. Lateral. 130. Cheliceraeand eye region.
131-133. Genital area. 131. Dorsal. 132. Ventral, cleared. 133. Left seminal receptacles, ventral. 134, 135. Carapace and
chelicera of mature females (Tamiami Trail, Dade Co., Florida). 136-139. Male. 136. Lateral. 137. Chelicerae and eye
region. 138, 139. Left palpus. 138. Ventral. 139. Conductor, ventral, and tip, mesal.

Figures 140-148. Tetragnatha caudate Emerton. 140-144. Female. 140. Lateral. 141. Chelicerae and eye region. 142-
144. Genital area. 142. Dorsal. 143. Ventral, cleared. 144. Left seminal receptacles, ventral. 145-148. Male. 145. Lateral.
146. Chelicerae and eye region. 147, 148, Palpus. 147. Ventral. 148. Conductor, ventral, and tip, mesal.

Scale lines. 0.1 mm; except Figs. 129, 130, 134-137, 140, 141, 145, 146, 1.0 mm.

312 BiiUctiu Museum of Comparative Zoology, Vol. 149, No. 5

and ponds. In the Everglades the species
has been fonnd in tall grass.

Distribution. From Great Slave Lake
to Panama, Cuba; absent from south-
western United States (Map 4). The
northernmost locality are juvenile speci-
mens from 5 mi. SE of Fort Providence,
15 Aug. 1965 (J. and W. Ivie, AMNH).

Tetragnatha straminea Emerton
Plates 4, 7d; Figures 149-157; Map 4

Tetragnatha struinitiea Emerton, 1884, Trans. Con-
necticut Acad. Sci., 6: 335, pi. 39, figs. 15, 17, 20,
21, 9 , 6 . One male and two female syntypes from
New Haven, Connecticut in the Museum of Com-
parative Zoology, examined. Seeley, 1928, Bull.
New York State Mus., 278: 136, figs. 49-.54, 9,
S. Roewer, 1942, Katalog der Araneae, 1: 994.
Comstock, 1940, Spider Book, rev. ed., p. 428, fig.
427, 9. Kaston, 1948, Bull. Connecticut Geol.
Natur. Hist. Surv., 70: 271, figs. 855-8.56, 868-
871, 9,6. Chickering, 1959, Bull. Mus. Comp.
Zool., 119: 494, figs. 41^6, 9,6. Bonnet, 19.59,
Bibliographia Araneonmi, 2: 4356.

Variation. Total length of females, 7.1
to 12.7 mm; total length of males, 5.7 to
10.1 mm.

Diagnosis. The lateral eyes are much
farther apart than the medians. Only the
dorsum of the abdomen is silver (Figs.
149, 150, 154, 155).

The male of T. straminea can be sep-
arated from the other four species having
similar eye placement by the paddle-
shaped tip of the conductor (Fig. 157).
The conductor has a narrow neck just be-
low the tip (Plate 7d; Figs. 156, 157), un-
like that of the similar T. shoshone.

The females can be separated from T.
caudata, T. hranda, T. vermiformis, T.
shoshone, and T. paUescens by several
features of the abdomen. Unlike T. cau-
data, the long abdomen of T. straminea
lacks a tail. Unlike T. caudata, T. .shosho-
ne, T. vermiformis, and T. paUescens, the
upper, silvery half of each side is clearly
delineated from the ventral half by a
straight line. Also, the seminal recepta-
cles are more variable in position than
those of other species. Although they

might be confused with those of T. pal-
lescens, in T. straminea these structures
are about their length apart (Fig. 153),
nor is the connection swollen.

Natural History. Specimens have been
collected from sweeping ferns and bogs
in Ontario; sweeping honeysuckle {Lo-
nicera sp.) in West Virginia; cedar bogs
in Ohio; rocky gorges in New York; and
bottomland forests and swamp foliage in
Florida.

Distribution. Northern Alberta, Ontar-
io, Utah, Florida, Cuba (Map 4). The
northernmost collection came from Fort
Murray, Alberta, July 1953, 6 (G. E.
Ball).

Tetragnatha shoshone new species
Plates 7e, f; Figures 158-166; Map 4

Holoti/pe. Male from Laketown, Rich County, Utah,
30 July 1978 (G. F. Knowlton) in the Museum of
Comparative Zoology. The name is a noun in ap-
position after the Indian tribe.

Description. Female. Carapace, ster-
num, legs yellowish. Abdomen evenly
covered with tiny silver pigment spots all
around, except for cardiac area and a nar-
row ventral band to spinnerets (Fig. 158).
Eyes subequal in size, except for anterior
laterals, 0.8 diameters of others. Laterals
very slightly farther apart than medians
(Fig. 159). Total length, 9.4 mm. Cara-
pace, 2.6 mm long, 1.7 mm wide. First
femur, 5.2 mm; patella and tibia, 6.0 mm;
metatarsus, 5.2 mm, tarsus, 1.4 mm. Sec-
ond patella and tibia, 3.9 mm; third, 1.9
mm; fourth, 4.0 mm.

Male coloration and eyes like female.
Chelicerae as illustrated (Fig. 164). Total
length, 6.8 mm. Carapace, 2.5 mm long,
1.4 mm wide. First femur, 5.8 mm; pa-
tella and tibia, 7.0 mm; metatarsus, 6.5
mm; tarsus, 1.8 mm. Second patella and
tibia, 4.8 mm; third, 2.2 mm; fourth, 4.5
mm.

Variation. Canadian specimens have
black pigment on the abdomen: those
from Great Slave Lake, two pairs of black

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 313

Figures 149-157. Tetragnatha straminea Emerton. 149-153. Female. 149. Lateral. 150. Chelicerae and eye region. 151-
153. Genital area. 151. Dorsal. 152. Ventral, cleared. 153. Left seminal receptacles, ventral. 154-157. Male. 154. Lateral.
155. Chelicerae and eye region. 156, 157. Left palpus. 156. Ventral. 157. Conductor with embolus, ventral, and tip, mesal.

Figures 158-166. Tetragnatha shoshone n. sp. 158-162. Female. 158. Lateral. 159. Chelicerae and eye region. 160-162.
Genital area. 160. Dorsal. 161. Ventral, cleared. 162. Left seminal receptacles, ventral. 163-166. Male. 163. Lateral. 164.
Chelicerae and eye region. 165, 166. Palpus. 165. Ventral. 166. Conductor with embolus, ventral, and tip, mesal.

Scale lines. 0.1 mm; except Figs. 149, 150, 154, 155, 158, 159, 163, 164, 1.0 mm.

314 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

patches; those from Ontario, wavy gray Illinois. McHenry Co.: Pistakee Bay, 26

lines on the sides of the dorsum. The lat- Aug. 1968, 9,6 (J. Carico). Nebraska.

eral eyes are a variable distance apart, Scottsbluff Co.: Morrill, on bridge of

always closer than those of T. pallescens Glendo Dam overflow, 30 July 1968, S

and T. straminea. If close, the lateral (A. Moreton). Colorado. Boulder Co.:

eyes are separated about 1.3 times the Valmont, E. of Boulder, 31 Aug. 1939, 6

distance of the medians. (U. Latham). Idaho. Bear Lake Co.: N.

Totallengthoffemales, 8.6 to 12.3 mm; shore of Bear Lake, 30 July 1952, 9 9,

total length of males, 5.4 to 8.8 mm. The 6 6 (B. Malkin); Mud Lake near Hot

smallest specimens are those from Utah Springs, 7 Aug. 1949, 6 6 (W. J. Gertsch).

and Idaho, the largest from the periphery Twin Falls Co.: Thousand Springs, June

of the range: Ontario and California. 1938, 9,6 6 (W. Ivie). Utah. Rich Co.:

Diagnosis. The lateral eyes are always Laketown, 30 July 1978, 9 9, c5 (G. F.

slightly farther apart than the medians Knowlton); Bear Lake, 25 Aug. 1929, 9,

(Figs. 159, 164). 6 ; SE shore Bear Lake, 3 July 1978, 6

Although Tetragnatha shoshone can (G. F. Knowlton); Granite, 26 Aug. 1938,

be confused with T. extensa, T. labor- 6 (G. F. Knowlton). Salt Lake Co.: Salt

iosa, T. pallescens, and T. straminea. Lake City, 10 July 1947, c5 (W. Ivie). Sev-

males can be separated by the shape of ierCo.: Richfield, July-Aug. 1930, 6 (W.

the conductor. T. .shoshone lacks a mes- J. Gertsch). Utah Co.: Utah Lake at

ally-facing beak as conductor tip (Plate mouth of Provo River, 25 June 1941,

7f), present in T. laboriosa and T. palles- 9 9, 6 6 ; W. side of Utah Lake, 24 Sept.

cens. It also lacks the laterally facing, 1939, 9 9, 6 6 (S. Mulaik, W. Ivie).

pointed tip of T. extensa; it has a wider Washington. Grant Co.: Dry Falls, Grand

neck below the conductor tip than does Coulee, 4 May 1940, 99,66 (M. Hatch);

T. straminea (Plate 7e, f; Figs. 165, 166). 20 June 1954, 9 9,6 (B. Malkin). Pend

Females differ from T. extensa and T. Oreille Co.: Diamond Lake, 13 Aug.

laboriosa in having the lateral eyes 1934, 6 (M. Hatch). California. Siskiyou

slightly farther apart than the medians Co.: Tule Lake, 15 July 1940, 9, c^ (L. W.

(Figs. 158, 159). T. shoshone can be dis- Saylor).
tinguished from T. pallescens by the

seminal receptacles: the posterior is larg- Tetragnatha branda new species

er than and almost touching the anterior P'^te 7g, h; Figures 167-175; Map 4

(Fig. 162), whereas those of T. pallescens Uolotupe. Male from sweeping salt marsh, Branford

are some distance from each other and Harbor, Connecticut, July 1975 (C. N. Olmstead)

separated by a lobe of the ducts. in the Museum of Comparative Zoology. The spe-

Natural History. This species has been ^'^'' "'^"'^ '' ^" arbitrary combination of letters,

collected sweeping vegetation near large Description. Female. Carapace, legs

lakes. yellow-white. Sternum, labium dusky.

Distribution. From Great Slave Lake, There is a median gray longitudinal band

Ontario, Indiana to the Pacific coast (Map on the carapace. Abdomen silvery; sides

4). with some black pigment overlying sil-

Paratypes. Northwest Territories: 5 ver, and venter with a ventral black band,

mi. S. of Great Slave Lake, 22 July 1944, Posterior eyes 0.7 diameters of anterior

6 (T. Kurata). Ontario. Goose Island, medians. Anterior laterals 0.5 diameters

Lake Nipissing, 26 July 1931, 9 9 (T. Ku- of anterior medians. Anterior medians are

rata). British Columbia. Victoria, 1931, slightly more than their diameter apart,

6. 1.5 from laterals. Posterior median eyes

Indiana. Steuben Co.: Center Lake, are 3 diameters apart, 2.5 from laterals.

Angola, 23 Aug. 1968, 9,6 (J. Carico). The lateral eyes are much farther apart

DoLicHOGNATHA AND Tetragnatha North OF MEXICO • Levi 315

(^^^

Figures 167-175. Tetragnatha branda n. sp. 167-171. Female. 167. Lateral. 167. Chellcerae and eye region. 169-171.
Genital area. 169. Dorsal. 170. Ventral, cleared. 171. Left seminal receptacles, ventral. 172-175. Male. 172. Lateral. 173.
Chelicerae and eye region. 174, 175. Left palpus. 174. Ventral. 175. Conductor, embolus, ventral, and tip, mesal.

Figures 176-184. Tetragnatha vermiformis Emerton. 176-180. Female. 176. Lateral. 177. Chelicerae and eye region.
178-180. Genital area. 178. Dorsal. 179. Ventral, cleared. 180. Left seminal receptacles, ventral. 181-184. Male. 181.
Lateral. 182. Chelicerae and eye region. 183. Palpus, ventral. 184. Conductor with embolus, ventral, and tip, mesal.

Scale lines. 0.1 mm; except Figs. 167, 168, 172, 173, 176, 177, 181, 182, 1.0 mm.

316 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

than the medians (Fig. 168). The al)cl()-
men is elongate and has a tail overhang-
ing the spinnerets (Fig. 167). Total
length, 8.8 mm. Carapace, 2.0 mm long,
1.4 mm wide. First femur, 3.3 mm; pa-
tella and tibia, 3.7 mm; metatarsus, 2.7
mm; tarsus, 1.2 mm. Second patella and
tibia, 2.7 mm; third, 1.4 mm; fovn-th, 3.1
mm.

Male. Coloration like female, but with
dark pigment. Secondary eyes slightly
larger in relation to anterior medians than
in female. Anterior median eyes slightly
more than their diameter apart, slightly
less than two diameters from laterals.
Posterior medians slightly more than two
diameters apart, the same distance from
laterals (Fig. 173). Total length, 6.8 mm.
Carapace, 2.2 mm long, 1.3 mm wide.
First femur, 3.8 mm; patella and tibia, 4.5
mm; metatarsus, 2.4 mm; tarsus, 1.3 mm.
Second patella and tibia, 3.4 mm; third,
1.7 mm; fourth, 3.7 mm.

Variation. Total length of females, 8.8
to 13.8 mm; total length of males, 6.8 to
11.2 mm. All have a relatively short tail.

Diagnosis. The lateral eyes of this
species are much farther apart than the
medians (Figs. 168, 173). The genitalia
(Figs. 169-171) are most like those of T.
vermiformis, but the species differs by
having a tail, the abdomen overhanging
the spinnerets.

This species can be confused with T.
caudata, but the male palpus lacks con-
ductor pleats, and the conductor has a
distal, lateral notch (Plate 7g, h; Figs.
174, 175).

The female differs from T. caudata by
having adjacent, spherical seminal recep-
tacles (Fig. 171).

Natural History. All specimens have
been collected from coastal salt marshes.

Distribution. New England to Missis-
sippi (Map 4).

Paratopes. Connecticut. New Haven
Co.: Branford Harbor, July 1975, 39 (C.
N. Olmstead, MCZ). Georgia. Mcintosh
Co.: Sapelo Island, 27 July 1957, 9 (E. P.
Odum, MCZ); July-Aug. 1960, 49, 4 c^.

salt marsh (S. Schmittner, AMNH). Flor-
ida. Levy Co.: 12 May 1951, 9, 2(5 (H. K.
Wallace, FSCA); Monroe Co.: Cape Sa-
ble, 4 June, 9 (H. V. Weems, FSCA).
Mississippi. Jackson Co.: Ocean Springs,
January 1905, 9 (J. H. Comstock, CUC).

Tetragnatha vermiformis Emerton
Plate 7i; Figures 176-184; Map 4

Tetragnatha vermiformis Emerton, 1884, Trans.
Connecticut Acad. Sci., 6: 333, pi. 39, figs. 12-14,
9,6. One male leetotype and one female para-
lectotype, plus two female paralectotypes of a dif-
ferent species from Beverly, Massachusetts in the
Museum of Comparative Zoology, examined.
Seelev, 1928, Bull. New York State Mus., 278:
138, figs. 55-58, 9,6. Roewer, 1942, Katalog der
Araneae, 1: 994. Kaston, 1948, Bull. Connecticut
Geol. Natur. Hist. Surv., 70: 272, figs. 877-878,
6. Chickering, 1957, Bull. Mus. Comp. Zool.,
116: 349, figs. 103-105, 6 (not 9); 1959, Bull.
Mus. Comp. Zool., 119: 495, figs. 47-50, 6 (not
9).

Eucta vermiformis: — Bonnet, 1956, Bibliographia
Araneorum, 2: 1808.

Note. The female specimens illustrated
by Chickering had been misidentified.

Variation. Total length of females, 6.7
to 12.3 mm; males, 5.0 to 9.1 mm.

Diagnosis. The lateral eyes are farther
apart than the medians, and the abdomen
is entirely silver (Figs. 176, 177, 181,
182).

Males can readily be distinguished
from the males of the four other species
having similar eye placement by the lack
of pleats on the long conductor of the pal-
pus. The shape of the conductor tip (Fig.
184) separates T. vermiformis from the
similar European T. (Eugnatha) striata.
The absence of a tail distinguishes it from
T. branda (Plate 7i; Figs. 176, 181, 183,
184).

The female has a much shorter, less
womi-shaped abdomen than is seen in
the other species with similar eye place-
ment. The abdomen is entirely silver,
with fewer spots on the venter. Some-
times there are also black marks on the
sides of the abdomen. Unlike T. caudata,
T. pallescens, and T. straminea, the sem-
inal receptacles of each pair are narrow

DoLicHOGXATHA AND Tetragnatha North OF MEXICO • Levi 317

and elongate, placed paraxial, the pair
more than twice their length apart and
more sclerotized laterally than toward
the median (Figs. 178-180).

Natural History. Nothing is known of
the habits of this uncommon species. The
very similar European T. striata is found
on reed grass (Phragmites) of larger lakes
and builds near the open water. It is best
collected by using a boat (Wiehle, 1963).

Distribution. Southern Canada, north-
west to eastern states to Canal Zone (Map

4).

Records. Ontario. La Seine River, July,
9 (N. Banks); Lac Seul, Aug. 1919, 6
(Waugh); Mallorytown Landing, 9 , c5 .

Massachusetts. Falmouth Co.: Woods
Hole, Aug. 1881, 6 (J. H. Emerton). Nan-
tucket, 9 Sept., 9, 6 (J. H. Emerton).
Rhode Island. Kingston, 6 (L. G. Wor-
ley). New York, Long Island: Quogue,
Oct. 1948, 9, (5 (R. Latham). Georgia.
Honev Isl. Prairies, Okefenokee Swamp,
31 May 1912, 9. Florida. Alachua Co.:
Newnan's Lake, Nov. 1931, 9,6 (T. H.
Hubbell). Dade Co.: Tamiami Trail, off
Route 27, 19 March 1968 (A. M. Chick-
ering). Lake Co.: Leesburg, March 1954,
6 (M. Statham). Monroe Co.: Everglades
Nad. Park, Shark River area, 1960, 9.
Michigan. Walnut Lake, 6 (A. M. Chick-
ering). Bay Co.: Aug. 1942, 9 (A. M.
Chickering). Calhoun Co.: Albion, July
1930, 6 (A. M. Chickering). Eaton Co.:
Olivet, Sept. 1933, 6 (A. M. Chickering).
Macomb Co.: Selfridge Field, June 1944,
9, S (B. Malkin). Ogemaw Co.: Aug.
1937, 6 (A. Peacock). Indiana. Steuben
Co.: Angola at Center Lake, 25 Aug.
1968, 6 (J. Carico). Nebraska. Holt Co.:
Goose Lake, 8 Sept. 1977, 6 (W. Rapp).
Washington. Cowlitz Co.: Toutle River,
24 July 1932, 6 (T. Kincaid).

Panama Canal Zone. Barro Colorado
Isl. 6 Feb. 1958, 9, cJ (A. M. Chickering).

LITERATURE CITED

Brignoli, p. 1979. Contribution a la connaissance
des Uloboridae palearctiques (Araneae). Rev.
Arachnol., 2: 275-282.

Bryant, E. B., and A. F. Archer. 1940. Notes on
Epeira pentagona Hentz. Psyche, 47: 60.

Chickering, A. M. 1957a. Notes on certain
species of Tetragnatha (Araneae, Argiopidae)
in Central America and Mexico. Breviora Mus.
Comp. Zool., 67: 1-4.

. 1957b. The genus Tetragnatha (Araneae,

Argiopidae) in Jamaica, B.W.I, and other neigh-
boring islands. Breviora Mus. Comp. Zool., 68:
1-15.

. 1957c. The genus Tetragnatha (Araneae,

Argiopidae) in Panama. Bull. Mus. Comp. Zool,
116: 301-354.

. 1959. The genus Tetragnatha (Araneae,

Argiopidae) in Michigan. Bull. Mus. Comp.
Zool., 119: 475-499.
-. 1962. The genus Tetragnatha (Araneae,

Argiopidae) in Jamaica, W.I. Bull. Mus. Comp.
Zool., 127: 423-450.

Crome, W. 1954. Beschreibung, Morphologic und
Lebensvveise der Eucta kaestneri sp. n. (Ara-
neae, Tetragnathidae). Zool. Jahrb. Abt. Svs-
tem., 82: 425^50.

FoRSTER, R. R. 1955. Spiders of the family Ar-
chaeidae from Australia and New Zealand.
Trans. Roy. Soc. New Zealand, 83: .391-403.

HoMANN, H. 1950. Die Nebenaugen der Araneen.
Zool. Jahrb. Abt. Anat., 71: 1-44.

. 1971. Die Augen der Araneae. Z. Morphol.

Tiere, 69: 201-272.

Legendre, R. 1970. Arachnides, Araignees, Ar-
chaeidae. Faune de Madagascar, Paris, 32: 1-
50.

. 1977. Les Araignees de la famille des Ar-

chaeidae et leur distribution geographique ac-
tuelle. Bull. Biol. France Belgique, 111: 231-
248.

Le Sar, C. D., and J. D. Unzicker. 1978. Life
history, habits and prey preferences of Tetra-
gnatha lahoriosa (Araneae: Tetragnathidae).
Environ. EntomoL, 7: 879-884.

Levi, H. W. 1959. The spider genus Latrodectus
(Araneae, Theridiidae). Trans. Amer. Microsc.
Soc, 78: 7^3.

. 1967. Adaptations of respirator.' systems of

spiders. Evolution, 21: 571-583.

. 1974. The orb-weaver genera AranieUa

and Nuctenea (Araneae: Araneidae). Bull. Mus.
Comp. Zool., 146: 291-316.

. 1980a. The orb-weaver genus Mecijnogea,

the subfamily Metinae and the genera Pachy-
gnatha, Glenognatha and Azilia of the subfam-
ily Tetragnathinae north of Mexico (Araneae:
Araneidae). Bull. Mus. Comp. Zool., 149: 1-
75.
-. 1980b. Orb-webs: primitive or specialized?

Proc. 8th Intern. Arachnol. Congr., Vienna, pp.

367-370.
Locket, G. H., and A. F. Millidge. 1953. British

Spiders. Ray Soc, London, vol. 2.
LuczAK, J., and E. D^browska-Prot. 1966. Ex-

318 Bulletin Museum of Comparative Zoology, Vol. 149, No. 5

perimental studies on the reduction of the
abundance of moscjuitoes by spiders. Bidl.
Acad. Pol. Sci.. CI. 2. 14: 315-320.
-. 1968. Studies on Tctragnatha montana Si-

mon. Bull. Brit. Spider Study Group, 38: 7-9.
LUCZAK, J., AND K. Tarwid. 1970. Influence de

la predation des araignees sur les mousticjues.

Bull. Mus. Nation. d'Hist. Natur., Paris, 2 Ser.,

Suppl. 1, 41: 237-242.
Opell, B. D. 1979. Revision of the genera and

tropical American species of the spider family

Uloboridae. Bull. Mus. Comp. Zool., 148: 443-

549.

Seeley, R. M. 1928. Revision of the spider genus
Tetragnatha. Bull. NY State Mus., 278: 99-
149.

Simon, E. 1894. Histoire Naturelle des Araignees,
1: 489-760.

WiEHLE, H. 1963. Tetragnathidae. Tierwelt
Deutschlands, 49: 1-76.

. 1967. Meta-eine semientelegyne Gattung

der Araneae. Senckb. Biol., 48: 183-196.

WiGGLESWORTH, V. 1972. Insect Respiration. Ox-
ford Biology Readers, Oxford Univ. Press, Lon-
don, p. 15.

INDEX

Valid names are printed in italics. Page numbers refer to main references, starred page numbers to
illustrations.

Afiamalu, 277

andina, Tetragnatha, 291

antillana, Tetragnatha, 291

aptans, Tetragnatha, 292

Arundognatha, 282

banksi, Tetragnatha, 296

branda, Tetragnatha, 297*, 314, 315*

caudata, Tegragnatha, 297*, 310, 311*

convexa, Tetragnatha, 302

dearmata, Tetragnatha, 293*, 300, 303*

Dolichognatha, 277

earmra, Tetragnatha, 293*, 294, 301*

almora, Tetragnatha, 292

elongata, Tetragnatha, 293*, 300, 305*

eremita, Tetragnatha, 292

Eucta, 282

Eugnatha, 282

extensa, Aranea, 298

extensa, Tetragnatha, 302

extei\sa, Tetragnatha, 293*, 298, 303*

festina, Tetragnatha, 292

galapagoensis, Tetragnatha, 291

grallator, Tetragnatha, 300

guatemalensis, Tetragnatha, 293*, 296, 301*

haitensis, Tetragnatha, 292

harrodi, Tetragnatha, 300

illinoiensis, Tetragnatha, 308

intermedia, Tetragnatha, 296

laboriosa, Tetragnatha, 295*, 299*, 308, 309*

lacerta, Eucta, 310

lacerta, Tetragnatha, 310

Landana, 277

laudativa, Tetragnatha, 296

limnocharis, Tetragnatha, 302

manitoba, Tetragnatha, 298

marianna, Tetragnatha, 302

munda, Tetragnatha, 302

Nocholasia, 277

nietneri, Dolichognatha, 277

nitens, Eugnatha, 291

nitens, Tetragnatha, 291, 293*, 299*

numa, Tetragnatha, 308

pallescens, Tetragnatha, 297*, 308, 311*

pallida, Tetragnatha, 308

pallidula, Eugnatha, 308

Paraebius, 277

pelusia, Eugnatha, 291

peninsulana, Tetragnatha, 291

pentagona, Dolichognatha, 281*, 282

pentagona, Epeira, 282

pentagona, Nicholasia, 282

pinea, Tetragnatha, 304

pinicola, Tetragnatha, 304

Prolochus, 277

rusticana, Tetragnatha, 298

seminola, Tetragnatha, 292

seneca, Tetragnatha, 296

shoshone, Tetragnatha, 297*, 312, 313*

steckleri, Tetragnatha, 292

straminea, Tetragnatha, 297*, 312, 313*

Tetragnatha, 282

tuberculata, Cyrtophora, 282

tuberculata, Dolichognatha, 282

vermifonnis, Eucta, 316

vermiformis, Tetragnatha, 297*, 315*, 316

versicolor, Tetragnatha, 295*, 302, 307*

vicina, Tetragnatha, 291

viridis, Tetragnatha, 295*, 304, 309*

(U& ISSN UUdf-'il

uiletin OF THE

\useum of
Comparative
Zoology

! -US. COMP. ZOOU

SEP 1 4 1P81

-HARVARD

A Revision of the Nearctic Beridlnae
(Diptera: Stratiomyidae)

NORMAN E. WOODLEY

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBERS
31 AUGUST 1981

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Breviora 1952-

BULLETIN 1863-

Memoirs 1864-1938

JOHNSONIA, Department of Mollusks, 1941-

OccASiONAL Papers on Mollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. I. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Tere-
dinidae (Mollusca: Bivalvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts, 02138, U.S.A.

© The President and Fellows of Harvard College 1981.

A REVISION OF THE NEARCTIC BERIDINAE
(DIPTERA: STRATIOMYIDAE)

NORMAN E. WOODLEY'

Abstract. The Beridinae (Diptera: Stratiomyi-
dae) from the Nearctic Region are taxonomically re-
vised. Nine species in four genera are recognized.
Keys to all taxa, diagnoses for genera and species,
and descriptions of each species are presented.
Male and female genitalia are illustrated for the first
time, and maps of distributions are included.

Beris strobli Dusek and Rozkosny, Beris fuscipes
Meigen, and Exodontha dubia (Zetterstedt) are
found to occur in the Nearctic Region, and are thus
Holarctic.

The generic name Exodontha Rondani is resur-
rected from synonymy with Antissa Walker, and is
used for the Holarctic species of the "Antissini."
Hemiberis Enderlein is removed from synonymy
with Actina Meigen, and placed in synonymy with
Beris Latreille.

Beris annulifera var. luteipes Johnson is resur-
rected from synonymy with Beris annulifera (Big-
ot), and raised to specific status. Beris californica
James, a junior name, is synonymized with Beris
luteipes Johnson. Allognosta similis (Loew) is syn-
onymized with Allognosta obscuriventris (Loew).
Beris annulifera (Bigot) is synonymized with Beris
fuscipes Meigen. Exodontha grandis (James) is
synonymized with Exodontha dubia (Zetterstedt).

Neotypes are designated for Beris fuscitarsis Say
and Sargus dorsalis Say, the latter a synonym of
Allognosta fuscitarsis (Say). Lectotypes are desig-
nated for Beris quadridentata Walker, Metoponia
similis Loew, and Oplacantha annulifera Bigot.

Biological notes are presented for each species
based on information in the literature and on spec-
imen labels.

INTRODUCTION

The subfamily Beridinae is a group of
primitive Stratiomyidae, as evidenced by

' Museum of Comparative Zoology, Harvard Uni-
versity, Cambridge, Massachusetts 02138.

their seven undifferentiated abdominal
segments, a character shared by the other
subfamily considered primitive, the Chi-
romyzinae. The only species of the latter
subfamily in the Nearctic Region is the
introduced Inopus rubriceps (Macquart),
which occurs in the San Francisco Bay
area of California. This species is easily
recognized by its vestigial mouthparts
and its antennal flagellum which lacks
distinct annuli. The other subfamilies of
Stratiomyidae in North America have an
abdomen composed of five principal seg-
ments, the sixth and seventh being tele-
scoped within the abdomen. Thus the
Beridinae may be easily recognized in
the Nearctic Region. The key to genera
presented by Cole (1969) will also enable
the reader to identify these flies readily
at the generic level.

Melander (1904) provided the first syn-
opsis of the Nearactic Beridinae. Curran
(1927) improved upon this in his treat-
ment of the Canadian Stratiomyidae, in-
corporating new species that had been
described after Melander's paper. James
(1939) again reviewed the subfamily for
the Nearctic Region, and his paper estab-
lished the classification that has been
only slightly modified until now. McFad-
den (1972) synonymized several varieties
described by Johnson (1926a, 1926b) with
their nominate species, and reduced
Beris canadensis (Cresson) to synonymy
with Beris annulifera (Bigot).

Workers on the Palaearctic fauna, no-
tably Rozkosny and his colleagues (Du-

Bull. Mus. Comp. Zool., 149(6): 319-369, August, 1981 319

320 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

sek and Rozkosny, 1963; Rozkosny, 1973;
Nartshuk and Rozkosny, 1975, 1976) and
Naj^atonii and Tanaka (1969, 1972), were
the first to utilize male genitalia exten-
sively to stabilize species concepts. A
number of names were synonymized,
and it became easier to compare the
Nearctic fauna with the Palaearctic fauna.
Subsequently, I examined the genitalia
of the Nearctic species of Beridinae for
the first time, which has resulted in one
synonymy of a Nearctic species with a
Palaearctic one, the discovery of two Pa-
laearctic species previously unrecorded
from the Nearctic Region, and new syn-
onymies within the Nearctic fauna. In ad-
dition, 1 have examined nearly all types
of North American Beridinae (of the
types in existence for names used in
North America, Beris fuscipes Meigen is
the only name for which I have not seen
a type). This paper reports the results of
my study.

Nine species of Beridinae in four gen-
era are now known to occur in North
America, three of which are Holarctic. At
present it is not possible to discuss evo-
lutionary relationships between the gen-
era that inhabit the Nearctic Region; this
must await a worldwide study of generic
concepts within the subfamily. The Ber-
idinae are most diverse in southern South
America, Australia, and New Zealand,
and they remain poorly known in the lat-
ter two regions.

ACKNOWLEDGMENTS

I wish to thank a number of institutions
and their associated curators for loaning
material in their care: University of Al-
berta (Danny Shpeley), University of Ar-
kansas (E. Phil Rouse), Auburn Univer-
sity (Kenneth L. Manuel), California
Academy of Sciences (Paul H. Arnaud,
Jr.), Canadian National Collection (Her-
bert J. Teskey), Cornell University (L. L.
Pechuman), University of Georgia (Cecil
L. Smith), University of Idaho (William
F. Barr), Illinois Natural History Survey

(Donald W. Webb), University of Kansas
(George W. Byers), University of Min-
nesota (Philip J. Clausen), United States
National Museum of Natural Historv
(Wayne N. Mathis and Willis W. Wirthj,
North Dakota State University (E. U.
Balsbaugh, Jr.), Ohio State University
(Charles A. Triplehorn), Oregon Depart-
ment of Agriculture (Richard Westcott),
Oregon State University (Michael D.
Schwartz), Utah State University (Wil-
ford J. Hanson), Washington State Uni-
versity (William J. Turner), and the Uni-
versity of Wyoming (R. J. Lavigne).
Several people kindly helped with the
loan of type specimens: Roy Danielsson,
Brian H. Cogan, Kenneth G. V. Smith,
Curtis Dunn, G. Morge, Maurice T.
James, and Willis W. Wirth. Their help
is very much appreciated.

I wish to thank Maurice T. James,
Wayne N. Mathis, Robert E. Silberglied,
William J. Turner, and Willis W. Wirth
for critically reading and commenting
upon the manuscript. Ronald J. Mc-
Ginley helped in many ways, particularly
with advice concerning preparation of
the illustrations. Rudolf Rozkosny kindly
sent European material for comparisons.
Their help measurably improved the
quality of the manuscript and is grateful-
ly acknowledged.

METHODS

Over 4000 specimens have been ex-
amined in the course of this work. The
genitalia of all known species were ex-
amined and illustrated with the aid of an
ocular grid in a Leitz binocular dissecting
microscope. After being cleared with hot
10% KOH, and neutralized with weak
acetic acid followed by a water rinse, the
parts were stored in plastic microvials
and attached to the pin of the appropriate
specimen.

Measurements were made with an ocu-
lar micrometer. The frontal width was
measured at the anterior ocellus and di-
vided bv the width of the head taken at

Nearctic Beridinae • Woodley 321

the same point to arrive at the ratio pro- tions of the ducts are unsclerotized and

vided in the description. Lengths of the thus nearly invisible in cleared prepara-

entire insects were measured from the tions of the female genitalia. In Beri.sfus-

anterior-most portion of the head to the cipes Meigen and B. luteipes Johnson,

tip of the abdomen, but do not include however, in which the posterior portions

the antennae. of the ducts are sclerotized, they may be

Collection data have been summa- followed all the way to the aperture,

rized, and do not include precise locali- which in both species is quite small. The

ties for the United States localities except median portions of the ducts are unscler-

for species that are uncommon or rare, otized and rather long, leading to the

Common species are summarized by sclerotized anterior portions of the ducts

county only. Since females of the genus and the spermathecae, which are also

Exodontha Rondani are difficult to iden- sclerotized. The spermathecae are very

tify with certainty, the sexes of speci- similar in all species of Beridinae exam-

mens from each locality are indicated, to ined, and are three in number. The illus-

give the reader an idea of which localities trations of the female genitalia present a

are best documented. ventral view. They are most valuable as

Stippling indicates internal surfaces of specific level characters in Beris. Even

sclerotized structures in all drawings ex- though they are distinctive for other

cept Figures 48 and 53, where it indi- species, except in the genus Exodontha,

cates membrane. In some other figures they are not necessary for routine iden-

of the male post-genital segments, areas tification. They are illustrated primarily

around the cerci are more or less mem- to assist future workers who may wish to

braneous, however, though the stippling compare Nearctic species with those of

is meant to indicate the internal surface other regions.

of the eleventh sternum. In describing the male genitalia, I have

used the terminology utilized by Hanson

TAXONOMIC CHARACTERS ^l^^^^ which is evidendy derived from

that proposed by Michener (1944). Im-

Most of the taxonomic characters uti- portant structures include the gonostylus

lized in this study are unambiguous. It ( = dististylus, telomere), the gonocoxites

seems advisable to comment briefly on ( = synsternite, basistyle), and the aedea-

the terms used for the male and female gal complex, which is composed of a me-

genitalia, however, as a variety of terms dian aedeagus ( = penis) and lateral ae-

have been used for the same structures, deagal valves ( = penis valves, parameres).

even within the literature of the Stratio- The gonocoxites have a rather distinctive

myidae. shape in each species, with the posterior

The female genitalia have been found margin of the ventral bridge being partic-
to exhibit useful characters at the specific ularly diagnostic. It is often endowed
level, especially the so-called genital fur- with a median process, which has been
ca. This structure evidently represents called the median process of the synster-
the highly modified ninth abdominal nite by Rozkosny (1973). Dorsal views
sternite (Bonhag, 1951), contrary to the are illustrated for both the genital cap-
statement by Nagatomi and Iwata (1976) sule and the aedeagus, as well as a lateral
that the ninth sternite is absent in most view for the latter structure, which is fre-
lower Brachycera. In the Beridinae the quently strongly arcuate. A dorsal view
furca is fused to the ninth tergite. The of the post-genital segments is also in-
spermathecal ducts pass through the me- eluded, which is composed of the tenth
dian aperture of the furca into the genital and eleventh tergites and the cerci. The
chamber. In most cases the posterior por- shape of the tenth tergite is often char-

322 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

acteristic at the specific level, and in Beris
fuscipc's it possesses posterolateral sur-
styli.

SUBFAMILY BERIDINAE

Diag,no.sis. These flies can be separated
from other Nearctic Stratiomyidae by the
presence of preapical tergal grooves on
abdominal tergites two to six in all genera
except ExodoJitJia. Exodontha may be
assigned to the subfamily because it pos-
sesses a spur on the middle tibia (a char-
acter shared with Allognosta), which is
absent in all other Stratiomyidae outside
of the Beridinae. All Nearctic genera ex-
cept Allognosta possess at least four scu-
tellar spines, while taxa outside the Ber-
idinae possess no more than two. Other
characters which most members of the
subfamily exhibit include porrect anten-
nae, with a simple flagellum normally
composed of eight flagellomeres; palpi
usually distinct, two-segmented (one-
segmented and minute in Beris) and por-
rect; labellae of proboscis not strongly
sclerotized; wing venation not as strongly
crowded toward anterior margin of wing
as in other Stratiomyidae, the wings nor-
mally evenly set with microtrichia; M3
present or absent; M4 arises from the dis-
cal cell; abdomen usually with seven
more or less undifferentiated segments
(five in Exodontha); male genitalia visi-
ble externally, aedeagal complex trifid;
female cerci two-segmented.

Most North American species of Beri-
dinae are poorly known biologically. The
larvae of only two Nearctic species are
known and were briefly treated by
McFadden (1967). Label data, personal
discussions with other entomologists,
and morphology suggest that all Nearctic
species, with the possible exception of
the species of Exodontha, form conspe-
cific male swarms. The structure of the
male eyes conforms to that reported for
swarming Diptera (Downes, 1969; Mc-
Alpine and Monroe, 1968), and the in-
flated hind basitarsus found in males of
Actijia and Beris possibly has a function

related to swarming behavior. I have per-
sonally observed members of the Neo-
tropical beridine genus Oplachantha
Rondani in male swarms.

Key to the Nearctic Genera of Beridinae

1. Middle tihia with a single, small, apical

spur; abdomen ovate 2

Middle tibia without an apical spur; abdo-
men elongate 3

2. Scutellum without spines; abdominal ter-

gites with preapical transverse grooves;
M3 absent; smaller species, less than

6 mm. Allognosta Osten Sacken (p. 327)

Scutellum with spines; abdominal tergites

without preapical transverse grooves; M3
present; larger species, greater than

7 mm Exodontha Rondani (p. 346)

3. Palpi minute, one-segmented; males hol-

optic; scutellum normally with six or

more spines Beris Latreille (p. 336)

Palpi elongate, two-segmented, easily visi-
ble; males dichoptic; scutellum normally
with four spines Aetina Meigen (p. 322)

Genus ACTINA Meigen

Aetina Meigen, 1804: 116; t\'pe species Beris ni-

tens Latreille (Rondani, 1863: 87).
AUactina Curran, 1924: 24; type species Beris vir-

idis Say, by original designation.

Diagnosis. Members of the genus Ae-
tina may be recognized easily by their
four scutellar spines and well-developed,
two-segmented palpi. It is likely to be
confused only with Beris, which is also
elongate in form, but that genus possess-
es very minute, one-segmented palpi and
at least six scutellar spines. In addition,
males of Aetina are the only Nearctic
Beridinae of that sex which are dichoptic.
Other generic characters are: eyes pilose
in both sexes; antennae with first seg-
ment at least two times as long as second,
flagellum elongate with eight distinct
flagellomeres; face receding, oral margin
not produced; palpi slender in male,
more robust in female; legs with hind
femora much longer than middle and
front femora, almost twice as long in
males; hind tibiae and femora clavate;
hind basitarsi of males noticeably inflat-
ed, second hind tarsomere also slightly

Nearctic Befudinae • Woodlctj 323

inflated; middle tibiae without spurs; nae; face slightly diverging toward oral

wings with R2+.3 arising at or slightly be- margin, blackish with slight metallic re-

yond r-m, R4 present, M, and M2 petiolate flections, grayish pollinose except for

to widely separate at discal cell, M3 nor- very narrow median line; occiput dark

mally absent; abdomen of male parallel- pollinose except for cerebrale, which is

sided, fifth and sixth segments slightly concolorous with frons; frons (except for

wider, female abdomen slender but small bare area just above pollinose por-

ovate; both sexes with tergal grooves on tion), face, upper cerebrale, and outer

segments two to six; male genitalia with edge of occiput with long black hairs,

spinose membrane between aedeagal each hair slightly bent medially and

complex and genital capsule; gonostyli about the length of the first two antennal

with internal median tooth. segments combined, but pile just above

There is only one Nearctic species of antennae slightly shorter than second an-

Actina. Elsewhere, there is one Euro- tennal segment; genae with darkish pile

pean species, several from the eastern about two-thirds length of frontal pile;

Palaearctic, and five species in the Ori- antennae black; first segment 2.5 times as

ental Region (James, 1975). Species from long as second, slender, slightly expand-

the Australian Region presently ascribed ed apically; flagellum about 1.5 times the

to Actina may not be congeneric, but ex- length of both basal segments combined,

elusion must await a study of the fauna of conically tapered (Fig. 1); longer hairs of

that region. antennae black; eyes with short, dark pi-

losity; ocellar tubercle slightly promi-

Actina viridis (Say) nent; palpi brownish to black, second

segment verv narrow at base, expanded

Beris viridis Say, 1824: 368. apicallv; proboscis vellow, with verv pale

Actina viridis var. obscuripes Johnson, 1926a: 90. -i rr^'i 1 . ' n. '

pile. 1 horax shiny metallic green, some-

Type Material. The type of Beris viri- times ranging to bluish, occasionally with

dis Say is believed to be destroyed (Lin- bronzy or coppery reflections, becoming

droth and Freitag, 1969). The material less shiny and more brownish ventrally

Say examined was from Pennsylvania, and just below wing; humeri and postalar

and probably consisted of a series of fe- calli more blackish brown, the latter

males. sometimes brownish; scutellum with

The holotype male oi Actina viridis four yellow spines; thoracic pilosity quite

var. obscuripes Johnson is presently variable in color and length: some hairs

housed in the Museum of Comparative short and usually mostly pale, longer

Zoology at Harvard University, Cam- hairs subequal to frontal hairs or a little

bridge, Massachusetts (MCZ 27272). The longer, ranging from almost wholly pale

wings are slightly folded, but it is other- to mostly dark; pleural pile nearly as

wise in excellent condition, long, middle of mesopleuron and hypo-

Diagnosis. There is only one species pleuron bare; legs with coxae brownish,

o^ Actina in the Nearctic Region, A. vir- remainder yellowish except for front tar-

idis (Say), which may be recognized on si, apical three tarsomeres of middle tarsi,

the basis of the generic characters. apices of hind femora, apical halves of

Description. Male. Head black, frons hind tibiae, and apical three tarsomeres

slightly converging toward antennae, of hind tarsi brownish, basal two tarso-

0.19 to 0.22 width of head, shiny green to meres of hind legs whitish, the extreme

bluish, silvery pollinose just above an- apices pale brownish; legs ranging in

tennae, dull velvety black along inner dark specimens to wholly brownish black

eye margins from level of anterior ocellus except for extreme apices of front and

to about two-thirds of the way to anten- middle femora, extreme bases of all tibi-

324 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

ae, extreme bases and ventral surfaces of with basal two segments yellow, flagel-

middle and hind basitarsi as well as ven- liim more slender, basal flagellomere not

tral surface of second tarsomeres of hind as wide as in male, the entire flagellum

legs, which are yellowish; legs with tapering more gradually (Fig. 2); basal

abundant long pile, pale to dark (always flagellomere mostly yellow, inner por-

with at least a few dark hairs), on poste- tions of the next two yellowish, ranging

rior surfaces of front femora; middle fem- in some specimens to basal three fla-

ora with a few hairs, shorter than those gellomeres and inner portions of four to

on front femora, longer ventrally; hind six yellowish; hairs on basal segments

femora with a few longer hairs dorsally black; palpi yellow; pile of eyes ex-

at base; posterior surfaces of fore and tremely short and sparse. Thorax with

middle tibiae with a few longer hairs, short, pale pile, about the length of that

usually at least two, often six to eight
halter yellow, stem often brownish

on head but becoming a little longer on
pleura; mesonotum bright metallic green
wings hyaline, veins brownish, stigma to blackish green, humeri and postalar
brown and contrasting strongly with calli yellowish, sometimes brownish;
wing membrane; the membrane evenly legs wholly yellow except for tarsomeres
set with microtrichia except at extreme of front legs and distal three tarsomeres
base and basal regions of second basal of middle and hind legs brownish, rang-
and anal cells. Abdomen brownish to ing in some specimens to yellow with
brownish black, often with metallic re- brownish coloration on apices of hind
flections on lateral margins of tergites femora, apical two-thirds of front tibiae,
and most of sternites; medial portions of hind tibiae except for short basal regions,
tergites pollinose, sparse on sixth tergite front tarsi, apical third of first tarsomere
and absent on posterior segments; pile and entire tarsomeres two to five of mid-
long laterally, subequal to mesonotal die tarsi, apices of basal two tarsomeres
pile, somewhat shorter and sparser ven- and entire tarsomeres three to five of
trally; pile of pollinose areas may be hind tarsi. Abdomen lighter in color,
dark, pile of other areas pale, but in very brownish to brownish yellow, often with
dark specimens a few dark hairs may be metallic reflections, especially on basal
present on the lateral portion of the first three tergites, and often with lateral and
segment; tenth tergite and cerci yellow; outer posterior margins of tergites yel-
genitalia (Figures 11-13, 15) with gon- lowish; sternites ranging from wholly
ostyli possessing an internal medial brownish to wholly yellow, without me-
tooth; aedeagal complex with lateral tallic reflections; cerci yellow; pile of ab-
valves parallel-sided, slightly widened domen shorter than in male, wholly pale;
apically, their inner margins diverging genitalia (Fig. 14) with posterolateral
apically (Fig. 12). Length 4.3 to 6.4 mm. processes of furca long, converging; pos-
Female. Differs from male as follows: terior portions of spermathecal ducts un-
head smaller, with upper frons shinier, sclerotized; median aperture of furca
almost impunctate, wider (0.25 to 0.31 large. Length 4.2 to 5.9 mm.

width of head), and the velvetv black ^. ^ ., ^. ,.. ,^ ^j .. .,

, ' . n' r Distribution (Map 1). Nearctic; north-
spots along eye margins smaller; race n v..- i, /-^ i i • 1.1.11 j

.^, , 1. 1 1 • • 1 r ern British Columbia east to Labrador,
with larger central shmv region; pile 01 1.1. 1. tti. u m \^ • 1.4. 1.1
I, ^ 1 1 : /f I1 ^1 south to Utah, New Mexico, east to north-
head sparser, much shorter (less than the ^
1 i.u f 1.U J 1. 1 1. ern Ueorgia.
length or the second antennal segment

except on genae where it is slightly Ion- Locality Records (1149 males; 703 fe-

ger than that segment), and mostly pale, males). CANADA: Alberto: Cooking

but a few dark hairs sometimes present Lake; Edmonton; Gull Lake; Jumping

around antennal bases; antennae longer. Pond Creek, 20 mi. W of Calgary; Slave

Nearctic Beridinae • Woodlcy 325

Lake; Wabamun. British Columbia: Fort
Nelson; Gagnon Rd., 6 mi. W of Terrace,
220 ft.; Terrace; Trinity Valley; Vernon.
Labrador: Goose Bay. Manitoba:
Awenie; Garberry; Gedar Lake; Minne-
dosa; Ninette; 5 mi W. of Shilo; Trees-
bank. ISlew Brunswick: Hwy. 27, 8 mi. S
of McAdams; Barber D.; E. Sussex; Pen-
obsquis Camp, Rte. 1; St. Jacques; St.
John. Newfoundland: Mts. E of Codroy;
Lomond. Nova Scotia: Truro. Ontario:
Bell's Gorners; Black Rapids; Brittania;
Goldwater; Grozier; 6 mi. S of Devlin;
Dryden; Eagle; Finland; Ft. Frances; Go
Home Bay, 8 mi. W of Bala; Grimsby;
Guelph; Hallville; Jordan; Kingsville;
Low Bush, Lake Abitibi; Marmora; May-
nooth; Mer Bleu; Moose Factory; Nor-
mandale; Ogoki; Orillia; Osgoode; Otta-
wa; Petawawa; Pinewood; Pt. Pelee; Port
Hope; Rainy River; Rondeau; Simcoe;
Stittsville; Sturgeon Bay; Tillsonburg;
Tobermory. Quebec: Abbotsford; Anti-
costi; Aylmer; Beechgrove; Brecken-
ridge; Burbridge; Fabre; Fairy Lake;
Farnham; Gaspe; Gracefield; Harrington
Lake, Gatineau Park; Hull; Kazubazua;
Knowlton; Laniel; La Verendrye Prov.
Park, Rte. 58, mi. 139; Megantic; Mistas-
sini Post; Montreal; Mt. Orford, 1200 ft.;
Mt. Ste. Hilaire, 500-700 ft.; Nomi-
ningue; Old Ghelsea, King Mt., 1500 ft.;
Queens Park; Roundtop Mt., 1300 ft.,
Sutton; Rupert House; Ste. Anne des
Monts; Valtetreau. Saskatchewan: Big
River; Glaslyn; Kenosee. UNITED
STATES: Arkansas: Garland Go.; Mis-
sissippi Go.; Washington Go. District of
Columbia: Washington. Georgia: Glarke
Go.; Dawson Go.; DeKalb Go.; Houston
Go.; Lumpkin Go.; Rabun Go.; Towns
Go.; Union Go. Illinois: Adams Go.;
Boone Go.; Ghampaign Go.; DeWitt Go.;
Douglas Go.; DuPage Go.; Jackson Go.;
Lake Go.; LaSalle Go.; McHenry Go.;
McLean Go.; Macoupin Go.; Mason Go.;
Morgan Go.; Rock Island Go.; Tasewell
Go.; Union Go.; Vermilion Go. Indiana:
Marion Go.; Pulaski Go.; Tippecanoe Go.
Iowa: Boone Go.; Story Go.; Woodbury

Go. Kansas: Douglas Go.; Johnson Go.;
Leavenworth Go.; Lyon Go.; Riley Go.
Kentucky: Greenup Go.; Nelson Go.
Maine: Hancock Go.; Piscataquis Go.
Maryland: Montgomery Go.; Prince
Georges Go. Massachusetts: Berkshire
Go.; Essex Go.; Middlesex Go.; Norfolk
Go.; Worcester Go. Michigan: Arenac
Go.; Gharlevoix Go.; Ghippewa Go.; Del-
ta Go.; Hillsdale Go.; Ingham Go.; Kala-
mazoo Go.; LaPeer Go.; Livingston Go.;
Roscommon Go.; Washtenaw Go.; Wayne
Go. Minnesota: Anoka Go.; Garver Go.;
Glay Go.; Glearwater Go.; Gook Go.;
Grow Wing Go.; Dakota Go.; Fillmore
Go.; Freeborn Go.; Goodhue Go.; Hen-
nepin Go.; Houston Go.; Koochiching
Go.; Lake Go.; Lake of the Woods Go.
Lyon Go.; Marshall Go.; Mille Lacs Go.
Norman Go.; Olmstead Go.; Pine Go.
Pipestone Go.; Pope Go.; Ramsey Go.
Red Lake Go.; Rice Go.; Roseau Go.; St
Louis Go.; Wabasha Go.; Wadena Go.
Washington Go.; Winona Go. Mississip-
pi: Lafayette Go. Missouri: Jackson Go.;
Phelps Go. Nebraska: Boone Go.; Gass
Go.; Gherry Go.; Lancaster Go.; Sharpy
Go. New Hampshire: Grafton Go. New
Jersey: Bergen Go.; Burlington Go.;
Gloucester Go.; Ocean Go.; Sussex Go.
Neiv Mexico: Otero Go. New York: Glin-
ton Go.; Gortland Go.; Erie Go.; Lewis
Go.; Niagara Go.; Ontario Go.; Orange
Go.; Oswego Go.; Suffolk Go.; Tompkins
Go. North Carolina: Buncombe Go.; Gra-
ham Go.; Haywood Go.; Jackson Go.;
Johnston Go.; Macon Go.; Pender Go.;
Swain Go.; Yancey Go. Ohio: Delaware
Go.; Franklin Go.; Hocking Go.; Madison
Go.; Ottawa Go.; Summit Go.; Tuscara-
was Go. Pennsylvania: Dauphin Go.;
Erie Go.; Indiana Go.; Lebanon Go.
Rhode Island: Washington Go. South
Dakota: Grant Go. Tennessee: Gampbell
Go.; Hamilton Go.; Knox Go.; Montgom-
ery Go.; Sevier Go. Texas: Dallas Go.
Utah: Millard Go. Vermont: Bennington
Go.; Galedonia Go.; Orleans Go.; Wind-
ham Go. Virginia: Arlington Go.; Augus-
ta Go.; Fauquier Go.; Fairfax Go.; Page

326 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

Co.; Roanoke Co. West Virgitiia: Hamp- scription of the bright green coloration of
shire Co.; Jefferson Co. Wisconsin: Dane the head and thorax and the four yellow-
Co.; Door Co.; Polk Co.; Sauk Co.; Wash- ish scutellar spines leaves little doubt
burn Co. that he was describing this common

Flight Period. Specimens have been species,
collected from 25 March (only two March This species is quite variable. Color,

records) to 30 July. The predominant particularly of the legs, varies greatly,

months of activity are May and June The legs range from largely yellow to al-

throughout the range, although in south- most wholly blackish brown. Darker

ern states a number of specimens have specimens are found in the northern

been collected in April. areas of the range, and at high altitudes

Biology. The larval stages of Actina in more southern areas. In fact, the dark-

t)/nV//.S' are unknown, but the lai"va of Ac- est specimens I have examined were

find nitens (Latreille), a closely related from Clingman's Dome, Great Smoky

European species, is being described by Mountains National Park, Tennessee.

Rozkosny (in preparation). Adults have There is also a gradual trend toward

been taken at black lights and in Malaise smaller size toward more northern lati-

traps. They have also been collected on tudes. It seems likely that the darker

a number of plants, presumably by specimens are a result of environmental

sweeping: alfalfa (Minnesota, New York), conditions, perhaps cold temperatures,

aspen (Alberta), Barharea (Ontario), ce- Johnson's Actina viridis var. obscuripes

dar (Quebec), choke cherry (Manitoba), is based on dark specimens from Anti-

Equisetum (Manitoba), Geranium mac- costi, Quebec, and the name was synon-

ulatum (Virginia), Ledum (Ontario), lo- ymized by McFadden (1972). I agree

cust (Georgia), Norway pine (Minnesota), with his decision as there is no sharp dif-

Prunus virginiensis (Illinois), Prunus sp. ference between any light and dark forms

(Nebraska), Ribes sp. (Manitoba), Salix with respect to geographic location such

sp. (Alberta, Ontario, Nebraska), and Ta- that a subspecific designation is warrant-

raxacum officinale (Ontario). Males have ed. Wing venation in the species is also

been collected in conspecific swarms in variable. The bases of Mi and M2 at the

wooded areas in Kansas (McGinley, per- discal cell may be petiolate, or widely

sonal communication). separate, or anywhere between these two

Remarks. Actina viridis is the most extremes. I have observed specimens

common member of the subfamily in collected at the same locality on the same

North America, and is one of the most day with the entire range of variation. M3

common Stratiomyidae in the Nearctic is occasionally present in some speci-

Region. It cannot easily be confused with mens as a small stub, sometimes only on

any other species. Since the lectotype one wing. One otherwise typical speci-

designation presented elsewhere in this mens was noted with R4 absent,
paper has removed Beris quadridentata Because of the variable nature of this

Walker from synonymy with this name, species, I checked genitalia from a num-

there are no synonymous specific names, ber of localities as well as from light and

only one varietal one. Consequently, dark forms to try to detect any differences

there seems to be no reason for designat- that might indicate more than one

ing a neotype for this species at present, species was present. Although the shape

even though the original type material is of the gonostyli is somewhat variable,

destroyed. Any such action should be left particularly with respect to the size of the

to future authors, if additional species of lobe above the median tooth, I detected

the genus are discovered in the Nearctic no differences that indicated the pres-

Region. Say's mention in the original de- ence of more than one species. The shape

Nearctic Beridixae • Woodleij 327

of the lateral valves of the aedeagal com-
plex is particiilarK' diagnostic, and is
quite constant. I have examined the ae-
deagal complex o{ Actina nitens as well
as a species from Taiwan for comparison,
and both were distinctly different from
Actina viridis. Thus, at present, it can be
concluded that one rather variable species
of Actina occurs in the Nearctic Region.

Genus ALLOGNOSTA Osten Sacken

Allognosta Osten Sacken, 1883: 297; type species
Beris fuscitarsis Say (Coquillett, 1910: 505).

Diagnosis. Members of the genus Al-
lognosta may be separated from other
Beridinae in North America by the lack
of spines on the scutellar margin. All oth-
er genera in the Nearctic Region have at
least four scutellar spines. Other features
exhibited by the genus include: eyes
bare to sparsely pilose; males holoptic,
females dichoptic; antennae with simple,
eight-segmented flagellum; palpi large,
two-segmented; middle tibiae with one
small apical spur; abdomen flattened,
ovate, the seventh segment somewhat re-
duced; tergites two to six with preapical
grooves; wing with II4 and r-m present,
Mg absent, Mj and M2 separate at discal
cell.

Enderlein (1921), apparently unaware
of Coquillett's designation of Beris fus-
citarsis Say as the type species for Allog-
nosta, designated the Palaearctic species
Metoponia vagans Loew as the type
species. This is, of course, invalid, as Co-
quillett's designation was earlier.

There are three valid species of Allog-
nosta in North America, found primarily
in the eastern half of the continent. Al-
though there are no synonyms of the ge-
neric name, only A. brevicornis was de-
scribed in the genus. Early authors
described species in Beris or Sargus, and
two others were described in Metoponia.
This last name is based upon a type
species in the Chiromyzinae, and it is not
surprising that members of Allognosta
were described in that genus since the

Chiromyzinae also have the sixth and
seventh abdominal segments rather large
and undifferentiated, and lack scutellar
spines.

Allognosta occurs in the eastern Pa-
laearctic Region, and is particularly rich
in species in the Oriental Region, where
there are currently 25 nominal species
(James, 1975). Species from Africa pres-
ently placed in Allognosta may belong in
other genera, but I have not examined
any specimens of the two species de-
scribed from that continent.

Key to the Nearctic Species of Allognosta

1. Males (holoptic) 2

Females (dichoptic) 4

2. Hairs of face silvery, at most as long as first

antennal segment; ventral portion of
mesopleuron usnally with shin>- area de-
void of pollinosity; medial portion of ven-
tral bridge of gonocoxites weakly
produced, with only two small lobes (Fig.

23) fuscitarsis (Say)

Hairs of face usually dark, and usually lon-
ger than the first antennal segment; ven-
tral portion of mesopleuron without bare,
shiny area; medial portion of ventral
bridge of gonocoxites stronglv produced
(Figs. 19, 28) : 3

3. Antennae short, the flagellum strongly ta-

pering toward apex (Fig. 3); mesonotum
somewhat shiny, usually bronz\- black;
process of ventral bridge of gonocoxites
strongly bilobed (Fig. 19); gonostylus

rounded at apex brevicornis Johnson

Antennae longer, the flagellum more slen-
der and tapering more gradualh- toward
apex (Fig. 5); mesonotum less shin\- due
to coarser punctation, although it may
have metallic reflections; process of ven-
tral bridge of gonocoxites only feebly
emarginate at middle (Fig. 28); gono-
stylus pointed at apex

obscuriventris (Loew)

4. P. X reddish to yellowish; frons below

.nsverse sulcus wholly pollinose

brevicornis Johnson

Pleura dark, nearly concolorous \\ ith meso-
notum; frons below transverse sulcus
normalK' with lateral bare areas 5

5. Mesopleuron with a ventral, shiny, bare

area; face with hairs pale, no longer than
the first antennal segment; disc of abdo-
men often yellowish, contrasting with lat-
eral margins fuscitarsis (Say)

Mesopleuron without a ventral, shin\-, bare
area, completely pollinose; face with

328 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

hairs usualh blackish, h)ii.yc'r than first
antennal segment; disc of ahdoinen usu-
ally dark obscuriventris (Loew)

Allognosta brevicornis Johnson

Allognosta brevicornis Johnson, 1923: 71.

Type Material. The holotype specimen
is presently housed in the Museum of
Comparative Zoology (MCZ 14928).
It is a male from Norwich, Vermont. It is
missing the right flagellum of the anten-
nae, the right middle leg, and both hind
legs; the right wing and abdomen are
glued to a small paper card attached to
the pin below the specimen.

Diagnosis. Females are distinguished
easily from other Nearctic species of the
genus by the reddish and yellowish pleu-
ra. Males may be recognized by their
short antennae and distinctive genitalia,
especially the strongly bilobed medial
process of the ventral bridge of the gon-
ocoxites (Fig. 19). The small size is some-
what distinctive, as is the wholly dark ab-
domen.

Description. Male. Head brownish to
black, frons and face densely silvery gray
pollinose, rest of head grayish pollinose
except for genal area, cerebrale, and the
somewhat prominent ocellar tubercle,
which are shiny; pile of face brownish, a
little longer than the first antennal seg-
ment, that on genae longer than the first
two antennal segments combined, pale to
blackish; antennae about 0.7 times length
of head, first two segments brownish to
black, second sometimes yellowish; fla-
gellum brownish black, sometimes with
first flagellomere and inner portions of
flagellomeres two to five yellowish; long-
er hairs of antennae black; palpi with first
segment yellowish to brownish, second
dark brown, both with dark hairs, first
segment with a few pale hairs; proboscis
yellowish. Thorax with mesonotum and
scutellum dark brownish black with faint
bronzy reflections, humeri and postalar
calli paler; pleura brownish, sclerites di-
rectly beneath wing more yellowish; pro-
thorax and pleura pollinose except for

most of sternopleuron and pteropleuron;
mesonotal pile mosdy dark, pleural pile
with a mixture of pale and dark hairs; legs
with front and middle coxae yellowish,
hind coxae brownish yellow, femora yel-
low except for apices of hind femora eva-
nescently brownish; tibiae yellowish,
grading evenly to brownish on apical
halves, tarsi brown except for extreme
apices of tarsomeres one to four of all tar-
si, and basal three-fourths of hind basi-
tarsi, yellowish; halter brown, stem yel-
lowish; wing light brownish, evenly set
with microtrichia, stigma weakly con-
trasting with rest of wing. Abdomen
brown, dorsally thinly pollinose; longer
hairs mostly dark dorsally, mostly pale
ventrally; hypopygium concolorous with
abdomen; genitalia (Figs. 16-19) with
gonostyli large, scoop-shaped, the inner
apical surface strongly concave (Fig. 19);
medial process of ventral bridge of gon-
ocoxites strongly produced, bilobed, api-
ces of lobes with short spicules (Fig. 19);
aedeagus shorter than lateral valves,
which diverge slighdy at apex (Fig. 17).
Length 4.2 to 4.4 mm.

Female. Differs from male as follows:
head brownish to blackish, gray pollinose
on frons below transverse sulcus and
face, faintly so on dorsolateral border of
occiput, rest of head shiny, finely punc-
tate; pile of head pale, shorter than first
antennal segment except on genae,
where it is a litde longer than that seg-
ment; frons gradually narrowing toward
antennae, 0.40 to 0.50 width of head at
anterior ocellus; first two antennal seg-
ments yellow, flagellum usually with in-
ner portions of flagellomeres two to five
extensively yellowish; palpi with first
segment usually yellowish. Thorax with
entire prothorax, humeri, postalar calli,
and pleura of mesothorax yellowish, ex-
cept for basal three-fourths of sternopleu-
ron which is brownish, although the
mesopleuron occasionally has evanescent
brownish tinges; mesonotal pile pale,
more appressed and shorter than in male,
rest of thoracic pile also pale; legs with

Nearctic Beridinae • Woodley 329

coxae, femora, and tibiae yellow, some-
times with tibiae brownish as in male;
tarsi with basal halves of all front and
middle tarsomeres often yellowish, oth-
erwise as in male; halter with knob yel-
low to brown; genitalia (Fig. 32) with fur-
ca reduced, the median aperture very
large; lateral processes present and
strongly diverging anteriorly; spermathe-
cal ducts unsclerotized posteriorly.
Length 3.9 to 4.9 mm.

Distribution (Map 2). Quebec and
New Brunswick south to Tennessee and
North Carolina, and west to British Co-
lumbia.

Locality Records (6 males, 41 females).
CANADA: Alberta: Elk Island. British
Columbia: Kleanza Creek, 14 mi. E of
Terrace; Lac la Hache. New Brunswick:
Doaktown, Hwy. 8. Ontario: Maynooth;
One Sided Lake; Ottawa. Quebec: Bol-
ton Pass, Knowlton, 800 ft.; Duncan
Lake, near Rupert; King Mtn., Old Chel-
sea. UNITED STATES: Maine: Salis-
bury Cove; Hancock Co., Bar Harbor;
Penobscot Co., Passadumkeag. Massa-
chusetts: Essex Co., Gloucester. Minne-
sota: Itasca Co., T-57, R-24, S-26. New
Hampshire: Cheshire Co., Jaffrev. New
York: Uphill Brook, 3200 ft; Essex Co.,
Lake Tear, 4300^600 ft.; Greene Co.,
2500 ft. North Carolina: Highlands, 3800
ft.; Graham Co., Robbinsville. Tennes-
see: Monroe Co., 20 mi. E of Tellico
Plains, 2080 ft. Vermont: Bolton Mtn.; E.
Charleston; Windsor Co., Norwich.

Flight Period. Collection dates range
from 26 May to 9 August, with most spec-
imens being taken in June or July.

Biology. The immature stages are un-
known. Adults have been collected on
"liliaceous undergrowth-hemlock forest"
(British Columbia), "ex Spiraea'^ (Ontar-
io), and by sweeping Kalmia angustifolia
(Quebec).

Remarks. AUognosta brevicornis re-
mains the least known species of the ge-
nus in North America. Specimens, partic-
ularly of males, are rare in collections.
Consequently, little can be said concern-

ing the range of variation within the
species, but it is easily recognizable. It is
the only member of the genus known to
occur west of the Rocky Mountains.

AUognosta fuscitarsis (Say)

Beris fuscitarsis Say, 1823: 29.
Sargus dor salts Say, 1824: 377.
Sarfius pallipes Wiedemann, 1830: 41.
Beris brevis Walker, 1848: 127.
Beris lata Walker, 1848: 127.

Type Material. The type (sex not stat-
ed, and not determinable from the origi-
nal description) of Beris fuscitarsis Say
was described from Pennsylvania. The
scutellum was broken; thus Say could not
have known that it was spineless, and
thus he placed the species in Beris. The
specimen is believed to be destroyed
(Lindroth and Freitag, 1969), and to pre-
serve the stability of the present usage of
this name, it seems advisable to desig-
nate a neotype. I hereby designate a male
specimen in the Museum of Comparative
Zoologv as neotvpe of this species; it is
labeled "Lehigh Gap VII.9.07 Pa/ Col-
lected by C. T. Greene/ Collection C. W.
Johnson/ NEOTYPE Beris fuscitarsis
Say, 1823: 29 des. N. E. Woodley 1980
MCZ # 32553/ AUognosta fuscitarsis
(Say) det. Woodley 1979." The neotype
label is red. The specimen fits the origi-
nal description fairly well. Although the
description is brief, the pale disc of the
abdomen with "brown incisures" and
coloration of the legs of the neotype spec-
imen agree with it. The neotype is from
Pennsylvania and agrees with my diag-
nosis of this species, which is given be-
low. The neotype is missing the right
front leg, but is otherwise in good con-
dition.

The type material of Sargus dorsalis
Say is also lost. It seems quite likely that
it was conspecific with A. fuscitarsis as
suggested by previous authors. Say's de-
scription of the abdomen and legs espe-
cially seems to fit A. fuscitarsis, but
could be interpreted as fitting light-col-
ored specimens of A. obscuriventris
(Loew). The latter interpretation, if fol-

330 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

lowed, would cause Sargns dorsalis to some sort of transcription error, and was
\i-a.ve^riox\\.y over Metoponiaobscuriven- really referring to A. fuscitarsis, but it
tris Loew, a name that has seen long seems unlikely in that case for him to
standing use. To preserve the present syn- have attributed the name to Sargus, since
onymy, it seems best to designate a neo- he had earlier combined Say's Sargus
type for Sargus dorsalis. I am hereby dorsalis with Berts (Wiedemann, 1828:
designating a male specimen in the Mu- 540). If, however, Wiedemann was de-
seum of Comparative Zoology as neo- scribing a specimen that he thought was
type, labeled: "Lexington, Ky. May 19- a new species, it was probably in Say's
24, 1972 Malaise Trap/ P. H. Freytag Col- collection and is now destroyed, as he
lector/ NEOTYPE Sargus dorsalis Say, mentioned that it was in the Philadelphia
1824: 377 des. N. E. Woodley 1980 MCZ museum. Due to the uncertainty sur-
# 32554/ Allognosta fuscitarsis (Say) det. rounding the name, I prefer not to des-
Woodley 1979." The specimen fits the ignate a neotype for it, particularly since
Say description quite well, except that Wiedemann attributed the name to Say.
the fourth and fifth abdominal tergites are It may well be a uomen nudum.
pale yellow on the disc. Say, in mention- The types of Beris brevis and Berts
ing that the first three tergites were pale lata, both described by Walker (1848),
centrally, probably was referring to the are in the British Museum. I have ex-
first four, as the first and second tergites amined both specimens, and they are
are fused and the suture is obsolete. The clearly Allognosta fuscitarsis, as has
neotype is from Kentucky, which was cit- been noted previously by various au-
ed as the type locality by Say. He mistak- thors. I have added determination labels
enly placed the species in Sargus, prob- to both. The specimen of Beris brevis
ably because the scutellum lacked spines was not previously labeled as a type, but
and the third antennal segments were has a locality label stating "Trenton" and
broken off. Say stated that he examined another label reading "New York, pres.
both male and female specimens, but by E. Doubleday." These agree with the
mentioned that the sexes were similar, information published by Walker with
He probably saw only males, as females his description. I have added a red label
of this species have widely separated indicating that this is the type, and the
eyes, which he probably would have not- fact that I added the label,
ed. Males are indicated by his statement Diagnosis. Males and females may be
"stemmata approximate on vertex. " The separated from other species of Nearctic
neotype male fits my diagnosis of A. /i(<5- Allognosta by the unique shiny, bare
cttarsis presented below. area on the ventral portion of the meso-
The name Sargus pallipes Wiedemann pleuron. The male genitalia are diag-
remains somewhat enigmatic. Wiede- nostic, particularly the posteromedian
mann attributed the name to Say, al- portion of the ventral bridge of the gono-
though the latter described only one coxites, which is not produced as in the
Nearctic Stratiomyidae with the name other two species (Fig. 23); this is usual-
pallipes, and that clearly belongs to the ly visible without dissection. Other diag-
genus Nemotelus Geoffroy. It is curious nostic features include the long basal
that the names Beris fuscitarsis and flagellomere of the antenna (Fig. 4), and
Nemotelus pallipes appear on the same the short, pale facial hairs. The transpar-
page of Say's work (1823: 29), and that ent wing with contrasting stigmal area is
the type locality of both is cited as "Penn- also characteristic.

sylvania," the same locality given by Description. Male. Head black, frons

Wiedemann for Sargus pallipes. It seems and face densely silvery pollinose except

at least possible that Wiedemann made for a tiny, shiny triangle dorsally on frons;

Nearctic Beridinae • Woodley 331

ocellar tubercle shiny black, somewhat
prominent; rest of head grayish polli-
nose, becoming less dense on genal re-
gion and on cerebrale; face with pale pile
about the length of the first antennal seg-
ment, genal region with pale pile some-
what longer than the first two antennal
segments combined; eyes bare; antennae
(Fig. 4) about 0.8 times length of head,
with first two segments and first two fla-
gellomeres yellow (first segment some-
times brownish), apical flagellomeres
brownish black, sometimes yellowish on
inner portions of three to five; longer
hairs of antennae black; palpi brownish
black, with a mixture of pale and dark
hairs; proboscis yellowish. Thorax with
mesonotiun and scutellum shiny black,
humeri and postalar calli usually paler,
brownish to yellowish; prothorax and
pleura brownish black, with thin silvery
pollinosity except for shiny, bare areas on
extreme dorsal, ventral, and posterior
margins of mesopleuron, a large postero-
medial patch and ventral portion of ster-
nopleuron, and most of pteropleuron;
longer hairs of thorax pale; legs yellow
except for dark brownish coloration on
apical two-thirds of fore tibiae (especially
dorsally), fore tarsi, apical four tarso-
meres of middle tarsi, hind tibiae except
extreme bases, apical three tarsomeres of
hind tarsi, and rather diffuse anterior
areas on the hind femora (rarely entire
hind femora dark); ranging in some spec-
imens to light brown, with yellow areas
on fore tibiae and hind femora reduced;
hind coxae brownish; wings almost clear,
evenly set with microtrichia, with eva-
nescent infuscation anteriorly; stigma
dark, contrasting strongly with the rest of
wing; halter with stem yellow, knob yel-
low with brownish base to wholly brown.
Abdomen creamy yellow except for lat-
eral margins and narrow posterior mar-
gins of tergites one to four, lateral margin
and apical half of tergite five, and re-
maining tergites dark brown, often with
tergite five wholly brown; sternites one
to four with lateral margins evanescently

brownish, remaining sternites brownish;
pile of abdomen mosth' dark dorsally,
pale ventrally; hypopygium brownish,
becoming yellow on gonostyli and cerci;
genitalia (Figs. 20-23) with medial region
of ventral bridge of gonocoxites not pro-
duced, feebly bilobed (Fig. 23); gonostyli
arcuate, slender, rounded apically; ae-
deagus slightly shorter than lateral valves,
which diverge slightly apically, the three
lobes of the complex well separated
(Figs. 21, 22). Length 4.3 to 6.1 mm.

Female. Differs from male as follows:
head brownish black, silvery gray polli-
nose below transverse sulcus of frons ex-
cept areas on each side slightly above
and lateral to antennae, which are nearly
bare and shiny; upper frons subparallel
to slightly convergent toward antennae,
0.33 to 0.41 width of head at anterior
ocellus, diverging below transverse sul-
cus, finely punctate, middle with elon-
gate, oval, shallow depression; antennal
flagellomeres three to five usually yel-
lowish internally, rest brownish to black.
Thorax with thin dorsal strip and entire
ventral half of mesopleuron bare and
shiny. Abdomen ranging in coloration
from that described for male to wholly
dark brown, usually with pale markings
when present restricted to medial por-
tions of tergites; tenth tergite and cerci
yellowish; genitalia (Fig. 31) with furca
emarginate laterally, the median aperture
truncate anteriorly; spermathecal ducts
unsclerotized posteriorly. Length 4.1 to
6.3 mm.

Distribution (Map 4). Nova Scotia and
northern Maine west to southern Mani-
toba, south to Arkansas and northern
Florida.

Locality Records (383 males; 577 fe-
males). CANADA: Manitoba: 5 mi. W of
Shilo. New Brunswick: Fredericton.
Nova Scotia: Lockeport; Truro. Ontario:
Anacaster; Blackburn; Cornwall; Emo;
Finland; Fort Frances; Grimsby; Jordan;
Kelly Lake; Mer Bleu, 5 mi. E of Ottawa;
Niagara Falls; Orillia; Ottawa; Picton; Pt.
Pelee; Simcoe; Smith's Bay, near Picton;

332 Bulletin Mitsciiin of Coiiiparativc Zoology, Vol. 149, No. 6

Stittsville. Quebec: Belanger; Brecken-
ridge; Fabre; Heiiiiningford; Knowlton;
Laval; Missisavoir River, Bolton; Mon-
treal; Richelieu; Wakefield. UNITED
STATES: Alahama: Lee Co. Arkansas:
Conway Co.; Mississippi Co. Connecti-
cut: Fairfield Co.; Hartford Co.; Litch-
field Co.; New Haven Co. District of Co-
lumbia: Washington. Florida: Alachua
Co. Georgia: Clarke Co.; Floyd Co.; Ful-
ton Co.; Oconee Co.; Rabun Co.; White
Co. Illinois: Adams Co.; Bond Co.; Bu-
reau Co.; Carroll Co.; Champaign Co.; Jo
Daviess Co.; Kane Co.; Lake Co.; La
Salle Co.; McHenry Co.; McLean Co.
Macoupin Co.; Madison Co.; Ogle Co.
Tazewell Co.; Union Co.; Vermilion Co.
Winnebago Co. Indiana: Brown Co.
Marion Co.; Randolph Co.; Tippecanoe
Co. Iowa: Johnson Co.; Scott Co.; Story
Co.; Woodbury Co. Kansas: Douglas Co.
Kentucky: Bullitt Co.; Fayette Co.; Nel-
son Co. Maine: Aroostook Co.; Hancock
Co.; Oxford Co.; Penobscot Co.; York Co.
Maryland: Anne Arundel Co.; Montgom-
ery Co.; Prince Georges Co. Massachu-
setts: Barnstable Co.; Berkshire Co.; Es-
sex Co.; Hampden Co.; Hampshire Co.;
Middlesex Co.; Suffolk Co.; Worcester
Co. Michigan: Berrien Co.; Cheboygan
Co.; Clinton Co.; Ingham Co.; Livingston
Co.; Wayne Co. Minnesota: Aitkin Co.;
Clay Co.; Dakota Co.; Dodge Co.; Fill-
more Co.; Goodhue Co.; Hennepin Co.;
Houston Co.; Itasca Co.; Lac Qui Parle
Co.; Lyon Co.; Olmstead Co.; Pine Co.;
Pipestone Co.; Polk Co.; Ramsey Co.;
Roseau Co.; Winona Co. Mississippi: La-
fayette Co. Missouri: Boone Co.; Wayne
Co. Nebraska: Antelope Co.; Cuming
Co. New Hampshire: Grafton Co.; Sulli-
van Co. New fersey: Bergen Co.; Bur-
lington Co.; Essex Co.; Hunterdon Co.
Somerset Co. New York: Broome Co.
Cayuga Co.; Clinton Co.; Herkimer Co.
Jefferson Co.; Lewis Co.; Madison Co.
Monroe Co.; Nassau Co.; Ontario Co.
Otsego Co.; Seneca Co.; Steuben Co.
Sullivan Co.; Tioga Co.; Tompkins Co.
Westchester Co.; Wyoming Co. North

Carolina: Buncombe Co.; Graham Co.;
Haywood Co.; Macon Co.; Mitchell Co.;
Wake Co.; Watauga Co. North Dakota:
Cass Co. Ohio: Champaign Co.; Erie
Co.; Franklin Co.; Greene Co.; Hocking
Co.; Lawrence Co.; Logan Co.; Medina
Co.; Pickaway Co.; Seneca Co.; Summit
Co.; Wayne Co. Pennsylvania: Alleghe-
ny Co.; Bedford Co.; Bucks Co.; Centre
Co.; Dauphin Co.; Delaware Co.; Hun-
tingdon Co.; Philadelphia Co.; West-
moreland Co. Rhode Island: Richmond.
South Dakota: Grant Co. Tennessee: An-
derson Co.; Union Co. Vermont: Orleans
Co.; Rutland Co.; Windham Co.; Windsor
Co. Virginia: Fairfax Co.; Giles Co.;
Montgomery Co.; Prince George Co.;
Shenandoah Co. West Virginia: Preston
Co. Wisconsin: Dane Co.; Polk Co.; Rusk
Co.

Flight Period. Collection dates range
from 20 April to 1 September. Most spec-
imens have been collected in June and
July, with a fair number in August and
May. This species seems to have a fairly
long flight period even in one region,
since I have seen almost the entire range
of collection dates from specimens col-
lected at a single locality.

Biology. The larvae oi Allognosta fus-
citarsis evidently live in decaying organ-
ic matter (McFadden, 1967). During a
study of reclaimed strip mine areas in
Ohio, large numbers of larvae were col-
lected in Berlese samples (Triplehorn,
personal communication). The record of
A. fuscitarsis larvae from stems of aquat-
ic plants (McFadden, 1972) seems unusu-
al. One adult specimen from Maryland
was annotated "reared-horse manure."
Adults have been taken at light, by
sweeping, and in Malaise traps. They
have been taken from quite a few differ-
ent plants; labels indicate collections
from: alfalfa (New York, Georgia), Bar-
barea vulgaris (Massachusetts), blue-
grass (Minnesota), Catalpa leaves (North
Carolina), elm (New Jersey), Eupato-
rium maculatum (New York), and vetch
(Tennessee). The record of A. fuscitarsis

Nearctic Beridinae • Woodley 333

taken on Kalmia (McFadden, 1972) is er-
roneous; the specimen referred to was
actually Allognosta brevicornis. Male
specimens from New York were exam-
ined that had a label that indicated they
were collected from conspecific swarms.
It seems probable that all members of the
genus have diis behavior.

Remarks. Allognosta fuscitarsis is the
most common member of the genus in
North America, and is present in most
collections. Malaise traps have turned up
rather long series in a single day, also at-
testing to the commonness of the species.

The species has previously been char-
acterized by the pale color of the middle
portions of the abdominal tergites. This
has been found to be quite variable in
females, which has led to some previous
misidentifications. I have seen series col-
lected on the same day from the same lo-
cality which exhibit a range of variation
from totally dark brown to rather exten-
sively pale. The species is much more
easily identified reliably by the shape of
the antennae, the bare area on the meso-
pleuron, and the genitalia of both sexes.
Males are less variable, and I have not
seen specimens with totally brown ab-
domens. However, I have seen speci-
mens of A. ohscuriventris (Loew) with
extensively pale tergites, which could
lead to their being confused with this
species. Allognosta fuscitarsis is evi-
dently quite closely related to the Jap-
anese A. flavimacidata Nagatomi and
Tanaka. The latter species also has pale
medial areas on the tergites (which are
variable according to Nagatomi and
Tanaka, 1969), and the male genitalia are
extremely similar in the two species. I
initially thought they might be conspe-
cific, but I subsequently examined a fe-
male paratype from the United States
National Museum of Natural History,
through the kindness of Willis Wirth. The
female genitalia are quite different, and
the species is larger, has darker legs, a
dark first antennal segment, and no bare
area on the mesopleuron.

Allognosta obscuriventris (Loew)

Metoponia ohscuriventris Loew, 1863: 299.
Metopouia similis Loew, 1863: 299. NEW SYN-
ONYMY.

Type Adaterial. The female type of Me-
toponia obscuriventris, from "D.C.," is
housed in the Museum of Comparative
Zoology (MCZ 12535), and is in perfect
condition.

There are two specimens labeled as
types of Metoponia similis Loew, also
in the Museum of Comparative Zool-
ogy (MCZ 12534). One specimen bears
Loew's handwritten label "similis, m.,"
and the other bears a label "New York,
Schaum," the type locality and collector
included in Loew's original description.
There is no indication in the original de-
scription as to how many specimens
Loew saw. Both of the above specimens
are males, the only sex which Loew de-
scribed. One specimen has bluish reflec-
tions on the mesonotum, a primary char-
acter used by Loew to characterize M.
similis, and I hereby designate it as lec-
totvpe. It is labeled "Loew Coll./ similis,
m./ Type 2 12534/ LECTOTYPE Meto-
ponia similis Loew, 1863 des. N. E.
Woodley, 1979/ Allognosta obscuriven-
tris (Loew) det. Woodley 1980." The
third antennal segments are missing, and
the left wing is somewhat crumpled, but
it is otherwise in good condition. I am
designating the other specimen as para-
lectotype, and it is labeled "New York,
Schaum/ Loew Coll./ Type 12534/ PARA-
LECTOTYPE Metoponia similis Loew,
1863 des. N. E. Woodlev 1979/ Allognos-
ta fuscitarsis (Say) det. Woodley 1978."
The entire head and left front leg are
missing from the specimen.

Diagnosis. Male specimens of Allog-
nosta obscuriventris (Loew) are best
identified by the shape of the genitalia,
particularly the long but shallowly emar-
ginate medial process of the ventral
bridge of the gonocoxites (Fig. 28), and
the large gonostyli which are somewhat
expanded apically, and sharply pointed.

334 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 6

Females are best identified by their dark
pleura and the wholly pollinose ventral
portion of the niesopleuron. Other diag-
nostic characters include the elongate an-
tennal flagellum with the first flagello-
mere being short (Fig. 5), the rather
poorly contrasting stigmal area of the
wing, and the long and usually dark facial
hairs.

Description. Male. Head black, frons
with small medial longitudinal groove;
ocellar tubercle slightly prominent; frons
and face strongly grayish white pollinose,
rest of head much more thinly pollinose;
vertex, frons, face, and genal regions pi-
lose, pile of face about 1.5 times as long
as first antennal segment, longer on gen-
ae, mostly brownish but some pale hairs
usually present on genae; eyes with ex-
tremely short, sparse hairs; antennae
about 0.7 times length of head, dark
brown with apex of second segment and
first and second flagellomeres yellow,
sometimes first and second antennal seg-
ments wholly yellow; first flagellomere
shorter than the following three com-
bined; longer hairs of antennae black;
palpi with first segment yellowish to
black, second velvety black, both seg-
ments with a mixture of pale and dark
hairs; proboscis yellowish. Thorax with
mesonotum shiny black, often with some-
what bluish or greenish metallic reflec-
tions, densely punctate; humeri light to
dark brown, postalar calli brownish to
yellowish; pleura concolorous with
mesonotum, except that area below wing
base is more brownish; pleura pollinose
except for bare areas on extreme dorsal
and posterior part of mesopleuron, entire
sternopleuron except for a dorsal band,
and entire pteropleuron and metapleu-
ron; thoracic pile erect, long and pale, a
few darker hairs sometimes present on
mesonotum; legs with coxae yellowish to
brown; remainder brownish with yellow
coloration on apical one-fifth of front fem-
ora, extreme apices of middle and hind
femora, basal one-third of front tibiae, ex-
treme bases of middle and hind tibiae,

basal two tarsomeres of middle legs, and
basal three tarsomeres of hind legs; rang-
ing in some specimens to more exten-
sively yellow, only faint medial portions
of hind femora and most of hind tibiae
brownish, with tarsomeres colored as
above except for yellowish bases of front
basal tarsomeres; wings very slightly in-
fuscated anteriorly, decreasing posterior-
ly, stigma distinct but weakly contrasting
with wing membrane; halter yellowish
brown to dark brown. Abdomen ranging
from yellowish, except for lateral margins
of tergites and sternites one to five and
segments posterior to five brown, to to-
tally brownish, the latter condition being
found in most specimens; abdomen thin-
ly pollinose (less so ventrally), except
along dorsolateral margins, with first ab-
dominal segment rather strongly polli-
nose both dorsally and ventrally; pilosity
short and sparse both dorsally and ven-
trally, longest along lateral margins; hy-
popygium becoming yellowish on gon-
ostyli and cerci; genitalia (Figs. 24-28)
with median process of ventral bridge of
gonocoxites produced, shallowly emar-
ginate (Fig. 28); gonostyli expanded api-
cally, sharply pointed; aedeagal complex
(Figs. 25, 26) with lateral valves apically
reflexed, slightly longer than aedeagus.
Length 4.1 to 5.7 mm.

Female. Differs from male as follows:
head brownish to black, frons 0.31 to 0.44
width of head, margins parallel to slightly
convergent toward antennae; frons sepa-
rated into two portions by a transverse
sulcus which is obsolete medially; a lon-
gitudinal groove separates the lower
frons into two halves, each of which is
surrounded by a border of silvery gray
pollinosity, medial portions impunctate
and shiny to sparsely pubescent with
very short hairs, and slightly convex; up-
per frons rather evenly punctate, with a
vaguely defined medial depression which
may extend longitudinally almost the
length of the frons; upper frons clothed
with short, pale hairs subequal in length
to the first antennal segment; face silvery

Nearctic Beridinae • Woodley 335

gray pollinose; ocelli vaguely prominent;
occipital margin rather sharp; vertex pol-
linose except lateral to ocelli, and above
and posterior to the middle of the eye.
Thorax with hairs shorter than in male.
Abdomen sometimes a little less densely
pollinose than in male; genitalia (Figs.
29, 30) with median aperture of furca
pointed anteriorly, large; anterolateral
margins ranging from nearly straight to
strongly emarginate. Length 4.0 to 5.4
mm.

Distribution (Map 3). Southern Mani-
toba east to southern Quebec, south
through Kansas, Mississippi, to Florida.
A record from Colorado needs substanti-
ation.

Locality Records (194 males, 206 fe-
males). CANADA: Manitoba: 5 mi. SW
ofShilo. Ontario: Black Rapids; Cayuga;
Chatham; Crozier; Eagle River; Emo;
Jock River; Leamington; Marmora; Mer
Bleu, 5 mi. E of Ottawa; Ft. Pelee; 10 mi.
SE of Renfrew; Stittsville; Tillsonburg;
Vernon. Quebec: Abbotsford: Aylmer;
Fabre; Fairy Lake; Harrington Lake, Ga-
tineau Park; Kazabazua; Knowlton; Lan-
iel; Montreal; Old Chelsea. UNITED
STATES: Colorado: Boulder Co., Val-
mont Butte, 5300 ft. Connecticut: Hart-
ford Co.; Litchfield Co. District of Co-
lumbia: Washington. Florida: Alachua
Co.; Calhoun Co.; DeSoto Co.; High-
lands Co.; Marion Co.; Orange Co.; Pi-
nellas Co.; Polk Co. Georgia: Clarke Co.;
Fulton Co.; Liberty Co.; Lynn Co.; Ra-
bun Co. Illinois: Kappa; LaRue; Spring
Grove; Champaign Co.; Cook Co.; Jo
Daviess Co.; Lake Co.; McHenry Co.;
McLean Co.; Macoupin Co.; Peoria Co.;
Union Co.; Vermilion Co. Indiana: Tip-
pecanoe Co.; Wells Co. Iowa: Camp
Dodge; Boone Co.; Story Co. Kansas
Miami Co.; Pottawatomie Co. Kentucky
Breathitt Co.; Fayette Co. Maryland
Prince Georges Co. Massachusetts
Sherborn; Essex Co.; Hampden Co.
Middlesex Co.; Norfolk Co. Michigan
Branch Co.; Clinton Co.; Eaton Co.,
Gladwin Co.; Ingham Co.; Ionia Co.;

Livingston Co.; Midland Co.; St. Clair
Co.; Wayne Co. Minnesota: Carver Co.;
Clay Co.; Fillmore Co.; Goodhue Co.;
Hennepin Co.; Houston Co.; Mille Lacs
Co.; Nicollet Co.; Olmstead Co.; Pine
Co.; Pipestone Co.; Ramsey Co. Missis-
sippi: Lafayette Co. Missouri: Callaway
Co. Nebraska: Antelope Co. New Hamp-
shire: Grafton Co. New Jersey: Burling-
ton Co.; Camden Co.; Gloucester Co.
New York: Erie Co.; Greene Co.; Her-
kimer Co.; Steuben Co.; Tompkins Co.
North Carolina: Graham Co.; Johnston
Co.; Pender Co.; Wake Co. Ohio: BuUer
Co.; Clinton Co.; Delaware Co.; Erie
Co.; Franklin Co.; Greene Co.; Tuscara-
was Co. Pennsylvania: Cumberland Co.;
Dauphin Co.; Lancaster Co.; Philadel-
phia Co. Tennessee: Hamilton Co.;
Montgomery Co.; Sevier Co.; Shelby Co.
Virginia: Chain Bridge; Maywood;
Chesterfield Co.; Fairfax Co.; Giles Co.;
Montgomery Co.; Shenandoah Co. Wis-
consin: Dane Co.; Pierce Co.; Polk Co.;
Sauk Co.

Flight Period. Specimens have been
collected on dates ranging from 15 Feb-
ruary (Florida) to 11 August. Collection
dates, even from northern areas of the
range, predominate in May and June,
with only a few records from July and
August. More southern regions have col-
lection dates predominating in March
and April, tapering off in May.

Biology. The immature stages of this
species are not known. Adults have been
collected in Malaise traps and by sweep-
ing vegetation. Plants from which speci-
mens have been collected include Abies
balsamea (Ontario), alfalfa (Minnesota),
raspberries (New York), and Vibernum
edulae (Manitoba).

Remarks. This species is less common
than A. fuscitarsis, but not nearly as rare
as A. brevicornis. The identit\' of the
species has not been well established in
the past, due to the variable nature of its
color. Specimens with metallic reflec-
tions on the mesonotum and pale medial
areas on the abdominal tergites have

336 Bulletin Museum of Comparative Zoolof^y, Vol. 149, No. 6

been known as A. siiyiilis (Loew). The proved to he useful in separating these

lectotype specimen is evidently teneral, two taxa.

and specimens with this paler coloration

may well he freshly emerged, and may Genus BERIS Latreille

darken with age. Variation in the abdom- „ ■ i * n icno ^i-r * • c* *•

* , r 1 r "^'^i* Latreille, 1802: 447; type species Strattomys

inal coloration is prevalent in temales ot sexdentata Fabricius {=Musca chahjhata For-

A. fuscitarsis, and dark specimens have ster), by monotypy.

been misidentified as A. obscuriventris. Hfrnt/jcm Enderlein, 1921: 209; type species Bern

The mesopleural pollinosity is a much r'''^''''^^'\^''f" '^'''''^' ^^.?^r!f /^l';i?.T;lwl'^^"^'

,. ,, , ^, r V ^. bv original designation. NEW SYNONYMY.
more reliable character tor the separation

of A. obscuriventris from A. fuscitarsis, Diagnosis. Members of the genus Ber-
as noted in the diagnosis for each species, is may be distinguished from all other
Some variation has also been noted in the North American Beridinae by their ex-
genitalia of both sexes, but particularly in tremely small, one-segmented maxillary
the females. The male gonostyli vary to palpi. Other generic features include me-
some degree in size; the range is shown dium size with elongate body form, and
in Figures 27 and 28. The usual size is rather uniformly distributed pilosity;
nearer to the larger example. The overall eyes pilose in both sexes; males holoptic;
width of the genital capsule is somewhat females dichoptic, with wide frons with
variable, as is the length of the space be- nearly parallel margins; antennal flagel-
tween the gonocoxal apodemes and the lum more or less conical, usually eight-
main part of the gonocoxites. The overall segmented (seven in B. strobli); oral
form of the genitalia is distinctive how- margin somewhat produced ventrally;
ever, and I have examined a fair number thorax metallic greenish in Nearctic
of individuals in the course of determin- species; scutellum usually with six
ing the status of A. similis. I have no rea- spines, sometimes eight, progressively
son to suspect that more than one species smaller laterally, sometimes irregular;
is involved. The female genital furca is legs with hind pair elongate, the basitar-
also variable. In particular, the anterolat- sus of hind leg inflated in male; middle
eral margins of the sternite are variable, tibiae without apical spur; wings with R4
ranging from nearly straight to deeply present, R2+3 arising basal to r-m, M, and
emarginate. The most constant feature M2 sometimes petiolate to separate at dis-
seems to be the shape of the median ap- cal cell, M3 absent; abdomen brownish,
erture, which is pointed anteriorly. How- tergites two to six with preapical trans-
ever, the one female specimen examined verse grooves, that of tergite six faint,
from Colorado has the aperture with the The new synonymy of Hemiberis En-
anterior margin rather straight, as in A. derlein with Beris rather than with Ac-
fuscitarsis. More material needs to be tina is based upon the lectotype desig-
collected from this region before this nation for Beris qiiadridenfata Walker,
variant can be interpreted properly. I which is discussed below,
have figured two specimens with female The Nearctic species of Beris have re-
genitalia representing the extremes ob- mained rather poorly understood until
served (Figs. 29, 30). A more extensive now. The genitalia of both sexes, which
study may indicate that the female geni- offer the best diagnostic characters, have
talia of this species are more variable not been previously examined in detail
than presently believed, and are not use- nor figured for our species. I examined
ful for separating A. obscuriventris from these structures, and they have indicated
A. fuscitarsis in every case. They seem the presence of a third species never be-
to be reliable in most cases, however, and fore recorded from North America, and
corroborate other characters that have one synonymy with a Palaearctic species.

Nearctic Beridinae • Wnodley 337

In addition, examination of the holotypes
has revealed one further synonymy. Thus
there are three Nearctic taxa known at
present, two of which are also widely dis-
tributed in the Palaearctic Region. The
genus as a whole is more diverse in the
Old World, being found in the Palaearc-
tic and northern Oriental Regions.
Species referred to Beris in the Southern
Hemisphere belong in other genera. De-
spite examination of numerous Beridinae
from the Southern Hemisphere, I have
never seen specimens which are conge-
neric with true Beris, which is apparent-
ly restricted to the Northern Hemisphere.
Most of the southern species have only
four scutellar spines and all have well-
developed two-segmented palpi.

The life histories of the North Ameri-
can species are poorly known. The larvae
have never been collected (or at least as-
sociated with adults), and the only
knowledge of the immature stages for our
species is based upon European work.
Males probably form conspecific swarms,
for this behavior has been recorded an-
ecdotally for British species (Verrall,
1909). Males have large, holoptic eyes
and larger ommatidia dorsally, although
they are not clearly divided from the
smaller, lower ones. This is commonly
believed to be an adaptation coupled
with swarming behavior (McAlpine and
Monroe, 1968; Downes, 1969). The en-
larged hind basitarsus of male Beris may
also function in some connection either
with swarming or associated mating be-
havior.

Key to the Nearctic Species of Beris

1. Males (holoptic) 2

Females (dichoptic) 4

2. Tenth tergite with posterolateral surstyli

(Fig. 33); aedeagus long, arcuate, the ae-
deagal valves subequal in length to the
aedeagus (Fig. 34); posteromedial area of
ventral bridge of gonocoxites feebly pro-
duced (Fig. 36) fuscipes Meigen

Tenth tergite truncate posteriorly, without
surstyli (Figs. 37, 41); aedeagal valves ei-
ther longer or shorter than the aedeagus
itself (Figs. 38, 42) 3

3. Antennal flagellum seven-segmented (Fig.

9); medial region of ventral bridge of gon-
ocoxites strongly produced (Fig. 44); ae-
deagus much shorter than aedeagal

valves (Fig. 42)

strobli Dusek & Rozko.sny

Antennal flagellum long, eight-segmented
(Fig. 7); medial region of ventral bridge
of gonocoxites feebly produced (Fig. 40);
aedeagus longer than aedeagal valves
(Fig. 38) luteipes Johnson

4. Antennal flagellum eight-segmented (Figs.

6, 8); spermathecal ducts sclerotized pos-
teriorly (Figs. 45, 46); median aperture of
ninth sternite reduced, very small, most

of furca sclerotized ,5

Antennal flagellum seven-segmented (Fig.
9); spermathecal ducts unsclerotized pos-
teriorly (thus not shown in Fig. 47); me-
dian aperture of ninth sternite large, en-
compassing most of that surface

strobli Dusek & Rozkosny

5. Antennal flagellum quite conical, tapering

sharply toward apex (Fig. 6); sclerotized
portion of spermathecal ducts ver\' long,
twice as long as the rest of the genital
segment (Fig. 45); median hind margin of
furca posterior to median aperture bi-
lobed; widespread in mountainous and
boreal regions except on the Pacific

coast fuscipes Meigen

Antennal flagellum usually long, tapering
more gradually toward apex (Fig. 8);
sclerotized portion of spermathecal ducts
shorter, only slightly longer than the rest
of the genital segment (Fig. 46); median
hind margin of furca posterior to median
aperture tnmcate; found in British Co-
lumbia, Alberta, Washington, Oregon,
California, western Nevada, Idaho, and
possibly Wyoming luteipes Johnson

Beris fuscipes Meigen

Beris fuscipes Meigen, 1820: 8.

Beris brcvicornis Heyden in Loew, 1846: 284.

Beris quadridentata Walker, 1848: 127. NEW SYN-
ONYMY.

Oplacantha annulifera Bigot, 1887: 21. NEW SYN-
ONYMY.

Actina canadensis Cresson, 1919: 174.

Beris annulifera var. brunnipes Johnson, 1926b:
109.

Beris sachalinensis Pleske, 1926: 408.

Beris fuscotibialis Pleske, 1926: 409.

Beris sijchuanensis Pleske, 1926: 411.

Beris mongolica Pleske, 1926: 414.

Beris petiolata Frey, 1960: 80.

Tijpe Material. The type of Beris fus-
cipes Meigen is housed in the Museum

338 Bulletin Museum of Coiuparative Zoology, Vol. 149, No. 6

National d'Histoire Natiirelle, Paris. It The female holotype of Actina ccina-

was not available for study. densis Cresson is housed in the Academy

Loew (1846) first used the name Beris of Natural Sciences, Philadelphia, Penn-
brevicornis in connection with specimens sylvania. It is labeled as Type 6252, con-
sent to him by "Hr. v. Heyden." Loew trary to the published number of 6196.
obviously used the name as a synonym of The specimen lacks the right front leg,
B. fitscipes. It is apparently only a manu- but is otherwise in good condition,
script name, never having been formally The male holotype of Beris annulifera
proposed; thus no type material exists. var. brunnipes Johnson is in the Museum

Two syntypes of Walker's Beris quad- of Comparative Zoology (MCZ 7494). It

ridentata are housed in the British Mu- lacks the right antennal flagellum, but is

seum. The male is probably conspecific otherwise in excellent condition,

with Actina viridis (Say), but is badly Diagnosis. Males of Beris fuscipes

damaged. Thus I hereby designate the Meigen may be readily recognized by the

female specimen as lectotype, which presence of posterolateral surstyli (Fig.

bears the following labels: "9 Type/Ber- 33) on the tenth tergite, which may or

is quadridentata, 9 . Walker. (Type). = B. may not be cui^ved. This character is eas-

viridis. Say (O.S.)./ one of Walkers series ily visible without dissection. Females

so named EAW; [verso reads] Beris quad- are readily recognized by their terminalia

ridentata Walk./ LECTOTYPE Beris with the posterior portions of the sper-

quadridentata Walker 1848: 127 des. N. mathecal ducts being long and sclero-

E. Woodley 1980/ Beris fuscipes Meigen tized, and the medial region of the ninth

det. Woodley 1980." The specimen lacks sternite posterior to the small median ap-

the right antennal flagellum, the distal erture bilobed. Both sexes have antennae

portion of the fifth tarsomere of the right with a moderately short, conical flagel-

front leg, the left middle leg, and the fifth lum which is wide basally and which ta-

tarsomere of the right hind leg, but is oth- pers strongly toward the apex (Fig. 6).

erwise in good condition. The male spec- The maxillary palpi, although very small

imen is designated as paralectotype, and as in all Beris, are larger in this species

bears identical labels except that on the than in other North American forms. The

first two labels "d" replaces "9 ," and the differences are slight, however, and are

fifth label reads "Actina viridis (Say) det. difficult to appreciate without compari-

Woodley 1980." son with the other two species.

The female syntypes of Oplacantha Description. Male. Head black, frons

annulifera Bigot are also presently and face shiny, upper occiput pollinose;

housed in the British Museum. I hereby entire head including eyes pilose, color

designate one as lectotype, which bears of pile variable, ranging from mostly pale

the labels "Opl. annulifera Big. Col. Big/ with scattered dark hairs to almost wholly

Octacantha [sic] annulifera Big. AMERI- dark, usually a mixture of both; liairs at

CA ex Bigot Coll: BM 1960-539./ LEC- oral margin subequal in length to the first

TOTYPE Oplacantha annulifera Bigot two antennal segments combined, those

1887 des. N. E. Woodley 1979/ Beris fus- above antennae subequal to the length of

cipes Meigen det. Woodley 1979." The the first segment; antennae black, very

specimen lacks the left middle leg, the rarely with inner portion of first flagello-

right hind leg, and the second segment mere yellowish, shorter than length of

of the right cercus. The second specimen head; first two segments subequal, the

that I examined, with labels identical to second tapering to about twice width of

those of the lectotype, is hereby desig- base at apex; flagellum 1.3 to 1.7 times

nated as paralectotype, and is labeled ac- lengtli of first two segments combined,

cordingly. It is conspecific with the lee- basal widtli wider than apex of second

totype. segment, tapering strongly to a rather

Nearctic Beridinae • Wondley 339

sharp apex, slightly concave on inner sur-
face; pile of antennae black; proboscis
yellow; palpi minute, elongate oval; ocel-
li yellowish, ocellar tubercle slightly
prominent. Thorax with mesonotum
shiny green, often with bluish reflec-
tions, sometimes medial portion of meso-
notum and scutellum bluish, very rarely
brassy green; scutellum normally with
six spines; pleura brownish black, upper
portions with greenish reflections; tho-
racic pile mostly pale, sometimes with
scattered dark hairs on mesonotum,
length on mesonotum subequal to first
two antennal segments combined; legs
with front and hind coxae brownish
black, middle coxae yellowish, occasion-
ally dark; rest of legs yellowish except
front and middle legs with apical halves
of basitarsi and entire tarsomeres two to
five dark brown, and hind legs with
brownish subapical region on femora,
tibiae and basitarsi slightly brownish,
apical four tarsomeres dark brown; rang-
ing in darker specimens to wholly brown
except for yellowish extreme apices of all
femora, basal one-thirds of front and mid-
dle tibiae, and extreme bases of hind tib-
iae, also bases of hind femora evanes-
cently yellowish; wings brownish, evenly
set with microtrichia, stigma darker; hal-
ter pale yellow with darker stem. Abdo-
men brown, evenly pollinose dorsally ex-
cept for lateral margins, becoming less
pollinose on apical tergites, ventrally
shiny; pile pale, longest laterally; geni-
talia (Figs. 33-36) with gonostyli ovate,
not arcuate, often yellowish; medial pro-
cess of ventral bridge of gonocoxites fee-
bly produced, truncate (Fig. 36); aedea-
gal complex (Figs. 34, 35) very long,
strongly arcuate, lateral valves subequal
in length to aedeagus; tenth tergite (Fig.
33) with well-developed surstyli, straight
or curved; cerci yellowish. Length 5.0 to
6.9 mm.

Female. Differs from male as follows:
head with frons and face brownish black
to black; frons 0.23 to 0.35 width of head
at anterior ocellus, finely punctate, occa-
sionally with faint longitudinal striae be-

low ocelli; a small circular depression
present above antennae; pile of head
pale, that of face shorter than length of
first antennal segment; antennae brown-
ish black to black, with inner portions of
flagellomeres two to four slightly yellow-
ish, ranging in pale specimens to apical
portions of first and second segments, en-
tire first and second flagellomeres, and
inner portions of flagellomeres three to
six yellow; pile of first two antennal seg-
ments ranging from mostly dark to mostly
pale. Thorax with pale pile, shorter than
in male; legs wholly yellow except for
front and hind coxae, extreme apices of
all basitarsi, and entire tarsomeres two to
five of all legs brownish; ranging to as
dark as in darkest males, except that ex-
treme bases of all femora and all basitarsi
are yellowish. Abdomen with entire first
tergite, basal half of second tergite, and
extreme bases of tergites three to five pol-
linose, remainder shiny; pile pale, short-
er than in male; genitalia (Fig. 45) with
sclerotized posterior portions of sper-
mathecal ducts very long, twice as long
as the genital segment itself; median ap-
erture of genital stemite small; medial
hind margin of genital sternite posterior
to median aperture bilobed; cerci yellow-
ish, first segment slightly thicker than
second. Length 4.7 to 6.9 mm.

Distribution (Map 5 shows Nearctic
portion). Holarctic; ranging from Great
Britain east to Japan in the Palaearctic
Region, and in the Nearctic Region essen-
tially montane and boreal, from Alaska
south through the Rocky Mountains to
southern Arizona and New Mexico; from
British Columbia east across the northern
United States and Canada to Labrador,
Maine, south to Tennessee.

Localiti/ Records (190 males, 262 fe-
males). UNITED STATES, ALASKA:
Anchorage; Big Delta; Brooks Lake; Cur-
ry; Eagle River; Independence Mine;
Kenai Peninsula, 15 mi. SE of Anchor-
age; Salcha River, Alaska Highway mile
1481; Seward; Shaw Creek, Richardson
Highway mile 289. CANADA: Alberta:

340 Bullcliii Miisciini of Comparative Zoolo'^ij, Vol. 149, No. 6

Athabasca; Banff; Bilhy; Blairmore;
Cooking Lake, near Edmonton; Edmon-
ton; Gorge Creek; Gnll Lake; Jnmping
Pond Creek, 20 mi. W of Calgary; Kan-
anaskis; McMurray; 8 mi. E of Morley;
Nordegg; Opal; Pincher; Seebe; Waba-
mnn; Waterton. British Columbia: Can-
im Lake; Gronnd Hog Basin, Selkirk
Mountains; Liard Hot Springs, Alaska
Highway mile 496, 1500 ft.; Nass River,
Aiyansh; Salmon Arm; 6 mi. W of Ter-
race; Hudson Bay Territory: no further
data. Labrador: Parroquet Island;
Straights of Belle Isle. Manitoba: Min-
nitonas; Ninette; 5 mi. SW of Shilo. New-
foundland: Mtns. E of Codroy; St. Antho-
ny; Stephenville. Ontario: Britannia;
Finland; Kearny; Low Bush, Lake Abiti-
bi; Macdiarmid, Lake Nipigon; May-
nooth; Normandale; One Sided Lake;
Ottawa; Simcoe; Thornhill; Vermilion
Bay. Quebec: Bradore Bay; Duncan
Lake, near Rupert; Forestville; Gaspe;
Hull; Knob Lake, 54°47', 66°47'; Laniel;
Megantic; Mistassini; Old Chelsea; Ru-
pert House. Saskatchewan: Kenosee.
Yukon: Dawson; 14 mi. E of Dawson,
1500 ft.; Dempster Highwav mile 87;
Sheldon Lake, 13r37', 62°56', 3500 ft.
UNITED STATES: Arizona: Colter's
Ranch, White Mts.; Graham Co. Colo-
rado: Blackhawk; Buckhorn Creek; Green
i\It. Falls, 8000 ft.; Sierra Blanca, 11,500
ft.; Costilla Co.; Grand Co.; Gunnison
Co.; Larimer Co.; Rio Grande Co.; Teller
Co. Idaho: Latah Co. Maine: Penobscot
Co. Michigan: Cheboygan Co.; Chippe-
wa Co. Minnesota: Anoka Co.; Clear-
water Co.; Cook Co.; Lake Co.; Pope Co.;
Ramsey Co. Montana: Avalanche Lake,
Glacier Park; Gallatin Co.; Missoula Co.
New Hampshire: Pinkham Notch; White
Mts.; Carroll Co.; Coos Co. New Mexico:
Therma; Otero Co.; Sandoval Co.; Taos
Co. New York: Essex Co.; Greene Co.
Oregon: Baker Co. Tennessee: Indian
Gap, 5200 ft.. Great Smoky Mountains Na-
tional Park; Sevier Co. Utah: Cache Co.;
Grand Co.; Uintah Co. Vermont: Cale-
donia Co. Washington: Columbia Co.

Wiscoixsin: Florence Co.; Vilas Co. Wy-
oming: Albany Co.

Flight Period. Collection dates range
from 12 May to 11 August. Almost all
specimens were collected in June and
July.

Biology. The larva of Beris fuscipes
was briefly described by Lenz (1923),
and was collected under bark. Larvae
have not been collected in North Amer-
ica to my knowledge. Adults apparently
occur in situations similar to those of oth-
er species. I have seen series of B. fus-
cipes and B. strobli collected at the same
locality on the same date. Specimens
have been taken by sweeping, in Malaise
traps, and at light. One label also reads
"in marsh clearing" (British Columbia).
Records attributed to this species refer-
ring to their being collected on flowers
(McFadden, 1972) actually refer to B. lu-
teipes. Males probably form conspecific
swarms, for reasons discussed above.

Remarks. 1 have been unable to ex-
amine the type oi Beris fuscipes. It was ap-
parently last examined by Verrall (1909),
who stated that the specimen was a male.
It needs to be reexamined, however, par-
ticularly with reference to the genitalia. I
have also not examined the various types
for synonyms based on Palaearctic ma-
terial, for these have been adequately
treated by Rozkosny (1973) and Nartshuk
and Rozkosny (1975, 1976), who have
cleared up much of the synonymy that
has accumulated for the Palaearctic
species. These workers have also provid-
ed figures of the male genitalia which
have proved very useful in determining
the status of the Nearctic species of Ber-
is.

Beris fuscipes has long been called B.
annidifera in North America. The work
cited in the previous paragraph led me to
discover the synonymy of this name with
B. fuscipes. I examined European spec-
imens of the species and determined that
our material was conspecific, and con-
firmed the synonymy by examining the
original Bigot type material. The types

Nearctic Beridinae • Woodley 341

were stated as being from Georgia, but
labels on the specimens indicate only
"America," and I have not seen modern
material from Georgia, although it may
occur in the northern part of the state.

I have also examined the type material
for Beris quadridentata Walker. The
male specimen, upon which previous
synonymy with Actina viridis (Say) has
been based, is badly damaged. I there-
fore have designated the female, which
is in much better condition, as lectotype.
It is conspecific with Beris fuscipes
(Walker states that the female has six scu-
tellar spines, indicating a Beris, and I
have examined the terminalia to confirm
the identity of the lectotype). This has
resulted in two changes in synonymy.
The genus Hemiberis Enderlein, which
was based upon B. quadridentata as type
species, becomes a junior synonym of
Beris rather than oi Actina as it has been
in the past. And Beris quadridentata be-
comes a synonym oi B. fuscipes. Because
the lectotype designation does not affect
any name presently in use, and is very
unlikely to do so in the future, I felt it
was more advisable to designate a spec-
imen as lectotype that could be positively
identified. The damaged male lacks most
of the abdomen, including the genitalia.
Thus, if another species oi Actina is dis-
covered in the future from the Nearctic
Region, it would be difficult to determine
the identity of B. quadridentata if the
male was the lectotype, even though at
present it can be assigned to Actina vir-
idis (simply because there is only one
Nearctic Actina). Therefore, I feel the
above solution is most likely to result in
future stability of all names concerned.

The female type of Actina canadensis
Cresson was also examined, and I concur
with McFadden (1972), who synony-
mized it with B. annulifera, that it is not
specifically distinct. I likewise could not
detect the "steel blue vittae" first report-
ed by Curran (1927). I also agree with
McFadden that Beris annulifera var.
hrunnipes Johnson is merely a variant

not warranting subspecific recognition,
and I have confirmed that synonymy by
examining the holotype.

Beris fuscipes is the most widespread
species of the genus in the Nearctic Re-
gion, being found in boreal and moun-
tainous regions across the continent. It is
fairly variable, particularly with respect
to leg coloration. The coloration of the
legs was briefly studied by McFadden
(1972), who determined that it did not
vary in any way distinctly correlated with
geographical regions. It is interesting to
note that two paratypes, one each of B.
annulifera var. luteipes and B. a. var.
hrunnipes, were collected from the same
locality in New Hampshire. Thus it is not
feasible to maintain subspecific taxa for
this species.

There has been some variation detect-
ed in the shape and size of the male tenth
tergite. In some specimens, particularly
from Washington and Idaho, the tergite
is small and the surstyli are quite straight,
rather than being curved as in most spec-
imens. These variations are minor how-
ever, and other features of the male gen-
italia are quite constant and distinctive.

The female terminalia were illustrated
by Nagatomi and Iwata (1978). The prep-
aration was not dissected, however, so
that the structures of the genital furca are
not well depicted. These structures are
quite diagnostic for the species of Beris
in North America, and are illustrated
here for the first time.

Beris luteipes Johnson

Beris annulifera var. luteipes Johnson, 19261): 109.
NEW STATUS.

Beris californica James, 1939: 546. NEW SYNON-
YMY.

Type Material. The holotype of Beris
annulifera var. luteipes Johnson, from
Seattle, Washington, is presently housed
in the Museum of Comparative Zoology
(MCZ 7493). The specimen is a male,
and is lacking both third antennal seg-
ments, the left middle leg, both hind

342 BuUvtin Museum of Comparative Zoology, Vol. 149, No. 6

legs, and most of the left wing. The fe- width to the second segment, occasion-
male allotype is conspecific, and is also ally slightly wider; longer pile of anten-
from Seattle. All of the paratypes are B£?r- nae black; proboscis yellow; palpi mi-
is fuscipes Meigen. nute, elongate-oval; ocelli yellow, ocellar

The holotype of Beris californica tubercle somewhat prominent. Thorax
James, from Del Norte Co., California, is with mesonotum shiny metallic green to
presently housed in the collection of the blackish green, sometimes with brassy
Abteilung Taxonomie der Insekten des reflections; pleura blackish, dark brown-
Institutes fur Pflanzenschutzforschung ish below wing base, often with greenish
der Akademie der Landwirtschaftswis- reflections, particularly on mesopleuron
senschaften der DDR, Eberswalde, East and sternopleuron; scutellum concolor-
Germany. The male specimen is missing ous with mesonotum, normally with six
the left middle leg, but is otherwise in spines; pile ranging from dark with a few
excellent condition. scattered pale hairs to entirely pale,

Diagnosis. Males may be separated about length of first two antennal seg-
from other species of Beris occurring in ments combined, to somewhat longer;
North America by their genitalia, espe- legs with front and hind coxae blackish,
cially the combination of the truncate middle coxae mostly yellowish; rest of
tenth tergite (Fig. 37), the long aedeagus legs wholly yellow except for brownish
(Figs. 38, 39), and the feebly produced black coloration on apical third of basi-
process of the ventral bridge of the gon- tarsi and entire tarsomeres two to five of
ocoxites (Fig. 40); females may be rec- front and middle legs, and last four tar-
ognized by their short but sclerotized someres of hind legs; ranging in some
posterior portions of the spermathecal specimens to yellow except for wholly
ducts, which are only about as long as the brownish apices of front and middle fem-
ninth sternite, the truncate medial hind ora, hind femora entirely brownish, dark-
margin of the genital sternite posterior to er at apices, front and middle tibiae
the small median aperture (Fig. 46), and brown except for basal one-thirds, hind
the elongate, slender antennal flagellum tibiae dark brown except for basal one-
(Fig. 8). In much of its range it is the only fifths, and all tarsi wholly brown; wings
species of Beris. brownish, stigma darker, evenly set with

Description. Male. Head black, frons microtrichia; halter ranging from wholly
and face shiny, upper portion of occiput yellowish to wholly brownish. Abdomen
pollinose, entire head including eyes pi- brown to blackish brown, dorsally sub-
lose; pile black, occasionally with inter- shining, evenly and finely pollinose, ven-
mixed pale hairs below antennae; hairs trally shiny; pile ranging from wholly
at oral margin subequal to, to consider- pale to blackish dorsally and laterally, the
ably longer than, the first two antennal remainder pale, longest laterally; geni-
segments combined, those above anten- talia concolorous with abdomen, but
nae subequal to the first antennal seg- often with gonostyli and cerci yellowish;
ment, occasionally slightly longer; anten- gonostyli ovate, not strongly arcuate; pos-
nae black, occasionally with extreme teromedial region of ventral bridge of
apex of second segment, and inner por- gonocoxites truncate and feebly pro-
tions of flagellomeres one to five yellow- duced (Fig. 40); aedeagal complex (Figs.
ish; length ranging from shorter than the 38, 39) with aedeagus longer than lateral
length of head to subequal to it, first two valves; tenth tergite truncate posteriorly,
segments subequal, second widened api- without surstyli (Fig. 37). Length 5.3 to
cally; flagellum long and slender, 1.8 to 6.6 mm.

2.2 times length of the first two segments Female. Differs from male as follows:

combined (Fig. 7), base subequal in head with frons and face brownish black

Nearctic Beridinae • Woodley 343

to black, shiny, sometimes with bluish or Hot Springs, 5 mi. S of Lakelse; Mission
greenish reflections; irons 0.28 to 0.37 City; Osoyoos, Anarchist Mtn.; Terrace;
width of head at anterior ocellus, finely Tyee, 27 mi. E of Prince Rupert,
punctate, with small round depression UNITED STATES: California: Del
above antennae, rarely with faint longi- Norte Co.; Los Angeles Co.; Nevada Co.;
tudinal striae below ocelli; pile of head San Bernardino Co.; San Mateo Co.; Sis-
pale, very rarely darker below antennae, kiyou Co.; Trinity Co. Idaho: Clearwater
that efface at most the length of the first Co.; Latah Co. Nevada: Ormsby Co. (W
antennal segment; antennae black with of Carson City). Oregon: Benton Co.;
apex of second segment and inner por- Clackamas Co.; Curry Co.; Douglas Co.;
tion of basal flagellomere yellow, ranging Hood River Co.; Jefferson Co.; Marion
to having yellow coloration on apex of Co.; Tillamook Co.; Washington Co.
first segment, entire second segment, en- Washington: Chelan Co.; Clallam Co.;
tire first flagellomere, and inner portions Clark Co.; King Co.; Lewis Co.; Mason
of flagellomeres two to six; hairs of first Co.; Okanogan Co.; Pacific Co.; Skaman-
two antennal segments ranging from en- ia Co.; Snohomish Co.; Spokane Co.;
tirely pale to entirely black; flagellum 2.0 Whitman Co.; Yakima Co.
to 2.5 times length of first two antennal One further locality seems doubtful:
segments combined (Fig. 8). Thorax with Wyoming: Yellowstone Park, Clematis
mesonotum shiny green, more often with Creek. Further collecting may substanti-
bluish reflections than in male, pleura ate the presence of B. luteipes that far
often more extensively brownish; pile east, but in light of what is currently
usually pale, shorter than in male; legs known of its distribution, it seems un-
with hind coxae sometimes yellow, fem- likely. Thus this single male specimen,
ora and tibiae yellow, apex of hind fem- which I have examined, may be misla-
ora may be evanescently brownish. Ab- beled.

domen dark to light brown; tergites with Flight Period. Specimens of this species
most of segments one and two, and basal have been collected as early as 20 May,
portions of three, four, and five pollinose, and as late as 2 August, with the vast ma-
rest shiny; pile pale, shorter than in male; jority of records from June and July,
genitalia yellowish, posterior portions of Biology. The immature stages of this
spermathecal ducts sclerotized but short, species remain unknown. Adults have
about the length of the genital segment been collected in Malaise traps, both un-
itself (Fig. 46); median aperture of genital baited and baited with CO2. I have col-
sternite small, medial hind margin behind lected females in low vegetation along a
aperture truncate; cerci yellowish, first small, open stream at moderate elevation
segment thicker than terminal segment, in Washington. Label data indicate col-
Length 4.7 to 6.6 mm. lections from "marshy lake and stream

Distribution (Map 4). This species is margin," "plants in swamp," and "plants

endemic to the Nearctic Region, and is along river." The species has also been

found from British Columbia south to collected from flowers of Herac/eu/?? /a«-

California and extreme western Nevada, atum and "Umbelliferae" (both from

east to Alberta and Idaho. A record from British Columbia). The latter record was

northwestern Wyoming needs confirma- cited by McFadden (1972) under Beris

tion. annidifera ( = B. fuscipes); I have exam-

Locality Records (95 males, 103 fe- ined the specimen seen by him, and it

males). CANADA: Alberta: Waterton. was misidentified.

British Columbia: Bevan, Vancouver Is- Remarks. This species has been known

land; Crowsnest; Duncan, Coldstream, as Beris californica since it was first dis-

Vancouver Island; Hope, Silver Lake; covered to be specifically distinct by

344 Bttllcfin Muscidu of Comparative Zoolof^y, Vol. 149, No. 6

James (1939). Examination of the holo- hut unfortunately are not visible without

type of Bcris (itiniilifera var. luteipcs dissection.

Johnson, originally clescril)ed as a vari- Coloration of the legs, pilosity, and to
ety, but subsequently recognized as a some extent the mesonotum, is variable
subspecies by James (1965), indicated in this species. The length of the anten-
that it was a distinct species, conspecific nal flagellum is somewhat variable, and
with B. californica rather than with B. is sexually dimorphic (Figs. 7, 8), that of
fuscipes. This was subsequently con- the males being shorter and more com-
firmed by examination of the holotype of pact. In general, the palest specimens,
B. californica. Beris luteipes thus be- and those with the longest antennal fla-
comes the valid name for the taxon by gella, have been recorded from Califor-
priority. McFadden (1972) synonymized nia.
the variety with B. annulifera (=B. fus-
cipes), and indeed Johnson's original se-
ries was mostlv that species, but the ho-
lotype male is clearly conspecific with B. ^eris strobli DuSek and RozkoSny

californica. The holotype was collected geris chalybeata var. obscura StrobI, 1909: 47

at Seattle, Washington, which is outside (preoccupied by obscura Meigen, 1820: 4).

the range of Beris fuscipes. I can only Beris strobli Dusek and Rozkosny, 1968: 294 (new

conclude that McFadden did not exam- name for oZ^snm/ Strobl). ,^^, .^,,
., 11. . 1 . .. . 1 . Bens lattfascics Nagatomi and lanaka, 1972: 100.
ine the holotype to substantiate his syn-
onymy. Type Material. The holotype female of

Beris luteipes is the only species in the Beris chalybeata var. obscura StrobI is

genus that is endemic to the Nearctic Re- in the StrobI collection, presently housed

gion. It must be pointed out, however, at the Institutes fiir Pflanzenschutzfor-

that the structure of the genitalia indi- schung in Eberswalde, East Germany, in

cates a close relationship with Beris hep- the care of G. Morge. It is in poor con-

tapotomica Pleske (see Nartshuk and dition, missing the left antenna, the fla-

Rozkosny, 1975) which is known from the gellum of the right antenna, the two api-

USSR, and Beris sp. A of Nagatomi and cal tarsomeres of the left front leg, the

Tanaka (1972). I have not examined spec- apical tarsomeres of the right middle leg,

imens of either of these taxa. Nartshuk the entire hind legs, the left wing, and

and Rozkosny have noted the similarity the abdomen,

of B. heptapotomica to Beris sp. A. Diagnosis. Males can be told easily by

Females from the area of distributional the structure of the genitalia: the gono-
overlap with B. fuscipes can be difficult styli are large and strongly arcuate, and
to determine with certainty. The length the medial process of the ventral bridge
of the antennal flagellum is somewhat of the gonocoxites is well developed and
variable, but it is usually not as wide ba- bilobed (Fig. 44). Females are difficult to
sally as that of B. fuscipes, nor is it gen- determine positively, except by use of
erally as short. The range of overlap be- the genitalia, which have the spermathe-
tween B. luteipes and B. fuscipes is not cal ducts unsclerotized posteriorly, and
yet well understood, and if the record of the median aperture of the genital sternite
the former from Wyoming that now ap- very large (Fig. 47). Both sexes have com-
pears doubtful is substantiated the region pact, short, conical antennal flagella, which
may be larger than is presently anticipat- taper gradually toward the apex, and are
ed. Thus, it is desirable to check the gen- composed of seven segments (Fig. 9), al-
italia of females not associated with though this is difficult to detect. The pal-
males to identify them positively. The pi are very small, and also quite difficult
genitalia are relatively invariable, and to observe; this character can be used to
are quite diagnostic at the specific level, some extent to separate females from B.

Nearctic Beridinae • Woodley 345

fuscipes, which has hirger (although still

minute) palpi.

Description. Male. Head brownish
black to black, frons and face shiny, up-
per occiput pollinose; entire head includ-
ing eyes pilose, pile mostly black, but
pale hairs may be intermixed below an-
tennae; hairs at oral margin about the
length of the first two antennal segments
combined, those of frons about the length
of the first segment; antennae (Fig. 9)
shorter than the length of head, black, first
two segments sometimes brownish; scape
and pedicel subequal in length; flagel-
lum compact, blunt conical, about 1.7
times as long as the first two segments
combined, composed of seven flagello-
meres; pile of antennae black; proboscis
yellow; palpi very minute, more or less
spherical; ocelli yellowish, ocellar tuber-
cle somewhat prominent. Thorax with
mesonotum somewhat shiny, metallic
greenish; pleura brownish to blackish,
dorsal sclerites sometimes with greenish
reflections; scutellum concolorous with
mesonotum, normally with six spines;
pile of thorax yellowish to brownish,
about half the length of antennae; legs
brownish black with apices of femora and
basal one-fourth to one-third of all tibiae
yellow; ranging in some specimens to
mostly yellow except front and hind cox-
ae, apices of all basitarsi, and entire distal
four tarsomeres of all legs brownish;
wings brownish, stigma darker, evenly
set with microtrichia; halter yellow,
sometimes with apex of stem and base of
knob brown. Abdomen brownish, ter-
gites subshiny, very finely pollinose;
sternites shiny; pile pale, longest later-
ally; genitalia (Figs. 41-44) brown, gon-
ostyli large, arcuate, yellowish; postero-
medial portion of ventral bridge of
gonocoxites strongly produced, bilobed
(Fig. 44); aedeagal complex (Figs. 42, 43)
arcuate, aedeagus much shorter than lat-
eral valves; tenth tergite (Fig. 41) trun-
cate posteriorly, without surstyli. Length
5.3 to 6.2 mm.

Female. Differs from male as follows:
head brownish to brownish black; pile of

head mostly pale, short, at most the
length of the first antennal segment; frons
at anterior ocellus 0.29 to 0.37 width of
head, slightly to distinctly longitudinally
striate below ocelli, slightly depressed
just above antennae; antennae brownish
to black, with apex of second segment
and flagellum with inner portions of fla-
gellomeres one and two (sometimes rang-
ing as far as the sixth), yellowish; hairs of
antennae pale to brownish, those at apex
of flagellum black. Thorax with meso-
notum sometimes more brilliantly metal-
lic than in male; pile shorter, pale; legs
rarely as dark as in most males, often
wholly pale yellow with apical tarso-
meres brownish. Abdomen brownish,
shiny, terminal segments sometimes
slightly paler; cerci with first segment
thick, second much thinner, deflated
when dry, thus appearing laterally com-
pressed; genitalia (Fig. 47) with sper-
mathecal ducts unsclerotized posteriorly
(thus not shown in Fig. 47); median ap-
erture of genital sternite very large; me-
dial hind margin of genital sternite poste-
rior to aperture feebly bilobed. Length
4.9 to 5.9 mm.

Distribution (Map 4 shows Nearctic
portion). Holarctic; from western Europe
east to Japan in the Palaearctic Region
(Nartshuk and Rozkosky, 1976: 132, map);
from Alaska east to Ontario, Wisconsin,
and Minnesota in the Nearctic Region.

Localitij Records (24 males, 29 fe-
males). UNITED STATES: ALASKA: 15
mi. SE of Anchorage; Big Delta, Lake
Boleo; Richardson Highway mile 315;
Salcha River, Alaska Highway mile 1481;
Shaw Creek, Richardson Highway mile
289; Unalakleet. CANADA: Alberta: Bil-
by; Edmonton; Gull Lake; Jumping Pond
Creek, 20 mi. W of Calgary; McMurray;
Wabamun. British Columbia: Ad in, 3000
ft.; Ft. Nelson; Liard Hot Springs, Alaska
Highway mile 496, 1500 ft. Manitoba: 5
mi. SW of Shilo. Northwest Territories:
Ft. McPherson; Norman Wells. Ontario:
Macdiarmid, Lake Nipigon; One Sided
Lake. Yukon: Dawson; LaForce, 132°20',
62°41', 3300 ft.; North Fork Crossing,

346 Bulletin Musciini oj Comparative Zoology, Vol. 149, No. 6

Peel Pit. Road mile 43, 3500 ft.; Rampart
House; Swim Lakes, 133°, 62°13', 3200 ft.
UNITED STATES: Minnesota: Cool
Co., Grand Marais. Wisconsin: Door Co.

FliffJit Period. Collection dates range
from 28 May to 28 July. All specimens
except the one May record were taken in
June and July.

Biology. The immature stages of this
species have not been collected. Adults
have been collected from vegetation, no-
tably "swept ex Equisetum sp." (Mani-
toba) and from Salix (Yukon). The former
record was previously attributed to Beris
fiiscipes (as B. atinulifera) by McFadden
(1972), but reexamination of that speci-
men has shown it to be this species.

Remarks. This species has not previ-
ously been reported from the Nearctic
Region. This is surprising, since some
male specimens have been previously
determined as Beris annulifera {=Beris
fuscipes), yet they completely lack the
surstyli of the tenth tergite which are eas-
ily visible externally on male specimens
of B. fuscipes. Specimens of B. strobli
are not common in North American col-
lections, probably because the species is
primarily an inhabitant of far northern
areas that are seldom collected.

I have examined the type of B. chahj-
heata var. obscura Strobl. While it is so
badly damaged now that it is difficult to
be sure of its identity, I am following the
synonymy established in the literature by
Dusek and Rozkosny (1968). Through the
kindness of Rudolf Rozkosny, I was able
to examine other European specimens of
B. strobli, and was able to confirm the
conspecificity of Nearctic material. Nart-
shuk and Rozkosny (1976) proposed the
synonymy of B. latifascies Nagatomi and
Tanaka, described from Japan. I have not
yet examined the holotype of that taxon.

This species is likely to be confused
only with Beris fuscipes, with which it is
widely sympatric in northern regions.
Males are easily determined by their
genitalia, which are visible externally.
But females are difficult to identify with

certainty unless specimens of both
species can be compared. With single fe-
male specimens it is necessary to exam-
ine the internal genitalia. But external
characters sometimes of use include the
frons, which tends to be wider in B.
strobli than in specimens of B. fuscipes,
and definite longitudinal striations are
usually present on the frons below the
ocelli. These two characters do overlap
to some degree in the two species, how-
ever. The antennal flagellum is distinc-
tive for each species, but this is difficult
to appreciate without having seen both.
The flagellum is only seven-segmented
in B. strobli, which has not been noted
in the past. This is difficult to ascertain
without clearing the flagellum. Speci-
mens of both Beris fuscipes and B. strob-
li have been collected at the same local-
ity on the same date, so overlap of ranges
definitely occurs at the local level. Thus
for positive identification of females, it is
best to examine the genitalia.

Beris strobli probably does not occur
in sympatry with Beris luteipes, except
possibly in southeastern British Colum-
bia or southwestern Alberta. They have
not as yet been recorded from the same
locality.

Genus EXODONTHA Rondani

Exodontha Rondani, 1856: 169; type species Exo-
dontha pedemontana Rondani {=Beris dubia
Zetterstedt), by original designation.

Hexodonta, emend.

Acanthomyia Schiner, 1860: 49; type species Beris
dubia Zetterstedt, by original designation.

Scoliopelta Williston, 1885: 152; type species Sco-
liopelta luteipes Williston, by monotypy.

Diagnosis. Members of the genus £.v-
odontha may be recognized by their rath-
er large, robust form and the lack of
preapical transverse grooves on the ab-
dominal tergites. Other characters in-
clude eyes densely pilose, holoptic in
males, widely separated in females; an-
tennae with basal two segments sub-
equal, flagellum with eight flagello-
meres; palpi two-segmented; thorax quite
convex, scutellum with three to four pairs

Nearctic Beridinae • Woodley 347

of spines; middle tibia widi one small
apical spur; wing with R4 and r-m pres-
ent, Mj and M2 petiolate to separate at
discal cell, M3 present, but abbreviated,
not reaching wing margin; stigmal area
elongate, narrow; abdomen broad, the
first five segments large, with six and sev-
en telescoping into the abdomen.

Exodontha is not likely to be confused
with any other North American genus of
Beridinae because of its large, robust
size. AUognosta is the only other genus
with a spur on the middle tibia, but it is
much smaller, possesses preapical trans-
verse grooves on the abdominal tergites,
and has no spines on the scutellum. The
genus Exodontha, as recognized here, is
a genus with two species, both of which
are found in the Nearctic Region. One of
these is also found in northern areas of
the Palaearctic Region. James (1973) syn-
onymized Exodontha with Afitissa Walk-
er, a genus based on the type species
Clitellaria cuprea Walker from Australia.
I have examined a number of the Austra-
lian and Neotropical forms assigned to
the "Antissini," and while they may well
be closely related to Exodontha, I cannot
concur with the above synonymy. The
form of the antennae and male genitalia
is quite different in these Southern
Hemisphere forms; thus at present it is
best to consider Exodontha generically
distinct. It is apparent from the generic
diagnosis that this genus is rather diver-
gent from the other Beridinae. The lack
of transverse preapical tergal grooves on
the abdomen, and the reduced sixth and
seventh abdominal segments are excep-
tional in this regard. A revision of the
genera of Beridinae is needed to clarify
the relationships of these genera, and
such studies may well indicate that the
genera related to and including Antissa
and Exodontha should be removed from
the subfamily. Hardy (1932) and James
(1973) consider the group to be a com-
ponent of the diverse subfiimily Clitel-
lariinae, and it has occasionally been ac-
corded subfamilial rank (White, 1916;

Daniels, 1978). I prefer to assign Exo-
dontha to the Beridinae at present simply
because it has traditionally been placed
in that subfamily in recent revisions and
catalogs (James,' 1965; McFadden, 1967;
Rozkosny, 1973) that deal with the Nearc-
tic and Palaearctic faunas, and because
its affinities have not otherwise been ad-
equately documented.

Key to the Species of Exodontha

1. Legs usually with femora and tibiae dark
medially; gonostyli of males with dorso-
medial process well developed, widely
diverging from the main lobe (Figs. 51,
52); aedeagus subequal in length to the
aedeagal valves (Fig. 49); ventral patch of
spinules on aedeagal complex large (Fig.
50); western and northern in distribu-
tion dubia (Zetterstedt)

Legs usually wholly yellow; gonostyli of
males with dorsomedial process poorly
developed, hardly diverging from the
main lobe (Figs. 57, 58); aedeagus shorter
than the aedeagal valves (Fig. 54); ventral
patch of spinules on the aedeagal com-
plex small (Fig. 55); eastern, Appalachian
distribution luteipes (Williston)

Exodontha dubia (Zetterstedt)

Beris dubia Zetterstedt, 1838: 512.
Exodontha pedemontana Rondani, 1856: 169.
Scoliopeha grandis James, 1938: 156. NEW SYN-
ONYMY.

Type Material. A male specimen in the
Zetterstedt collection at the Entomolog-
ical Museum, Zoological Institute, Lund,
Sweden, is labeled as the lectotype of
Beris dubia (by Rozkosny). The speci-
men is missing both third antennal seg-
ments, the left front leg beyond the tro-
chanter, the last tarsomere of the right
front leg, and has a little mold on it.

I have not examined the type of Exo-
dontha pedemontana Rondani, and do
not know where it is presently housed.

The holotype female of Scoliopeha
grandis James is presently housed in the
James Entomological Collection at Wash-
ington State University, Pullman, Wash-
ington. It was originally said to be in the

348 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 6

collection of" Oregon State University
(James, 1938). The holotype is missing
the right antennal flagellum, the left front
tarsus, the last tarsomere of the right front
leg, both middle legs (which are glued to
the locality label), and the right wing.

Diagnosis. This species is character-
ized primarily by the male genitalia, as
outlined in the key. The shape of the
gonostylus with its large dorsomedial
process is particularly useful because it
may be viewed without dissection. Fe-
males are very difficult to identify on a
morphological basis; at present they may
be identified most reliably by where they
were collected, i.e., from the western and
northern parts of the North American
continent. The leg coloration is frequent-
ly but not always dark, while the legs of
£. liiteipes are nearly always wholly yel-
low.

Description. Male. Head black, frons
and face strongly pollinose; face slightly
concave toward middle, with irregular,
oblique, longitudinal striations; entire
head pilose, hairs of ocellar tubercle
blackish, pile of eyes brownish, rest of
head with pale yellowish hairs, some-
times a few darker hairs present on face;
length of pile fairly uniform, about 1.5
times as long as the first antennal seg-
ment, that on lower face longer, and long-
est on genae where it is almost as long as
the antennal flagellum; antennae 0.6 to
0.7 length of head; flagellum elongate
conical, about twice as long as both basal
segments combined; antennae dark
brown to black, apical margin of second
segment may be yellowish, inner portion
of flagellum occasionally very faintly
lighter; longer hairs on antennae a mix-
ture of pale and dark; palpi black, basal
segment may be brownish; proboscis
brownish yellow. Thorax blackish, often
with metallic greenish reflections, dense-
ly but finely punctate; humeri, postalar
calli, lateral part of mesonotal suture, and
sometimes extremities of scutellum
brownish yellow to yellowish; pleura
concolorous with mesonotum but usually

without metallic reflections, sometimes
brownish around sutures, the sclerites
below the wing base usually brownish;
almost entire thorax clothed with pale
whitish yellow hairs, sometimes darkish
on mesal portion of mesonotum and scu-
tellum, averaging a little longer in length
than both basal antennal segments com-
bined; posterior part of pteropleuron and
almost entire hypopleuron bare; legs
dark brown except for yellowish color-
ation on apices of all femora, bases and
extreme apices of all tibiae, and basal
two-thirds of basitarsi and basal halves of
tarsomeres two and three of all legs;
wings light grayish brown, evenly set
with microtrichia, stigma darker but not
strongly contrasting with rest of wing;
halter yellow, stem sometimes brownish.
Abdomen dark brown to blackish, finely
punctate; basal rows of pits on tergites
two to five rather confused, individually
indistinct; fifth tergite with fine, trans-
verse sculpturing; pilosity mostly dark,
short medially, longest laterally, sternites
with appressed, short, pale yellowish
pile; hypopygium yellowish brown, tenth
tergite rounded posteriorly (Fig. 48);
genitalia (Figs. 49-52) with ventral
bridge of gonocoxites without medial
process; gonostyli large, with dorsome-
dial process large, diverging strongly
from main lobe (Figs. 51, 52); aedeagus
subequal in length to the aedeagal
valves, which are attenuated posteriorly
(Fig. 49); spinules of aedeagal complex
forming a large patch (Fig. 50). Length
7.4 to 8.8 mm.

Female. Differs from male as follows:
head may be brownish; lower frons and
face pollinose, sharply delimited from the
finely punctate upper frons by a transverse
line; width of frons at anterior ocellus 0.40
to 0.47 width of head; upper frons with a
very shallow medial longitudinal depres-
sion; frontal and occipital eye margins
very narrowly pollinose; pilosity of head
shorter than in male, not longer than the
length of the first antennal segment, pale,
except darkish on eyes. Thorax with pile

Nearctic Beridinae • Woodley 349

shorter than that of male, wholly pale, fused until now, in part due to their rar-
semiappressed on mesonotimi; legs may itv. All specimens have been known
be lighter in color than in male, in the under the name Exodontha luteipes (Wil-
palest specimens they may be wholly liston), except for two female specimens
yellow. Both thorax and abdomen with that were identified as E. gra?ic/i.s (James)
more extensive brown areas than found (the holotype and one additional speci-
in males, occasionally almost wholly men). It has been discovered during the
brown. Female terminalia with no dis- course of this work that specimens from
tinctive specific features, essentially the the northwestern United States and Can-
same as figured for E. luteipes (Fig. 56). ada were conspecific with Exodontha du-
Length 7.8 to 11.3 mm. hia (Zetterstedt), a well known but rare

Distribution (Map 2 shows Nearctic species previously known only from the

portion). Holarctic; ranging from Scan- Palaearctic Region. I have examined the

dinavia and mountainous regions in cen- Zetterstedt type which confirmed this

tral Europe east to Japan in the Palaearc- identity.

tic Region (Nartshuk and Rozkosny, 1975: The holotype female of Scoliopelta

88, map) and from the Pacific Northwest grandis James was also examined, and I

east to Quebec in the Nearctic Region. have concluded that it is a junior syn-

Locality Records (10 males, 10 fe- onym of £. dtibia. It is unfortunate that
males). CANADA: British Columbia: the specimen is a female, as females are
Hope Mts. (19); Robson (26, 29); St. very difficult to identify reliably in this
Marys (19); Skeets Falls (IS). Quebec: genus. The female genitalia have not
Mistassini Post (Id). UNITED STATES: proved diagnostic as they have in other
Idaho: Kootenai Co., 10 mi. N of Harri- Beridinae. The brownish color, the
son on Lake Coeur d'Alene, 2100 ft. (1 9); length of Mg, and the separation of M,
Shoshone Co., Wallace (3c?, 29); Valley and Mg at the discal cell, all of which were
Co., 1-10 mi. S of Smith's Ferry (1(5). characters used to characterize E. gran-
Oregon: Baker Co., Lower Goose Creek, dis, are variable within this species. Con-
36 mi. SE of Union, 4000 ft. (1(5); Clatsop sequently, I feel that this taxon is best

Co., Cannon Beach (19). Washington
Asotin Co., Fields Spring State Park (19)
Grays Harbor Co., 2 mi. S of Queets (IS)

relegated to synonymy under E. dubia
until evidence is discovered to the con-
trary. It should be pointed out that

Whitman Co., Kamiak Butte (19). McFadden (1972) also suspected this

Flight Period. Dates of collection range synonymy,
from 3 June to 12 August. As mentioned above, the coloration of
Biology. McFadden (1967) found and this species is variable. It is possible that
reared larvae of what is almost certainly this is simply a reflection of the maturity
this species (from Banff National Park, of specimens at capture. The female gen-
Alberta), which were identified as E. hi- italia are not specifically distinct; they
feipes, since the identity of western spec- are illustrated for E. luteipes only, for the
imens was not known at that time. They sake of comparison with other genera
were found in "moist rotting wood under treated in this paper. At present, females
large boulders on a mountainside at an are best identified by where they were
elevation of approximately 6000 ft." collected, but due to the rarity of both
Adults occur in forested areas, but little species in the genus, their distributions
is known of their habits. Specimens have are incompletely known. To give some
been collected by sweeping and in Ma- indication of the reliability of the locality
laise traps baited with CO2. records cited in this paper, I have includ-
Remarks. The identities of the Nearc- ed the sexes of the specimens examined
tic members of this genus have been con- from each locality in that section.

350 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

Exodontha luteipes (Williston)

Scoliopelta luteipes Williston, 1885: 154.

Type Material. The lectotype of Sco-
liopelta luteipes Williston is presently
housed in the Snow Entomological Mu-
seum at the University of Kansas, Law-
rence, Kansas. The male specimen is in
perfect condition. The lectotype was des-
ignated by McFadden (1972; the second
original label reads "Aug.," not "A. Fig."
as indicated by McFadden).

Diagnosis. Male genitalic characters
outlined in the key are the only reliable
characters useful for determining mem-
bers of this species. The small dorsome-
dial process of the gonostylus may be
viewed without dissection. Female spec-
imens may be identified most reliably by
their collection sites; the species has an
Appalachian distribution. A character of
limited usefulness is the leg coloration.
The legs are wholly pale in all specimens
examined except one, whereas they are
frequently darker in E. dtibia, often hav-
ing the femora and tibiae with extensive
dark brown coloration.

Description. Male. Similar to E. dubia,
but differing as follows: head with eye
pilosity sparser, subequal in length to the
first antennal segment; antennae with
basal two segments brownish, yellowish
at apex of second, inner portions of flag-
ellomeres three to five with distinct light
brownish area. Thorax with mesonotum
with metallic reflections rather more
bronzy; lateral areas paler, brownish yel-
low, irregularly more extensive than in
E. dubia, extending along lateral margins
and often onto scutellum; pleura brown-
ish, with the areas which are brownish in
E. dubia quite yellowish, including
sclerites below wing base; pile of thorax
more golden yellow; legs wholly orangish
yellow, femora may be faintly darker.
Abdomen brownish, sometimes with an
almost purplish cast; basal rows of pits
on tergites two to five with individual
pits distinct; pilosity pale, golden col-
ored; genitalia (Figs. 53-55, 57-58) with

gonostyli only slightly bilobed, the dor-
somedial process reduced (Figs. 57, 58);
aedeagus shorter than aedeagal valves,
which are truncate apically (Fig. 54); spi-
nules on ventral surface of aedeagal com-
plex less extensive (Fig. 55); tenth tergite
truncate posteriorly (Fig. 53). Length 7.1
to 8.1 mm.

Female. Extremely similar to E. dubia;
frons 0.37 to 0.47 width of head; color-
ation more brownish than in males, es-
pecially the thorax; legs as in male, ex-
cept in one specimen they are similar in
coloration to those of E. dubia (see be-
low). Genitalia (Fig. 56) very similar to
those of E. dubia. Length 7.0 to 9.3 mm.

Distribution (Map 2). Known from the
Appalachian region of eastern North
America, from Vermont and New Hamp-
shire south to North Carolina.

Localiti/ Records. (14 males, 8 fe-
males). UNITED STATES: Massachu-
setts: Hampden Co., Chester {76, 19).
New Hampshire: White Mts. (2d, 19);
Grafton Co., Franconia (16). New York:
Essex Co., Upper Ausable (1<5); Orange
Co., Bear Mtn. (1 9); Schuyler Co., Texas
Hollow (19). North Carolina: Macon
Co., Highlands (Horse Cove) (1(5); Tran-
sylvania Co., Pisgah Forest, Looking
Glass Park {16, 59). Vermont: Camel's
Hump (19); Rutland Co., Rutland (2 9);
Windsor Co., Woodstock (1<5).

Flight Period. Specimens have been
collected from 19 July to 30 August.

Biology. Nothing is known about this
species, either for the immature stages or
adults.

Remarks. The name Exodontha lu-
teipes has long been used for nearly all
North American specimens of this genus.
Examination of the lectotype of this
species has confirmed the application of
this name to this taxon, which was found
to be distinct from specimens from Can-
ada and the western United States, which
are properly assigned to E. dubia for the
first time in this paper. Exodontha lu-
teipes is apparently restricted to the Ap-
palachian region, and is not known to be

Nearctic Beridinae • Woodley 351

sympatric with E. diihia. One specimen,
a female from Chester, Massachusetts, in
the collection of the Museum of Compar-
ative Zoology, has leg coloration that is
similar to that found in E. ditbia, how-
ever, and it may be that species. It would
appear that if the two species were to
have overlapping distributions, it would
most likely be in the northern part of the
range of E. luteipes. Thus a Massachu-
setts locality for E. dubia seems at least
possible. Specimens of both species are
quite scarce, and a full understanding of
their distributions and identities must
await the collection of more material, es-
pecially males.

LITERATURE CITED

Bigot, J. M. F. 1887. Dipteres nouveaux ou peu
connus. SP partie, XXXIX: Descriptions de
nouvelles especes de Stratiomydi et de Co-
nopsidi; 32*^ partie, XL: Descriptions de nou-
velles especes de Myopidi. Soc. Ent. Fr. (Ann.
ser. 6), 7: 20^6, 203-208.

BONHAG, P. F. 1951. The skeleto-muscular mech-
anism of the head and abdomen of the adult
horsefly (Diptera: Tabanidae). Trans. Amer.
Entomol. Soc, 77: 131-202.

Cole, Fr\.\k R. 1969. The Flies of Western North
America. Berkeley and Los Angeles, Universit>'
of Cahfornia Press, 693 pp.

COQUILLETT, D. W. 1910. The type-species of the
North American genera of Diptera. Proc. U.S.
Natl. Mus., 37: 499-647.

Cresson, E. T. 1919. Dipterological notes and de-
scriptions. Proc. Acad. Nat. Sci. Phila., 71: 171-
194.

CuRRAN, C. H. 1924. The generic postion oi Beris
viridis Sav (Stratiomyidae, Diptera). Can. Ento-
mol., 56:24.

. 1927. Synopsis of the Canadian Stratio-

mvidae. Proc. and Trans. Roy. Soc. Can. (Ser.
3)^ 21(5): 191-228.

Da.viels, G. 1978. A catalog of the type specimens
of Diptera in The Australian Museum. Rec.
Aust. Mus., 31: 411-471.

DowNES, J. A. 1969. The swarming and mating
flight of Diptera. Annu. Rev. Entomol., 14:
271-298.

DuSek, J., AND R. RozkoSny. 1963. Revision mit-
teleurop'aischer Arten der FamiHe Stratiomyi-
dae (Diptera) mit besonderer Beriicksichtigung
der Fauna der CSSR. I. Acta Entomol. Bohe-
moslov., 60: 202-221.

. 1968. Beris strobli nom. nov. (Diptera,

Stratiomyidae). Reichenbachia, 39: 29.3-298.

Enderlein, G. 1921. Uber die phyletisch iilteren
Stratiomyidensubfamilien (Xylophaginae, Chi-
romyzinae, Solvinae, Beridinae, und Coeno-
myiinae). Mitt. Zool. Mus. Berlin, 10: 151-214.

Frey, R. 1960. Orientalische Stratiomyiiden der
Subfamilien Beridinae und Metoponiinae
(Dipt.). Notul. Entomol., 40: 7.3-85.

Hanson, W. J. 1958. A revision of the subgenus
Melanonemntelus of America north of Mexico
(Diptera: Stratiomvidae). Kansas Univ. Sci.
Bull., 38: 1351-1391.

Hardy, G. H. 1932. Notes on Australian Stratio-
myiidae. Proc. Roy. Soc. Queensl., 44: 41-49.

James, M. T. 1938. A second species of Sco/r'ope/fa
(Diptera, Stratiomyidae). Pan-Pacific Entomol.,
14: 1.56-157.

. 1939. A review of the Nearctic Beridinae

(Diptera, Stratiomvidae). Ann. Entomol. Soc.
Amer., 32: 543-548.

. 1965. Family Stratiomyidae, pp. 299-319.

In A. Stone et al. (eds.), A catalog of the Diptera
of America north of Mexico. U.S.D.A. Agricul-
ture Handbook 276. Washington, D.C., iv -I-
1696 pp.

. 1973. Family Stratiomyidae, No. 26. In \

catalog of the Diptera of the Americas south of
the United States. Sao Paulo, Museu de Zoolo-
gia, Universidade de Sao Paulo, 9.5 pp.
-. 1975. Family Stratiomyidae, pp. 14^2. In

M. D. Delfinado and D. E. Hardy (eds.), A cat-
alog of the Diptera of the Oriental Region. Vol-
ume II. Suborder Brachycera through division
Aschiza, suborder Cyclorrapha. Honolulu, Uni-
versity Press of Hawaii, 459 pp.

Johnson, C. W. 1923. New and interesting species
of Diptera. Occas. Pap. Boston Soc. Nat. Hist.,
5: 69-72.

. 1926a. The svnonymv of Actina viridis

(Say). Psyche, 33: '88-91.

1926b. A note on Beris annulifera (Bigot).

Psyche, 33: 108-109.

Latreille, P. A. 1802. Histoire naturelle, gene-
rale et particuliere, des Crustaces et des In-
sectes, Tome troisieme, (Vol. 95), 468 pp. In C.
S. Sonnini (ed.), Histoire naturelle par Buffon,
nouvelle edition accompagnee de notes. Dufart,
Paris, "An X."

Lenz, F. 1923. Stratiomyiiden aus Quellen. Ein
Beitrag zur Metamorphose der Stratiomviiden.
Arch. Naturgesch., 89(Abt. A, 2 Heft): 39-62.

LiNDROTH, C. H., A.ND R. Freitag. 1969. North
American ground-beetles (Coleoptera, Carabi-
dae, excluding Cicindelinae) described by
Thomas Say: designations of lectot\'pes and
neotypes. Psyche, 76: 326-361.

LOEW, H. 1846. Benierkungen iiber die Gattung
Beris und Beschreibung eines Zwitters von
Beris nitens. Stettin. Entomol. Ztg., 7: 282-289.

. 1863. Diptera Aniericae septentrionalis in-

digena. Centuria quarta. Berliner Entomol.
Ztschr., 7: 275-326.

352 Btillctiit Museum of Comparative Zoology, Vol. 14fJ, No. 6

McAlpine, J. F., AND D. D. Monroe. 1968.
Swarming of loncluieid flies and other iiiseets,
with descriptions of four new species of Lon-
chaeidae (Diptera). Can. Entoniol., 100: 1154-
1178.

McFadden, M. W. 1967. Soldier fly larvae in
America north of Mexico. Proc. U.S. Natl. Mus.,
121: 1-72

. 1972. The soldier flies of Canada and Alas-
ka (Diptera: Stratiomyidae). I. Beridinae, Sar-
ginae, and CliteUariinae. Can. Entomol., 104:
531-562.

MeigE.N, J. VV. 1804. Klassifikazion und Beschrei-
bung der europiiischen zweifliigeligen Insec-
ten (Diptera Linn.). Braunschweig, Reichard,
Erster Band. Abt. I, pp. i-xxviii + 1-152, pis.
1-8; Abt. II, pp. i-iv + 153-314, pis. 9-15.

. 1820. Systematische Beschreibung der be-

kannten europaischen zweifliigeligen Insekten
Aachen, Forstmann, Vol. 2, x + 365 pp., pis.
12-21.

Melander, a. L. 1904. Notes on North American
Stratiomvidae. Can. Entomol., 36: 14-24, 53—
54.

MiCHENER, C. D. 1944. Comparative external mor-
phology, phylogeny, and a classification of the
bees (Hvmenoptera). Bull. Amer. Mus. Nat.
Hist., 82: 151-326.

Nagatomi, a., and K. Iwata. 1976. Female ter-
minalia of lower Brachycera — I (Diptera). Beitr.
Entomol., 26: 5-47.

. 1978. Female terminalia of lower Brachy-
cera—II (Diptera). Beitr. Entomol., 28: 26.3-
293.

Nagatomi, A., and A Tanaka. 1969. The Jap-
anese Actina and Allognosta (Diptera, Stratio-
mvidae). Mem. Fac. Agric. Kagoshima Univ., 7:
149-176.

. 1972. The Japanese Ben's' (Diptera, Stratio-

m\'idae). Mem. Fac. Agric. Kagoshima Univ., 8:
87-113.

Nartshuk, E. p., and R. RozkoSny. 1975. New
distributional data on Palearctic Beridinae with
redescriptions of some little known species
(Diptera, Stratiomyidae). Scripta. Fac. Sci. Nat.
Ujep Brun., Biol. 2, 5: 81-90.

. 1976. Taxonomic and distributional notes

on some Palearctic Beridinae (Diptera, Stratio-
myidae). Acta. Ent. Bohemoslov., 73: 128-134.

OSTEN Sacken, C. R. 1883. Synonymica concern-
ing exotic dipterology. No. II. Berliner Ento-
mol. Ztschr., 27: 295-298.

Pleske, T. 1926. Etudes sur les Stratiomyidae de
la region palearctique (Dipt.) — Revue des es-
peces palearctiques des sousfamilles Sarginae
et Berinae. Eos, 2: .385-420.

RONDANI, C. 1856. Dipterologiae italicae prodro-
mus. Vol. 1. Genera italica ordinis dipterorum
ordinatim disposita et distincta et in familias et
stirpes aggregata. Parma, 228 pp.

. 1863. Diptera exotica revisa et annotata.

Modena, 99 pp.

RozkoSny, R. 1973. The Stratiomioidea (Diptera)
of Fennoscandia and Denmark. Fauna Ento-
mol. Scand., 1: 1-140 + catalog.

Say, T. 1823. Descriptions of dipterous insects of
the United States. Jour. Acad. Nat. Sci. Phila.,
3: 9-54, 73-104.

. 1824. Appendix. I. Natural history. 1. Zo-
ology. E. Class Insecta, pp. 268-378. In W. H.
Keating, Major Long's second expedition. Vol.
2. Philadelphia, Carey & Lea, 459 pp.

SCHINER, J. R. 1860. Vorliiufiger Commentar zum
dipterologischen Theile der "Fauna Austria-
ca," mit einer niiheren Begriindung der in der-
selben aufgenommenen neuen Dipteren-Gat-
tungen. I. Wien Ent. Monatschr., 4: 47-55.

Strobl, p. G. 1909. Die Dipteren Von Steier-
mark. — II. Nachtrag.-Mitt. Naturvv. Ver. Steier-
mark, 46: 45-293.

Verrall, G. H. 1909. Stratiomyidae and succeed-
ing families of the Diptera Brachvcera of Great
Britain. Vol. 5, 780 pp. In G. U. Verrall (ed.),
British Flies. London, Gurney & Jackson.

Walker, F. 1848. List of the specimens of dipter-
ous insects in the collection of the British Mu-
seum. London, British Museum, Pt. 1, pp. 1-
229.

White, A. 1916. A revision of the Stratiomyidae of
Australia. Proc. Linn. Soc. New South Wales,
41: 71-100.

Wiedenlann, C. R. W. 1828. Ausseroeuropaische
zweifliigelige Insekten. Hamm, Vol. 1, xxxii +
608 pp.

■. 1830. Ausseroeuropaische zweifliigelige

Insekten. Hamm, Vol. 2, xii + 684 pp.

WiLLISTON, S. W. 1885. On the classification of
North American Diptera. Entomol. Amer., 1:
152-155.

Zetterstedt, J. W. 1838. Dipterologis Scandi-
naviae. Sect. 3: Diptera, pp. 477-868. In J. W.
Zetterstedt (ed.), Insecta Lapponica. Lipsiae,
vi + 1140 pp.

Nearctic Beridinae • Woodley 353

Map 1. Distribution of Actina viridis (Say).

354 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

Map 2. Distributions of Beridinae: •, Allognosta brevicornis Johnson; A, Exodontha luteipes (Williston); ■, Nearctic
distribution of Exodontha dubia (Zetterstedt).

Nearctic Beridinae • Woodley 355

Map 3. Distribution of Allognosta obscuriventris (Loew).

35(-) Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

Map 4. Distributions of Beridinae: •, Allognosta fuscitarsis (Say); A, Nearctic distribution of Beris strobli DuSel< and
RozkoSny; ■, Sen's luteipes Jofinson.

Nearctic Beridinae • Woodley 357

Map 5. Nearctic distribution of Sen's fuscipes Meigen.

358 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

e^^

Figures 1-10. Antennae of Beridinae. 1. Actina viridis (Say), male. 2. Actina viridis (Say), female. 3. Allognosta brevi-
cornis Johnson. 4. Allognosta fuscitarsis (Say). 5. Allognosta obscuriventris (Loew). 6. Beris fuscipes Meigen. 7. Beris
luteipes Johnson, male. 8. Sens luteipes Johnson, female. 9. Sens strobli DuSek and RozkoSny. 10. Exodontha dubia
(Zetterstedt).

Nearctic Beridinae • Woodley 359

11 AT

10AT

ma

Figures 11-15. Male and female genitalia of Actina viridis (Say). 11. male post-genital segments, dorsal view. 12. male
aedeagal complex, dorsal view. 13. male aedeagal complex, lateral view. 14. female genital furca, ventral view. 15. male
genital capsule, dorsal view.

Abbreviations: aed, aedeagus: AT, abdominal tergite; av, aedeagal valve; c, cercus; gc, gonocoxites; gs, gonostylus; it,
internal tooth; ma, median aperture of genital furca; pp, posterolateral process.

3(S() Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

16

17

Figures 16-19. Male genitalia of Allognosta brevicornis Johnson. 16. post-genital segments, dorsal view. 17. aedeagal
complex, dorsal view. 18. aedeagal complex, lateral view. 19. genital capsule, dorsal view.

Abbreviation: pvb, process of ventral bridge of gonocoxites.

Nearctic Beridinae • Woodley 361

20

21

Figures 20-23. Male genitalia of Allognosta fuscitarsis (Say). 20. post-genital segments, dorsal view. 21. aedeagal
complex, dorsal view. 22. aedeagal complex, lateral view. 23. genital capsule, dorsal view.

362 Bttllctin Museum of Comparative Zoology, Vol. 149, No. 6

Figures 24-28. Male genitalia of Allognosta obscuriventris (Loew). 24. post-genital segments, dorsal view. 25. aedeagal
complex, dorsal view. 26. aedeagal complex, lateral view. 27. gonostylus, variant, same view as for genital capsule. 28.
genital capsule, dorsal view.

Nearctic Beridinae • Woodley 363

Figures 29-32. Female genital furcae of Allognosta spp., ventral views. 29. A. obscuriventris (Loew). 30. A. obscuri-
ventrls (Loew), variant. 31. A. fuscitarsis (Say). 32. A. brevicornls Johnson.

364 Bulletin Museum of Comparative Zoology, Vol 149, No. 6

33

34

Figures 33-36. Male genitalia of Beris fuscipes Meigen. 33. post-genital segments, dorsal view. 34. aedeagal complex,
dorsal view. 35. aedeagal complex, lateral view. 36. genital capsule, dorsal view.

Abbreviation: ss, surstylus.

Nearctic Beridinae • Woodley 365

37

Figures 37-40. Male genitalia of Beris luteipes Johnson. 37. post-genital segments, dorsal view. 38. aedeagal complex,
dorsal view. 39. aedeagal complex, lateral view. 40. genital capsule, dorsal view.

366 Bulletin Musciiw of Comparative Zoology, Vol. 149, No. 6

Figures 41-44. Male genitalia of Beris strobli DuSek and RozkoSny. 41. post-genital segments, dorsal view. 42. aedeagal
complex, dorsal view. 43. aedeagal complex, lateral view. 44. genital capsule, dorsal view.

Nearctic Beridinae • Woodley 367

47

Figures 45-47. Female genital furcae of Beris spp., ventral views. 45. B. fuscipes Meigen. 46. S. luteipes Johnson. 47.
B. strobli DuSek and RozkoSny.

Abbreviation: sd. spermathecal duct.

368 Bulletin Museum of Comparative Zoology, Vol. 149, No. 6

dmp

51

48

Figures 48-52. Male genitalia of Exodontha dubia (Zetterstedt). 48. post-genital segments, dorsal view (note: stippling
indicates membrane). 49. aedeagal complex, dorsal view; 50. aedeagal complex, lateral view. 51. gonostylus, dorso-
lateral view. 52. genital capsule, dorsal view.

Abbreviation: dmp, dorsomedial process.

Nearctic Beridinae • Woodley 369

Figures 53-58. Male and female genitalia of Exodontha luteipes (Williston). 53. male post-genital segments, dorsal view
(note: stippling indicates membrane). 54. male aedeagal complex, dorsal view. 55. male aedeagal complex, lateral
view. 56. female genital furca, ventral view. 57. male gonostylus, dorsolateral view. 58. male genital capsule, dorsal view.

^U9 ISSfM UU^/-41UU;

bulletin OF THE

seum O
Comparative
Zoology

/v|i;s i-dfo

MUS. COMP. ZOOU,

H

Rodents from the Deseadan Oligocene

of Bolivia and the Relationships

of the Caviomorpha

BRYAN PATTERSON AND ALBERT E. WOOD

HARVARD UNIVERSITY

CAMBRIDGE, MASSACHUSETTS, U.S.A.

VOLUME 149, NUMBER 7
12 JANUARY 1982

PUBLICATIONS ISSUED

OR DISTRIBUTED BY THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD UNIVERSITY

Brevior\ 1952-

bulletix 1863-

Memoirs 1864-1938

JOHNSONIA, Department of Mollusks, 1941-

OccASiONAL Papers on ISIollusks, 1945-

SPECIAL PUBLICATIONS.

1. Whittington, H. B., and E. D. I. Rolfe (eds.), 1963. Phylogeny and
Evolution of Crustacea. 192 pp.

2. Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Tere-
dinidae (Mollusca: Bividvia). 265 pp.

3. Sprinkle, J., 1973. Morphology and Evolution of Blastozoan Echino-
derms. 284 pp.

4. Eaton, R. J. E., 1974. A Flora of Concord. 236 pp.

Other Publications.

Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine.
Reprint.

Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification
of Insects.

Creighton, W. S., 1950. The Ants of North America. Reprint.

Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam-
malian Hibernation.

Peters' Check-list of Birds of the World, vols. 1-10, 12-15.

Proceedings of the New England Zoological Club 1899-1948. (Complete
sets only.)

Publications of the Boston Society of Natural History.

Price list and catalog of MCZ publications may be obtained from Publi-
cations Office, Museum of Comparative Zoology, Harvard Universit\-, Cam-
bridge, Massachusetts, 02138, U.S.A.

© The President and Fellows of Harvard College 1982.

RODENTS FROM THE DESEADAN OLIGOCENE OF BOLIVIA AND
THE RELATIONSHIPS OF THE CAVIOMORPHA

BRYAN PATTERSON' and ALBERT E. WOOLF

CONTENTS

Abstract 372

Introduction 373

Acknowledgments 379

Abbreviations 379

Taxonomy and Morpholog>' 380

Infraorder Caviomorpha 380

Superfamily Octodontoidea 380

Family Octodontidae 380

Migraveramus new genus 380

M. beatus new species 380

Family Echimyidae 382

Sallamys Hoffstetter and Lavocat 382

S. pascuali Hoffstetter and Lavocat 383

Superfamily Cavioidea 395

Family Dasyproctidae 395

Incamys Hoffstetter and Lavocat 395

/. bolivianus Hoffstetter and Lavo-
cat 395

Undescribed Genus, Dasyproctidae

aff. Neoreomys 430

Cephalomys bolivianus Lavocat 430

Dasyproctidae inc. sed. 431

Family Dinomyidae 431

Branisamys Hoffstetter and Lavocat 431
B. luribayensis Hoffstetter and La-
vocat 431

Caviomorpha inc. sed. 448

Luribayomys masticator L.a.vocat 448

Palmiramys waltheri Kraglievich 448

Discussion 449

Conflicting Theories of the Origin of the

Caviomorpha 449

Eocene Geography 451

Distribution of Hystricognathous Rodents in

the Eocene 453

The Composition of the Deseadan and Fa-

vum Rodent Faunas 460

' Late Professor of Vertebrate Paleontology,
Emeritus, Harvard University, Cambridge, Massa-
chusetts.

^ Professor of Biology, Emeritus, Amherst Col-
lege, Amherst, Massachusetts. Present address: 20
Hereford Avenue, Cape Mav Court House, New
Jersey 08210.

The Evidence from Parasitic Nematodes — 471
Comparisons of Caviomorpha and Old

World Hystricognathi 475

General 475

Hystricognath Mandible 476

Pter\goid Fossa 478

Jugal Fossa 479

Hystricomorphy 481

Heaviness and Suture of Malar 481

Middle Ear, Malleus and Incus 482

Openings at Base of Skull; Cephalic Arter-
ies 484

Alisphenoid 487

Squamosal-Mastoid-Parietal-Occipital

Junction 487

Clef^ Between Bulla and Squamosal 488

Maxillar\ in Orbitotemporal Fossa 488

Interparietal 488

Lacrimal Foramen 489

Tibia 489

Brain 489

Molars, Premolars and Deciduous Teeth 491

Multiserial Incisor Enamel 503

Ectoparasites 505

Other Features 506

Summar>' 507

The Classification of the Suborder Hystrico-
gnathi 509

The Superfamilies of the Caviomorpha _- 509
Other Taxonomic Problems of the Hystri-
cognathi 512

General 512

Infraorder Franimorpha 513

Cylindrodontidae and Tsaganomyidae 514

Bathyergidae 516

Hystricidae 518

Thryonomyoidea 519

Classification of the Suborder Hystrico-
gnathi 522

Conclusions 523

Appendix 1: Referred Specimens of Sallamys
pascuali, Incamys bolivianus and Brani-
samys luribayensis 528

Appendix 2: Validation oi Acarechimys 529

Literature Cited 529

Addendum 538

Index 539

Bull. Mus. Comp. Zool., 149(7): 371-543, January, 1982 371

372 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Abstract. The extensive Princeton collections of
fossil caviomoriDh rodents from the early Oligocene
Deseadan of Bolivia are described. Much new in-
formation is reported concerning the teeth and
skulls of Sallamys pascuali, Incamys boliviatius,
and Braiiisainys luribaycnsis, only one species ot
each genus being recognized. These are referred to
the Echimyidae, Dasyproctidae, and Dinomyidae,
respectively. An octodontid, Migraveramus beatus,
n. gen., n. sp., is described. Lavocat's description
o( Cephalomys bolivianus is briefly reviewed. Two
undescribed dasyproctid genera are reported. Lu-
ribayomys masticator and Palmiramys waltheri
(the latter from Uruguay) are considered to be i^o-
mina vana.

The two theories for the origin of the cavio-
morphs are reviewed. The first states that they are
descended from African thryonomyoids that crossed
the South Atlantic as waif migrants. The second de-
rives them from Middle American franimorph
ancestors that reached South America, also as waif
migrants, either directly or through the Antilles.

Eocene hystricognathous rodents have so far
been reported only from North America. There is
good negative evidence that none reached Europe,
but they presumably did reach Asia when that con-
tinent was connected with North America, both ear-
ly and late in the Eocene.

The early Oligocene South American and Egyp-
tian rodent faunas are compared and contrasted, and
compared with localized collections of Recent ro-
dents from northeastern Brazil. The differences be-
tween and general composition of the various De-
seadan local faunas are comparable to those
between individual Recent sites. The African Oli-
gocene rodents seem not so diversified as are the
modern Australian Pseudomyinae. These latter
have achieved their current diversity in about 4.5
m.y. It is proposed that rodents reached Africa no
more than this length of time before the lowest fos-
sil levels of the Fayum. The Deseadan rodents are
as diversified as are those of the early Miocene of
East Africa, and had presumably been in South
America about as long before the Deseadan as the
thryonomyoids had been in Africa before the early
Miocene. This would place the arrival of rodents in
South America in the middle Eocene.

The Eocene geography of southern Central
America and the Antilles was such that migration
of ancestral caviomorphs from Middle to South
America, either directly or through the Antilles,
would have been entirely possible. It is probable
that there was a single invading stock, although
there is evidence supporting a distinct origin for the
Erethizontidae.

Parasitic nematodes have been cited as support-
ing a thryonomyoid ancestry of the Caviomorpha.
The evidence is reviewed, and it is concluded that
there is no valid evidence to support such a hy-
pothesis. In every case of supposed close relation-
ship of nematodes parasitic on Old and New World

forms, there is reason to (juestion the validity of the
conclusions.

The morjihologic evidence used by Lavocat to
support the derivation of the Oligocene South
American Caviomorpha from the contemporaneous
Thryonomyoidea of Africa is reviewed. It is con-
cluded that, although there are many similarities
between the Old and New World fonns, the evi-
dence indicates that the African fomis could (mor-
phologically) have been derived from the South
American ones, but that the reverse is structurally
impossible. It is concluded that both groups were
derived, independently, from Holarctic Eocene
members of the Infraorder Franimorpha. The lice
and fleas of hystricognaths support the concept that
the Hystricognathi are a natural group, but do not
allow one to choose between the two hypotheses
for the origin of the caviomorphs.

The suprafamilial classification of all hystrico-
gnath rodents (Suborder Hystricognathi) is re-
viewed. Among the Caviomorpha, we now follow
a growing consensus that the Families Dasyprocti-
dae, Dinomyidae and Cuniculidae should be trans-
ferred from the Superfamily Chinchilloidea to the
Superfamily Cavioidea. These families form a
group (Series B) distinct from the Families Eocar-
diidae, Caviidae and Hydrochoeridae (Series A).
The Echimyidae are an even more complex group
than previously believed, the genus Chaetomys
(Subfamily Chaetomyinae) being transferred from
the Erethizontidae to the Echimyidae. The two
subfamilies Capromyinae and Plagiodontinae, for-
merly included by us in the Family Capromyidae,
are likewise here transferred to the Echimyidae.
This combines all caviomorphs that ever retain dnV
throughout life in the Family Echimyidae.

The Superfamily Erethizontoidea is retained in
the Caviomorpha, in spite of evidence that it is
quite distinct from the rest of the infraorder. It was
presumably an early offshoot of the basal stock of
the Caviomorpha.

The Family Neoepiblemidae is added to the
Chinchilloidea as a second family. The Family
Heptaxodontidae is left incertae sedis as to super-
family.

The Eocene to Oligocene Infraorder Franimor-
pha, including the Families Reithroparamyidae
(new rank), Protoptychidae, Cylindrodontidae, and
two genera incertae sedis as to family, is considered
to have been ancestral to all other hystricognaths.
In the Old World, Eocene franimorphs are still un-
known, the Chapattimyidae of Pakistan and India
being shown to be ctenodactyloids.

The Tsaganomyidae (new rank) of north Asia are
considered to be cylindrodont descendants, as are
also the Bathyergidae. The two families differ from
the cylindrodonts in being fully hystricognathous,
and are united in the Superfamily Bath\ ergoidea.
They are both fundamentally protrogomorphous.

No evidence is found for or against deriving the
Thryonomyoidea from the Hystricidae. If the two

Oligocene Rodents of Bolivia • Patterson and Wood 373

are united at the infraordinal level, the infraorder
should be called Hystricomorpha rather than
Phiomorpha because of the long priority and gen-
eral recognition of that term.

Three families of Thryonomyoidea are recog-
nized: Thryonomyidae, Petromuridae and My-
ophiomyidae. Stromer's genus Neosciuromys, placed
in synonymy by Lavocat, is re-established.

INTRODUCTION

The Tertiary vertebrates of Bolivia are
poorly known. Simpson (1940), in his
summary of reported extra-Argentinian
Tertiary mammals of South America,
could cite only three Bolivian records, all
of uncertain age: a tooth referred, per-
haps incorrectly, to Nesodon; parts of an
armadillo, the type of Dasyptodon ata-
vus Castellanos; and a jaw fragment of a
haplodontheriine toxodontid, the type of
Pachtjnodon validus Burmeister. Discov-
eries of Deseadan faunas in the Cordille-
ra Oriental of Bolivia, not far from La Paz,
are therefore of the greatest interest.

The first find was made in the Salla-
Luribay Basin, Provincia Loaza, Depar-
tamento La Paz, by G. Bejarano in 1962.
Dr. L. Branisa began field work there in
1964. With encouragement from the late
Professor Glenn L. Jepsen, he undertook
intensive work in this and other localities
of younger ages for Princeton University.
Field, preparational, and other expenses
connected with the acquisition by
Princeton of the materials brought to-
gether by Branisa during 1964-1966
were met by a generous bequest from the
late Gordon Barbour, a long-time friend
and benefactor of the University, for the
express purpose of furthering vertebrate
paleontological work in Bolivia. As a re-
sult, Princeton possesses an excellent
collection of Deseadan mammals from
Salla-Luribay. The discovery was an-
nounced by Baird et al. (1966). From
1965 to 1972 Hoffstetter visited this area
and made collections (Hoffstetter, 1968,
1976), and Hartenberger and Villarroel
have worked at the locality since then.
More recently there have been other ex-
peditions in the area.

At Salla-Luribay, which is some 90 km
SE of La Paz, about 450 m of Tertiary
continental sediments rest unconform-
ably upon folded Devonian rocks. The
lower part of the Tertiar>' sequence con-
sists of chocolate colored sandstone and
conglomerate beds; the upper of well
stratified, red and reddish-brown clays
and silts with four calcareous beds dis-
tributed through the sequence (Hoffstet-
ter, 1976: 4, 6). The lower portion was
named the Luribay Conglomerate (Ahl-
feld and Branisa, 1960: 132) and the
higher the Estratos de Salla, or Salla Beds
(Evernden et al., 1966). The vertebrate
remains occur in the lower part of the
Estratos de Salla, the upper part of that
unit and the underlying Luribay Con-
glomerate apparently being unfossilifer-
ous (Hoffstetter, 1968; a notoungulate
skull possibly from the conglomerate has
been mentioned by Hoffstetter, 1976: 8).
The thickness of the fossil bearing por-
tion of the Salla has not been reported,
although Hoffstetter (1976: 8) stated that
the fossils are found "depuis la base jus-
qu'au Seme niveau calcaire . . . ." We
tried to measure the thickness of the fos-
siliferous zone from the geologic section
given by Hoffstetter (1976: Fig. 3), but it
would appear that the section is not to
scale since, in the figure, the 300 m thick
Salla Beds measure 9.5 cm, and the 150
m thick Luribay Conglomerates measure
6 cm. Hoffstetter did not state whether
the section represents the maximum
thickness of the fomiation. If it does not,
and if the Salla and the underlying Lu-
ribay Conglomerate are to scale in the
figure, the fossiliferous portion would be
about 150 m thick. Ahlfeld and Branisa
(1960: 137) note a resemblance between
this sequence and the Tertiar>' Chacarilla
Formation of the Altiplano.

Subsequent to the Salla-Luribay dis-
covery, a second, much less extensive
and less fossiliferous locality was found
at Lacayani, 50 km NW of Luribay and 30
km SE of La Paz (Hoffstetter et al, 1971).
The Tertiary deposits there are uncon-

374 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

formable upon folded late Cretaceous
and Devonian rocks, and are "some tens
of meters" in thickness. The lower por-
tion consists of alternating clays, sand-
stones and yellowish brown conglomer-
ates. This is separated by a 50 cm thick
cinerite from the upper, reddish clays,
which contain the mammalian remains.

Rodents occur at both localities. Hoff-
stetter and Lavocat (1970) very briefly
described three new genera from Salla:
Sallamys, Incamijs, and Branisamys.
From Lacayani, Hoffstetter et al. (1971)
have listed Scotamys, a genus known
from Patagonia, and an eocardiid. Har-
tenberger (1975), listing these, has also
mentioned "Octodontoidea nov. gen. 1
and 2" from Salla, and described Villar-
roelomys, also from this locality, which
we believe (p. 444) to be referable to
Branisamys. He has further recorded,
again from Salla, a species of Cephalo-
mys, another genus known from Patago-
nia. Lavocat (1976) described Sallamys,
Branisamys and Incamys in detail, con-
cluded that the type of Villarroelomys
probably was dm4 of Branisamys (a con-
clusion that we had already reached), de-
scribed a second species of Incamys, a
new species of Cephalomys (C. bolivi-
anus) and Luribayomys, a new genus
that he left incertae sedis. He was vague
(1976: 59) as to whether Hartenberger's
Cephalomys was or was not the same as
C bolivianus. The Princeton collection
contains the jaw of a new octodontid ge-
nus, additional material of Sallamys, ex-
tensive and excellent material of Inca-
mys, and better material of Branisamys
than was available to Hoffstetter and La-
vocat (1970) or to Lavocat (1976). These
specimens are described in detail below.

There can be no doubt as to the Pro-
vincial (or Land Mammal) Age of these
Bolivian faunas; each contains mamma-
lian species inseparable from Patagonian
Deseadan ones or assignable to Patago-
nian Deseadan genera. The two differ
from each other and from Patagonian lo-
cal faunas — as these latter do among

themselves — regarding presence, ab-
sence or differing abundances of various
faunal elements. It is not at all surprising
to encounter such differences between
the Bolivian and the Patagonian faunas,
separated as they are by nearly 30° of lat-
itude, but it is rather remarkable that all
Deseadan local faunas, whether Bolivian
or Patagonian and regardless of degree of
geographic proximity, should differ to the
extent that they do. The rodents. Table
8 (p. 461), are no exception. Each of the
better known local faunas — Cabeza Blan-
ca. La Flecha, Salla — contains one species
that is predominant numerically and one
or two that are moderately common,
whereas the rest are represented only by
one or two specimens. In no case is the
predominant species the same in more
than one locality. The question is: are
these differences due primarily to age or
primarily to the sampling of different
ecologies?

Both views have been advanced. Har-
tenberger (1975) has tentatively suggest-
ed, as a working hypothesis, that Rinco-
nada de los Lopez (Scarritt Pocket) is the
earliest local fauna, Salla intemiediate,
and La Flecha and Cabeza Blanca the lat-
est. This was based on the primitive na-
ture oi Platypittamys, on the presence of
a species in the Salla local fauna (his Oc-
todontoidea nov. gen. 1) that "on dental
characters could be considered a descen-
dant of Platypittamys,' and on the Salla
species of Cephalomys, stated to be
small and more primitive than those oc-
curring at Cabeza Blanca and La Flecha.
Wood (1949) had previously considered
Rinconada de los Lopez as being earlier
than the others, also on the basis of Platy-
pittamys. The alternative, ecological hy-
pothesis is based on the local faunas as
wholes. In no instance (with the possible
exception of Hartenberger's octodont) is
it demonstrated that a species in one lo-
cal fauna is ancestral to a species in
another (Wood and Patterson, 1959: 300).
As Hoffstetter et al. (1971) have put it:
"... les faunes . . . appartiennent au De-

Oligocene Rodents of Bolivia • Patterson and Wood 375

seadien . . . sans qu'on puisse exclure la
possibilite de legers decalages chrono-
logiques, de toute fagon minimes."

In agreement with these authors, we
doubt that anyone would be prepared to
assert that all Deseadan local faunas were
exactly synchronous. Granting this, were
the differences in age between them
minimal in the geological sense or suffi-
ciently long to pennit evolution to or to-
ward the species level? Support for the
latter possibility must, it would seem,
come from the rodents since other faunal
elements do not, so far as known, favor it.
The most suggestive evidence, as yet un-
documented, is of course provided by
Hartenberger's remark that his Salla Oc-
todontoidea gen. nov. 1 could be de-
scended from Platypittamys (a claim he
has not made concerning the relation-
ships of the Cephalomys present at Salla
to those occurring at La Flecha and Ca-
beza Blanca). It is possible, however, for
more primitive and more advanced, yet
closely related, species to coexist in the
same local fauna. Deseadomys loomisi,
type and only known specimen from Ca-
beza Blanca, is larger and more advanced
as regards metalophid reduction than D.
arambourgi, type from La Flecha. If
knowledge rested there it could very
plausibly be argued that the latter was
ancestral to the fonner, but negating this
possibility is a specimen from Cabeza
Blanca that can be referred with confi-
dence to D. arambourgi (Wood and Pat-
terson, 1959: 305). As a minor item of evi-
dence suggesting a close approximation
in time between Cabeza Blanca and Sal-
la, we may point to the presence at the
fonner of an isolated P* that we are un-
able to distinguish from corresponding
teeth of Incamys bolivianus, the com-
monest rodent at Salla (see Table 8 and
p. 429).

On the age difference side of the ar-
gument, it seems probable that the De-
seadan rodents, not yet fully beyond the
first flush of their initial adaptive radia-
tion but nonetheless the most abundant

group of Deseadan fossils, were evolving
rapidly (more so than other faunal ele-
ments whose ancestors had been in
South America for millions of years be-
fore the rodents arrived there) and should
therefore be the best indicators of tem-
poral differences within that Age. This
may indeed be so, but the documentation
of it is not yet satisfactory. However, the
upper premolars are three-crested in all
or nearly all (see below, p. 495) of the
Patagonian genera in which they are
known, and four- or five-crested in the
ones from Salla. We believe the three-
crested pattern to be the primitive one.
This suggests, to Wood at least, that the
Salla deposits may be later than those in
Patagonia. We are open minded as re-
gards the two hypotheses, although we
must confess to differing preferences:
Patterson, in agreement with Hoffstetter
(1976: 11-12), inclining toward ecologi-
cal differences; Wood toward age ones.
(The composition of Deseadan rodent
faunas is further discussed below, pp.
460-471.)

As concerns the age of the Deseadan as
a unit, welcome new evidence is at hand.
Marshall et al. (1977) have reported that
a basalt above Deseadan sediments at
Pico Truncado has an age of 33.6 ± .4
m.y.b.p., and another, at Cerro Blanco,
apparently directly conformable upon
Deseadan sediments, one of 35.4 ± .4
m.y.b.p. These authors report that the ba-
salts record a period of volcanic activity
"around 34-35 mybp," and this "is ten-
tatively accepted as an approximate ter-
minal date for known Deseadan." They
regard the 33.6 m.y.b.p. date as being a
little post-Deseadan. It is gratifying to
see the previous consensus as to early
Oligocene age confinned by the new evi-
dence; our suggestion (1959: 388) that
the Deseadan was early but not earliest
Oligocene may not have been too wide
of the mark. How long the Age may have
lasted remains uncertain. We utilize the
earlier Tertiary epoch boundaries of
Berggren and Van Couvering (1974: Fig.

376 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

PA MSF ^^

ATL-

MTF
MTL

ATL HY

Figure 1. Nomenclature of parts of caviomorph upper cheek teeth. Anterior ends of all teeth to the left. A. Protostei-
romys medianus, LW " ^ B. Protosteiromys medianus, Lh/P. C. Incamys bolivianus, LM^^

Abbreviations: ATL— anteroloph; HF— hypoflexus; HY— hypocone; MC— nnetaconule; ME— metacone; ML— metaloph;
MSF— mesoflexus or mesofossette; MTF — metaflexus or metafossette; MU — mure; NEL— neoloph; NF— neofossette;
PA— paracone; PC— protoconule; PF— paraflexus or parafossette; PL— posteroloph; PR — protocone; PRL— protoloph.
Flexi are unworn valleys that become fossettes with wear.

1). They date the beginning of the Oli-
gocene at 37.5 m.y.b.p.

Most unfortunately, no radiometric
dates are available for the Mustersan.
The duration of the hiatus, undoubtedly
long, between this Age and the Deseadan
can only be guessed at; a reasonable es-
timate is that it lasted throughout most or
all of late Eocene time and perhaps into
earliest Oligocene as well. Absence of ro-
dents— and of primates — in Mustersan
and earlier deposits is the basis for the
view that the earliest members of these
orders to reach South America arrived at

some time during the Mustersan-Desea-
dan hiatus. Just when they did so and
whether or not they arrived at essentially
the same time are questions that will re-
main unanswered until the hiatus is
filled. (One small local fauna of very pe-
culiar facies, the Divisadero Largo,
which contains no rodents or primates, is
pre-Deseado and post-Musters.) The new
Bolivian evidence of course does not di-
rectly contribute to detemiining the time,
within the hiatus, when the rodents
reached South America.
Hoffstetter and Lavocat (1970) did not

Oligocene Rodents of Bolivia • Patterson and Wood 377

MED

END

MED MSFD

MED MLD MSFD END MTFD

ALD
MLD

A

AFD

ALD

PRD /ECTD \ \APD\\HFD

HFD HYD ALD PRD

ECTD

PRD

PLD

HYD

MED MLD END MED MLD END

MSFD/ MTFD I / MSFDJ MTFD

PLD

HFD

Figure 2. Nomenclature of parts of caviomorph lower cheek teeth. Anterior ends of all teeth to the left. A. Migraveramus
beatus, RP^-M, , reversed. B. Incamys bolivianus, LP4. C. Incamys bolivianus, LN^j.

Abbreviations: ACD — anteroconid; AFD — anterofossettid; ALD — anterolophid; ECTD — ectolophid; END — entoconid;
HFD — hypoflexid; HYD — hypoconid; MED — metaconid; MLD — metalophid or metalophulid II; MSFD — mesoflexid;
MTFD — metaflexid or metafossettid; PLD — posterolophid; PRD — protoconid. Flexids are unworn valleys that become
fossettids with wear.

allocate their three new genera from Sal-
la to families. Hartenberger (1975) placed
Sallamijs in the Octodontoidea, Incamys
in the Chinchilloidea, and Branisamys,
with a query, in the Dinomyidae. We had
independently arrived at broadly similar
conclusions (cf Patterson and Pascual,
1972: 278), and present evidence in the
following pages that Sallamys was a
member of the Echimyidae, Incamys of
the Dasyproctidae, and Branisamys of
the Dinomyidae — without query. A new
genus, Migraveramus, is referred to the
Octodontidae. One specimen of Lavo-
cat's hypodigm of Incamys bolivianus
represents, we believe, an undescribed
dasyproctid genus. Salla thus adds a fam-
ily to the six previously represented in

the Deseadan. In 1959 we inadvertently
cited seven families on page 389, thus
contradicting our correct statement on
page 284 (with Lavocat, we are unable to
agree that Hartenberger's Villarroelomys
is referable to the Hydrochoeridae). La-
vocat (1976) also placed Sallamys in the
Echimyidae and Incamys in the Dasy-
proctidae, but erred, we believe, in also
including Branisamys in that family.
Hoffstetter and Lavocat (1970), on the
basis of some points of similarity in cheek
tooth structure between these fomis and
the approximately contemporaneous Fa-
yum rodents, favor descent of cavio-
morphs from (as yet unknown) African
Eocene thryonomyoids, which they be-
lieve to have been rafted across a then

378 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

narrower South Atlantic Ocean. This sug- tion of the rodents, and that, therefore, all

gestion has been further elaborated upon hystricognathous rodents are related, and

by Lavocat (1973, 1976). We do not find are more closely related to each other

the evidence presented in support to be than any are to other groups of rodents,

any more convincing than we did a few There is some documentary support for

years ago (Wood and Patterson, 1970). this assumption, but it cannot be taken to

This and related topics are treated at have been demonstrated. However, if

some length in the concluding sections. hystricognathy arose independently, sev-

Cheek tooth tenninology is as previ- eral times, as parallelisms, it means that

ously employed by us (1959: 286-288, there is no justifiable basis, at the present

Fig. 1). The terminology is repeated, and time, for grouping rodents at the subor-

somewhat enlarged, in Figs. 1-2; it differs dinal level, and that no attempts at de-

in important respects from that adopted by temiining relationships between larger

Lavocat (1976), stemming from his belief taxonomic groups of rodents should be

that the more complex Deseadan cavio- made until there is full documentation

morph teeth are primitive and that the less for the evolutionarv' sequences involved

complex ones are specialized, whereas (see below, pp. 476-477).

we take the opposite point of view. For example, at the present time, there

Measurements of hypsodont rodent is no final proof that the various families

teeth (both cheek teeth and incisors) may of caviomorphs have a common South

vary with the age of the animal and with American ancestor. We believe that they

the precise manner in which the mea- did have such a common ancestry, but,

surements have been made. They were since seven families were already pres-

made, for this study, as previously de- ent in the Deseadan, when the cavio-

scribed (Wood and Patterson, 1959: 288- morphs first appear, the conclusion that

289). We have been encouraged, how- they were the descendants of a single in-

ever, by the fact that, while the coeffi- vading stock is an unproven assumption

cient of variation for our measurements on our part.

is often high (Tables 3-4; Wood and Pat- As we classify them (see below, pp.
terson, 1959: Tables 3-5), it is usually 522-523), the Old World hystricognaths
less than 10. The measurements that give are divided into three groups that, in our
the highest variation are transverse di- opinion, cannot at present be demonstrat-
ameters of upper cheek teeth (Table 3), ed to have had a common ancestry later
where the curvilinear eruption of the than the Eocene franimorphs. The Hys-
teeth often produces an elongate buccal tricomorpha contain a single family and
shelf; the anteroposterior diameter of Mg , the Bathyergomorpha contain two fami-
due to variation in the measurement of lies, one Asian and one African. The Af-
the long posterior slope because of loss rican Thryonomyoidea perhaps should
of the alveolar rim behind the tooth; and be referred to the Hystricomorpha and
the actual measurements (as opposed to perhaps deserve recognition as a separate
the ratio) of the incisors, due to diametric infraorder, the Phiomorpha. Lavocat
enlargement of these teeth with age. The (1973, 1976) united all the Old World
reasons behind the variability of the hystricognaths in the Infraorder Phio-
tooth measurements are discussed in morpha. To avoid the probability of con-
more detail below (p. 416). fusion between our usage and that of

We have reviewed the classification of Lavocat, we have avoided using the in-

the Suborder Hystricognathi in the final fraordinal terms Phiomorpha, Bathyer-

part of our discussion. In this paper, we gomorpha, and Hystricomorpha, and

have made the explicit assumption that have, instead, used the suprafamilial or

hystricognathy arose once in the evolu- familial tenns, everywhere except in our

Oligocene Rodents of Bolivia • Patterson and Wood

379

final section, Classification of the Subor-
der Hystricognathi.

The manuscript of this paper had been
largely completed at the time of the death
of the senior author. All of the first half^ —
Taxonomy and Morphology — was in or-
der; the various sections of the Discus-
sion had been written, although some
were not in their final fomi and, in some
instances, no final decision had been
reached as to their sequence. Extensive
notes by Patterson were also available. In
general, it is Wood's belief that the state-
ments in the discussion are the opinions
of both authors, unless the contrary is ex-
plicitly stated. Some sections that had to
be prepared essentially or entirely by
Wood are so indicated. Any errors that
may have developed in the final prepa-
ration of the paper must be assumed to
be the sole responsibility of the junior
author.

ACKNOWLEDGMENTS

We are deeply indebted to the late Pro-
fessor Glenn L. Jepsen for pemiission to
study the Salla collection in Princeton
University, and to Drs. Vincent J. Maglio
and Donald Baird for continuing courte-
sies in this connection. We have received
assistance in a number of matters from:
Professor Robert W. Fields, Geology De-
partment, University of Montana, Mis-
soula; Dr. J.-L. Hartenberger, Labora-
toire de Paleontologie, Universite de
Montpellier; Dr. P. E. Vanzolini and Dr.
Lindolpho R. Guimaraes, Museu de
Zoologia da Universidada de Sao Paulo;
and Dr. John H. Wahlert, then of the
American Museum of Natural History.
Dr. J. J. Raab, DDS, of Cape May Court
House, N.J., made the X-ray photograph
of the specimen oi Sail amy s pascuali de-
scribed below (p. 383), and Dr. Luiz Fer-
nando da Castro Leite, C. D., of Sao Pau-
lo, made those of Chaetomys discussed
on p. 394. Professor Elwyn L. Simons and
Dr. Grant E. Meyer have generously con-
tributed information about the Yale Fa-
yum collections. Dr. Hartenberger has

provided us, through the courtesy of Dr.
Lavocat, with the numbers of specimens
of various taxa of rodents included by the
latter in his study of the Salla-Luribay ro-
dents, data not given in the published re-
port (Lavocat, 1976). Hartenberger stated
{in lit., 10 July 1976) that his own large
collections of Salla rodents were still un-
identified. We are grateful to Dr. R. W.
Wilson of the Museum of Natural Histo-
ry, The University of Kansas, who has
critically reviewed the manuscript and
who has made a number of suggestions
on points that needed clarification. We
have attempted to carry out all of his sug-
gestions.

The senior author had the benefit of
numerous discussions witli Dr. Ernest E.
Williams of the Museum of Comparative
Zoology, covering many aspects of the
present paper. The major concrete sug-
gestion here incorporated is that men-
tioned on p. 461, but there were many
others of lesser fundamental import.

This study has been aided by awards
to Wood of Grants 148 and 150 from the
Marsh Fund of the National Academy of
Sciences, and to Patterson from the Fun-
dagao de Amparo e Pesquisa do Estado
de Sao Paulo. The drawings of Miocene
thryonomyoids from Southwest Africa
(Fig. 28G-K) were made by Wood during
his tenure of a National Science Foun-
dation Senior Post-Doctoral Fellowship
in 1967-68. Illustration and publication
have been aided by the Gordon Barbour
fund of Princeton University-. Continuing
support has been received by Wood from
the Biology Department, Amherst Col-
lege.

ABBREVIATIONS

Abbreviations used for the names of
museums:

ACM — Pratt Museum of Amherst Col-
lege

AMNH — American Museum of Natural
History

380

Bulletin Museum of Coinpamtive Zoology, Vol. 149, No. 7

AMNH Mam — American Museum of
Natural Histor>', Dept. of Mammalogy

FMNH— Field Museum of Natural His-
tory

GB — Coleccion Geobol, La Paz

MACN — Museo Argentino de Ciencias
Naturales

MCZ — Museum of Comparative Zoology

Munich — Institut fiir Palaontologie und
historische Geologic, Universitat Miin-
chen

MZUSP — Museu de Zoologia da Univer-
sidade de Sao Paulo

PU — Princeton University, Department
of Geology

UCMP — University of California (Berke-
ley), Museum of Paleontology

YPM — Yale Peabody Museum

TAXONOMY AND MORPHOLOGY

Order RODENTIA Bowdich 1821
Suborder HYSTRICOGNATHI Tullberg

1899
Infraorder CAVIOMORPHA Wood and

Patterson 1955 (in Wood, 1955, as a

suborder)
Superfamily OCTODONTOIDEA Simpson

1945
Family OCTODONTIDAE Waterhouse

1839

Migraveramus^ new genus

Diagnosis. Lower cheek teeth low
crowned lingually and of medium height
buccally, fully crested; dm4 replaced at
normal stage of life; P4-M2 fully tetralo-
phate, with crests nearly perpendicular
to anteroposterior axis; an anteroposte-
rior crest crossing lingual part of antero-
fossettid; lower incisor with flat anterior
face; coronoid process passing alveolar
border at middle of M3; angular process
leaving body of ramus beneath middle of
M2; mental foramen in front of P4, open-
ing anteriorly and ventrally.

' The pluperfect of migrare, to migrate, plus mus,
a rodent, in indication of our belief that this genus
is closely related to the caviomorphs that first mi-
grated to South America.

Type Species. Migraveramus beatus.
Distribution. Deseadan Oligocene of
Bolivia.

Migraveramus beatus^ new species
Figure 3

Diagnosis. As for the genus; measure-
ments as given in Table L

Type. PU no. 21948, partial right ramus
with P4-M2 and the base of the incisor,
Branisa field no. 6515.

Hypodigm. Type only.

Horizon and Locality. Deseadan early
Oligocene, lower part of Estratos de Sal-
la; Salla-Luribay Basin, Bolivia.

Description. The three anterior lower
cheek teeth (P4-M2; Fig. 3A) are all four-
crested. For temiinology of parts of teeth,
see Fig. 2. The anterofossettid is crossed,
on all, by an anteroposterior crestlet of
varying prominence, arising from the an-
terolophid and reaching the lingual tip of
the metalophid (Fig. 3A). The metalo-
phid does not unite directly with the
metaconid, but with a crest running buc-
cally from it. On M2, the connection has
just been established (Fig. 3A). The lin-
gual part of the anterofossettid basin is
very shallow, and would be eliminated
with but little wear. The protoconid is set
off from the ectolophid by an angulation
on P4. A distinct anteroconid is present
on this tooth (Fig. 3A). There is no trace
of a mesoconid or mesolophid on any of
the teeth. The hypolophid is connected
with the hypoconid by a narrow isthmus
arising from the front of the latter. It is
marked by a swelling midway between
the ectolophid and entoconid. The pos-
terolophid shows indications of there
having been a hypoconulid. The antero-
fossettid is closed lingually; the meso-
flexid and metaflexid are open on all
three teeth, the mesoflexid being the
deeper; the hypoflexid is the deepest of
all. The buccal cusps are somewhat
higher than the lingual ones; there is lit-
tle difference in height between the an-

^ Named for the Little Angel.

Oligocene Rodexts of Bolivia • Patterson and Wood 381

Figure 3. Teeth and jaw of Migraveramus beatus, gen. et sp. nov., PU no. 21948, type. A. RPi-Mj, xlO. B. Cross section
of Rl,, beneath diastema, xlO. C. Lateral view of right lower jaw, x5.

terior and posterior cusps of each side.
The base of the enamel is nearly horizon-
tal around the teeth. The anterior root of
dm4 is still present in front of P4 (Fig. 3C).

The incisor has a flat anterior face, with
the enamel extending fairly far around on
both sides (Fig. 3B). The pulp cavity, be-
low the diastema, is small and oval.

The mandible is badly damaged. No
features of any significance were ob-
served on the medial side. The mental
foramen (Fig. 3C) is about half way up
the side of the mandible, a short distance
in front of the anterior root of P4, and
faces forward and downward. The ante-
rior tip of the masseteric crest lies on the
upper third of the mandible, beneath the
rear of P4, whence it curves posteroven-
trad and increases rapidly in prominence.
Only the beginning of the coronoid pro-
cess is visible, but it passed the alveolar
border near the front of M3 (Fig. 3C).

Relationships. The mandible agrees
with that of the Deseadan octodontid
Platypittamys (Wood and Patterson,

1959: 298) in the absence of any distinct
scar for the M. masseter medialis, pars
anterior, suggesting that this muscle was
poorly developed.

The pattern of the lower molars agrees
with that of Platypittamys alone among
Deseadan rodents — there are four trans-
verse crests, the two anterior and the two
posterior being widely separated by the
deep mesoflexid (Wood, 1949: Fig.
3C,D). As in Platypittamys (Wood, 1949:
Fig. 3C; M2), the metaconid is somewhat
isolated from the metalophid. The

Table 1. Measurements, ln mm, of teeth of
Migraveramus beatus, TYPE, PU NO. 21948, right.

?< anteroposterior

2.38

M2 anteroposterior

2.48

width, metalophid

1.74

width, metalophid

2.53

width, hypolophid

2.14

width, h\polophid

2.49

M, anteroposterior

2.40

I, anteroposterior

2.16

width, metalophid

2.02

transverse

1.50

width, h>polophid

2.26

ratio

.70

382 Bulletin Museum oj Comparative Zoologij, Vol. 149, No. 7

anteroposterior bar across the anterofos- tral stock of the cavionioriDhs may have

settle! also appears in Platypittanujs had tetrah)phate premolars as well as mo-

(Wood, 1949: Fig. 3D). The pattern of P4 lars, agreeing in this with the Eocene and

of the two genera, however, is quite dis- Oligocene North American forms men-

tinct, that of Migravercunus being fully tioned above. This is supported by the

molarifomi, whereas P4 of Platypittamys fact that, in a completely unrelated group

has no trace of a posterolophid, and the of rodents, a somewhat analogous pattern

anterior half is not so molarifomi as in of P4 occurs, clearly as the result of sec-

Migraveranius. ondary reduction of complexity {Melia-

The cheek teeth of Migraveramus are krouniomys. Wood, 1974a: 79-80, Fig.

appreciably higher crowned than those of 33C).
Platypittamys. We conclude that the octodontids are

The incisor (Table 1) is considerably the most primitive of the Deseadan ro-

wider than that of Platypittamys (width/ dents, and that the nearest known ap-

length ratio = .70 as opposed to .39 and proach to the ancestral condition is rep-

.57 in the two specimens of Platypitta- resented by Migraveramus.
mys given by Wood, 1949: Table 2). The

general shape of the incisors, is, how- ^ ., ^^, ,,. .^.r., . ^ ^ .^^^

ever, identical. Family ECHIMYIDAE Gray 1825

Wood and Patterson (1959: 295) agreed We believe that this family should be
with Wood (1949: 45) that Platypittamys dated from Echimyna Gray 1825, the first
and its relatives (the primitive octodon- usage covering the modern family in es-
tids, of which Platypittamys was then sential concept though not in spelling,
the only Deseadan representative) were rather than from Waterhouse 1839.
"ancestral, at least structurally, to all ro-
dents . . . included in the Caviomorpha, ^ ux 1 in— n-i-. <->!-> o^n/iv/i ma 1- a xi_

..1 .1 .11 .. r . 1 t. Subfamily HETEROPSOMYINAE Antho-

with the possible exception of the Ere- iqi7

thizontidae." We still take this position, o // u xx * i.x _i . ^ An-,r,

If 1 . . 4.U i-u- 4.U I Sa/Zamys Hoffstetter and Lavocat 1970

and reel certain, among other things, that '

the teeth of Migraueramu* are complete- Emended Diagnosis. Infraorbital fora-

ly different from those of any thryonomy- men and area of origin of M. masseter

oid and demonstrate the origin of the Ca- medialis, pars anterior large (Hoffstetter

viomorpha from a tetralophate ancestry, and Lavocat); cheek teeth brachydont,

of the type found during the middle and uppers higher lingually than labially,

late Eocene in North American rodents, lowers higher labially, both about as in

including franimorphs as well as para- Paramyidae and Reithroparamyidae; dnV

myids and sciuravids. The same pattern replaced; dm^ possibly present; P^ vari-

is found in the latest Eocene and Oligo- able, trilophate or incipiently tetralo-

cene ischyromyids. phate, with or without mure, paracone

The significance of the peculiar pattern and buccal tip of anterocone distinct in

of P4'* of Platypittamys is still uncertain, early stages of wear; M^~^ notably wider

Incidentally, it should be mentioned that than long, tetralophate with lophs direct-

the "recently discovered but still unde- ed transversely, metafossette shallow; P4

scribed lower Gray Bull paramyid" cited with fully formed hypolophid, small

by Wood and Patterson (1959: 296) in this metalophid spur, metaconid distinct in

connection, is Franimys amherstensis early stages of wear; Mi_3 basically tri-

from the latest Paleocene Clark Fork lophate with minute vestiges of metalo-

beds, the earliest known hystricognath phid; anterolophid, hypolophid and lin-

rodent. The fully molarifomi premolar of gual portion of posterolophid transverse;

Migraveramus suggests that the ances- dni4 replaced in normal manner, very

Oligocene Rodents of Bolivia • Patterson and Wood 383

complex, with two labial, one anterola-
bial, two lingual flexids.

Type Species. Sallamys pascuali Hoff-
stetter and Lavocat 1970.

Distribution. Deseadan Oligocene of
Bolivia.

Sallamys pascuali Hoffstetter and Lavo-
cat 1970
Figures 4-5

Diagnosis. As for the genus. Tooth
measurements as given in Table 2.

Type. Anterior part of skull, preserved
back to the glenoid on the left side, with
articulated jaw and including a complete
set of cheek teeth, in "ma collection per-
sonelle, deposee a ITnstitut de Paleon-
tologie du Museum Nationale d'Histoire
Naturelle, a Paris" (Hoffstetter, 1976:
13); listed without museum number by
Hoffstetter and Lavocat (1970: 174) and
without museum reference as "SAL 101 A
(Crane) et SAL lOlB (Mandibule)" by
Lavocat (1976: 23).

Referred Material. Eleven specimens
in the Princeton collections, listed in Ap-
pendix 1 (p. 528).

Horizon and Localities. Deseadan, ear-
ly Oligocene, lower part of Estratos de
Salla, Salla-Luribay Basin, Bolivia. PU
nos. 20906-20911 were found at Pampa
Tapial in a grayish-white, gypsiferous
clay some 10 m above a 30-50 cm thick
bed of greenish-gray limestone, within
an area 20 by 20 m; PU nos. 20912-20913
are from dark red to reddish-brown rocks
in the lower part of the Salla succession;
and PU nos. 20982 and 21727 are from
Tomolo (so spelled by Branisa) midway
between Salla and Aroma (L. Branisa,
field notes). We doubt whether Tomolo
is the same as the locality identified by
Hoffstetter as Toloma (1976: Fig. 2),
since Hoffstetter stated (1976: 14) that he
found no rodents at Toloma.

Description. Beyond stating that the
infraorbital foramen was large and that
the area of origin of M. niasseter medi-
alis, pars anterior was extensive, Hoff-

stetter and Lavocat (1970: 173) gave no
account of cranial structure. The skull
was, however, described in detail by
Lavocat (1976: 24-26).

The material at our disposal of course
reveals ver\' little of the skull structure.
The maxillary fragment shows no trace of
an alveolus for P^. There is, however, a
slight elevation on the curved surface of
the maxilla anterad and slightly mesiad
of the anterobuccal root of P^ (Fig. 4A; X).
When viewed with the light shining from
the correct angle, this area is seen to have
quite a different luster from that of the
adjacent areas of bone, and to be bifid
posteriorly. This area may represent the
closed alveolus of a shed dm^, a tooth
now known to occur in some Deseadan
caviomorphs. An X-ray photograph shows
that this area is clearly differentiated
from the adjacent bone, but that there is
no identifiable tooth structure.

The zygoma arises from the maxilla an-
terior to P as in echimyids generally, and
the ventral surface of the maxilla ascends
abruptly anterior to P"*. Medial to this
tooth is a concavity (Fig. 4A; IF), which
is the posterior end of a very elongate in-
cisive foramen that reaches back to the
middle of P, farther than Lavocat (1976:
26) was able to trace the foramen, since
this area is still buried in matrix in the
type (Lavocat, 1976: PI. 1, Fig. 3). The
foramen extends farther posterad than in
any of the Santa Cruz rodents figured by
Scott (1905), or than in any of the other
known Deseadan caviomorphs except
Incamys (Fig. 14A). There are similari-
ties to that of Metaphiomys, although it
is not sunk in a palatal depression, as is
the case in the Fayum fonn (Wood, 1968:
Fig. 6A). The dorsal view of the maxillarv'
fragment shows the passage of the sphe-
nopalatine foramen (Fig. 4B; SPF), run-
ning anteromesially from the orbit to the
nasal passage. The lateral wall of the na-
sal passage, anterior to the sphenopala-
tine foramen, contains a groove that runs
as far forward as the bone is preserved
(Fig. 4B; NLC?). This is believed to be

384 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Figure 4. Teeth and bones of Sallamys pascuali. C-E. x5; others, xio. Dotted areas are matrix. A-B. Maxillary
fragment containing RP*-M', PU no. 20906. A. Occlusal view, anterior end to the right. B. Dorsal view, anterior
end to the left. C. Lateral view of lower jaw, PU no. 20907, with anterior part of diastema, mental foramen, angular
process and incisor added from PU no. 20911. D-E. Anterior and posterior ends of PU no. 20912. D— anterior view,
broken through posterior root of P,; E— posterior view, broken just behind M^; F-H. Cross sections of lower incisors,
seen from the front. F. PU no. 20913, LI,, below diastema. G. PU no. 20909, Rl,, below front of dm,. H. PU no. 20912,
LI,, below front of M,.

Abbreviations: A — angular process; AR — anterolingual root of M.2; I — projection of position of incisor; IF — posterior
end of incisive foramen; NLC?— ?nasolacrimal canal; R— posterior root of P,; SPF— sphenopalatine foramen; X— possible
unerupted dm^ within the bone.

the nasolacrimal canal, which apparently
descended vertically from the lacrimal
foramen, above M^ and then turned
abruptly forward. This groove is separat-
ed from the sphenopalatine foramen by
a wall of bone in PU no. 20906 (Fig. 4A).
If this is the nasolacrimal canal, its course
was quite different from that in Incamys,
in which it seems to have followed a
straight line from the lacrimal foramen to
the nasal passage. In Branisamys, the ca-
nal curved more ventrally than in the oth-
er genera, and extended forward as a dis-
tinct passage (Figs. 24A, 25B). Three
small nutritive foramina are present in
the floor of the orbit, just anterior to the
sphenopalatine foramen. There is a shal-
low groove on the floor of the infraorbital

foramen, presumably marking the course
of the infraorbital nerve and blood ves-
sels.

The masseteric crest of the mandible
arises high on the horizontal ramus be-
neath the posterior part of P4 (Fig. 4C),
and becomes very large as it rims down-
ward and backward. Anteriorly, the inci-
sor lies mesiad of the roots of the cheek
teeth (Fig. 4D); but, farther back, the in-
cisor curves laterad and crowds the pos-
terior root of M2 (Fig. 4E) and all the roots
of M3. This is accompanied by a spiral
twisting of the incisor (Fig. 4D-E). The
exact location of the mental foramen
caused us some uncertainty. In PU no.
20907, the break at the anterior end of the
bone passes through what looks like a

Oligocene Rodents of Bolivia • Patterson and Wood 385

small mental foramen verv high on the
bone. However, in PU nos. 20908, 20911
and 20913, there clearly is a fair-sized fo-
ramen, just in front of P4, slightly above
the middle of the lateral side of the man-
dible, and at the level of the upper edge
of the incisor. There are several nutrient
foramina on the medial surface of the jaw
below the molars and above the incisor.
In the contemporary Xylechimys and
Deseadomys aramhourgi (but not in D.
loomisi) the masseteric crest begins be-
neath Ml, and in D. aramhourgi there
are two mental foramina, the posterior
beneath the middle of P4 (no such fora-
men is present in any of our specimens
of Sallamys pascuali) and the anterior
beneath the middle of the diastema,
somewhat farther forward than in D. ar-
amhourgi (Wood and Patterson, 1959:
Fig. 6).

The cheek teeth, notably the uppers,
increase in width posteriorly, to Mg^. For
terminology of the parts of the cheek
teeth, see Figs. 1-2 (pp. 376-377). Accord-
ing to the illustrations given by Hoffstetter
and Lavocat (1970) and by Lavocat (1976:
PI. 1, Fig. 5), Mg^ are slightly smaller than
M/ (Lavocat, 1976, unfortunately gave
no measurements except the length of
the tooth row for any of the Salla rodents,
stating, on page 22, that he had not in-
cluded other measurements because
Hartenberger was working on a statistical
study). The teeth are brachydont, al-
though the uppers show some unilateral
hypsodonty, the height of crown on the
lingual side being about twice that on the
labial (Fig. 5B). The base of the enamel
is essentially horizontal around the tooth.
As indicated below, in our opinion the
teeth are clearly four-crested (Fig. 5A).
Lavocat (1976: 29-30) describes these
teeth as four-crested, but believed that he
could find the remnants of a fifth, the
mesoloph (see below, p. 386). On a sub-
sequent page, however, he concluded
that Sallamys showed "clairement un
morphologic pentalophodonte" (1976:

43), which seems to us to be stretching a
point.

In the lower teeth, the crown is highest
at the hypoconid, but the disparity' in
height of various parts of the crown is less
pronounced than in the uppers. The mo-
lars are wider relative to length than
in Deseadomys, resembling those of
Platypittamys in this respect and also in
the very transversely directed lophs and
lophids. Except as noted below, our de-
scriptions of the teeth are based on the
Princeton specimens.

The contour of P* is oval, and the tooth
is notably smaller than \P (Fig. 4A). An-
teroloph, protocone and posteroloph
sweep around the anterior, lingual and
posterior faces of the tooth, and a slight
vertical depression on the lingual face
(too faint to show on Fig. 4A except as an
irregularity of the enamel-dentine bor-
der) may perhaps mark an exceedingly
incipient separation of a hypoflexal divi-
sion. Lavocat (1976: 29) did not see such
a depression, and none shows on his ste-
reophoto {op. cit.: PI. 1, Fig. 5). The la-
bial extremity of the anteroloph closely
approaches the paracone, from which it
is separated by a shallow notch; with a
little more wear the two would have
united. The paracone is large, elongate
transversely and independent, the pro-
toloph being incomplete. As a result of
this, the paraflexus and the mesoflexus
are confluent, fonning a deep U-shaped
fossa in the center of the tooth. This is
partially divided by a thin, low ridge —
the serial homolog of the mure of the
molars — connecting a spur from the pro-
tocone to one from the posterior crest and
isolating a deep central basin. This ridge
is absent on the type (Lavocat, 1976: PI.
1, Fig. 5). The posterior crest is rather
wide in its labial half, where it contains
a small fossette, the serial homolog of the
molar metafossette. The tooth is thus in-
cipiently tetralophate and, for an echi-
m\id, is advanced for its time in this fea-
ture. Hoffstetter and Lavocat state that, in

386 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

the type, there is a "vestige" (rudiment,
in our opinion) of the metaloph on the
anterior face of the posteroloph, but this
is not shown in their figure. Their figure
does, however, show a small crest on the
posterior side of the protoloph (1970:
Fig.), which we believe to be homolo-
gous to what we interpret as the mure
(Fig. 4A). However, Lavocat (1976: 20)
does not mention either structure, and
his stereophoto (op. cit.\ PI. 1, Fig. 5)
shows neither one.

In this respect, P of PU no. 20906 is at
once more and less advanced than in con-
temporary octodontoids. It is more ad-
vanced than that of the type (Hoffstetter
and Lavocat, 1970: 173, Fig.) in the com-
plete incorporation of the anterocone in
the anteroloph, the presence of the mure,
the fossette in the posteroloph, and the
incipient lingual division. It is less ad-
vanced than those of Platypittamys and
Deseadomys arajubourgi in the isolation
of the paracone; more advanced than
either in the incipient division of the
posteroloph; and more advanced than
Platypittamys in the possession of a
mure. What all this indicates, we suspect,
is merely that octodontoids, and indeed
all caviomori^hs, were actively undergo-
ing molarization of the premolars during
Deseadan time and that the process was
accompanied by much individual and
populational variation. This definitely in-
dicates, to us, that the ancestral cavio-
morph stock did not have fully molarized
upper premolars.

The first upper molar of PU no. 20906
(Fig. 4A) is worn in a manner slightly dif-
ferent from that of the corresponding
tooth of the type. The metafossette has
been worn away, and the paraflexus has
been reduced to a very shallow groove,
although it is deeper than in the type
(Lavocat, 1976: PI. 1, Fig. 5). The hypo-
flexus is still much more widely open
than is shown by Hoffstetter and Lavocat
(1970: Fig.) or by Lavocat (1976: PL 1,
Fig. 5). The deep mesoflexus turns pos-
teriorly at its lingual extremity, revealing

that, as in M^ of the type, the metaloph
had shifted its moorings from the mure to
the posteroloph. In the type, but not in
the Princeton specimen, the fomier site
of metaloph attachment is marked by a
minute projection on the mure. Lavocat
(1976: 30) discussed this projection as
"un portion linguale rudimentaire du
mesolophe," an interj^retation that seems
to us to be without justification. The
crestlet is aligned with and points toward
the metacone, which should not be true
of a mesoloph. It lies well behind the
middle of the mure, whereas a mesoloph
normally lies at the center of the mure.
Lavocat pointed out that there is a low
ridge, below the wear surface of M* of the
type, that connects the metacone and
what he considers the rudiment of a me-
soloph. This low ridge is not present in
the Princeton specimen. The presence of
this ridge would seem to us to demon-
strate that the entire assemblage is a
metaloph, which would disprove Lavo-
cat's assertion that this ridge strengthens
the evidence for the existence of a me-
soloph.

The isolated left upper molar, PU no.
20982 (Fig. 5A), is probably LM^ rather
than LM^ because (by comparison with
Hoffstetter and Lavocat, 1970: 173, Fig.)
the relationships of the three measure-
ments are closer to those of M^ than to
those of M^. Moreover, Lavocat's illustra-
tion of the type (1976: PI. 1, Fig. 5) shows
a complete metaloph on M^ and an inter-
rupted one on ML The Princeton tooth is
clearly a molar because of the wear facet
on the anterior face (Fig. 5B). The width
of the posterior half (Table 2) shows that
it cannot be M^. The absence of a wear
facet on the posterior face is because this
tooth was newly erupted at the time of its
owner's death, and the next succeeding
tooth had not yet come into contact with
it. This tooth is less worn than any tooth
figured by Hoffstetter and Lavocat (1970)
or by Lavocat (1976), and clearly shows
(Fig. 5A) the unworn pattern, essentially
as described in the previous paragraph.

Oligocene Rodents of Bolivia • Patterson and Wood 387

Figure 5. Cheek teeth of Sallamys pascuali, Protacaremys prior, and Prospaniomys priscus. H., x5; others, xio. A-B.
Sallamys pascuali, PL) no. 20982, LM'. A. Occlusal view. Dotted area is matrix. B. Anterior view, showing unilateral
hypsodonty. C. S. pascuali, PL) no. 20907, RP4-M2. D. S. pascuali, PL) no. 20909, Rdm4-M2. Unerupted Mj with no enampi
cap or, at most, an incomplete one. Grooves are in dentine. E. S. pascuali, PU no. 21727, RNVa- F- Protacaremys prior,
FMNH no. P 13295, RM,. G. Protacaremys prior, AMNH no. 29707, Rdm^-M^. H. Prospaniomys priscus, AMNH no. 29697,
Ldm^-Mj.

There are only four crests, with not the
slightest suggestion of a fifth. The third
(the metaloph) is finnly attached to the
posteroloph and isolated from the hypo-
cone, although there is a crestlet from the
hypocone, running posteriorly across the
mesoflexus, at right angles to the direc-
tion of what Lavocat (1976: 30) described
as a rudimentary mesoloph. The tooth is
supported by three roots, two buccal and
one lingual. The lingual side of the tooth
has about twice the height of the buccal.
The last two upper molars are not rep-
resented in the Princeton collection. The
published figures and description show
that the third loph, the metaloph, has the
usual attachment to the mure, and, sur-
prisingly, that the hypoflexus of M^ was
converted by wear to a hypofossette ear-
lier than on M^ Evidently there was in-

dividual variation in depth of flexus as
well as in angle of wear.

The last upper molar, as figured by
Lavocat (1976: PI. 1, Fig. 5) seems to us
to have a structure completely homolo-
gous to that of the anterior molars, with
a complete metaloph followed by a com-
plete posteroloph. The latter bends to-
ward, but does not reach, the metaloph.
This identity of structure among the mo-
lars seems to us clearly to demonstrate
the accuracy of our interpretation. We
continue to feel that there is no possible
basis for assuming, as does Lavocat, that
the four-crested upper teeth of Sallamys
can be interpreted as "structures penta-
lophodontes" (Lavocat, 1976: 30).

Lavocat's discussion of Wood's inter-
pretation of the crests in Sallamys (La-
vocat, 1976: 30) seems rather confused,

388 Bulletin Museum uf Comparative Zoology, Vol. 149, No. 7

since he comments on the crest consid-
ered by Wood to be the mesoloph. But
Wood (1974b: discussion, Fig. 1) clearly
indicated that, in his opinion, there is no
mesoloph in any caviomoq^h, a position
that we still maintain.

Lavocat (1976: 31) stated that the large
size of the protocone and small size of the
hypocone of the upper molars of Salla-
niys is a primitive condition, "plus
proche du plan trigonodonte" than are
the teeth of Platijpittanujs. We disagree
with this conclusion; none of the Ca-
viomorpha are very close to the trigono-
dont condition. The rodents that invaded
South America must, we feel, already
have had a well-developed hypocone.
The difference in size of the protocone in
Sallamys and Platypittamys arises from
the greater lingual hypsodonty (an ad-
vanced character) of the fornier, which
would pennit a rapid enlargement of the
protocone with wear. How this would oc-
cur is clearly shown in Fig. 5A.

According to Lavocat (1976: 31), "le
sinus interne des molaires se fenne lin-
gualement, avec I'usure, alors que les
vallees externes sont encore bien ou-
vertes labialement." However, his ster-
eophotograph of the type (1976: PI. 1,
Fig. 5) shows that there is no sequence
of closure of the valleys that is consistent
from tooth to tooth, and it seems evident
that there is great individual variation
both in the openness of the hypoflexus,
paraflexus and metaHexus (see, for ex-
ample. Fig. 5A, metaflexus) and the angle
at which the wear surface cuts the teeth.
This latter is highly variable because of
the lingual hypsodonty, which resulted
in rotational eruption of the upper teeth,
albeit less pronounced than in Incamys
(see below, pp. 416-418).

As Hoffstetter and Lavocat stated, the
talonid of P4 is completely molariform; a
hypolophid as large as those of the molars
is present, its lingual extremity joining
the posterolophid to isolate a metafosset-
tid. We have previously been in some
doubt as to the identification of the pos-

terior lophid of P4 of echimyids and oc-
todontids, and have tentatively identified
it as the serial homolog of the molar hy-
polophid (Wootl, 1949; Wood and Patter-
son, 1959: 307). Whatever the situation
may be in the previously described
fomis, it would seem that, in Sallamys at
least, the posterior crest should be called
the posterolophid.

The above-cited molarization of the tal-
onid is an advanced character for a De-
seadan octodontoid; neither Platypitta-
mys (Wood, 1949: Fig. 3C-D) nor
Descadomys arambourgi (Wood and Pat-
terson, 1959: Fig. 5A) shows any trace of
it. Anteriorly, in the type of Sallamys
pascuali (Lavocat, 1976: PI. 1, Fig. 6), the
rather stout metaconid has not quite
fused with the protoconid, and a short
metalophid spur (identified by Lavocat,
1976: 31, as more probably a mesolophid)
projects posterolingually from the proto-
conid; the whole is rather similar to the
trigonid area of D. arambourgi. In the
somewhat more worn PU no. 20907 the
protoconid and metaconid are united, but
the metalophid spur is still clearly visi-
ble, running to the posterior margin of
the protoconid (Fig. 5C). The position
and relationships of the crest in this tooth
seem to us to demonstrate that it cannot
be a mesolophid. With a little more wear,
this crest would lose its identity, and the
tooth would become almost wholly mo-
larifonn. The ectolophid bears an en-
largement that looks somewhat like, but
is not, a mesoconid (Fig. 5C). This fea-
ture is not present in P4 of the type (La-
vocat, 1976: PL 1, Fig. 6).

The lower molars are rather simple in
structure, consisting of three essentially
transverse lophids (the antero-, hypo-,
and posterolophids) separated by deep
meso- and metaflexids, and of anterola-
bially directed, moderately acuminate
protoconid and hypoconid separated by
a deep hypoflexid (Fig. 5C-E). This last
is by far the deepest of the folds, but, due
to the oblique manner of wear, all three
folds were evidently obliterated at about

Oligocene Rodents of Bolivia • Patterson and Wood

389

Table 2. Measurements, in mm, of teeth of Sallamys pascuali.

Lower teeth

PU nos.

20907

20908

20909

20910

20911

20912

20913

21727

21950

R

R

R

R

L

L

L

R

R

N

X

P4-xM,-, anteroposterior

8.70

1

P4 anteroposterior

L92

d2.32

1/1

width metalophid

L41

dl.21

1/1

width hypolophid

L84

dl.47

1/1

M, anteroposterior

L84

1.92

1.89

1.98

2.00

2.10

6

1.96

width metalophid

L87

2.01

1.69

1.88

2.03

1.97

6

1.91

width hypolophid

1.97

1.96

2.03

2.06

2.15

2.20

6

2.06

M2 anteroposterior

1.81

ca. 2.0 ca. 1.9

1.93

2.14

2.17

2.38

7

2.05

width metalophid

2.22

ca. 2.3

2.32

2.18

4

2.24

width hypolophid

2.21

ca. 2.4

2.39

2.21

2.48

5

2.34

M3 anteroposterior

2.39

1

width metalophid

1.98

1

width hypolophid

1.74

1

Ii anteroposterior

1.66

1.13

1.53

1.16

1.73

1.51

6

1.45

transverse

1.07

0.75

1.04

0.77

1.27

1.07

6

1.00

ratio

.62

.67

.68

.66

.73

.71

6

.68

Upper teeth

PU no. 20906

PU nos. 20906

20982

R

R

L

P* anteroposterior

1.45

M' anteroposterior

1.88

1.84*

width protoloph

2.12

width protoloph

2.65

2.33*

width metaloph

2.04

width

metalopl

1

2.21

2.22*

d = deciduous tooth (dm4).
* possibly M^.

the same time, the occkisal surfaces end-
ing up as featureless quadrangles sur-
rounded by a continuous wall of enamel.
Even in the highly worn teeth of PU nos.
20908 and 20911, the anterior enamel,
although greatly thinned by interdental
wear, is still continuous across the front
of the tooth. An unworn Mj (PU no.
20909, Fig. 5D) shows that there was no
trace of anything that could possibly be
imagined as a mesolophid; that all crests
were thin and continuous upon eruption;
and that there are minute irregularities,
which can only be vestiges of the poste-
rior ami of the metaconid, near the center
of the upper half of the posterior face of
the anterolophid. This seems to show on
M3 of the type (Hoffstetter and Lavocat,
1970: 173, Fig.; Lavocat, 1976: PI. 1, Fig.
6), but to have been eliminated by wear
on M3 of PU no. 21727 (Fig. 5E). M3 nar-
rows posteriorly.

An unerupted Mg is present in PU no.
20909. It lacks the entire enamel cap, and
shows the tooth pattern only in dentine.
The pattern seems to have been identical
to that of Ml of the same specimen (Fig.
5D), and again there clearly was no me-
solophid. There is a series of radiating
furrows in the dentine, which may have
marked the attachment of the enamel
cap, or which may be shrinkage cracks in
the dentine. More worn specimens of M2
(Fig. 5C,E) likewise show a pattern in-
distinguishable from that of Mj.

The incisor in PU no. 20913 (Fig. 4F)
is crowded against the posterior root of
M2, and the curvature is such that the
crowding would have been even greater
beneath M3. This is confimied by the
edentulous fragment of a left ramus, PU
no. 20913, in which the roots of M3 are
small, crowded and exceedingly short,
because of interference by the incisor.

390 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

These observations suggest that Mg may
have been very variable in this species,
and in process of reduction.

A Httle worn, essentially complete dm4
is present in PU no. 20909 (Fig. 5D); it
is a complex tooth. Appreciably longer
than any of the molars, it tapers forward
rather rapidly. The trigonid and talonid
slope down toward the ectolophid, which
is worn. A possible explanation of this
state of affairs, unusual in so lophate a
rodent tooth, is that dm^ was present, as
in Incamys and Branisamijs, and that
this tooth and dm"* fomied a wedge when
newly erupted, with dm^ opposing the
anterior third of dm4, and dm'* opposing
the posterior two thirds. There are two
buccal, one anterobuccal and two lingual
flexids. The identities of the various lo-
phids separated by these flexids (and
hence those of the flexids) can be deter-
mined by comparison of dm4 in various
Deseadan and Colhuehuapian cavio-
morphs, as discussed below (pp. 500-501).
The talonid is molarifonn. The large hypo-
conid was separated by notches from the
posterolophid and hypolophid when the
tooth was freshly erupted, to judge from
its very constricted connections in the
specimen. The metaflexid is large and
deep, and would convert to a fossettid
with extreme wear. Anterior to the point
of divergence of the transverse hypolo-
phid, the sinuous, centrally situated ec-
tolophid gives off a labially directed lo-
phid that parallels the angulate labial
extension of the hypoconid. The deep
flexid between them is the hypoflexid,
and the anterior lophid is probably the
serial homolog of the molar crest be-
tween protoconid and ectolophid that has
here parted company with the protoco-
nid. The flexid between this neomorph
labial lophid and the protoconid is also
a neomorph; it is a little shallower than
the hypoflexid, and a minute cuspule is
situated at its buccal extremity, probably
a remnant of the ancestral marginal crest
in this area. Immediately anterior to the
point of junction with this neomorph la-

bial lophid, the ectolophid runs anteroin-
ternally to fuse with one crest running
posterointernally from the protoconid
and with another shorter one running
posteroexternally from the metaconid.
The latter is the serial homolog of the
metalophid, while the fonner is what we
have elsewhere (Wood and Patterson,
1959: 288) called the neolophid. The an-
teroconid is a large cusp that still retains
much of its individuality. A slight notch
separates it from the metaconid, a much
deeper one from the protoconid, and
there is a posterior extension, directed
toward the ectolophid, which may be the
initial stage in a connection of the an-
teroconid with this crest. The anterola-
bial flexid, lateral to the anteroconid,
deepens toward the center of the tooth,
and would presumably convert to a fos-
settid at a late stage of wear.

The lower incisor extends back to a
point slightly behind and pressed against
the roots of Mg. Anteriorly, it lies mesiad
of the roots of the cheek teeth (Fig. 4D),
and is oriented in the usual rodent man-
ner, with the enamel cap ( = anterior face)
down. However, farther back the tooth
rotates (Fig. 4E) so that, beneath the rear
of M2 or the front of M3, the incisor is
twisted, with the enamel cap directed
mesiad. Where this takes place, the tooth
presses against the posterior root of M2
(Fig. 4E) and all three roots of Mg (PU no.
20913), greatly reducing their lengths.
The incisor is rather small (Table 2), slen-
der and oval in outline, the posterior por-
tion being very little narrower than the
anterior face. The anterior face is smooth
and gently convex (Fig. 4E-H). The
enamel extends over approximately half
of the lateral face and over about a third
of the medial. Unfortimately, it was im-
possible to obtain measurements of the
rear end of the incisor of the juvenile PU
no. 20909, due to breakage, so that we
have been unable to detemiine the rate
of diametric growth of the tooth.

Relationships. As we have previously
stated (1959: 300), the Deseadan mem-

Oligocene Rodents of Bolivia • Patterson and Wood 391

bers of the Octodontidae and Echimyi-
dae are very similar to each other in mo-
lar structure, and are separated formally
only because their descendants diverged.
From the Colhuehuapian on, the two
families are distinguished by the reten-
tion of dnV in the Echimyidae and their
replacement, ontogenetically, by P/ in
the Octodontidae. So far as known, there
was no retention of dm/ in Deseadan
echimyids. Of the molar characters that
we previously gave as separating the ear-
lier members of the two families, Salla-
mys agrees with octodontids in two: the
decidedly transverse alignment of the
lophs and lophids and the width/length
relationships. It agrees with echimyids in
the somewhat deeper flexi and flexids
and most importantly, in the loss of the
metalophids. In this feature Sallaiiiijs is
almost as advanced as Deseadomys loo-
misi, more advanced than D. arambour-
gi, Xylechimys, or the Colhuehuapian
Protacaremys (Fig. 5F-G), Prospani-
omys (Fig. 5H), and Protadelphomys.
Loss or interruption of the metalophid is
very rare in the molars of the earlier oc-
todontids. Out of a rather large series of
the Santacruzian Sciamys principalis in
the Ameghino collection, only two show
metalophid irregularities (MACN nos. A
1886 and A 4113). We therefore conclude
that this last feature is the most critical
one so far recognized for separating the
Echimyidae and Octodontidae during
the Deseadan, and hence refer Sallamys
to the Echimyidae with some confidence.
Lavocat recognized the similarity of the
Deseadan echimyids and octodontids,
and thought that "il s'agit pratiquement
ici de saisir la separation entre deux ra-
meaux issus d'une meme souche . . ."
(1976: 72). He concluded (1976: 73) that
there are many reasons for placing Sal-
lamys in the Octodontidae; that it is by
no means evident that the octodontids
and echimyids were as clearly separable
in the Oligocene as we had indicated (we
think we made it clear in 1959, pages 300
and 389, that the Deseadan members of

the two families were very closely relat-
ed); and that he left Sallamys in the
Echimyidae because Wood (1974b: 26)
had so placed it and the evidence for
shifting it to the Octodontidae was not
conclusive. The two molar characters in
which Sallamys agrees with the earlier
octodontids (but also, it may be noted,
with some Recent heteropsomyines)
should no longer be considered as having
diagnostic significance. As an aside, this
is an argument against strict numerical
taxonomy.

Hoffstetter and Lavocat (1970: 174) see
a resemblance to Paraphiomys in the
lower molars of Sallamys. This exists,
but is completely offset by the wholly
different dm4 in Paraphiomys pigotti and
P. stromeri (cf. Fig. 5D and Lavocat,
1973: PI. 26, Figs. 7-10) as well as in the
Fayum P. simonsi (Wood, 1968: Fig. 5F),
transferred below (p. 520) to Neosciuro-
mys. Furthemiore, it must be noted that
these species oi Paraphiomys, like all the
Petromuridae, have a cingulum anterior
to the protoconid of the molars, a feature
lacking in caviomorphs (see p. 496). As
indicated below (p. 521), we retain Para-
phiomys in the Petromuridae, rather than
place it, as does Lavocat, in the Thryon-
omyidae.

Within the Echimyidae, Sallamys ap-
pears to fit rather comfortably into the
Heteropsomyinae (Fig. 6), displaying, in-
cipiently at least, a number of the molar
characters of that group (Patterson and
Pascual, 1968: 6). It differs from the Adel-
phomyinae in the more transverse lophs
and lophids and a greater degree of uni-
lateral hypsodonty, and from Prospani-
omys, the earliest member of the My-
ocastorinae (if Patterson and Pascual
were correct in deriving Myocastor from
the Spaniomys group of echimyids), in
the loss of the metalophids and the more
complex dm4 (cf. Fig. 5D,H).

In their definition of the Adelphomyi-
nae, Patterson and Pascual (1968: 5) stat-
ed that there was "no unilateral hypso-
donty." This is true of the later fomis, but

392 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

MYOCASTORINAE

PLAGIODONTINAE

CAPROMYINAE

HETEROPSOMYINAE

SANTACRUZIAN spaniomys

ACARECHIMYS

COLHUEHUAPIAN

DESEADAN

PROSPANIOMYS

PROTACAREMYS

OCTODONTIDAE

Figure 6. Phylogeny of the early Echimyidae.

the earliest, Deseadomys, shows some
degree of it in the upper teeth. The state-
ment should be amended to read: "slight
initial unilateral hypsodonty, not subse-
quently increasing."

Sallamys displays no detectable ten-
dencies toward the types of molar struc-
ture encountered in the Echimyinae and
Dactylomyinae, subfamilies so far unrep-
resented, or at any rate unrecognized, in
the Tertiary.

The Colhuehuapian Paradelphomys
(Patterson and Pascual, 1968: Fig. 3)
shows some resemblances to the dacty-
lomyines in the obliquity of the lophids,
the isolation of the posterolophid, the ex-
tension of the mesoflexid nearly to the
anterior extremity of the tooth, and the
simple trigonid of dm,. Whether or not
this is in any way indicative of affinity is
at present quite uncertain and will re-
main so until later Tertiary dactylo-
myines are discovered. The possible re-
lationships are suggested in Fig. 6.

On the basis of the few early forms
then known, we tentatively sketched a
phylogeny of the Echimyidae (1959: 302,
Fig. 3) that suggested an early dichotomy

into Adelphomyinae and the ancestry of
the remainder of the family. The discov-
ery of Sallamys suggests that this may
have been an oversimplification, and that
the actual situation was even more com-
plex than we had supposed. Sallamys is
about as advanced in molar structure as
is Deseadomys, although in somewhat
different ways, and is less advanced in
the premolar. It is even more advanced
over the later Protacaremys and Pro-
spaniomys as regards crown height and
metalophid reduction, and possesses a
much more complex dm4 than does Pro-
spaniomys (cf Fig. 5D,H). It is more ad-
vanced than the poorly known Colhue-
huapian Protadelphomys in metalophid
reduction and than Paradelphomys in
lower milk molar complexity.

As pointed out below (p. 501), the adel-
phomyines and the Spaniomys group
have a relatively simple dm4, comparable
to those of octodontids, whereas the
known heteropsomyines have a complex
dm4 nearly comparable to that of Ceph-
alomys. The latest Pleistocene or early
Recent Antillean heteropsomyines have
a simple dm4, which suggests either that

Oligocene Rodents of Bolivia • Patterson and Wood 393

ECH'MYINAE DACTYLOMYINAE

CHAETOMYINAE

?
? ?

^ADELPHOMYINAE

STICHOMYS ADELPHOMYS

PARADELPHOMYS

XYLECHIMYS

DESEADOMYS

Figure 6. Continued.

not all early members of the subfamily
had a complex one, or (a less probable
hypothesis) that secondary simplification
occurred in the ancestry of these forms.

We interpret these bits of evidence to
mean that the heteropsomyines began
their radiation at least as early as did the
adelphomyines, and that the Spaniomys
group may well be of equal antiquity and
not an offshoot from the Heteropsomyi-
nae, despite the close similarity between
Protacaremijs and Prospaniomys in mo-
lar structure.

Sallamys cannot have been ancestral to
any Colhuehuapian or Santacruzian het-
eropsomyine now known. With a few eas-
ily made mental rearrangements of crests
and valleys it would be possible to visu-
alize it as being in or near the ancestry of
various later fomis, but in the absence of
known intermediates such exercises
would be vain and probably misleading.
There is a resemblance, however, be-
tween Sallamys and Chasichimys of the
Arroyo Chasico (Pascual, 1967).

Due to conservative inertia as much as
anything else, we have hitherto main-
tained the Capromyidae as a distinct fam-

ily descended from the Echimyidae.
Morphologically, this is not really justi-
fied, and, in general agreement with au-
thors such as Ellerman (1940), we now
place the two Antillean subfamilies Cap-
romyinae and Plagiodontinae in the
mainland family. Patterson and Pascual
(1968:1, n. 2), in commenting on this pos-
sibility, stated that the subfamilies of the
Echimyidae and Capromyidae would in
that case become tribes. We do not take
this step here because the capromyines
and plagiodontines, while agreeing in
possessing hypselodonty (the capro-
myine Hexalobodon alone excepted) and
cement, differ in cheek tooth structure
quite as much as do the other subfami-
lies. Their divergence, if they diverged
from a common stock, may or may not
have taken place in the Antilles; these
subfamilies may or may not be relict sur-
vivors of the original caviomorph migra-
tion to South America (see below, p. 456);
at present there is no evidence that bears
on these questions.

The occurrence of macroscopic depos-
its of cement is sporadic in the Cavio-
morpha — as in rodents generally — and is

394 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

not necessarily indicative of higher affin- ly negative as regards erethizontid affin-
ities. In adcUtion to the above-mentioned ities and one strongly positive as regards
Antillean forms, it is present in the De- echimyid ones. In contrast to the known
seadan Litodoutomys, in its possible rel- erethizontids and in agreement with all
ative the Santacruzian Olenopsis, in other living caviomorphs, Chaetomys
chinchillids from the Deseadan on, in lat- lacks a posterior carotid foramen, and
er eocardiines and in caviids and hydro- hence an internal carotid artery. Eller-
choerids, and in Pliocene and later octo- man (1940: 175, Fig. 42) shows what
dontids, abrocomids, myocastorines and could be interpreted as a small foramen
dinomyids. A thick deposit of cement is anterointernal to the foramen lacerum
merely an enlargement of the universally posterior. In the MZUSP skull there is a
present microscopic layer of cement that depression in this area that does not com-
covers unerupted mammalian teeth, and municate with the cranial cavity. Reten-
is associated with hypsodonty and hyp- tion of this artery is a basic character of
selodonty in rodent lineages in which the the New World porcupines, as Bugge
other components of the cheek teeth are (1974a: 71; 1974b: 70) has recently
either separated or only narrowly con- stressed. In agreement with all post-De-
nected. seadan echimyids and in contrast to all

The peculiar genus C/iflefom(/s is now- other caviomorphs, dm4^ are retained

adays universally accorded a subfamily of throughout life. There can be no doubt of

its own, usually in the Erethizontidae. this in the specimen of C. siibspinosiis

There have been few informed com- examined (MZUSP unnumbered). The

ments on the taxonomic position of this anterior cheek teeth are lower in crown

subfamily, due to the rarity of specimens height than are the first molars, and the

of Chaetomys. Through the kindness of flexi and flexids are more nearly obliter-

Dr. Paulo Emilio Vanzolini, one of us ated, both features being due to wear.

(B.P.) recently had the opportunity to ex- They were clearly the first of the series

amine a skeleton of Chaetomys suhspi- to have erupted, and X-rays have re-

nosus in the MZUSP. vealed no trace of replacing teeth be-

If one looks mainly, or only, at the neath them. We believe reference of the

structure of the feet, placing the Chae- Chaetomyinae to the Echimyidae to be

tomyinae in the Erethizontidae is a logi- fully justified.

cal step, the resemblance between the With this transfer of Chaetomys, the

two groups in this respect being striking- Erethizontidae becomes a very compact

ly close. The cheek teeth of Chaetomys, group, and the Echimyidae an even more

however, are very different in structure complex one. However, the Echimyidae

from those of the erethizontids. On the now include all caviomorphs in which

basis of tooth structure, Schaub (in Steh- dm4* are retained.

lin and Schaub, 1951: 369) placed Chae- As a subfamily of the Echimyidae, the

tomys in the Echimyidae, following Mil- Chaetomyinae may be defined as fol-

ler and Gidley (1918: 445) who, however, lows: cheek teeth generally resembling

as usual, gave no reasons for their action, those of the Echimyinae, but uppers with

Later (1958: 742), Schaub placed Chae- five lophs, metaloph fomiing an indepen-

tomys in the echimyid subfamily Echi- dent central crest; malar robust, very

myinae along with Heteropsomys and deep anteriorly, with large postorbital

Spaniomys, as opposed to the only other process nearly meeting well-developed

subfamily he listed, the Dactylomyinae, postorbital process of frontal, and with

again stressing the cheek-tooth structure, little or no trace of jugal fossa on lateral

This familial assignment is strongly surface; parietal crests well developed,

supported by two characters that we re- only slightly convergent posteriorly; bul-

gard as fundamental, both of them strong- la with prominent porus acusticus exter-

Oligocene Rodents of Bolivia • Patterson and Wood 395

nus directed anteroextemally; paroccipi- problematical than that of the other

tal process blunt, barely projecting; subfamilies.

coronoid process of mandible greatly re- x •, <^A\/i<->in>i-A 1/ i- ■ u

duced; feet erethizontid-like. i^i^T^ CAVIOIDEA Kraghevich

Our present very tentative concept of DASYPROCTIDAE, DINOMY-

echn^y.d phytogeny ,s summed up in CUNICULIDAE (pp. 51 1 , 523)

F,g. 6. For so long as five of he eight DASYPROCTIDAE Smith 1842

subfamilies are unknown or . "recog- J Hoffstetter and Lavocat 1970

nized in the Tertiary, matters will remain ^

in an unsatisfactory state; the evolution Emended Diagnosis. Skull basically

of all five may well have gone on for the dasyproctid, with large orbits and bullae;

most part in the northern half of the con- snout elongate, ventral side horizontal;

tinent or (for the Capromyinae and Pla- anterior part of fossa of origin for M. ma.s-

giodontinae) in the Antilles. Our placing seter medialis weakly marked, posterior

of the Dactylomyinae next to the Adel- part deep and corrugated; nasals parallel-

phomyinae rests on nothing more than sided, tubular anteriorly; large interpre-

the few resemblances between the living maxillary foramen; incisive foramina

forms and Paradelphomijs noted above enormous, elongate and laterally expand-

(p. 392); these hardly constitute compel- ed; lateral wall of nasolacrimal canal very

ling evidence. Whether or not the Echi- thin; interorbital fenestra between fron-

myinae arose from or had a common an- tal and maxillary; anterior suture of malar

cestry with the Heteropsomyinae is with maxillary straight; no postorbital

wholly conjectural. The placement of the process or jugal fossa on malar; orbito-

Capromyinae next to the Myocastorinae sphenoid large, quadrate; no separate in-

rests on somewhat fimier ground. The terparietal; mastoid continuous with bul-

two groups have long been associated by la, inflated and septate; large fenestra

a number of authorities, usually in a fam- between presphenoid and palatine,

ily of their own. George and Weir (1974: opening into orbitotemporal fossa; dental

102) conclude that on karyotypic evi- fomiula 1/ Pi^ Mg^, dnij^; incisors slen-

dence the two groups are closely related, der, proportionately small; cheek teeth

more closely than either is to the heter- tetralophate, marked lingual hypsodonty

opsomyine Proechimys, the only other in uppers; anteroloph fully separate in

echimyid whose karyotype was known to earliest stages of wear, becoming joined

them. Woods (1972: 189) reported that with paracone buccally; other lophs unite

myological evidence does not permit a early and converge posterointenially on

decision as to whether Mijocastor is an hypocone; posterolophid isolated in ear-

echimyid or an octodontid (dental evi- ly stages of wear of lower teeth; other lo-

dence clearly favors echimyid affinities), phids unite early and converge anteroex-

Gorgas, in a comparative study of the ro- ternally on protoconid; metalophid short,

dent stomach and alimentary canal (1967: connected to protoconid.

367), has found that Mijocastor and the Type Species. Incamys bolivianus

capromyines stand apart from other echi- Hoffstetter and Lavocat 1970.

myids-and also from each other-in Distribution. Deseadan Oligocene,

some features. It seems to us that all this Bolivia and Patagonia,

is not incompatible with an early origin i^camys bolivianus Hoffstetter and La-

of myocastorines from the basal echimyid vocat 1970

stock, with the capromyines perhaps aris- Figures 7-19

ing from some member of the myocas- , i ,■■ u ce ^ ». .. a t o,.^^of iQ7n.

=> -^ . Incamys bolivianus Hottstetter and Lavocat 19 /U:

torine group. The position ot the Chae- 173, Fig.; Lavocat, 1976: 50-54; PI. 3, Figs. 1-7;

tomyinae within the family is even more PI. 4, Figs. 1-3, 7.

396 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Si/noin/ui. Incamys pretiosus Lavocat 1976: 54-59;
PI. 4, Figs. 4-6; PI. 5, Figs. 6-7; PI. 6, Figs. 3-5.

Diagnosis. As for the genus. Tooth
measurements as given in Tables 3-4.
The statistics seem to us to demonstrate
that our sample represents a single vari-
able species.

Type. Anterior portion of the skull,
back as far as M^, with associated left ra-
mus, including a complete set of upper
and lower cheek teeth, deposited in Mu-
seum Nationale d'Histoire Naturelle,
Paris; cited without identifying number
by Hoffstetter and Lavocat (1970: 173)
and without museum identification as
SAL 117A (skull) and SAL 117B (ramus)
by Lavocat (1976: 50).

Synonymized Species. Lavocat (1976:
54) described a second species of Inca-
mys, I. pretiosus, which he differentiated
from the genotypic species only by say-
ing "molaires superieures sans meso-
lophe, metalophe developpe." If we are
correct in not recognizing a mesoloph in
any caviomorph, this becomes a mean-
ingless distinction. Hoffstetter and La-
vocat (1970: Fig.) and Lavocat (1976: PI.
4, Figs. 1, 3) show the metaloph in In-
camys either joined to (M^~^) or about to
join (P^) the combined protoloph and
mure. In only one (PU no. 20963, Fig.
17F) of the 50 specimens of upper cheek
teeth in the Princeton collection does the
metaloph do this, although in some other
specimens it might make such a junction
with further wear (Fig. 17D). Lavocat' s
specimens that he referred to I. pretiosus
(1976: PI. 4, Fig. 4; PI. 6, Fig. 3) agree
with the majority of our specimens.
There is very considerable variability in
all of the teeth of /. bolivianus, a point
concurred in by Lavocat (1976: 46),
enough so that we at first suspected the
possibility of there being two species at
Salla, but which we now feel means that
there was only a single, highly variable
species. Certainly the tooth measure-
ments (Tables 3^) indicate a high de-
gree of size variability, but it seems to be
continuous. It is partly due, as discussed
below (p. 416), to the rotational eruption

of the teeth, and there does not seem to
be any association between size and mor-
phological variants. Lavocat (1976: 41)
suggests that there may have been a third
species of Incamys at Salla, but he gives
no data. Assuming that the scale given for
Lavocat's stereophotos is accurate, the
upper teeth of /. pretiosus (1976: PI. 6,
Fig. 3) fall within the size range of those
of /. bolivianus (Table 3), and their pat-
tern is identical to that here considered
to be typical of /. bolivianus (Fig. 17B).
We therefore conclude that /. pretiosus
is a synonym of /. bolivianus.

Referred inaterial. 140 specimens in
the MNHN, Princeton, and MACN col-
lections, listed in Appendix 1, p. 528.

A maxilla, SAL 118, containing RP*-
M^, differs markedly from all other spec-
imens that have been referred to Inca-
mys (Lavocat, 1976: PI. 4, Fig. 2), and
represents, we believe, an undescribed
genus, as discussed below (p. 430).

Horizon and Localities. Deseadan, ear-
ly Oligocene, lower part of Estratos de
Salla; Salla-Luribay Basin, Bolivia. PU
nos. 20937^0 and 20945-6 are from To-
molo (see above, p. 383, for discussion of
Tomolo and Toloma), midway between
Salla and Aroma; nos. 20941-3 and
20947-9 were collected in the vicinity of
Salla Fami; nos. 20950-2 are from dark
red to reddish-brown rocks in the lower
part of the Salla succession; nos. 20953-
81 were found at Pampa Tapial, in a gray-
ish-white, gypsiferous clay some 10 m
above a 30-50 cm thick bed of greenish-
gray limestone within an area of 20 by 20
m that also yielded specimens of Salla-
mys pascuali. Nos. 20916-36 and 20949
are without precise locality data. The par-
tial skull, PU no. 20944, was found at lo-
cality "V-33," the precise location of
which is not known to us (L. Branisa,
field notes). MACN no. A 52-113 is pre-
sumably from Cabeza Blanca, Chubut,
Argentina.

Description. The juvenile Princeton
skull has suffered some distortion — par-
tially corrected in the figures — and was
enclosed in exceedingly hard matrix,

Oligocene Rodents of Bolivia • Patterson and Wood 397

n

X

CO

o

c

Q.
to"

C

CO

o

■Q
(0

o

c

3

o
S

■>
«

o
Q

a>

3

a

CO
(0

o

a

O
Q.

E

CO

CO

O

E

CO

3

tn

I

O
CO

(fl
</>

0)

o
o

!5

k_
o

CO

o

Q.

I
Q.
O
Q.

in

3
C

X

0)
(/>

3
O

o •=

(A

3
O
CO

<n

3
k_

o
o.

I

<
O

o
o
o

.i

O

■D
O

ts

n

E
I

CO

09

c
p

s

■Q
•Q

398

Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

which unfortunately tended to be at its
hardest along suture lines and around fo-
ramina. Removal of all of the matrix has
not been possible, with the result that
some details are still obscure. The skull
of PU no. 21945, however, was in a softer
matrix, was less distorted, and has pro-
vided an almost complete understanding
of the skull structure. Moreover, the fa-
cial region (PU no. 20944) supplements
the skulls in various features. These
three specimens differ somewhat in size,
the facial region representing the largest
individual. These differences seem to be
due to age. The skull of PU no. 21726 is
that of an immature animal in which the
milk molars are still in place and M^ has
just come into wear. Series of living das-
yproctids reveal that skull growth contin-
ues until the full adult dentition is in use.

The skull as a whole is strongly arched,
agreeing in this with the type (Lavocat,
1976: 57), and is about as in Mijoprocta,
with large orbits and very large bullae.
The region in front of the cheek teeth
fonns approximately a quarter and the re-
gion posterior to the front margin of the
glenoid cavity approximately a third of
the total length (Figs. 10, 12, 14). The
skull roof is much flatter transversely
than in later dasyproctids, in this respect
resembling Cephalomys, and is even
slightly concave in the nasal and frontal
regions. In contrast to other members of
the family, there is a marked interorbital
narrowing of the dorsal surface of the
skull, and the braincase is flatter and rel-
atively wider than in later dasyproctids.
The sagittal crest is short and feebly ex-
pressed (Figs. 7-8).

The rostrum is moderately long and
rather slender, comparable to the rostra
of living dasyproctids in general propor-
tions, but less tapering and more nearly
parallel-sided, so that the dorsal root of
the zygoma stands out about at right an-
gles to the sides of the rostrum (Fig. 8),
a resemblance to Cephalomys. The ven-
tral profile of the snout is nearly straight,
as in living dasyproctids, and there is a

suggestion of the characteristic abrupt
downturn of the premaxilla immediately
behind the incisors (Fig. 10).

The infraorbital foramen is preserved
in toto only on the right side of PU no.
21945. The foramen was a large, vertical-
ly oriented oval, rather different from
those of the modern dasyproctids. Be-
cause both the dorsal and ventral roots of
the lateral wall of the foramen are con-
siderably elongate anteroposteriorly in
contrast to the slender central section of
the lateral wall, the infraorbital foramen
is nearly circular as seen from the side
(Fig. 12). The dorsal and ventral roots ex-
tend about equally far forward, so that the
dorsoventral axis of the foramen would
have been vertical, rather than inclined
forward as Lavocat believed (1976: 52).
There is a prominent elevation (below
the ridge that marks the course of the in-
cisor) that occupies part of the ventral
border of the masseteric fossa. A similar
elevation occurs in living dasyproctids,
but not in Neoreomys.

The nasals, not preserved in either of
Lavocat's partial skulls, are very much
longer than in Dasyprocta or Myoprocta.
They extend back to a point above dm^
in the juvenile or M^ in the adult, and
slightly behind the lacrimal foramen,
about as in Cephalomys and Neoreomys,
although their posterior ends are in front
of the lacrimal foramen in the Santacruz-
ian genus. Very long nasals are charac-
teristic of the Paramyidae and Reithro-
paramyidae (Wood, 1962a: Figs. 2A, 13A,
48A) and we accordingly interpret the
condition in Incamys (and also in Bran-
isamys, see below, p. 433) as primitive.
Platypittamys has somewhat shorter na-
sals, while those of living dasyproctids
are shorter still. The anterior end of each
nasal is tubular, the tubular part fonning
perhaps half the length of the bone. In
living dasyproctids, the combined nasals
have a tubular anterior end, but there is
no groove between the bones as there is
in Incamys (Figs. 7-9). The nasals of
Neoreomys are much less tubular than in

Oligocene Rodents of Bolivia • Patterson and Wood

399

CD

L.

£

m

in

X

A

^

O

£

g>

•2

'k.

<o

«

£

G

H^

o

Q.

T3

3

C

to

a>

^«

k-

£

o

qj

'k-

CO

£

CO

c

CD

(0

E

§

o

^

O)

>>

c

N

gi

♦-

£

'i-

O)

o

'k_

k_

o

••—

o

CO

r

en

CO

CO

Q.

O

^

g

'^

C/j

0)

2

c

. .

CO

w

o

1

i

QJ

0)

m

■>

>~

<u

n)

■w

k.

0)

♦-

CO

c

CO

<u

CO

>

E

CD

5

CO

•^

X

o

_-

c

3

O)

^

'k_

in

o

.«_

^_

O

o

S

CO

0)

(A

'>

</i

o

«

1

0)

o

2j

Q

2

<

■D

ih

o c

S

CO E

T

E 2

CM

1 °

6

c

^i

^

k'S

Q.

.!2 E

CO

o to

3

C 3

c

- CT

.<o

1- c/>

0) ^

>

Q. 3

*<^

Q. CO

o

3

*J

t: Ll.

§,

°C0

E

Q) C/)
CO

CO

CO — "

o

fj CO

c

1 -

J. Q-

. o

00

CO o

c o

0}

o ^

3

■:: O

O)

•2 to

i 1

i to

l^

<U 01

s8

•Q a

•< Q.

400 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

living dasyproctids. The posterior por- On the lateral side, the premaxillary
tions of the nasals in Incaniys are flat- part of the fossa for the origin of M. ma.s-
tened, even slightly concave transverse- seter medialis, pars anterior is not so
ly. The sides of the bones are nearly clearly defined as in Neoreotnys and liv-
parallel, but they are slightly pinched at ing dasyproctids, and does not have so
the middle. The nasals are generally sim- large an anterior extension (Fig. lOA), al-
ilar in outline to those of modern dasy- though it is relatively deep dorsoventral-
proctids. The nasofrontal suture bends ly. Lavocat (1976: 51) apparently con-
forward toward the midline, an irregular- fused the crest marking the anterior limit
ly pointed process of the frontals sepa- of the masseteric fossa (Fig. lOA; AM)
rating the posterior ends of the nasals with the premaxillary-maxillary suture on
(Figs. 7-9). the ventral lateral part of the snout. His
The ascending portions of the premax- description of the suture, "elle part de
illaries are long and somewhat expanded fagon rectiligne a 45 degres vers le haut
at their posterior ends. Their termina- et I'avant," agrees perfectly with the lim-
tions are curiously asymmetrical in two its of the masseteric fossa as we indicate
of the three specimens at hand (Figs. 7- them on Figs. lOA and 12, and is very
9). On both sides of PU no. 21726 they unlike the highly crenulate suture that
extend back farther than the nasals, we show on the same figures. A close ex-
whereas only the right one of PU no. amination of Lavocat's excellent stereo-
20944 does so. The left premaxillary of photo (1976: PI. 3, Fig. 1) shows that the
the fonner is of nomial shape, but the suture in this area is still covered by a
right in the course of development evi- thin layer of matrix.

dently expanded medially at the expense A striking feature of the Incamys skull

of the nasal, reaching farther toward the is the enomious size of the incisive fo-

midline and taking over the area occu- ramen, which occupies the greater part of

pied on the left side by a small, forwardly the palate anterior to the cheek teeth

extending wedge of frontal. The nasals (Fig. 14A,B). The anterior part of this ap-

and premaxillae reach back to the same erture seems to correspond to the nonnal

level in PU no. 21945, and the two pre- rodent incisive foramina, but the poste-

maxillae are symmetrical (Fig. 8A). The rior part is peculiar indeed. The anterior

fronto-maxillary suture is well in advance half of the septum between the foramina

of the rear of the nasal and premaxillary, is broad and rounded and reaches the

and is not so obliquely inclined as in liv- level of the palate in the usual manner,

ing dasyproctids and Cephalomys. In but the posterior half is a slender vertical

Neoreomys the entire anterior end of the septum that does not reach so far ventrad

frontal is an essentially transverse line. (Fig. 14B). The premaxillary-maxillary

On the ventral side of the bones, the in- suture reaches the sides of the foramina

terpremaxillary foramen is very large in near the level at which the change in the

PU no. 20944, much larger than in any septum occurs. The apparent difference

other dasyproctid we have seen, being between the shape of the suture in PU

about 1.6 mm in diameter (Fig. 14B); it no. 21726 (Fig. 14A) and that in the type

is relatively smaller, although still large, (Lavocat, 1976: PI. 3, Fig. 2) is merely

in the type (Lavocat, 1976: PI. 3, Fig. 2) because this area had not been cleared of

and in PU no. 20953, both of which are matrix in the latter. Near the rear of the

smaller and younger individuals, and this opening, the ventral margin of the sep-

area is missing in PU no. 21945. It is par- tum is about 3 mm dorsal to the palate

tially obscured by crushing in PU no. (Fig. 14B). None of the available speci-

21726. The size of the foramen is variable mens is complete, but by combining

in Recent dasyproctids. them all (Fig. 14A; Lavocat, 1976: PI. 3,

Oligocene Rodents of Bolivia • Patterson and Wood 401

(0

E

a>

k.

a

(0

as

a>

•^

^

X

CO

TO

■*

a>

t

ji

O)

o

1

o

CM

Q

1

d

J5

•1

c

X

M

Z>

CO

•A

Q.

E

,'\

0)

CO

^

';.\

3

a.

•'•'■'\

C

1

.■■..■••/I

.<o

1

X

M

S

2

o

Q.

>t ■'■•A

-Q

^V.* ''"'y

CO

"5

^*x^

^

CO

/

E

CO

/

CD

c

/

O

1

c

4:

,^

z

o

—

^

3

«

^

CO

X

15

r

CO

E

CO
Q.

o

CO

O

CO
CD

$

o

0)

O

■>

o.

«

O)

CO

c

^

O

T3

Q

C

o
CO

Cjj

CO

0)

1

3

X

Ol

s

L.

CO

§
1

-Q

402 BuUctiii Museum of Comparative Zoology, Vol. 149, No. 7

men
pital
eno-

2 " -5

kki

-O en

D "S ("

c t: 3

D) Q. C

c6

X

a.

™ O Q)

C ? 0)
Q) 1 O)

EO 3

o

«o y

o

h. 4-^

o

O — <n

v^

•V "> 3

■

•2 o

CO

J, nj (0

5 c

•*

if 1 s

X
0)

Ei°

t Z Q.

- ^ 1

■>

■D >- J.
OJ <0 <

"(5

E = ?

h_

X Q-

to

ifg^i

_l

sill

<

xi

S--io

re

E

p O ^ 0)
t *-; Q. CO
CO CO 1 (8

s;

°E>t-^

(0

0)
(0

— 1 "< (0 >-

a>

ca

"-^ S-?

. . • - (0

03

3 S Q> >

♦-

o

^ u -^ N

Q

1 -D "D •-

_J O O (0

—

CD -? ^O

0)

. E <u c

>

CO

2f t^

CsJ

O _j Q. -5"

t^

O »-

o ^ — o

C\J

~ (0 CO

d

(0 c *^ c-

5 1^.1

c

CO

3

(0 "- (0

C

O 1 o CO

■2

o

- j. -^ £

■Q

y-" o f

s.

a) to' >< 1,

(0 0) ■— >■

O

(8 C X3 K

C

F <i^ 2 •-
'=•<-<? CO

o

"5

C/)

-5 1 o

l202
IS - o

O 1= C Q.

d

.y -r CU g

a3 1 E (i>

^~

k-

3

•j: u. (0 Y

O)

1 - oH

Ll

2 -is -^ ■ -

< iS O.C

CO*

i/i o 1 E

c^ii 2

•2 1 o o

.|£ b;«

■Q CO C «

J3 > O CO

«^ O O Ql

Oligocene Rodents of Bolivia • Patterson and Wood 403

Fig. 2) it is clear that, in life, the septum
was continuous, as postulated by Lavocat
(1976: 51). The maximum width, across
both foramina and the septum, is 5.0 mm
or perhaps a little more. There is no trace
of the premaxillary-maxillary suture on
the septum between the incisive forami-
na in any of the specimens, although in
all of them a median suture is discern-
ible. We therefore suspect that all or most
of the septum may be composed of the
premaxillaries, as in other caviomorphs.
Just behind the premaxillary-maxillary
suture, on the lateral wall of the fenestra,
is what seems to be a large opening,
probably the anterior end of the nasolac-
rimal canal. The incisive foramina are
closed posteriorly slightly behind the
base of the upper incisors and opposite
the middle of P, at which point the sep-
tum descends abruptly to the ventral
margin of the opening.

We have not observed similar fenestrae
in any living dasyproctid, in all of which
the premaxillary-maxillary suture passes
across the palate at the rear of the incisive
foramina. However, in known fossil das-
yproctids there appears to be a backward
expansion of the foramina into the max-
illar\' (Cephalomys, Wood and Patterson,
1959: Figs. 21B, 22B; Neoreomys, Scott,
1905: PI. 65, Fig. 1; Scleromys, Scott,
1905: PI. 65, Fig. 13). In Neoreomys (PU
no. 15572), the suture crosses the middle
of the foramina, the posterior part of the
foramina being solidly roofed by the max-
illae. A posterior enlargement of the in-
cisive foramina is clearly visible in the
Santacruzian octodontids Sciamys (Scott,
1905: PI. 67, Figs. If, 5a, 6a, 7a) and
Acaremys (Scott, 1905: PL 67, Fig. 4a),
and echimyids (Scott, 1905: PI. 67, Fig.
10a), and also occurs in Recent mem-
bers of these families and in Abrocoma,
Chinchilla and Lagidiuryi. This may be
a primitive character for caviomorphs.
The octodontids are primitive cavio-
morphs, and it should be noted that the
enlargement is present in the North
American Uintan franimorph Protopty-

chiis (Wahlert, 1973). No trace of an en-
largement is present in the Deseadan di-
nomyid Branisamys (possibly indicating
an earh' reduction in the size of the open-
ing in this family); and the two skulls of
the primitive caviomorph Platypittamys
are so preserved as effectually to conceal
this area. This region is not preserved in
the partial skull of Sallatuys (Lavocat,
1976: PL 2, Figs. 1, 4). The foramina are
very short in the snout described by La-
vocat as Lurihayomys masticator (1976:
PL 5, Fig. 1).

An elongate incisive foramen, appar-
ently extending as far back as the level of
the anterior end of the cheek teeth, was
present in the early Oligocene African
Metaphiomys (Wood, 1968: Fig. 6A),
where its development was apparently
similar to that in Petromus (Tullberg,
1899: PL 6, Fig. 16). In these fonns, how-
ever, the incisive foramina are sunk into
a palatal depression (Wood, 1968: 51-52)
quite different from conditions in most
caviomorphs, although there are some
similarities to the well-developed gutter
that leads back from the foramina in Neo-
reomys (PU no. 15572). The foramen in
the African fomis is more nonnal in its
shape than in Incamys and the Santa-
cruzian octodonts. Such an enlarged in--
cisive foramen is not primitive for ro-
dents (e.g.. Wood, 1962a: Figs. 13C, 24C,
48B for paramyids and reithroparamyids;
Dawson, 1961: Pis. 2, 3 for Sciuravus).

The fossa for the origin of M. bucci-
nator extends forward, lateral to the in-
cisive foramina, as far as the premaxil-
lary-maxillary suture, approximately as in
Cephalomys, Neoreomys and Scleromys.
We disagree with Lavocat's statement
(1976: 52) that this fossa is quite different
from what is found in other caviomorphs.
In living dasyproctids the fossa extends
medially almost to the midventral line in
adult specimens, the fossae of the two
sides being separated only by a narrow
ridge. It would seem that the buccinator
is a more powerful muscle in Dasyprocta
and Myoprocta than it was in known ear-

404 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

lier genera. This suggests the possibility
that in the ancestry of these forms the
posterior part of the incisive foramina
may have become floored by medial ex-
tensions of the maxillae that enlarged
pari passu with the increase in the size
and strength of the buccinator muscle in
later members of the family.

In the living fomis, and even more in
Neoreomys, the maxillary part of the fos-
sa for M. masseter medialis, pars ante-
rior, dorsal to the incisor, is the deepest
part. This deepening is present to an ex-
aggerated degree in Incamijs, where the
area seems to be subdivided into a num-
ber of subsidiary pockets (Fig. 10). How-
ever, the absence of these from both
sides of the best preserved skull (Fig. 12)
suggests that they may have been formed
post-mortem either by crushing or by
breaking. The right zygoma is preserved
in PU no. 21945, revealing a fossa for the
origin of M. masseter superficialis, dis-
tinct from that of M. masseter lateralis
(Fig. 8B).

There is a large opening in the maxil-
lary on the dorsal surface of the ventral
root of the zygoma (Fig. 7). This, we be-
lieve, is too far laterad to be related to the
nasolacrimal canal, and we interpret it as
a nutritive foramen for the superior al-
veolar blood vessels (cf. Wood, 1974a:
35). Lavocat (1976: 54) mentioned "un
petit foramen qui est sans doute I'entree
du nerf alveolaire antero-superieur."
This foramen is faintly visible in his PI.
3, Fig. 1, where it can be seen to be as
large as in the Princeton material.

The orbital portion of the maxillary is
arched transversely over the roots of the
cheek teeth and does not meet the orbital
portion of the frontals above the molars.
In PU no. 21726 (Fig. lOA) there is an
oval vacuity here in which no trace of
bone was present, and in the older PU
nos. 20944 (Fig. 11) and 21945 (Figs. 12,
15A) there are corresponding but some-
what smaller openings. Such an opening
is likewise present in the type (Lavocat,
1976: PI. 3, Fig. 1). This is a puzzling fea-

ture that seems to have resulted either
from a lack of ossification in the ventral
part of the frontal or, less probably, from
universal post-mortem loss of delicate
bone. In living dasyproctids the ventral
portion of the frontal is very thin and
grows down over the maxillary; union of
the two bones is not entirely complete in
some of the younger individuals at hand,
and in these it is possible to pass a bristle
through the skull between the frontal and
the maxillary. Since the parts of the fron-
tals bordering these vacuities in the fos-
sils do not show signs of breakage, and
since the openings are present in all four
specimens (the type skull of /. pretiosus
would seem, from Lavocat's stereopho-
tos, not to have had this area cleared of
matrix), we conclude that the vacuities
are natural. We have temied the opening
the interorbital fenestra (Figs. 10, 12,
15A; INOF). Lavocat (1976: 52 and Fig.
3; I) identified this opening as the sphe-
nopalatine foramen, an interpretation
that seems highly improbable to us, al-
though the foramen possibly is included
in the fenestra. What we have identified
as the foramen (Fig. 10; SPF) lies farther
to the rear. Openings of this nature with-
in the orbit are highly unusual among ro-
dents, but they do occur (e.g., Myosciu-
rus, Jaculus; Ellennan 1940: Figs. 91,
158).

The palatal portions of the maxillary
send a sharply pointed median process
forward into the posterior part of the
enonnous incisive foramina. This and the
equally pointed anterior end of the pal-
atines are situated dorsal to the level of
the adjacent cheek teeth, and form the
roof of a prominent groove that extends
from the incisive foramina to the middle
of M^ (Fig. 14A), and ends at the posterior
palatine foramina.

The frontals are suturally separated for
most of their length as seems generally
to be true of caviomorphs, franimorphs,
and paramyids (Wood, 1962a: Figs. 13A,
31 A, 35A, 41 A, 48A, 57A). The same is
true of Old World hystricognaths. In PU

Oligocene Rodents of Bolivia • Patterson and Wood 405

o

a.
o

1
O

x
a

E

c

0)

E

CO

. to

q) CO
« .i

t

a>
o

CM

d

c

3
Q.

<o

C

■2
o

-Q
CO

o

c

3
CO

0)

3

O
(0

(0 TO (0

2 !8

k_ CD

O <D

kl >-

Q) CO

c -o

IS

OQ

5^
^- o

CO c

Q.
(0

c
o
_ ^
CO ■y

CO
Q.

CD

cc a-

"J)

CO Q.

<" ii;^

O C
■^ jO

o —

C X

0) CO

c/) a>

co Q.

E I

1-5

c

.2 ^

i- Q.

?5

CO Q.
C C/3
OO

■5 c
$ E

O CO

406 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 7

no. 21945 (Fig. 8A), the suture extends
the entire lengtli of the hones, as in My-
oprocta and Dasypwcta. The dorsal sur-
face is extreme!}' flat, with sHght longi-
tudinal depressions on either side ol the
midline (Fig. 8A), as in NeorcoDujs. As
noted above, the frontals narrow at the
middle of the orbits. Within the orbit, the
frontal is well preserved onlv in PU no.
21945 (Figs. 12, 15A). It was nearly as
concave as in living dasyproctids. The
orbital process of the frontal is essentially
rectangular (Fig. 15A). The vertical lac-
rimo-frontal suture is hidden in Figure
12 by the lateral wall of the infraorbital
foramen. At the posterior end, the frontal-
squamosal suture is highly crenulate, but
is also nearly vertical. Ventrally, the pos-
terior three-fifths of the frontal meets the
orbitosphenoid in a line that is generally
straight but irregular in detail (Fig. 15A).
The anterior part of the ventral border of
the frontal is formed by the interorbital
fenestra.

There is a small but pointed postorbital
process of the frontal, somewhat behind
the middle of the orbit (Figs. 8, 12), that
extends out over the fossa for M. tempo-
ralis, as in Dasypwcta and Myoprocta.
Behind this, and in the squamosal im-
mediately behind the frontal-squamosal
suture, is a smaller, rounded postorbital
process. Between these two processes, an
anterior slip of M. temporalis apparently
extended a short distance onto the dor-
sum of the skull, arising entirely from the
frontal (Fig. 8). There is no trace of the
posterior postorbital process, or of any
division of M. temporalis, in the skulls of
Neoreomys, Dasyprocta, or Myoprocta.
A similar condition, however, may be
seen in the skull of Myocastor (Ellerman,
1940: Fig. 20). M. temporalis in cavio-
morphs is divisible into three portions
which may be called the main, orbital
and posterior parts (Woods, 1972: 129-
130). The main part arises from the tem-
poral fossa on the dorsum of the skull; the
orbital from the frontal and squamosal
within the orbit; and the posterior from

the squamosal root of the zygoma {op.
cit.). It would seem, therefore, that the
orbital portion of M. temporalis must
have been unusually large and distinct in
Incamys, as would also seem to be the
case in Echimys (Tullberg, 1899: PI. 8,
Figs. 7-8) and in Proechimys (Woods,
1972: Fig. 2A,B). It is possible that this
is a primitive condition for caviomorphs.
In the franimorphs, no skulls are pre-
served so as to show whether or not the
orbital part of M. temporalis was visibly
distinct on the dorsal surface. However,
in Paramys (Wood, 1962a: 17; Figs.
2A,B; 13A',B), Thishcmys {op. cit.: Fig.
36A), and manitshines {Pseiidotomus, op.
cit.: 174; Ischyrotomus, op. cit.: 190;
Figs. 68A,C; 71B,C; Manitsha, op. cit.:
222), the conditions are basically similar
to those in Incamys except that, in this
last genus, the fossa for the orbital part of
M. temporalis is in the frontal; in the par-
amyids, it is in the alisphenoid or pari-
etal. Among the thryonomyoids, there is
possibly a rather similar situation in Par-
aphiomys stromeri (Lavocat, 1973: PI. 2,
Figs. 3-4; the specimen was not prepared
enough for us to feel at all certain), and
Diamantomys luederitzi (Lavocat, 1973:
PL 3, Figs. 1-2), but rather clearly not in
Paraphiomys pigotti {op. cit.: PI. 1, Figs.
3-4) and apparently not in Thryonomys
or Petromiis.

The right lacrimal is parti v preserved
in PU nos. 20944 (Fig. 11)' and 21945
(Fig. 12). There was a small expanse of
the bone on the dorsum of the skull (Fig.
8) and it forms the posterior part of the
dorsal root of the zygoma as in living das-
yproctids and in Neoreomys. It is uncer-
tain, due to breakage, just how far down
the zygoma the bone extended. It may
have reached the malar, as it does in Neo-
reomys. The opening of the nasolacrimal
canal is near the posteroventral border of
the bone. The canal curved forward and
slightly ventrad. In Neoreomys, the
opening of the canal is farther forward,
on the lateral buttress of the lacrimal. In
the living Dasyproctidae, the canal leads

Oligocene Rodents of Bolivla • Patterson and Wood 407

DC «

in

■ ■ OJ

S <»
o »-

Q) O)

■^ Q.

408 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

forward and downward from the lacrimal
foramen to the nasal passage, in a normal
manner. However, half of its length is
exposed by a large, oval opening in the
medial wall of the masseteric fossa.
There is a large opening in this position
in Neoreomys (Scott, 1905: PI. 64, Fig. 6),
but not, apparently, in Cephalomijs
(Wood and Patterson, 1959: Fig. 21C). As
preserved in Incamijs, the course of the
canal is a deep groove extending from the
lacrimal foramen nearly as far as the rear
of the upper incisor. Presumably, as also
indicated by Lavocat (1976: 52), there
was a thin lateral wall covering most of
this groove, although there may have
been an opening in it just behind the in-
cisor. However, the preservation is such
that we could not be certain whether or
not the lateral wall was originally present
in Incamijs, or how large the supposed
lateral opening might have been. The an-
terior limits of the lacrimal could not be
detemiined because all specimens were
broken in this area, but we find it impos-
sible to visualize any construction of the
lacrimal that would not have included a
laterally directed buttress near its ante-
rior end, supporting the dorsal root of the
zygoma, a structure that is very charac-
teristic of fossil and Recent dasyproctids,
and that is also present in caviids and
Hydrochoerus.

Small portions of both ends of the ma-
lar and the imprint of the medial surface
of most of the remainder are preserved in
PU no. 21945 (Fig. 12; IM). The ventral
surface of the arch is unifomily rounded,
as in Myoprocta and Dasyprocta. There
is no suggestion of the pronounced ven-
tral angulation seen in Paraphiomys
(Lavocat, 1973: PI. 1, Fig. 4). The anterior
end of the bone is very similar to that
of Myoprocta — much more so than like
Dasyprocta. The malar fomis a thin plate
that lies on the lateral surface of the max-
illa, which is exposed on the lateral side
ventral to this thin plate. The imprint of
the medial face of the zygoma shows the
malar-maxillary contact rising postero-

dorsally across the arch. The preserved
anterior portion of the malar is restricted
to the horizontal bar of the zygoma, and
does not extend far up the plate lateral to
the infraorbital foramen, in this resem-
bling both Myoprocta and Dasyprocta.
How high it extended on the arch and
whether or not it reached the lacrimal are
undetemiinable. The anterior malar-max-
illary suture is slightly irregular, but not
crenulate, and forms almost a straight
line. The slender posterior tip of the ma-
lar extends a short distance behind the
posterior end of the zygomatic process
of the squamosal, as in Myoprocta and
Dasyprocta. It is, however, more ex-
posed in lateral view (Fig. 12) than in
those genera.

The central part of the malar is missing.
The imprint in the matrix preserves al-
most the entire ventral limit of the arch,
and seems to preserve at least half of the
dorsal limit. It is possible that part of the
dorsal imprint has been broken away, al-
though there is no evidence that it has.
If the imprint of the central part of the
malar is as complete as it seems to be, the
bone was of essentially unifonn dorso-
ventral diameter for its entire length, re-
sembling Myoprocta and Dasyprocta in
this, but perhaps being a little less en-
larged anteriorly. There is absolutely no
similarity in any respect to the malar of
Paraphiomys (Lavocat, 1973: PI. 1, Fig.
4; PI. 2, Fig. 4) or Diamantomys {op. cit.:
PI. 3, Fig. 2), both of which fonn heavy
vertical plates.

The fused palatines extend forward to
the middle of P or dm^, as in Dasyproc-
ta. Lavocat (1976: 52) stated that the max-
illary-palatine suture reaches the level of
the posterior end of P^. A careful exami-
nation of his stereophoto (op. cit.: PI. 3,
Fig. 1) suggests that, if this area had been
completely prepared, the suture would
probably have the position shown in our
Fig. 14A. The maxillary-palatine sutures
diverge posteriorly, at first abruptly, then
very gently, and then abruptly again be-
hind M^. The posterior palatine foramina

Oligocene Rodents of Bolivia • Patterson and Wood 409

are completely enclosed by the palatines, cal contact, the suture turns backward
On the midline, the palate ends opposite nearly horizontally, the anterior portion
the middle of M^ (Fig. 14A) as in Ceph- of the orbitosphenoid lying dorsal to the
alomys. Lavocat (1976: 54) suggests that maxillary. Below the posterior portion of
there may have been a rudimentary me- the bone is the orbital opening of the
dian process at the midline of the poste- choano-orbital fenestra (Fig. 15A), that
rior margin of the palate of the type opens both into the pharyngeal passage
(1976: PI. 3, Fig. 2); one clearly was ab- between the pterygoid and the presphe-
sent in PU no. 21945 and in the slightly noid, and also into the pterygoid fossa,
younger PU no. 21726 (Fig. 14A). The The confluent optic foramina are entirely
palatine forms the lateral walls of the within the orbitosphenoid. The bone
pharyngeal passage behind the rear of fonns the median wall of the foramen la-
the hard palate, reaching almost to the cerum anterius (Fig. 15A).
front of the pterygoid fossa (Fig. 15B), The pterygoid fomis the anterior end
about as in Reithroparamys (Wood, of the pterygoid fossa and its median wall
1962a: Fig. 41B), the only franimorph in (Fig. 15B). The internal pterygoid pro-
which any of this area is known. The pal- cess, preserved in toto only on the right
atine-pterygoid suture is highly crenu- side of PU no. 21945, is a long, slender
late. In both of these respects, Incamijs process, reaching back to touch the an-
is very similar to Dasyprocta and My- teromesial tip of the bulla (Fig. 15B), as
oprocta. in the living dasyproctids. It is bounded

In PU no. 21945, the palatine-ptery- dorsally by the usual vacuity between it

goid suture very clearly lies behind the and the bulla.

middle of the "orifice choano-orbitaire" The alisphenoid, as in other post-

(choano-orbital fenestra), as stated by Eocene hystricognaths, is reduced in

Lavocat, and as shown in Fig. 15B, but height, extending, in the temporal fossa,

there is what might be another suture little if any above the level of the glenoid

along the lateral wall of the pharyngeal cavity (Figs. 12, 15A). It is separated from

passage anterior to this (Fig. 15B; ?). It is the frontal by a narrow strip of the orbi-

very doubtful that this really is a suture; tosphenoid, a significant difference from

irregularities in this area that are vaguely the African Paraphiomys (Lavocat, 1973:

suture-like can be seen in some individ- 31). Anteriorly, the alisphenoid forms the

uals of Dasyprocta. The orbital process lateral wall of the foramen lacerum an-

of the palatine extends along the front of terius. The dorsal border of the bone runs

the choano-orbital fenestra up to the or- back in essentially a straight line, below

bitosphenoid (Fig. 15A). We do not know the zygoma, to the front of the bulla. Ven-

whether this fenestra has any great sig- trally, the alisphenoid fonns the lateral

nificance or not. A similar opening is wall of the pterygoid fossa, reaching the

present in many sciurids, but absent in palatine just in front of the point where

others, varying from present to absent the pterygoid fossa begins (Fig. 15A).

within the genus Sciurus. There is a broad shelf of the alisphenoid.

The orbitosphenoid is a large bone. Its behind the pterygoid fossa, separating

dorsal margin is somewhat higher than the fossa from the large foramen lacerum

the dorsal limit of the alisphenoid, and it medium (Fig. 15B). This seems to be ho-

separates that bone from the frontal (Fig. mologous to the shelf in Faraphiomys

15A). The suture between the orbito- (Lavocat, 1973: PI. 2, Fig. 2) but differs

sphenoid and the maxillary is irregular, in that, in the African form, it is at the

especially in its anterior portion, just be- level of the dorsal portion of the ptery-

neath the rear of the interorbital fenestra goid fossa, whereas, in Incamys (Fig.

(Fig. 15A). After a short, generally verti- 15B), it is ventral to the fossa. The fossa

410 BuUciin Museum of Compdnitivc Zoo/o^'//, Vol. 149, No. 7

and the foramen lacerum medium join
dorsal to this shelf.

The pterygoid tbssa is short, wide, very-
deep, and perforate (Fig. 15B). It is much
wider than in Neoreomys or the living
dasyproctids. We have been imable to
deteniiine what, if any, dorsal border
there may have been to the pterygoid fos-
sa, the alisphenoid wall being very thin,
so that it is easy to cause damage inad-
vertently. Nor have we been able to de-
tect any indication, within the dorsal fe-
nestra, of a partition between the
alisphenoid canal and the passage for M.
ptenjgoideus internus. However, in view
of the damage sustained by the specimen
prior to burial and of the delicate nature
of the bony separation in various hystri-
cognaths, little significance can be at-
tached to this. We believe that the open-
ings in this region described here were
all that were present in life; however, the
thinness of the bone and the pre-fossil-
ization damage that had occurred to the
presphenoid, basisphenoid and basioc-
cipital of PU no. 21945 (Fig. 15B) shows
how much serious disintegration can oc-
cur, even in heavy bones on an otherwise
well preserved skull. Lavocat's material
of Incamijs apparently did not preserve
this area.

Along the medial wall of the fossa,
there is a large opening, separating the
internal pterygoid process from the pre-
sphenoid (Fig. 15B). This is the palatal
opening of the choano-orbital fenestra
discussed above. The function of this
opening is not known to us. Comparable
openings occur in some other h>strico-
gnaths, among them Neoreomys and the
living dasyproctids. A thin bar of the pre-
sphenoid separates this fenestra from
another one posteromedial to it, that lies
mainly within the basisphenoid, al-
though extending for a short distance into
the presphenoid. We apply the noncom-
mittal term sphenoidal fenestra to it, and
confess total ignorance as regards its
function (Fig. 15B; SF). We do not know

of any other rodents in which this fenes-
tra occurs.

The parietals have undergone crushing
and buckling in most specimens, but are
perfectly preserved in PU no. 21945 (Fig.
8A). They are, as shown by the dorsal
contour of the occiput (Fig. lOB), much
less arched transversely than those of
Neoreomys and the living dasyproctids.
In PU no. 21945 there seems to be a short
section of suture still visible between the
parietals (Fig. 8A); in other specimens,
there is no trace of such a suture. The
parietals resemble those of the Recent
dasyproctids in general outline (Fig. 7),
but the anterolateral comers extend for-
ward (Fig. 8A), more as in Neoreomys.
Faint, lyrate crests converging at the sag-
ittal crest mark the medial limits of the
weakly-developed temporal fossae. The
sagittal crest is small and weak, as in
Myoprocta. In contrast to Platypittamys,
but in agreement with the great majority
of caviomorphs (p. 488), there is no trace
of a separate interparietal. Near the lat-
eral margin of the parietal, directly above
the porus acusticus extemus, a fair-sized
vascular foramen leads forward into the
bone (Fig. 8A). The parietal extends, for
the middle half of the width of the occi-
put, back almost to the top of the lamb-
doid crests. Laterally, however, the pos-
terior margin of the parietal is bowed
forward (perhaps indicating the lateral
limits of the otherwise unidentifiable in-
terparietal) by the lateral process of the
occipital.

The squamosal is fundamentally simi-
lar to that of other members of the family.
Posteriorly it is excluded from the occip-
ital crest by the exoccipital. It is tightly
appressed to the dorsal portion of the
tympanic, but, in contrast to other dasy-
proctids, does not extend posteriorly be-
yond the porus. A small subsquamosal fo-
ramen is situated directly above the
porus acusticus externus (Fig. 12); on the
left side of PU no. 21945 this foramen is
partially divided into anterior and pos-

Oligocexe Rodents of Bolivia • Patterson and Wood 411

terior portions. Anterodorsally, the squa-
mosal extends to, and largely forms, the
posterior postorbital process (Figs. 8, 12).
The frontosquamosal suture bends back-
ward, below the posterior postorbital pro-
cess, and then runs almost vertically
down behind the frontal to the level of
the glenoid cavity. Below here, the fron-
tosquamosal suture becomes highly cren-
ulate, in contrast to the essentially
smooth suture in modern dasyproctids
and the slightly crenulate one in Neoreo-
mys. Finally, this portion of the squa-
mosal makes contact, ventrally, with the
alisphenoid, slightly below the level of
the glenoid (Fig. 15A). From here, the
alisphenoid suture runs almost directly
backward, below and medial to the gle-
noid cavity, until it reaches the front of
the bulla. It is uncertain how far forward
the squamosal extended on the zygoma;
certainly no farther than the middle of
the arch (Fig. 8).

The bulla is not preserved in the type;
it is apparently partially preserved but
still buried in the matrix in the type of I.
pretiosus (Lavocat, 1976: 58), although
we have been unable to recognize any
trace of it in Lavocat's stereophotos
(1976: PI. 5, Figs. 6-7; PI. 6, Figs. 4-5).
In the skulls of PU nos. 21726 and 21945
(Figs. lOA, 12, 14), it is finnly attached
and ver>' large, about one quarter the
length of the skull. It is roughly compa-
rable in size to those of Platypittamijs,
Chinchilla, Schistomys, Microcavia and
other small caviomorphs, and is relative-
ly larger in all dimensions than those of
other dasyproctids. Size apart, however,
there is a similarity in shape and other
features to the dasyproctid bulla. Widest
and deepest anteriorly, the bulla of hi-
camys extends from a point in advance of
the basioccipital-basisphenoid suture to
the back of the condyle. The ventral sur-
face is strongly arched, both anteropos-
teriorly and transversely. The anterior
face is very gently convex. The anterior
part of the medial surface is parallel to

the long axis of the skull; the posterior
part is inclined posterolaterally around
the condyle (Fig. 15B). The basisphenoid
and basioccipital press tightly against the
anterior half of the bulla. In the posterior
portion, where the bullae diverge, there
is a narrow trench between the bulla and
the basioccipital. The posterior portion of
the trench is occupied by the foramen
lacerum posterior (Fig. 15B). Partially
separated from the anterior end of this is
what seems to be another opening, lead-
ing forward. If this actually is a foramen,
it may be a vestigial carotid canal, or it
may have transmitted a vein. We incline
to the latter alternative since some indi-
viduals of Dasyprocta have a similarly
situated small opening, and Bugge (1974a:
54) found no trace of an internal carotid
in that form. No other specimen of De-
seadan caviomorph is sufficiently well
preserved (and prepared) to demonstrate
the presence or absence of this opening
(see pp. 484-486 for a discussion of the
carotid circulation in hystricognaths).

The porus acusticus externus is large
and extends posterodorsally, imparting
an oval outline to the opening (Fig. 13).
Posteriorly, its rim fonns a narrow ridge
bounded behind by a deep and narrow
groove. The tympanohyal articulation
and the adjacent stylomastoid foramen
are situated at the ventral extremity of the
groove. The anteroventral lip of the porus
acusticus is prominent and elevated on
PU no. 21945, as in Neoreomys. There is
a notch at the middle of the ventral wall
of the porus, from which a deep, initially
narrow, curving slit extends downward,
then enlarging (Fig. 13), as in Neoreomys
and young individuals of Dasyprocta,
giving a figure eight shape to the open-
ing. The anterior lip of the porus is some-
what thickened, but not so much so as in
Neoreomys, and is perforated by a mi-
nute foramen. The porus is completely
surrounded by the tympanic, as in ro-
dents generally. Following Van Kampen
(1905: Fig. 59), there has been some ten-

412 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Figure 13. Incamys bolivianus, PU no. 21945, parts of right ear. A. Lateral view of bulla from somewhat behind and
below, x5. B. Lateral view of porus acusticus externus, xlO.

Abbreviations: B — bulla; M — mastoid process; OC — occipital; POP — paroccipital process; SO — squamosal; TYHL — area
of articulation of tympanohyal.

dency (e.g., Grasse and Dekeyser, 1955:
1334) to regard the portion of the bulla
above the poms as being composed of the
periotic. Evidence of its tympanic nature
in rodents is, we believe, provided by
certain paramyids, especially Ischijroto-
mus oweni (Wood, 1962a: 207), in which
the bulla, found detached from the peri-
otic, completely surrounds the opening
of the porus. That part of the bulla dorsal
to the porus is not so large as in other
dasyproctids, especially posteriorly. A
prominent groove runs backward from
the anterior border nearly to the center
of the anterior margin of the porus (Fig.
10); comparable, although shallower and
more dorsally placed, grooves are present
in Neoreomys and modem dasyproctids.

Posteriorly, the bulla narrows, as in
other dasyproctids, and just reaches the
short paroccipital process. Breakage in
the right bulla of PU no. 21726 (Fig. 14A)
reveals the lower part of the crista tym-
panica. The groove lateral to it is larger
and deeper than in Dasijprocta, and is
partially divided by at least three cross
struts, features not present in the living
fomis. Within the matrix-filled tympanic
cavitv' of PU no. 21726 are small pieces
of bone; unfortunately, the extreme hard-
ness of the rock has prevented further
preparation, and it is impossible to de-
tennine whether these are exposed por-
tions of the auditory ossicles or fragments
of the bullar floor.

The mastoid process is large and in-

Oligocene Rodents of Bolivta • Patterson and Wood 413

(8 O

r <=

Q.Q.

O c

CM o

• ^.

O "^

== E

Z) o

Q- ^

cd -7

2 i

c 2

> =. (0
C '-' w

.2 c (0

o O J

«*^

<0 I- — i

o o — I
^ ;d o

°C: O

/f\ ^ ._

^ X (1)

^•'^ 2

3 = CO

(0
(0

o

E

(0
3
CT
(O

CO

if

0)

E

^ o

CO

C c

o

3 r;

•D

•i o

o

id (n

^S8

c

0)

U)

f^ 2

CO

^

• o

a.

o .

3

C CD

to

LL

. CM

U

< .

CO

. o

X C

<o"

'^^

3
C

(0 Q_

L.

E

0)

X

Q>

CO -

CO

</) O

3

CO i:

O

0) »-

« E

0)

3

■D 2

O
CO

CO

o Zi

3

o :i

o

CD °

a

T- C

1

X o

-k

LL O

o

1 a>

CL

o <"

. . <«

•o

^o

'o

t5

. o

CO

m -
1 •*

E

^ CD
(0 O

1
2

!l cm

o .

CO o

E

o c

3

Ez.

73
0)

■dQ-

E

c

CO O

E

* c

3

» O
> ^

O
(0

c

o

c

CI
(O

CO

c
o

a>

»■«

-C3
-Q

414 Bulletin Muscitin oj Comparative Zoology, Vol. 149, No. 7

flated, participating to a greater extent in dyles, narrows rapidly forward for half its
the lateral surface of the skull than in oth- length, and then maintains an even width
er dasyproctids, and is visible in poste- for the remainder of its course. The ven-
rior view (Fig. lOB). Due to the inflation tral surface of the basioccipital in PU no.
it merges smoothly with the posteroex- 21945 bears a strong median keel; its
ternal surface of the bulla. A thin strip of height is largely due to the crushing of
the exoccipital, narrower than in other the adjoining bone (Fig. 15B). The trench
members of the family, separates it from between the lateral margin of the basioc-
the squamosal. Breakage on the left side cipital and the bulla has been described
of PU no. 21945 reveals that the inflated above. The foramen lacerum medium
sinus of the mastoid was crossed by a lies in front of the bulla at its junction
number of septa. What may be the suture with the basioccipital and basisphenoid.
between the tympanic and the mastoid is The occiput is essentially vertical, as in
shown in Fig. 13A, but we are not certain Neoreomys. The supraoccipital is less
either that it is the suture or that we have arched than in that fonn or in the Recent
located it correctly in its entirety. dasyproctids, and extends over onto the
The ventral part of the skull between dorsal surface only for a very short dis-
the palate and the bullae is not preserved tance at the center (Figs. 7, 8A), in con-
in the type (Lavocat, 1976: PL 3, Fig. 2), trast to other dasyproctids but in agree-
and in Lavocat's type of Z. pre^iosu5 only ment with Platypittamys; on the sides,
the endocranial surface has been pre- however, it plays a larger part in the for-
pared, which, as Lavocat stated (1976: mation of the roof, agreeing in this with
56), makes comparisons difficult. This other dasyproctids and differing from
area is badly broken in PU no. 21726, but Platypittamys. The lateral portion, sep-
part of it is very well preserved in PU no. arating the mastoid from the squamosal
21945 (Fig. 15B). Even in this specimen, (Figs. 7, 8, 12; p. 487) starts out, at its
however, the ventral surfaces of the pre- mesial end, as a slender process between
sphenoid, basisphenoid and basioccipital the central part of the parietal and the
were damaged, before fossilization. dorsal tip of the mastoid. It then widens
The presphenoid is a long, slender considerably, separating the mastoid
bone, its posterior end being just behind from the squamosal and lateral part of the
the anterior limit of the pterygoid fossa, parietal (Fig. 8A). At about mid-height of
Its posterior portion fomis the median the squamosal, however, the process
wall of the pterygoid fenestra. There is a abruptly narrows (Figs. 10, 12, ISA) and
lateral brace, near the posterior end of extends as a very slender sliver of bone
the bone, running posterolaterally to the to just below the ventral limit of the squa-
alisphenoid and pterygoid, and forming mosal, at the posterodorsal limit of the
the anterior border of the sphenoidal fe- tympanic.

nestra. The presphenoid is more slender. The suture between the exoccipital
relatively, than in Neoreomys, and more and the mastoid follows a ridge that runs
comparable to that of Myoprocta. laterally and a little dorsally from the up-
Enough of the basisphenoid is pre- per part of the foramen magnum, as it
served to reveal that there was an angle does in other members of the family,
between it and the basioccipital on the Dorsal and ventral to this ridge are
ventral surface, comparable to that in oth- depressions for attachment of neck mus-
er dasyproctids. Otherwise, it is badly cles (the depth of the dorsal one is ex-
damaged in all available specimens. Its aggerated in PU no. 21726, Fig. lOB, due
anterior extremity surrounds most of the to breakage). There is an indication of a
sphenoidal fenestra. vertical crest running from the top of the
The basioccipital is widest at the con- foramen magnum to the occipital crest.

Oligocene Rodents of Bolivia • Patterson and Wood 415

Figure 1 5. Incamys bolivianus, x 5. A-B. PL) no. 21 945, anterior end to the left. A. Oblique view of left orbit and temporal
fossa. B. Ventral surface of basicranium. C-D. Posterior views of right upper molars, showing change in width of
occlusal surface with wear. C. PL) no. 20950, slightly worn M' or M-. D. PL) no. 20955, highly worn M-.

Abbreviations: AL— alisphenoid; B — bulla; BO — basioccipital; BS — basisphenoid; CC — possible carotid canal; F— vacuity
between internal pterygoid process and bulla; FLM — foramen lacerum medium; FLP — foramen lacerum posterior; FR —
frontal; INOF — interorbital fenestra; IP — internal pterygoid process; LA — lacrimal; LAP — opening in wall of lacrimal ca-
nal; M — mastoid; MX — maxillary; 00 — occipital; OF — optic foramen; OS — orbitosphenoid; P, PAL — palatine; PA — pari-
etal; PF — choano-orbital orifice; PS — presphenoid; PT — pterygoid; PTF — pterygoid fossa; PTF? — possible extension of
pterygoid fossa; SF — sphenoidal fenestra; SO — squamosal; ? — probably a pseudosuture.

416 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

The paroccipital process is short, squat,
and attached to the posterior surface of
the mastoid and bulla, somewhat as in
the octodontid Sciamys (Scott, 1905: PI.
67, Fig. 9); in other dasyproctids it is
longer and freely projecting.

The condyles are narrow, face ventral-
ly as much as posteriorly — as in Recent
dasyproctids — and lie closely alongside
the bullae for much of their lengths. In-
flation of the auditory region has brought
much more of the lateral surface of the
cranium into view from the occipital as-
pect than is the case in other members of
the family. The foramen magnum is es-
sentially circular and very large.

The mandibular fragments at our dis-
posal (Figs. 12, 16) show that the hori-
zontal ramus was approximately as deep
and wide as that of Cephalomys. The
posterior part of the dorsal border of the
diastemal surface sloped gently down-
ward and forward at almost the same an-
gle as in that fonn and in Neoreomys, not
so abruptly nor so deeply as in the living
dasyproctids. The masseteric crest be-
gins beneath the center of P4 and agrees
in size with those of other members of
the family (Fig. 16A,B). The low coronoid
(Fig. 16B) passes the alveolar level a lit-
tle behind M3, as in Cephalomys, and in
contrast to the more forward position in
later forms. The coronoid is clearly too
large in Lavocat's reconstruction (1976:
Fig. 31). The condyle is an elongate oval,
presenting dorsally, compressed lateral-
ly, and nearly on a level with the cheek
teeth (Fig. 12); the postcondyloid process
is moderately developed. The mental fo-
ramen is high and a little in front of P4.
It opens upward (Fig. 16B). A peculiar
feature is the remarkable number of mi-
nute vascular foramina along the course
of the incisors (Fig. 16D); none of Lavo-
cat's figures show this area. Similar fo-
ramina, although not in such rich profu-
sion, also occur in the living dasyproctids.
As in the living forms, the posteroventral
portion of the symphysis extends back
beneath P4. The genioglossal pit is very
weakly developed.

Cement is lacking on all cheek teeth.
The teeth are larger relative to skull size
than in the living dasyproctids, being
closer to Neoreomys in this respect. For
terminology of the parts of the cheek
teeth, see Figs. 1-2 (pp. 376-377). The up-
per teeth are inclined posteriorly, the low-
ers anteriorly. In both uppers and lowers
the major lophs and lophids unite with
wear into an S-shaped pattern, the S run-
ning from the lingual end of the anteroloph
to the labial end of the posteroloph in the
uppers and from the metaconid to the hy-
poconid in the lowers. This is especially
evident after an intennediate amount of
wear (Figs. 18A, 19G). The enamel over
unworn crowns is thinner than that on
the sides, as is shown by its thinness
around the edges of lophs and lophids of
slightly worn teeth.

The cheek teeth display unilateral hyp-
sodonty, especially notable in the uppers
which are hypsodont lingually and
brachyodont labially (Fig. 17E). The me-
sodont lower teeth show much less dis-
parity in height of crown, although the
enamel descends lower on the labial and
posterior faces (especially on the hypo-
conid) than on the lingual and anterior
ones. This pattern for the initial stages in
the development of hypsodonty is ex-
ceedingly common among the Rodentia,
and cannot be taken to indicate any spe-
cial relationships of Incamys with any
other genus or group of genera.

The unilateral hypsodonty of the upper
teeth results from the fact that there is no
post-eruptional growth of the labial mar-
gin of the crown, whereas there is long
continued growth of the lingual half of
the tooth. As a result, eruption causes a
tooth to rotate, in a transverse plane,
about the buccal roots. The center of ro-
tation is approximately at the point where
the roots clear the alveoli. From this it
ensues that there is extensive wear on the
lingual half of the crown, and little wear
on the extreme labial parts, which grad-
ually extend farther and farther labially
as rotation progresses. This is demon-
strated by the posterior views of two up-

Oligocene Rodents of Bolivia • Patterson and Wood 417

Figure 16. Lower jaws of Incamys bolivianus, x4. Dotted areas are matrix. A. Lateral view of PU no. 20933, completed
from PU nos. 20921, 20925, and 20941. B. Lateral view, PU no. 21726. C. Mesial view, PU no. 21726. D. Mesial
view, PU no. 20933.

418 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Figure 17. Left upper cheek teeth of Incamys bolivianus, anterior ends to the left, x5. Dotted areas are matrix.
A. PU no. 21726, dm3-^ M' ^. B. PU no. 21728, P^-M^ C. PU no. 20960, P^-M^ D-E. PU no. 20965, P^ D. Occlusal
view. E. Anterior view. F. PU no. 20963, P". G. PU no. 20968, M' '" -. H. PU no. 20981, dm^ I. PU no. 20980, dm^ and
root of dm^.

Abbreviations: IF — incisive foramen; R — roots of dm*; S — maxillary-palatine suture.

per molars. In the less worn tooth (Fig.
15C), the wear surface ends essentially
ventrad of the labial margin of the buccal
roots; in the worn tooth (Fig. 15D), the
wear surface extends 1 mm or more be-
yond the labial margin of the roots. The
situation just described is particularly
characteristic of M^"^. In both P and M^,
the absence of adjacent teeth on one side
has pemiitted the forward (or backward)
growth of the hypsodont lingual portion,
and the rotation of the tooth as eruption
proceeds is not in a simple transverse
plane. Therefore, these teeth do not in-
crease so markedly in width with wear as
do M*~^, although an increase does occur
(Table 3). The anteroposterior diameters
of P-M^ remain approximately constant
as wear proceeds. That of M^ increases in

length with wear, but we have not com-
puted a and V for this tooth, as we did
not have an adequate sample. Again, be-
cause of the size of the sample, we are
not sure whether this tooth is or is not
significantly smaller than the others (Ta-
ble 3). P, however, does not differ great-
ly in size from the molars.

Lavocat (1976: 42) indicated that P is
smaller than M'~^. This opinion is at
least partly due to the fact that, of the
three specimens he figured, two (the
type, SAL 117, Lavocat, 1976: PI. 4, Fig.
1 and SAL 118, Lavocat, 1976: PI. 4, Fig.
2) have P in process of eruption and the
greatest diameters of these teeth are still
intra-alveolar. The third specimen (SAL
163, Lavocat, 1976: PI. 4, Fig. 4), which
Lavocat referred to /. pretiosus, is an iso-

Oligocene Rodents of Bolivia • Patterson and Wood 419

Figure 18. Right upper cheel< teeth of Incamys bolivianus, anterior ends to the right, x5. A. PL) no. 20947, RP*-M^.
Part of zygomatic root added from PU no. 20938; anterior part of maxillary-palatine suture from PU no. 20954. B. PL) no.
20955, RP*-!^. C. PU no. 20943, RP*-M=. D. PU no. 20939, RP*. E. PU no. 20956, RP-M'.

lated P^, and thus its size cannot be com-
pared with that of M' or M^.

In the lower teeth, P4 is longer than the
rest, and M2 is the widest of the molars
(Table 4). The most variable factor in the
lower cheek teeth is the length of M3.
The length of the wear surface is to a
great extent dependent on age, and is
(otherwise) essentially meaningless. We
have therefore taken the greatest extra-
alveolar length of the crown as the
anteroposterior diameter of the tooth. In
cases where the entire alveolar border is
preserved, we believe this measurement
to be a homologous one. However, M3 is
implanted at an angle (Fig. 16B,C), so
that the crown of the tooth extends a con-
siderable distance behind the rear of the
alveolus, within that cavity. A recheck of
a number of specimens with unusually
large anteroposterior diameters for M3
showed that in most cases the unusually
long measurements resulted from the fact
that the posterior margins of the alveolus
were damaged or destroyed.

Because of the element of uncertainty
as to the length of M3 and the variation
of transverse diameters of M^~^ with age,
we have prepared Tables 3-4 both to in-
clude all specimens as originally mea-
sured and to exclude the anteroposterior
diameters of M3 where the location of the
posterior margin of the alveolus is uncer-
tain and the transverse diameters of all
specimens of M^~^ where the buccal mar-
gin of the crown extends 1 mm or more
beyond the labial margin of the roots
(Fig. 15D).

Lavocat (1976: 42^3) discussed the
homologies of the various parts of the up-
per cheek teeth of Incamys. We are in
agreement with him that the anterior
crest is an anteroloph uniting lingually
with the protocone, and that the second
crest is the protoloph uniting lingually
with the hypocone. On the labial sides of
the crowns of unworn teeth there are two
large, subequal cusps, continued lingual-
ly by ridges reaching about halfway
across the teeth (Fig. 17A, M^-^; 17B, F',

420 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 7

Table 3. Measurements, in mm, of upper teeth of Incamys bolivianus.

OR

P*-\P or dni^'-M*

dm^ anteroposterior
transverse

dm* anteroposterior
widtli protoloph
width nietaloph

P* anteroposterior
width protoloph
width nietaloph

M' anteroposterior
width protoloph*
u idth protol()ph#
width nietaloph*
width nietaIoph#

M^ anteroposterior
width protoloph*
width protoloph#
width nietaloph*
width nietaloph#

M^ anteroposterior
width protoloph
width nietaloph

I' anteroposterior
transverse
ratio

7

11.2-14.0

12.70

2

.61 -.68

.65

2

1.08-1.27

1.18

5

2.83^.47

3.14

4

2.81-3.68

3.30

2

2.95-3.47

3.21

21

3.04-4.11

3.59

±.29

8.05

21

3.28-4.50

3.77

±.27

7.16

20

3.25-4.36

3.85

±.36

9.35

22

2.61-3.30

2.91

±.17

5.84

16

3.40-5.8

4.15

±.56

13.49

13

3.40-4.60

3.98

±.37

9.30

13

3.32-5.8

4.03

±.73

18.11

10

3.32-4.24

3.69

±.26

7.05

14

2.68-3.48

3.00

±.22

7.33

15

3.40-5.68

4.13

±.76

18.40

12

3.40^.29

3.80

±.31

8.16

13

3.44-5.6

4.14

±.71

17.15

10

3.44-4.51

3.80

±.32

8.42

9

2.43-3..39

2.87

8

2.84-4.54

3.61

6

2.81-3.48

3.12

5

2.35-3.13

2.76

6

1.40-2.13

1.72

5

.60-.70

.64

* Including all specimens.

# Excluding PU nos. 20929, 20947, and 20955. The rationale for making computations omitting these
specimens is discussed in the text.

M2-3; Lavocat, 1976: PI. 4, Fig. 1, P). At early Oligocene (Fig. 30C), have a para-

the rear of the teeth is a crest that we and cone and inetacone (recognized as such

Lavocat agree is the posteroloph. by him) fundamentally like what we con-

The crux of our differences lies in the sider to be the paracone and metacone in

nature of the two labial cusps. We be- Incamys, but separated by one or more

lieve them to be the paracone and meta- intennediate crests (two in Fig. 30C), at

cone (Fig. lA-C; PA, ME), as they are least one of which must be the mesoloph.

subequal and occupy the nomial mam- A very similar pattern is present in the

malian positions for such cusps. Lavocat more advanced Metaphiomys schaubi

accepts the anterior as the paracone, but (Fig. 30D), also from the early Oligocene.

interi^rets the second as the labial end of In Incamys, there certainly is no trace of

the mesoloph, more specifically (1976: a crest that could be a mesoloph between

45) as a mesostyle, which would there- what we believe to be the paracone and

fore (although he does not say so) mimic metacone.

the nonnal metacone. Lavocat (1976: 42) Lavocat pointed out that some thryon-

stated that the structures of this area in omyoids such as Gaudeamus, with a re-

the thr\'onomyoids "sont morphologique- duced number of crests, are found in

ment absolument comparables." We can- association with the more primitive five-

not understand this statement, since the crested ones. He is entirely correct in

upper teeth of the most primitive known this, and Gaudeamus is also clearly spe-

thryonomyoid, Phiomys andrewsi oi the cialized in being somewhat higher

Oligocexe Rodents of Bolivia • Patterson and Wood 421

Table 4. Measurements, in mm, of lower teeth of Incamys bulivianus.

N

OR

X

a

v

P4-M3, alveolar

4

12.5-14.0

13.4

dm4-M3, alveolar

2

15.5-15.9

15.7

dm4 anteroposterior

3

ca. 3.8-5.21

4.45

width metalophid

3

1.94-2.43

2.10

width hypolophid

4

2.65-2.96

2.78

P4 anteroposterior

22

2.97-4.24

3.67

±.32

8.72

width metalophid

20

2.14-3.09

2.64

±.24

9.09

width hypolophid

21

3.03-3.61

3.28

±.20

6.09

M, anteroposterior

27

2.67-3.57

3.12

±.24

7.70

width metalophid

18

2.86^.87

3.38

±.26

7.69

width hypolophid

22

2.93^.85

3.44

±.23

6.69

Mj anteroposterior

25

2.65-3.51

3.19

±.25

7.84

width metalophid

19

2.78-4.15

3.55

±.35

9.86

width hypolophid

23

2.99-4.05

3.58

±.28

7.82

M3 anteroposterior

13

2.95^.78

3.26

±.28

8.59

revised ant. post.*

11

2.95^.50

3.17

±.20

6.31

width metalophid

12

2.85^3.68

3.29

±.24

7.30

width hypolophid

14

2.64^3.50

2.97

±.22

7.41

Ii anteroposterior

29

1.85-2.63

2.20

±.22

10.00

transverse

31

1.38-2.01

1.61

±.15

9.32

ratio

29

.61-.84

.73

±.05

6.85

* The anteroposterior measurements of two specimens (PU nos. 21952 and 22335) were omitted for this
calculation, as the alveolus was so damaged that anteroposterior measurements were not accurate.

crowned and notably more lophate than
the five-crested genera. Lavocat then ar-
gued (1976: 43) that we should agree that
four-crested genera in the Deseadan
should likewise be recognized as derived
from something like the contemporane-
ous five-crested fomis. But, in the Desea-
dan, the majority of the low crowned and
clearly cuspidate genera have four-crest-
ed upper teeth (Platypittamys, Desea-
doinijs and Sallamys; Protosteiromys is
the only exception), whereas the higher
crowned and lophate genera are fre-
quently five-crested (Cephalomys, Bran-
isamys and the unnamed genus dis-
cussed on p. 430; we do not know how
many crests there were in the upper teeth
of ScotaiJiys). Incamys, with an inter-
mediate stage in the development of hyp-
sodonty, has upper teeth that do not show
a fifth crest. As a result, Lavocat is forced
to argue that all of the primitive looking
Deseadan rodents are specialized, a con-
clusion that seems to us to make sense
only if one adopts the preconceived idea

that five-crested teeth are necessarily
primitive — a point to which we shall re-
turn below (pp. 493-495).

According to Lavocat (1976: 42), care-
ful observation of upper teeth oi Incamys
revealed a tubercle, distinct from the
posteroloph, at the labial extremity of
the latter, which he considered to be
the metacone. Unfortunately, he gave no
indication as to which specimens showed
such a structure, and we could find no
indication of its presence, either on his
stereophoto (1976: PL 4, Fig. 1) or on
any of the Princeton specimens (e.g.. Fig.
18B) — or, indeed, of anything other than
normal minor irregularities of the pos-
teroloph. Lavocat also reported what he
considered to be a variable and reduced
metaloph lying on the anterior flank of
the posteroloph. His figure of P^ of the
type (1976: PI. 4, Fig. 1) shows what may
be such a structure, well down in the
deep valley; we cannot determine
whether or not a similar structure is pres-
ent on NF of the same figure. This type

422 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

of structure, deep withiu a valley of a
tooth that is evolving toward hypsodonty,
is secondary, a neomorph, as indicated
clearly in Recent dasyproctids (Fig. 2()J),
and not a degenerating metaloph. SAL
118, which Lavocat referred to /. bolivi-
anus, does have a projection in this po-
sition on all teeth preserved, but this
specimen represents an undescribed
dasyproctid genus (p. 430).

The upper premolar is submolariform.
It is more advanced in structure and larg-
er relative to the molars than those of
contemporary octodontoids, smaller rel-
atively but somewhat more advanced
structurally than those of Cephalomys
and Scotamys, less advanced than that of
Branisamijs. In unworn or little worn
specimens (Fig. 18B-D), the anteroloph
slopes upward and a little forward from
the protocone and then curves posterior-
ly to join the paracone near its base. La-
vocat (1976: 45) noticed this condition in
the type, but concluded that the differ-
ence from the situation in the molars is
due to reduction, whereas we believe
that the premolars of Incamys were in
the process of molarization, and that the
anteroloph was developing.

A cuspule may or may not be present
at the labial end of the anteroloph; in one
specimen, PU no. 20929, there is a cus-
pule at the base of the paraflexal notch,
and in another, PU no. 20947 (Fig. ISA),
at the base of the center of the antero-
loph. The protocone is almost wholly
subordinated in the crest. The protoloph
is robust, thick, and abuts against the lin-
gual portion of the anteroloph. The hy-
pocone is sharply angulate lingually and
slightly concave posteriorly toward the
tip. It sends a blunt projection antero-ex-
ternally, the mure (interpreted by Lavo-
cat, 1976: 45, as the mesoloph), that
comes into contact with both protocone
and protoloph, and a smaller, pointed one
that touches the posteroloph. The latter
crest is expanded anteriorly at its labial
end as a smoothly curving bulge in a few
specimens (e.g., PU nos. 20960, 20943;

Figs. 17C, 18C). A metaloph is in the pro-
cess of developing, however, and this re-
gion of the tooth is in consequence ex-
tremely variable. Metaloph differentiation
ranges from nil, as in the teeth just men-
tioned, to well developed (PU no. 21728,
Fig. 17B), with all intermediate stages
represented. In some specimens, e.g., PU
no. 20965 (Fig. 17D), two small, angular
projections extend forward from the an-
teroloph. The metaloph of P*, when pres-
ent, fails to reach the mure, as is also the
case in the molars. The paraflexus is
wide, and both it and the mesoflexus are
much deeper in the middle of the tooth
than are the notches that separate the
paracone from the anteroloph and the
metacone from the posteroloph. These
notches are variable in depth; the former
may or may not be deeper than the latter.
The hypoflexus is well developed, ex-
tending up the crown for over half its
height. As wear progresses, the spurs
from the hypocone soon unite with the
protoloph and posteroloph, but the sep-
aration of the anteroloph from the proto-
loph, on the contrary, increases. A cleft
opens between these lophs, and the hy-
poflexus becomes confluent with the
paraflexus, remaining so until about half
of the tooth has been worn away. Anter-
oloph and protoloph then become joined
labially (Fig. 18A,E), the hypoflexus re-
maining open until a very advanced stage
of wear. Before lingual closure takes
place, a parafossette is isolated when
wear reaches a small isthmus that con-
nects the anteroloph to the protoloph
deep within the combined para- and hy-
poflexus. The mesoflexus is converted to
an anteriorly curving fossette either
slightly before or at about the same stage
as the labial union of ante ro- and postero-
loph. The grinding surface as a whole in-
creases greatly in width as wear pro-
gresses, due to the marked unilateral
hypsodonty (see above, p. 416), which is
comparable to that occurring in Cephalo-
mys. As in that genus, V for these values
is high (Table 3).

Oligocene Rodents of Bolivia • Patterson and Wood 423

Figure 19. Lower cheek teeth of Incamys bolivianus, x5. Dotted areas are matrix. A. PL) no. 20921, Rdm4-M,. B. PL) no.
21726, Ldm^-Mj, partly restored from right side. C. PU no. 20975, LP^-M,. D. PU no. 20918, LP^. E. PU no. 20930, RP4.
F. PU no. 20919, RP^-Mj. G. PU no. 20935, LP4-M2. H. PU no. 20937, RP^-Mj. I. PU no. 20948, RP4-M,. J. PU no. 21735,
LP4-M2. K. PU no. 20936, RM,.^. L. PU no. 20916, RM2.3.

Unworn or little worn upper molars
(Fig. 17A,B; Hoffstetter and Lavocat,
1970: Fig.; Lavocat, 1976: PI. 4, Fig. 1)
are approximately as long as wide at the
occlusal surface and are tetralophate,
sometimes with a minute projection from
the anterior slope of the posteroloph of
M^~^. As in P, the anteroloph is usually
separated from the combined protoloph
and mure by a deep hypoflexus, but, in
contrast to P*, the labial notch, the para-
flexus, between the two crests is very
shallow (Figs. 17B, M^; 17G, M^ or M^
18A,C, M^). The hypoflexus becomes pro-
gressively more extensive, vertically, in

the posterior part of the molar series. On
M' it extends up for approximately half
the height of the crown, on M^ rather
more, and on M^ it comes close to the
base of the enamel (PU no. 20947, Fig.
18A, is exceptional in having an exten-
sive hypoflexus on M^). The protoloph
sweeps lingually and posteriorly to the
hypocone, which is also connected from
the beginning of wear with the postero-
loph. On the anterior face of the latter
crest, a small projecting crestlet is pres-
ent on M^~^. The metaloph does not meet
the protoloph. The mesoflexus has the
deepest labial notch, and the metaflexal

424 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

notch is almost as shallow as the para-
flexal one. The metaflexus is short. Due
to the shortness of the metaloph, the
mesoflexus extends postero-internally to
the posteroloph, joining the metaflexus.
Moderate wear soon obliterates much of
the detail. Paraflexus, metaflexus and
crestlet disappear, the metaloph joins the
posteroloph, transitorily isolating a shal-
low metafossette in all except PU no.
20947 (Fig. 18A), and the crown surface
progressively widens. With further wear,
the mesoflexus converts to an anteriorly
curving fossette and a parafossette is iso-
lated when the deeply situated isthmus
connecting anteroloph and protoloph is
reached, as in P* (Fig. 18). In the terminal
stages, the hypoflexus converts to a fos-
sette, the parafossette disappears, the
mesofossette becomes irregularly pear-
shaped in outline, and the convexity of
the anterior faces and the concavity of the
posterior ones become decidedly pro-
nounced.

The upper cheek teeth resemble in
many ways those of Branisamijs (Figs.
23, 24). They differ in being higher
crowned and shorter relative to width, in
lacking neolophs, and in having shorter
metalophs that fail to reach the proto-
lophs. A basic resemblance also exists
between the upper molars of Incamijs
and Neoreomys (Wood and Patterson,
1959: Fig. 17A), and between the upper
cheek teeth of Incamys and Scotamys
(Wood and Patterson, 1959: Fig. 8). Re-
semblances to living dasyproctids are
discussed below (pp. 427-429).

The lower cheek teeth are four-crested
(Fig. 19B, M3; C, P4; D, P4; E, P4; F, P,;
H, P4; K, Mi_2; L, Ma-s). The anterior
part of P4 is longer than the posterior, as
Lavocat noted (1976: 46), and the anterior
part of the molars has a greater area than
does the posterior. On many molars,
there is a crest extending buccad from the
metaconid that ends freely (Fig. 19B, M3;
K, M,_.,; L, M2^3; Lavocat, 1976: PL 3,
Fig. 3, M2; Fig. 4, M,_,; Fig. 6, M,; Fig.
'^5 Mi_3). There seems to be no reason to

doubt that this is the posterior arm of the
metaconid that does not reach the proto-
conid to fomi a metalophid (or metaloph-
ulid II; Wood and Wilson, 1936); Lavocat
(1976: 47) considers this crest to be a me-
solophid, a crest with which it has not the
slightest resemblance either in position,
shape, or relationships.

P4 is tetralophate, longer than the mo-
lars, and broader across the talonid than
the trigonid. The anterolophid is trans-
verse. The metalophid is variable as to
size and degree of development (Fig. 19).
In little worn teeth it is an independent,
more or less transverse crest, but it soon
becomes connected by wear to the pos-
terior face of the metaconid. The poste-
rior ami of the protoconid is connected,
in some specimens early, in others rather
late, to the hypolophid (in which the ec-
tolophid is incorporated). The latter is
joined, after slight wear, to the postero-
lophid, which curves antero-externally to
the sharply angulate hypoconid. On
eruption, the posterolophid touches the
hypolophid (Fig. 19C), but the two crests
separate as wear proceeds, just as do the
anteroloph and protoloph of P. There is
a small spur from the protoconid extend-
ing toward, but not fusing with, the meta-
lophid, and there may or may not be one
extending posteriorly from the center of
the anterolophid. The trigonid may be
somewhat twisted labially with respect to
the long axis of the talonid, as in PU no.
20919 (Fig. 19F). The hypoflexid is deep
and wide, extending downward nearly to
the base of the enamel on the labial side.
The metaflexid is confluent with it, and
the metaflexidal notch is very shallow.
Development of the mesoflexid is vari-
able. In little worn specimens it crosses
the tooth. Thereafter it may be closed off
lingually at a rather early stage of wear or
may persist until a considerablv later one
(cf. PU nos. 20937 and 20948, Fig. 19H-
I). The anteroflexid is confluent with the
deeper and larger mesoflexid, due to the
separation of metalophid and protoconid.
In little worn specimens there is a shal-

Oligocene Rodents of Bolivia • Patterson and Wood 425

low notch between protoconid and an-
terolophid (cf. Neoreomys, Scott, 1905:
PI. 65, Fig. 9, and Scotainys, Wood and
Patterson, 1959: Fig. 9A). With deep
wear the combined anteroflexid-meso-
flexid would convert to a fossettid. The
hypoflexid ultimately converts to two fos-
settids (Fig. 19J).

The lower molars are basically tetra-
lophate, although the metalophid is in-
complete (Fig. 19). Trigonid and talonid
are approximately equal in width in Mi_2;
in M3 the talonid is narrower (Table 4).
The anterolophid is again transverse and
is joined to the obliquely-running hypo-
lophid in the unworn teeth; as in P4, a
small spur may or may not project pos-
teriorly from it toward the posterolophid
(Fig. 19B; M3). The short transverse
metalophid was apparently joined to the
metaconid in the unworn state (Fig.
19B,K; Lavocat, 1976: PI. 3, Fig. 4, Fig.
7 M3). The large, oblique posterolophid
approaches the hypolophid labially on
unworn teeth (Fig. 19B), but the cleft be-
tween them widens with wear. Lingual-
ly, the two crests are separated by the
shallow metaflexidal notch. The com-
bined hypoflexid and metaflexid is large,
deep, and extends nearly to the base of
the enamel on the labial sides of M2_3;
on Ml its extension is variable. The com-
bined meso- and anteroflexid is compar-
atively shallow and of variable shape.
There is a minute, ephemeral fossettid at
its anterolabial extremity in one speci-
men (PU no. 20916, Fig. 19L). The incli-
nation of the protoconid varies from lat-
eral to somewhat posterolateral. After
moderate wear the meso- and metaflexi-
dal notches disappear, the lingual wall of
the tooth becomes continuous, the hy-
poflexid widens, the metalophid loses all
individuality, and the combined antero-
and mesofossettid becomes slit-like; in
one specimen (PU no. 21726) a small an-
terofossettid is briefly isolated. With
deep wear the posterolophid and the hy-
polophid become joined by an isthmus,
isolating a rounded metafossettid, the

mesofossettid shortens, and the hypoflex-
id closes. The final stage revealed by our
material is a rounded hypofossettid in an
otherwise featureless crown surface, with
the enamel eliminated from considerable
parts of the periphery of the crown (Fig.
19J).

As in the case of the upper cheek teeth,
the lowers exhibit a close resemblance to
those of Branisamijs, differing chiefly in
their greater height of crown and shorter,
freely projecting metalophids. Again,
there is a basic resemblance to Neoreo-
mys in molar structure, and a less precise
one to Cephalomys. Comparisons with
living dasyproctids are made below (pp.
427-429).

As in Branisamys, dm^ was single root-
ed. It is a small, rather peg-like tooth,
wider than long, convex anteriorly and
concave posteriorly (Fig. 17A). Dm"*
(Figs. 14A; 17A,H,I) is much shorter
crowned than the molars, although the
degree of unilateral hypsodonty is essen-
tially the same. In the unworn state the
crown is longer relative to width than in
the molars, but as wear progresses the
length decreases and the width increases
until, shortly before being shed, the tooth
becomes square. The tooth is tetralo-
phate, a much longer spur extending for-
ward from the posteroloph than occurs in
the molars. The anteroloph and proto-
loph form stout crests, the fomier extend-
ing posteriorly for a little over half the
length of the tooth, the latter merging
with the sharply angular hypocone. The
metaloph fails to meet the protoloph, as
in the pennanent upper cheek teeth. The
posteroloph continues around the labial
face, where it joins the metaloph. From
this portion of the crest a small spur
(barely visible in PU no. 20980, Fig. 171)
projects lingually. The hypoflexus per-
sists to within a short distance of the base
of the enamel. In apparent contrast to the
molars, it does not appear to be confluent
with the paraflexus, the protocone and
anteroloph in PU nos. 20980 and 20981
(Fig. 17H,I) giving every appearance of

426 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

having joined at a very early stage of The incisors are slender, rather small,

wear. The paraflexus is ver\' large, deep and D-shaped in outline. The anterior

and expanded anteroposteriorly. The face is more rounded in juveniles than in

mesoflexus, due to the incomplete meta- adults (Fig. 14C-F), particularly in the

loph, sweeps posteriorly to join the large, case of the lower incisors. The thick

shallow and partially divided metafos- enamel extends much farther onto the lat-

sette. Deep wear is shown by PU no. eral than onto the mesial face. The up-

21726 (Fig. 17A). Three fossettes, two la- pers extend back into the maxilla to a

bial and one lingual, are all that remains point slightly in front of and well above

of the crown pattern; these are, respec- P. The lowers pass beneath the cheek

tively, the para-, meta-, and hypofossette. teeth at the level of Mg and temiinate

The tooth provisionally identified by posterior and lateral to M3. In all these

Scott (1905: 390; PI. 65, Fig. 3) as dm^ of respects they closely resemble the inci-

Neoreomys is similar to that of Incamys, sors of Myoprocta acouchy. The incisors

differing in the union of metaloph and are relatively much more slender than in

protoloph, and in the presence of a small Branisamys. The disparity in size be-

neoloph. We believe Scott's provisional tween the incisors of PU no. 21726 (Fig.

identification to have been correct. 14C,E) and those of PU nos. 20944 (Fig.

Four specimens of dm4, all deeply 14D) and 21735 (Fig. 14F) is due to age.
worn, are available to us. They reveal The first is a juvenile, the other two
that the tooth was long and narrowed an- adults, and rodent incisors increase in
teriorly. Two preserve the last remnant diameter as growth proceeds,
of the hypoflexid and three fossettids Relationships. Hoffstetter and Lavocat
(Fig. 19A). The posterior of these is clear- (1970: 173) called attention to a resem-
ly the metafossettid, and the intermedi- blance between the lower molars of In-
ate one is presumably the mesofossettid. camys and those of the Fayum thryon-
The anterior may be a feature of the de- omyoid Gaudeamus. Lavocat (1976)
ciduous dentition, analogous to the pos- placed Incamys in the Dasyproctidae,
terior fossettid of the trigonid of My- but gave no reasons, and his comparisons
oprocta. A less worn tooth (Lavocat, were largely with non-dasyproctid gen-
1976: PI. 3, Fig. 3) shows four fossettids. era, especially African thryonomyoids.
Lavocat considered it to be essentially The resemblance of Incamys to some
identical to one type of dm4 in the early specimens of Gaudeamus (although by
Oligocene thryonomyoid Phiomys an- no means to all — cf. Wood, 1968: Figs. 14,
drewsi (Wood, 1968: Fig. IG). Lavocat's 15) exists, but this is more than offset by
figures (1976: Fig. 2I,K) show that both major differences in the premolars and in
are elongate teeth, but to us his figures the upper molars. Affinities do not, we
strongly suggest that a more meaningful believe, lie in this quarter. As pointed out
comparison could be made with dm4 of above, there is a very close resemblance
Cephalomys (Fig. 26B). We find Lavo- between Incamys and Branisamys in
cat's description somewhat difficult to in- cheek tooth structure, but differences in
terpret, especially since the pattern the skull rather definitely point to a broad
shown in his drawing (1976: Fig. 21) rather than a narrow relationship be-
seems to differ in many respects from that tween the two.

shown by his stereophoto (1976: PI. 3, The known form to which Incamys is

Fig. 3). We continue to believe that in- most closely related is beyond doubt the

terpretation of uncertain tooth structures Santacruzian Scleromys (Scott, 1905:

in Deseadan rodents should be based on 401-403, PI. 65, Figs. 13-15), as Hoffstet-

comparisons with other South American ter (1968: 1096) originally believed. The

rodents and not with African ones. two resemble each other closely insofar

Oligocene Rodents of Bolivia • Patterson and Wood 427

as they can be compared. In both, the in-
cisors are relatively small, the patterns of
upper and lower cheek teeth are very
close at comparable stages of wear, and
the tooth rows are nearly parallel. The
snouts are similar as regards parallel-sid-
edness, position of the premaxillary-max-
illar>' suture in relation to the very large
incisive foramina, and the size and rela-
tions of the fossae for M. buccinator.
Scleromys differs from Incamys in hav-
ing cheek teeth that are somewhat longer
anteroposteriorly, a snout that is longer
relative to the length of the cheek-tooth
row, and a relatively wider palate that
lacks a groove running back from the in-
cisive foramina, and is somewhat less ex-
cavated posteriorly by the choanae.
These distinctions are such as might be
expected between earlier and later
closely related forms, and do not, in
themselves, rule out the possibility of
ancestor-descendant relationships.

As we have previously intimated
(1959: 328), we do not believe that the
Colombian Friasian species, schurmanni
and colombianus, placed in Scleromys
by Fields (1957), really belong to it. As he
pointed out, they resemble dinomyids in
various features, and we believe they
should be placed in that family. Sclero-
mys was a dasyproctid. The Colombian
species may for the present be referred,
with a large query, to the Patagonian
Friasian dinomyid Simplimus Ameghi-
no. The type species, S. indivisus, is larger
than the Colombian ones, but poorly
known. Nevertheless, the various cheek
teeth figured by Kraglievich (1930c: Figs.
2, 4) agree in structure with stages of
wear in colombianus revealed by Fields'
sections (cf. Kraglievich, Fig. 2 with
Fields, Fig. 14, P* at stage 2; his Fig. 4,
center, with Fields, Fig. 14, P at stage 5;
and his Fig. 4, left, with Fields, Fig. 14,
P4 at stage 1). The femoral and tibial frag-
ments, the calcaneum, and the astragalus
referred by Kraglievich (1930c: PI. 1) to
Simplitnus indivisus may represent an-
other dinomyid. They are disproportion-

ately large relative to cheek-tooth size,
much more so than in S. colombianus
and Drytomomys.

Stirton's record (1953: 61) of Sclero-
mys? from the Coyaima faunule of Co-
lombia is very dubious. Neither the
generic identification nor the Colhuehua-
pian age assigned to the deposit can be
regarded as established.

Most of the extinct dasyproctids so far
discovered were neither involved in nor
near to the ancestry of the surviving
fomis. Cephalomys, Litodontomys and
Olenopsis (if they are correctly placed
here) would appear to represent side
branches (Wood and Patterson, 1959:
Fig. 12). Neoreomys, the best known of
the extinct dasyproctids, is clearly a
member of a lineage pursuing an evolu-
tionary path that was not heading in the
direction of Dasyprocta and Myoprocta.
Incamys and Scleromys, however, ap-
pear to us to be close to the ancestry of
the living genera.

This is particularly evident in the
cheek teeth (Fig. 20). Worn uppers and
lowers of the living fomis display a short
hypoflexus and hypoflexid and an almost
bewildering array of small fossettes and
fossettids (Fig. 20E,F), patterns which at
first glance seem almost to defy interpre-
tation. Examination of freshly erupted or
unerupted teeth, however, reveals sim-
pler patterns (Fig. 20A-D,G-I) that ba-
sically resemble those of Incamys (Fig.
17B-D), and which clearly show the
S-arrangement of the lophs and lophids
that occurs in that fonn and in Scleromys
(see Hagmann, 1907: Figs. 25-26 for the
same pattern in Dasyprocta). The pro-
fusion of fossettes and fossettids that
characterizes worn Dasyprocta and My-
oprocta cheek teeth comes about not so
much from structures at the apex of the
tooth, as from variable projections and
pillars that arise at various depths within
the tooth, from the surrounding enamel
walls of the lophs and lophids and also,
independently it would seem, from the
bottoms of valleys (Fig. 20J and Hag-

428 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

mann, 1907: PI. 26, Fig. 9). A neoloph
and a neoloph id are present in the eheek
teeth of the living species, the former
arising from projections extending ante-
riorly from the anterior wall of the pos-
teroloph and lingualh' from its antero-la-
bial extremit>; in Iticaimjs the first of
these is sporadically present. (Here is
clear evidence, we believe, of a neoloph
arising within one lineage of a family af-
ter the same crest had come into exis-
tence in two other lineages, those of
Cephalomys and of Neoreotnys.) The
neolophid appears to arise from the base
of the tooth and to participate in the
grinding surface only after considerable
wear (Hagmann, 1907: PL 26, Fig. 9; re-
drawn here as Fig. 20J). Upper cheek
teeth of Dasyprocta and Myoprocta are
longer anteroposteriorly than those of In-
camys, but those of all three increase in
transverse diameter and decrease in
anteroposterior as wear proceeds; the
lowers also do this, but to a less marked
degree. No figured specimens of Sclero-
mys have unworn teeth, but the teeth
plainly show the S-configuration and are
intennediate in length between those of
Incamys and those of the living species
(Scott, 1905: PI. 65, Figs. 13-15). No oth-
er known extinct dasyproctid resembles
the living species as closely as do these
two in molar structure.

As concerns the skull, there are various
resemblances between Incamys and the
surviving fonns. These are described in
detail above. The more striking are the
general shape of the skull in lateral view,
the large orbits with their posteriorly sit-
uated postorbital processes, and the near-
ly parallel cheek tooth rows.

The resemblances are, of course, ac-
companied by an array of differences, as
is only to be expected when comparing
an early Oligocene fonn with living ones.
The majority of these differences seem to
be characters primitive for the Cavio-
morpha. In this category are: 1), the flat-
ness of the cranial roof; 2), the relative
shortness of the snout; 3), the more

abruptly projecting anterior root of the
zygoma; 4), the great posterior extension
of the nasals and ascending processes of
the premaxillaries; 5), the very large in-
cisive foramina, the groove running back
from them, and the associated wide sep-
aration of the fossae for the Mm. buccina-
tores; 6), the less arched occiput; and 7),
the relatively slight dorsal extension of
the supraoccipital at the center. Of these,
1) to 3) are present in Cephalomys, Bran-
isamys, and Plat y pit t amy s; 4) occurs in
Cephalomys and Branisamys, but not in
Platypittamys. Franimys and Protopty-
chus have this character very well devel-
oped, however; Platypittamys — and the
octodontoids generally — may have short-
ened these elements early. Character 5)
is shared with Cephalomys and also with
Protoptychus. Platypittamys is unknown
as regards this character, but Santacruz-
ian octodontoids possess it. Branisamys
does not, and this is perhaps an early spe-
cialization. Characters 6) and 7) occur in
Platypittamys, and are not known for the
other Deseadan genera. In features in
which Scleromys differs from Incamys
(p. 426), it is intermediate structurally be-
tween the latter and the living species.
Cranial evolution in the lineage leading
to Dasyprocta and Myoprocta has in-
volved a considerable lengthening of the
skull that has not been accompanied by
a proportionate increase in the length of
the cheek-tooth row. Cheek-tooth length
is nearly equal to snout length in Inca-
mys (99%), less in Scleromys (86%), and
still less in Dasyprocta (57%). The palate
of Scleromys is relatively wider and
longer than in Incamys, relatively nar-
rower and shorter than in Dasyprocta,
and, as in the latter, lacks the anterior
groove.

Incamys is distinguished, among
known dasyproctids, by the relatively
very large, posteriorly expanded bulla
and those features related to it, namely
the narrowness of the anterior portion of
the basioccipital, the slight degree of in-
flation posterodorsal to the tympanic,

Oligocene Rodents of Bolivia • Patterson and Wood 429

Figure 20. Molars of Dasyprocta and Myoprocta, x5. A-D. Dasyprocta noblei, MCZ no. 15937. A. Worn LM'. B. Worn
RM,. C. Unerupted LM^ D. Unerupted RMj. E-F. Dasyprocta punctata isthmica, MCZ no. 36361. E. LM^ F. RMj. G-l.
Myoprocta acouchy. G. MCZ no. 30808, erupting LM^. H. MCZ no. 30808, erupting RM2. I. MCZ no. 30805, unerupted
RM,. J. Vertical section of (unspecified) lower molar of Dasyprocta croconata, redrawn from Hagmann, 1907, PI. 26, Fig.
9. Anterior end to the left.

and, possibly, the short, non-projecting
paroccipital process. This is partly a func-
tion of size. A large bulla tends to be
characteristic of small rodents; in Platy-
pittamijs, for example, it is large although
not so expanded posteriorly. {Protopty-
chus has an even larger bulla that is more
posteriorly expanded and accompanied
by a much greater degree of inflation in
areas dorsal and posterodorsal to it; it
may well have been a saltatorial adapta-
tion.) The large size of the bulla would
not rule Incamys out as an ancestor of the
living dasyproctids, but the posterior ex-
pansion and the neighboring slight infla-
tion are features of which not a trace is to
be seen in Dasyprocta and Myoprocta.
These, the presence of a "posterior post-
orbital process," and the peculiar inter-
orbital fenestra incline us to suspect that
while Incamys was close to the Dasy-
procta-Myoprocta lineage, it lay a little
to the side of it. Whether Scleromys was
a descendant of Incamys or was more

nearly on the main line are questions that
only future discoveries can answer.

Knowledge of the upper dentition of
Incamys pennits the identification of an
isolated left upper cheek tooth from the
Deseadan of Patagonia that we described
earlier (1959: 362, Fig. 24) as PDasyproc-
tidae gen. et sp. indet. This is so close in
details of structure to P of /. bolivianus,
especially to that of PU no. 21728 (Fig.
17B), as to leave no doubt at all as to ge-
neric identity, and on the evidence we
cannot separate it specifically from the
Bolivian species. We had previously
been in some doubt as to the position of
this tooth in the series, suspecting that it
was too molarifonn in structure for a De-
seadan premolar on the basis of the Pata-
gonian evidence then available. The new
Bolivian evidence conclusively shows
that at least in the Dasyproctidae and
Dinomyidae quite molarifonn premolars
had evolved by Deseadan time. We thus
have two rodents of this Age in common

430

Bulletin Mu.seum of Comparative y.oolo^y. Vol. 149, No. 7

between Bolivia and Patagonia, now that
CeplialoDiys has been recorded from Sal-
la. The latter is the commonest mammal
in the Patagonian Deseadan, as Incamijs
may well be in the Bolivian.

Undescribed Genus, Dasyproctidae aff.

Neoreomys
Incamys bolivianus Lavocat, 1976, in part

{nee Hoffstetter and Lavocat, 1970)

An incomplete right maxilla with
erupting P"* and little worn M*~^, SAL
118, was identified as /. bolivianus by
Lavocat (1976: 43-44, 45; Fig. 4B; PI. 4,
Fig. 2). He noted various differences
from the mine run of specimens of that
species, but regarded them as probably
characterizing "une variete intraspeci-
fique." The distinctions strike us as being
of considerably higher taxonomic impor-
tance. Five, not four, lophs are present
on the molars (see Fig. 1, p. 376), and the
protoloph is joined to the apex of the pro-
tocone in the unworn P, both of which
rule out reference to Incamijs. Our series
of /. bolivianus, considerably larger than
that available to Lavocat, contains no
specimen similar to SAL 118, nor any that
are intennediate between it and typical
Incamijs.

The depths of the flexi and fossettes to-
gether with the early union of the proto-
loph and hypocone in the molars and the
presence of a hypocone on P* forbid ref-
erence to Cephaloniys. We are dealing,
it would appear, with a hitherto unknown
form. Of known dasyproctids it is closest
to Neoreomys in molar structure, the only
differences of any consequence being the
more continuous, less interrupted lophs
and the smaller size of the spur on the
anterior slope of the posteroloph, which
does not extend to the fourth crest, the
neoloph (cf. Lavocat's stereophotograph
with Wood and Patterson, 1959: Fig.
17A). The premolar differs in that the
lophs are less obliquely directed relative
to the long axis of the tooth row (cf Scott,
1905: PL 65, Figs. 1-2), and a well-de-
fined hypocone is present. The paracone

sends a short spur linguad that would
unite with the protoloph after very little
wear; this could be an individual pecu-
liarity.

On the scanty evidence available, we
have, in SAL 118, a fonn that may have
been close to the Neoreomys lineage, al-
though not on the direct line to that ge-
nus.

Cephalomys Ameghino 1897
Cephalomys bolivianus Lavocat 1 976

Cephalomys bolivianus Lavocat, 1976: 59-62; Fig.
3A; PI. 6, Figs. 1-2.

Cephalomys, the commonest Patago-
nian Deseadan genus, is represented,
with certainty, by one Bolivian specimen
in the collection deposited in the Mu-
seum Nationale. The type, SAL 164, con-
sists of the ventral portion of a rostrum
and most of the palate with the base of
the incisor and deeply worn P*-M^ of
both sides. As noted by Lavocat, the di-
mensions of the teeth fall within the size
range of C. plexus, the smaller of the two
previously known species. He gives
(1976: 59) as diagnostic features the
wedge shape ("en coin") of the rostrum,
the very slight degree of posterior diver-
gence of the cheek-tooth rows and the
absence of an infraorbital groove.

The first two of these characters — and
the smaller size — distinguish SAL 164
from C. arcidens, the type species. The
rostrum is unknown in C. plexus, but the
tooth rows are more divergent than in the
Bolivian specimen. The reported ab-
sence of the groove (presumably for the
infraorbital nerve and blood-vessels) may
be more apparent than real. This is shal-
low in C. plexus, in which it is bounded
laterally by a slight elevation on the side
of the maxilla (Wood and Patterson, 1959:
Fig. 22A). So far as can be judged from
Lavocat's illustration, the type of C. bo-
livianus may have suffered some abra-
sion in this area, sufficient perhaps to
have removed a comparable elevation
had one been present. We are by no

Oligocene Rodents of Bolivia • Patterson and Wood 431

means certain that C. holivianus is not a
synonym of C. plexus, but accept it pro-
visionally.

As noted above, the rarity of Cephalo-
mys in the Bolivian Deseadan is in strik-
ing contrast to its abundance at certain
Patagonian localities, precisely the re-
verse of the situation as regards Incamys.

Dasyproctidae inc. sed.

Genus indet., Lavocat, 1976: 68; PI. 5, Figs. 4-5.

Two mandibular fragments, one (SAL
166) with deeply worn Mj.g, the other
(SAL 167) with badly broken P4-M2,
were figured and described by Lavocat,
who noted some resemblances to Cepha-
lomys. There is indeed a certain similar-
ity to teeth of that fomi at comparable
stages of wear, particularly as regards the
preserved posterior portion of P4 (cf. La-
vocat, 1976: PI. 5, Fig. 5 and Wood and
Patterson, 1959: Fig. 18B). The molars of
SAL 166 are so deeply worn that the me-
soflexids are represented only by undu-
lations on the lingual surfaces, yet a mi-
nute remnant of the anterofossettid of Mg
and of the metafossettid of M3 still re-
main. In the Patagonian species of Ceph-
alomys all traces of these would have
been lost at much earlier stages of wear.

It is conceivable that these specimens
might be referable to C. holivianus, with
which they agree in size, but it is impos-
sible to be certain on such material.

Family DINOMYIDAE Alston 1876
Branisamys Hoffstetter and Lavocat 1970

Synonym. Villarroelomys Hartenberger 1975.

Emended Diagnosis. Skull, so far as
known, basically dinomyid. Snout heavy,
ventral side of rostrum elevated as in
some dinomyids, less so than in Dino-
niys; fossa for origin of M. masseter me-
dialis, pars anterior very deep, markedly
excavated dorsally; nasals wide anterior-
ly, tapering gradually posteriorly, longer
than in Dinomys, extending back over
M*. Posterior ends of nasals and premax-

illae fonning nearly straight line across
top of skull. Interpremaxillary foramen
half way back on mid- ventral surface of
premaxilla, small, round, not slit-like.
Premaxillary-maxillary suture crossing
palate at rear of short, broad incisive fo-
ramina. Strong vertical ridge near front of
lacrimal. Posterior palatine foramina be-
side M^ entirely within palatine. Ante-
rior end of palatine reaching forward to
posterior end of P. Posterior margin of
palate with small median process as in
Paramyidae and Reithroparamyidae.
Cheek-tooth rows divergent posteriorly.
Knob for M. masseter superficialis on
ventral surface of zygoma, as in Dinomys.
Dental fonnula Ij^ P/, Mg^, dnij^; inci-
sors with flat anterior faces, robust but
smaller and narrower relative to cheek
teeth than in later dinomyids, gently
rounded anteriorly. Base of upper incisor
above P. Single rooted dm'' present; dm^
and P4'* essentially molarifonn, P^ with
low metaloph; dm4 with highly complex
anterior portion. Cheek teeth low
crowned with incipient unilateral hyp-
sodonty, lingual in upper teeth, labial in
lowers. Cheek teeth slightly longer than
wide when little worn, becoming wider
than long with wear. Anteroloph of upper
cheek teeth isolated in early stages of
wear; remaining lophs uniting early and
converging posterointernally on hypo-
cone. Lower cheek teeth with posterolo-
phid isolated in early stages of wear, re-
maining lophids uniting early and
converging anteroexternally on protoco-
nid.

Type Species. Branisamys luribayen-
sis Hoffstetter and Lavocat 1970.

Distribution. Deseadan Oligocene of
Bolivia.

Branisamys luribayensis Hoffstetter and
Lavocat 1970
Figures 21-25

Synonym. Villarroelomys holivianus Hartenberger

1975.

Diagnosis. As for the genus. Tooth
measurements as given in Tables 5-6.

432 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 7

Figure 21. Dorsal view of skull of Branisamys luribayensis, PU no. 20914, x3.

Abbreviations: FR — frontal; FR? — probable process of frontal, with crack between it and frontal; MX — ascending process
of maxillary. Dotted areas are matrix.

Type. Portion of skull with root of dm^,
dm'*-M^ right, cited without museum
number or museum location by Hoffstet-
ter and Lavocat (1970: 172), and as SAL
102, without museum designation, by
Lavocat (1976: 33).

Type of Synonym. An incomplete
Ldm4, GB 014.

Referred Material. 31 specimens in the
MNHN and Princeton collections, listed
in Appendix 1 (p. 528).

Horizon and Locality. Deseadan early
Oligocene, lower part of Estratos de Sal-
la; Salla-Luribay Basin, Bolivia; all spec-
imens are from near Salla Fann.

Description. The type is a skull miss-
ing the snout, the zygomatic arches, the
occiput and part of the basicranium. Un-
fortunately, it has not been prepared
enough to show any of the basicranial
features or those of the orbit. Lavocat
(1976: 34) stated that the orbitotemporal
region is poorly preserved; his stereo-
photos (PI. 2, Figs. 1, 3) suggest that it is
still buried in matrix. Unfortunately, su-

tures are often impossible to identify on
the photos. The Princeton skull, PU no.
20914, therefore adds considerably to our
knowledge of the genus.

The rostrum is short — the length of the
diastema being approximately equal to
that of the cheek-tooth row (Figs. 22,
23A) — and very stout, comparable to
those of Neoreomys and Ditiomys, and
much more robust than those of Platy-
pittamys, Cephalomys, Incamys or the
living dasyproctids. Dorsally, the sides
are nearly parallel for most of their length
but are slightly convergent anteriorly
(Fig. 21); ventrally, they converge for-
ward from the anterior roots of the zygo-
matic arches (Fig. 23A). The ventral sur-
face is wide, gently rounded in the
premaxillary portion and slightly concave
in the maxillary portion anterior to the
cheek teeth; this is probably the fossa of
origin of M. maxillolabialis . As may be
seen in side view (Fig. 22), the under
side of the rostrum slopes steeply upward
relative to the occlusal surface of the

Oligocene Rodents of Bolivia • Patterson and Wood 433

Figure 22. Lateral view of skull, Branisamys lun'bayensis, PL) no. 20914, x3. Part of orbital region restored from right
side.

Abbreviations: AEF — anterior ethmoid foramen; F — foramen into nasolacrimal canal; L — lacrimal foramen; N — nutritive
foramina; OF — optic foramen; OSPH — orbitosphenoid; RLA — right lacrimal, reversed from opposite side; RMX — part of
right maxillary, reversed from opposite side; SPF — sphenopalatine foramen. Dotted areas are matrix.

cheek teeth, not so much as in Dinomys
but in striking contrast to the condition
in the dasyproctids. Most of the lateral
surface of the snout is occupied by the
very large fossa for M. masseter medialis,
pars anterior, which extends far forward
onto the premaxilla. The fossa is compa-
rable in extent, except anterodorsally, to
that of Dinomijs but is more deeply in-
cised and is bounded anteriorly and dor-
sally by a much more prominent bony
rim (Fig. 22).

As in Incamys, the nasals are very long
for caviomorphs, extending back to a
point above M*. In Branisamys these
bones are narrow. Bluntly pointed, they
increase somewhat in width posteriorly
for slightly less than half their lengths
and then decrease evenly toward the
frontals. In contour, therefore, they re-
semble the shorter and wider nasals of

Dinomys, and differ from the posteriorly
tapering ones of Neoreomys or the more
parallel-sided ones of Incamys and the
living dasyproctids.

The premaxillaries are large and ro-
bust. Ventrally, they are relatively wider
and the lateral margins are more conver-
gent anteriorly than in Dinomys and, as
already noted, they are more excavated
laterally for the reception of the anterior
portion of M. masseter medialis. The in-
cisive foramina are kidney-shaped (Fig.
23A), much broader in proportion to
length than in Dinomys or the living das-
yproctids; in contrast to Incamys and oth-
ers (p. 400), they do not extend poste-
riorly. The interpremaxillar>' foramen is
small, positioned as in Dinomys, but is
round, rather than elongate as in the liv-
ing genus or in Neoreomys. The premax-
illary-maxillary suture crosses the palate

434 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Figure 23. Skull and teeth of Branisamys luribayensis, and tooth of Dinomys. Skull, x3, teeth, x5. A. Ventral view of
skull, PU no. 20914. Dotted areas are matrix. B. LI', PU no. 20914, from the front. C. LP*, PU no. 20915. D. LP*-M^, PL)
no. 20914. E. Rl,, PU no. 21734, from the front. F. LP*-M', PU no. 21732. G. RM3, PU no. 21734. Dotted area is matrix.
H. Dinomys sp., LM', AMNH (Mam.) no. 100011. Dotted areas are cement.

at the rear of the incisive foramen, then
bends forward along the lateral margin of
the foramen to its middle, where it turns
transversely laterad across the palate,
turns dorsad on the side of the snout,
though with somewhat of a forward bulge

that is reminiscent of the paramyids and
reithroparamyids (e.g., Wood, 1962a:
Figs. 2B, 13B, 48B). The positioning of
the suture along the ventral surface is
very like that in Dinomys. The dorsal
process of the premaxillary extends back

Oligocene Rodents of Bolivia • Patterson and Wood 435

almost as far as the posterior end of the cisor, opens into the nasolacrimal canal,

nasal (Fig. 21). as can be seen by passing a flexible probe

As in Dinomys, there are paired gutters through the canal from the lacrimal fora-
along the palate, running from the inci- men. This opening is very much smaller
sive foramina to the posterior palatine fo- than in the later dasyproctids. There is a
ramina. The two gutters are separated, foramen, in Branisamys, just behind and
again as in Dinomys, by a median ridge, below the base of the incisor, from which
but one much lower than in the living a faint groove extends upward and back-
genus. The palate narrows forward (or ward toward the lacrimal foramen (Fig,
the cheek tooth rows converge) as in Di- 22; F). Breakage on the right side shows
nomys. The posterior palatine foramina that this foramen is continued forward in
lie in the gutters, as in Dinomys, but the the bone, toward the nasal passage, and
foramina are in the palatine, whereas in that it unites with another channel, run-
Dinomys they are on the maxillary-pala- ning in the same general direction, which
tine suture. Our description of this area appears to be the nasolacrimal canal,
differs slightly from that of Lavocat This foramen would seem, then, to be
(1976: 35). We cannot determine whether homologous to the opening in Dinomys
this is due to variability of this area in and to the much larger one of the later
Branisamys, or whether his description dasyproctids.

is in error, since we found it impossible The fragmentary snout (PU no. 21960)

to trace the maxillary-palatine suture on is broken, anteriorly, at about the mid-

his stereophotos (Lavocat, 1976: PI. 2, point of the upper incisors (Fig. 25B).

Figs. 1, 4). Posterior to the foramina, the The incisors, inclosed in the premaxillae,

maxillary fomis a shelf, lateral to the pal- have a large pulp cavity. Ventrally, the

atine and separating it rather widely from incisive alveolus is separated by a thin

the alveoli of the molars. lamella of bone from a rounded cavity,

There is a prominent rugosity for the the anterior extension of the nasolacri-

tenclon oi M. masseter superficialis at the mal canal (Figs. 24A, 25B; NLC). The

medial end of the zygoma. To judge from ventral portion of the premaxilla bears a

conditions in the Paramyidae and Rei- V-shaped dorsal groove that houses the

throparamyidae (Wood, 1962a: Figs. 24C, vomer (Fig. 25B). Between the incisive

31C, 48B), this is a primitive character, alveoli, there are at least two vertical la-

As in Dinomys, the passage for the in- mellae of bone, probably portions of the

fraorbital nerve and blood vessels is turbinals, above which the nasal cavity

crowded between the base of the incisor widens (Fig. 25B; NC).
and the root of the zygomatic arch. There Seen from the side (Fig. 24A), the most

are two fair-sized nutritive foramina in important feature of this fragmentary

the ventral wall of this passage (Fig. 22; snout is the nasolacrimal canal, which

N). The sphenopalatine foramen lies in passes forward and ventrally from the lac-

the posterior part of the orbital portion of rimal foramen, ventral and mesial to the

the maxilla and is touched posteriorly by base of the incisor and lateral to the nasal

the orbitosphenoid and bounded along passage. At about the middle of the frag-

its posterodorsal portion by the frontal, ment, the canal enlarges and curves up-

On the left side, a valley for a nerve or ward (Fig. 24A) to fomi a large, rounded

blood vessel runs ventrally and then pos- cavity just ventral to the incisive alveoli,

teriorly from this foramen (Fig. 22). No as seen from the front (Fig. 25B; NLC).

trace of this channel was detected on the Posteriorly, the snout is broken off just

right side. behind the cribriform plate, so that this

In Dinomys, a small opening in the latter is exposed from the rear. The plate

maxillary, just behind the root of the in- is about 8.5 mm wide and 7 mm high.

436 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

D

^ W K^ 1* ^

Figure 24. Snout and teeth of Branisamys luribayensis. Teeth, x5; other scales as indicated. A. Ventrolateral view of
snout, PL) no. 21960, oriented with dorsal surface of nasal horizontal, x2. B. PL) no. 21955, Rdm-'-M' and root of
dm-'; anterior end to the right. C. PU no. 21943, RP^-M^ anterior end to the right. D. Root pattern of PL) no. 21978,
edentulous lower jaw, x3. E. PU no. 21980, RP^. F. PU no. 21944, RM, 3. G. PU no. 21987, broken Rdm^ and M,. H. PU no.
21986, LM..

Abbreviations: Bl — base of incisor; FR — frontal; I' — position of upper incisor; LF — lacrimal foramen; MX — maxilla; NA —
nasal; NLC — nasolacrimal canal; NP — nasal passage; PMX — premaxilla.

Oligocene Rodents of Bolivia • Patterson and Wood 437

There are relatively few, rather large ol-
facton,' foramina. There is a well-devel-
oped lamina perpendicularis, on either
side of which is an elongate vertical
opening.

The nasofrontal sutures are irregular;
the frontals notch the nasals very slightly
at the midline, and send forward small
wedges between the premaxillaries and
nasals (Fig. 21), as in Cephalomys. The
Princeton specimen does not extend far
enough posteriorly to detemiine whether
or not a postorbital process was present,
but one clearly was present in the type
(Lavocat, 1976: 34; PL 2, Fig. 3). On the
right side of the skull of PU no. 20914
there is a piece of bone (Fig. 21; FR?)
lateral to the premaxilla and abutting an-
teriorly against the maxilla, from which
it is separated by an obvious suture. Pos-
teriorly, this section is separated from the
frontal by either a suture or a ver>' suture-
like crack. Unfortunately this area is
missing on the left side, as well as on the
type. If this area is part of the frontal, as
seems likely, then a projection from that
bone extended forward between the pre-
maxilla and maxilla as in Cephalomys,
Incamys, and the living dasyproctids. If,
on the other hand, this area is not part of
the frontal but some other ossification
(most likely a peculiarly shaped dorsal
process of the lacrimal), the anterior edge
of the frontal is almost exactly as in Di-
nomys and Neoreomys.

Much of the orbital portion of the fron-
tal is present on the right side of PU no.
20914, and agrees closely, so far as pre-
served, with the corresponding part in
Dinomys, being much less concave in
the anteroposterior direction than in the
living dasyproctids or Neoreomys. The
ventral margin of the frontal forms a near-
ly horizontal line, running back from the
posteroventral corner of the lacrimal to
the orbitosphenoid. A good-sized ante-
rior ethmoid foramen enters the frontal,
in a posterodorsal direction, above xM^.
This area has not been prepared in the
type (Lavocat, 1976: PI. 2, Figs. 1, 3).

The ventral half of the right lacrimal is
present; it is ver>' similar to that of Di-
nomys, being essentially flat and lacking
any trace of the vertical plate that, in das-
yproctids, connects the cranial and preor-
bital portions of the bone and buttresses
the dorsal root of the zygomatic arch. A
vertical ridge near the front of the lacri-
mal of Branisamys marks the anterior
margin of the lacrimal gland just above
the nasolacrimal canal. Such a ridge does
not occur either in Dinomys or in the
dasyproctids, unless, perhaps, the ridge is
homologous to the initial development of
the buttress of the dasyproctids.

The anterior end of the palatine, at the
midline, reaches the level of the rear of
P^, a condition that is more dasyproctid
than dinomyid. According to Lavocat
(1976: 35) the palatines of the type reach
only to the anterior quarter of \P. The
notch at the rear of the bony palate is
rounded in Lavocat's figure of the type
(1976: PI. 2, Fig. 4), and lies opposite the
middle of M^; in PU no. 20914 (Fig. 23A)
it clearly curved across the palate farther
posterad, and there are the remains of
what seems to have been a median tu-
bercle (Fig. 23A) as in the Paramyidae,
Reithroparamyidae, and Neoreomys, a
structure not present in Dinomys or liv-
ing dasyproctids.

The little that is preserved of the or-
bitosphenoid reveals that, anteriorly, it
reached the posterior margin of the
sphenopalatine foramen (Fig. 22), as in
Incamys and modern dasyproctids and in
contrast to the condition in Dinomys, and
that the optic foramina were close togeth-
er and confluent as in Dinomys, Incamys,
and modern dasyproctids, whereas the
two foramina are widely separated in
Neoreomys.

The posterior part of the frontals and
the parietals, together with the available
parts of the squamosal and bulla, have
been described by Lavocat (1976: 34-35).

As stated by Lavocat (1976: 35), the
mandible is heavy. It is, however, not
"extremement" robust as stated by Lavo-

438 Bulletin Museum of C-oiupurative Zoo/o^?/, Vol. 149, No. 7

Figure 25. Snout, jaws and teeth of Branisamys luribayensis. Teeth, x5; other scales as indicated. A. PL) no. 21951,
RPj-M... B. PL) no. 21960, snout brol<en about mid-length, seen from front. C. Lateral view of lower jaw, x3.
Largely based on left mandible, PL) no. 21978, but masseteric fossa and angle completed from PL) no. 21979; teeth
reversed from PL) nos. 21944 and 21951. D. PL) no. 21978, cross section of LI, below diastema. E. PL) no. 21944,
cross section of Rl, below anterior root of dm^. F. Medial view of lower jaw, <3. Based on same specimens as in C.
G. PL) no. 21987, cross section of Rl, below rear of M,, seen from the rear. H. PL) no. 22172, cross section of Rl, below
rear of Mo, seen from the rear.

Abbreviations: I — cross section of incisor;
V — vomer. x2.

NA — nasal; NC — nasal cavity; NLC — nasolacrimal canal; PMX — premaxilla;

Oligocene Rodents of Bolivia • Patterson and Wood 439

cat, nor anywhere near so heavy as it (Woods, 1972: 128, Fig. 3A) is the most

would appear in his figures (1976: Pi. 2, prominent.

Figs. 5-7), in which the jaw was photo- On the medial surface of the mandible,
graphed perpendicular to the occlusal the chin process, behind the symphysis,
surface. As in all caviomorphs, if the jaw is highly corrugated (Fig. 25F). There is
is oriented in its natural position, the oc- a distinct cavity, above the corrugated
clusal surface of the cheek teeth slopes area and in front of P4 , which may rep-
down at about 30° toward the midline, resent the fossa for an unusually large M.
When photographed perpendicularly to ge/u'og/os^us. This view of the jaw shows
the occlusal surface, the medial surface the position of the deep groove between
of the jaw, which was vertical in life, ap- the incisive alveolus and the angular pro-
pears as a broad medial shelf (Lavocat, cess.

1976: PI. 2, Figs. 5-7). Anteriorly, our At the time of its description, the type
specimens are all broken farther to the provided, for the first time among Desea-
rear than is SAL 105 (Lavocat, 1976: PI. dan caviomorphs, some evidence of the
2, Fig. 6) and show nothing of the length upper deciduous dentition (Hoffstetter
of the diastema (Fig. 25C,F). However, and Lavocat, 1970: Fig.; Lavocat, 1976:
his photograph looks as though the inci- 36, PI. 2, Fig. 4), but it is no longer
sor had been pushed to the rear in SAL unique, as the upper deciduous dentition
105, greatly shortening the diastema, of Incamys is now known (see above, p.
There is a deep and very narrow groove 425) and dm'* and the root of dm^ oi B rati-
on the ventral surface of the mandible isamys are present in PU no. 21955 (Fig.
between the incisive alveolus and the an- 24B). This specimen and the type pre-
gular process, which seems to differ from serve the single root of a very small tooth
what Lavocat saw, since he said (1976: in front of dm"* (for a discussion of the
35) that the groove was "large et pro- precise nature of the tooth in front of dm"*
fond." However, his stereophoto (1976: or P, see Wood, 1969; 1970b: 245-246).
PI. 1, Fig. 6) shows that the matrix had In the adult Princeton specimens there
not been removed from this region before is no trace of such an alveolus, the max-
the specimen was photographed. There ilia sloping upward from the anterior bor-
is a prominent chin process (Fig. 25C) on der of the alveolus of P* with no indica-
which there are a number of nutritive fo- tion of an anterior alveolus, unless what
ramina. There is no trace in any of the seem, in PU no. 20914, to be nutritive
Princeton specimens of a mental foramen foramina represent the last stage in the
(Fig. 25C; the small foramen beneath P4 filling of the alveolus with replacement
in PU no. 21978 faces backward and can- bone, as occurs in the European para-
not be a normal mental foramen), which myid Plesiarctomys (Wood, 1970b: 245,
would seem to indicate that the foramen Fig. 8), and, possibly, in Sallamys (p.
was farther forward than in Dinomys 383). However, there are definite nu-
(Ellennan, 1940: Fig. 41). Lavocat nei- tritive foramina in this position in, e.g.,
ther figures nor mentions a mental fora- Erethizon, in which neither dm^ nor P^
men. The dorsal border of the masseteric is ever present (Wood and Patterson,
fossa is clearly marked and is continued 1959: 378). The areas in front of P in PU
into the coronoid process, which passes nos. 20915, 21732, 21733, and 21943
the alveolar border by the talonid of M3. show no nutritive foramina and no sug-
There are a number of suggestions of gestion of an alveolus,
subdivision within the masseteric fossa The functional upper cheek teeth are
(Fig. 25C), among which the insertion of basically low crowned, although unilat-
M. masseter medialis, pars posterior eral hypsodonty has begun to develop on

440 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Table 5. Measurements, in mm, of upper teeth of Branisamys luribayensis.

20914

PU nos.

20915
L

21732
L

21733
R

21943
R

21955
R

21960

L N

P^-M^ anteroposterior

20.8

dm'' anteroposterior

width nietaloph

P"* anteroposterior

5.20

width protoloph

width metaloph

M* anteroposterior

4.34

width protoloph

ca. 5.9

width nietaloph

M^ anteroposterior

5.05

width protoloph

ca. 6.0

M^ anteroposterior

4.82

width protoloph

5.44

width metaloph

5.30

I' anteroposterior

4.9

transverse

3.32

ratio

.68

3.24

20.9

5.83

5.00

5.50

4.68

4.53

5.53

6.82

5.18

5.05

7.30

6.29

5.24

6.15

4.88

4.85

4.45

5.40

6.45

6.30

5.92

5.83

5.40

5.57
4.98
5.14

5.83

2

20.85

1

1

5

5.09

4

6.09

3

5.89

5

4.78

4

6.14

3

5.69

2

5.31

1

6.0

2

4.90

1

5.44

2

5.22

4.21

2

4.56

3.14

3

3.23

.75

2

.71

the lingual side. Lavocat (1976: 36) com-
pares this condition to that in the Mio-
cene Paraphiomys pigotti of Africa.
There is no question but that the two are
at comparable stages of evolution of uni-
lateral hypsodonty. However, the same
may be said for many other rodents (para-
myids, reithroparamyids, theridomyids,
prosciurines, ischyromyids and cylindro-
donts, to name a few selected at random
from the early Tertiary). All that this in-
dicates is that unilateral hypsodonty is
very common in the upper cheek teeth of
rodents — and lagomorphs as well (Wood,
1942: 4, Table p. 5). Unilateral hypso-
donty has absolutely no weight as an in-
dication of caviomorph-thryonomyoid re-
lationships short of the ordinal level. The
degree of lingual hypsodonty in Brani-
samys is no greater, proportionately, than
in most paramyids and reithroparamyids.
Enamel was continuous around the
crown until late stages of wear (Figs.
23D, 24C, rear of M^, and there does not
seem to have been very much interdental
wear. The teeth are somewhat wider than
long for most of their lifetime, but in the
early stages of wear the grinding surfaces
are rather longer than wide (Table 5).

This change takes place through the pro-
gressive widening of the wear surface
down the lingual slope of the protocone
and hypocone, but the change is not so
prominent as in Incamys.

All upper cheek teeth are five-crested
(Figs. 23C,D,F; 24B,C; Lavocat, 1976: PI.
2, Fig. 4), having a small fourth crest, the
neoloph, that attaches to the anterior face
of the fifth crest, the posteroloph, near its
middle. For terminology of the parts of the
cheek teeth, see Figs. 1-2 (pp. 376-377).
Each tooth overhangs posteriorly behind
the hypoflexus, which results in progres-
sive decrease in the size of the posteroin-
ternal corner as wear proceeds. There is
a slight tendency for the enamel to be
thicker on some of the anterior than on
some of the posterior faces of the lophs,
a tendency carried much farther in Di-
nomys. Cement is lacking on all teeth.

Little worn teeth are preserved in PU
nos. 20915, 21732, 21733, 21943 and
21955. P^ is almost, but not quite, fully
molariform. Protocone and anteroloph to-
gether fonn a stout crest, bulging antero-
lingually (Fig. 23C,F), that extends from
the posterolingual portion of the crown
around most of the anterior face, falling

Oligocene Rodents of Bolivia • Patterson and Wood 441

short of the paracone. Paraflexus and hy-
poflexus are confluent through much of
the depth of the crown, effectively iso-
lating this anterior crest until an ad-
vanced stage of wear (Figs. 23A, 24C).
The protoloph, thus separated from the
protocone, extends obliquely from the
paracone to the hypocone. In early wear
stages (Fig. 23F) the two cusps are sep-
arate. The mure, as such, has lost all in-
dividuality. The neoloph is a short crest
that joins the center of the posteroloph
after a little wear. Labially, it soon fuses
with the posteroloph to isolate a small,
ephemeral neofossette. The "metaloph"
is the last of the crests to come into wear.
Although we employ the same names for
structures on premolars as for the topo-
graphically corresponding ones on the
molars, it is nevertheless quite clear that
the metaloph has had different histories
in the two sets of teeth (p. 386). It is clos-
est to the wear surface labially where it
approaches the neoloph, from which it is
separated by a slight notch, and then
makes contact far dorsal to the grinding
surface with a bulge on the labial side of
the protoloph. Midway in its course, the
metaloph is slightly constricted from be-
low and from the sides, and these con-
strictions give it a fleetingly bicuspidate
structure. The combined para- and hy-
poflexus and the mesoflexus extend to
within short distances of the base of the
enamel. The small neoflexus, which is
open labially only in the earliest stages
of wear (Fig. 23F), is the shallowest of
the crown features, and disappears after
a small amount of wear (Fig. 24C). This
is the most advanced P yet reported from
Deseadan rodents. The premolars of PU
nos. 21732 and 21733 have different mea-
surements and are smaller than those of
the other specimens (Table 5) because
the teeth are not fully erupted.

The molars differ from P^ primarily in
the more buccal extension of the antero-
loph, in their greater length on eruption
(Fig. 24B) and in the more nearly central
position of the lingual opening of the hy-

poflexus, again on eruption (Fig. 24B).
The length rapidly decreases with wear.
The three posterior lophs are variable in
their connections. M' and M^ of the type
resemble P of PU no. 20915. In M^ of the
type, however, the central portion of the
metaloph inclines posteriorly to unite
with the neoloph, its normal course being
marked by a buccal cuspule. This cus-
pule was regarded by Hoffstetter and
Lavocat (1970) and by Lavocat (1976: 37)
as a mesostyle. We take it to be a minor
individual variant representing a second-
arily isolated remnant of the metacone.
Comparable metaconal isolations can be
seen in little worn molars oi Dinomys, as
individual variations (Fig. 23H). The val-
ley between the posteroloph and the neo-
loph is extremely shallow (Fig. 24B) and
rapidly removed by wear (Fig. 23F).

In M^ of PU no. 20914 (Fig. 23D), the
metaloph and neoloph unite in an irreg-
ular V that is isolated from the protoloph
and posteroloph by the combined meso-
flexus and neoflexus; slight additional
wear would result in a neoloph-postero-
loph union. In PU no. 21943 (Fig. 24C),
the posteroloph and neoloph have united,
leaving a nearly isolated metacone.
These variations are of no taxonomic sig-
nificance; minor and inconsequential
shiftings of loph connections occur with
some frequency among caviomorphs.
The molars of PU nos. 20914 and 21940,
older individuals than the type, reveal
that the combined paraflexus and hypo-
flexus closes from the buccal toward the
lingual side, and that the metafossette is
somewhat more persistent than the me-
sofossette. In the type, M^ is small in
comparison with the anterior teeth,
which is not the case in PU nos. 20914
and 21943; this may be due to incomplete
eruption of M^ in the juvenile type. Lavo-
cat's stereophoto (1976: PI. 2, Fig. 4) sug-
gests, however, that the eruption is al-
most complete but that the tooth is still
partly buried in matrix.

There are a number of lower jaws in
the Princeton collection, several being

442 Biillcfin Museum of Comparativv Zoolofiy, Vol. 149, No. 7

Table 6. Measurements, in mm, of lower teeth of Brani.sanuj.^ luribaijensis .

PU nos.

21734

21944

21951

21978

21980

21986

21987

22172

22173

R

R

R

L

R

L

R

R

R

N

X

P4-M3 anteroposterior

20.55

(alv.)

1

clm4 anteroposterior

ca. 7.15

1

width hypolophid

ca. 4.7

1

P4 anteroposterior

6.13

6.28

2

6.20

width metalophid

4.58

4.29

2

4.44

width hypolophid

4.91

5.01

2

4.96

M, anteroposterior

5.17

4.64

5.40

3

5.07

width metalophid

5.02

5.26

5.37

3

5.22

width hypolophid

5.21

5.47

5.43

3

5.38

M2 anteroposterior

5.56

4.80

5.67

3

5.34

width metalophid

5.50

5.46

5.55

3

5.50

width hypolophid

5.54

5.28

5.48

3

5.43

M3 anteroposterior

5.67

5.63

2

5.65

width metalophid

5.63

4.74

2

5.19

width hypolophid

4.93

4.30

2

4.62

I, anteroposterior

3.47

4.03

4.08

4.55

3.48

3.80

4.00

7

3.92

transverse

2.65

2.92

3.07

3.66

2.65

3.04

3.17

7

3.02

ratio

.76

.73

.75

.80

.76

.80

.79

7

.77

edentulous, but the dentition is well rep- and anterolophid are united lingually,

resented by PU nos. 21944 (Fig. 24F); even in unworn teeth (Fig. 24F,G), and

21951; 21987 (Fig. 24G); 21980 (Fig. there is thus an anterofossettid from the

24E); 21734 (Fig. 23G); and 21986 (Fig. beginning of wear. The metalophid is a

24H). The edentulous jaws clearly show large crest that shows no signs of reduc-

the root pattern (Fig. 24D), with four tion. Lavocat (1976: 38) identified what

well-separated roots (the largest beneath we temi the metalophid as "un mesolo-

the hypoconid) on P4-M2, but with only phide ou bras posterieur du protoco-

a single anterior root under dm4 and a sin- nide," which, apparently, he considered

gle posterior one under M3. to be the same thing. They are, of course,

Lower teeth were illustrated by Hoff- quite different. The mesolophid is a

stetter and Lavocat (1970; Fig.), and La- structure derived from the mesoconid, a

vocat described lower teeth in detail cusp on the ectolophid behind the pro-

(1976: 38-40), figuring three specimens toconid. The posterior ami of the proto-

(1976: PI. 2, Figs. 5-7). All pemianent conid has nothing to do with the ectolo-

teeth are tetralophate with the enamel phid, and runs directly linguad from the

extending farther rootward on the buccal posterior part of the protoconid. If it is

side, especially on the hypoconid, than large enough, and reaches the metaconid,

on the lingual side. They are to a consid- it is known as the metalophid (or meta-

erable extent mirror images of the uppers lophulid II). The presence of both a me-

(except that the latter have five crests), solophid and a posterior ami of the pro-

the anterior three lophids uniting at the toconid is clearly shown in the Egyptian

protoconid, with the posterolophid ini- Phiomijs (Wood, 1968: Fig. IC; ID, Mj;

tially set apart by confluence of the hy- IE; IG, Mj). Nothing of this sort has been

poflexid and metaflexid; these, however, reported in any caviomorph.
become separated at an intermediate There are two specimens of P4 in the

stage of wear (Fig. 23G). Hypolophid and Princeton collections; Lavocat apparent-

metalophid are separated lingually by a ly had three. The posterior half of the

rather deep mesoflexid, but metalophid premolar is identical to that of the molars

Oligocene Rodents of Bolivia • Patterson and Wood 443

(Figs. 24E, 25A). There is apparently
considerable variation in the anterior half
of the tooth. Lavocat stated that one spec-
imen, referred to by him (1976: 39) as
SAL 103, with P4 in process of eruption
and Ml , shows essentially the same struc-
ture in the anterior half of P4 as do the
molars. However, his photograph of SAL
103 (1976: PI. 2, Fig. 7) shows a jaw with
Mi_3 and the roots of dnxj. This photo-
graph is apparently of the jaw that he
identified as the one referred to the holo-
type (1976: 33), which he stated in-
cludes a left ramus with the alveolus of
dm4 and Mi_3 (M3 unerupted). This is
also apparently the lower jaw mentioned
by Hoffstetter and Lavocat (1970: 172) as
having P4-M3 with a length of about 20.8
mm. It clearly is not the right P4-M1 they
figured (1970: Fig., B.l.Dr.). The list of
referred specimens given by Lavocat
(1976: 33) does include one that agrees
with the 1970 figure: "P4 inferieure en
eruption et Ml SAL 106." This is pre-
sumably the specimen he identified
(1976: 39) as SAL 103.

In both Princeton specimens of P4
(Figs. 24E, 25A) there is a posterior ex-
tension from the anterolophid, suggest-
ing the rear of an anteroconid, which
dams or partly dams the anterofossettid.
There are also one or more lingual cusps
that close or nearly close the anterofos-
settid lingually. Lavocat's description of
this area in SAL 108 (1976: 39; unfortu-
nately the specimen was not illustrated)
is rather hard to follow, but probably in-
dicates a tooth essentially like the two
figured here. The anterior half of P4 of
SAL 105 (Lavocat, 1976: PI. 2, Fig. 6) is
completely disintegrated, and shows
nothing. It is possible that Lavocat's de-
scription of this specimen (1976: 39) was
written before the anterior part of the
tooth was destroyed. It is unfortunate that
Lavocat published no stereophotos of the
jaw with P4 in process of eruption (SAL
106?), which he stated (1976: 39) is fun-
damentally like the molars in its pattern.
This is in accord with the drawing pub-

lished by Hoffstetter and Lavocat in
1970, but we wonder whether a ridge
from the anterolophid might not be pres-
ent deep within the anterofossettid, far
enough below the surface in the partly
erupted tooth so that its presence was not
noted. If so, this would make it identical
to the other premolars.

Dm^ is somewhat broken and deeply
worn in the type (Hoffstetter and Lavo-
cat, 1970: Fig., B.I.; Lavocat, 1976: PL 2,
Fig. 4). Fortunately, a specimen in the
Princeton collection is only slightly worn
(Fig. 24B). Lavocat (1976: 36) stated that
the pattern was clearly the same as that
of the molars, which is only partially cor-
rect. The two central lophs (protoloph
and metaloph) are like those of the mo-
lars. The anteroloph seems to have had
a large central anterocone, and there is a
posteromesially directed hook from the
buccal tip of the anteroloph, being the
first stage in the development of an ac-
cessory crest. From the center of the pos-
teroloph there arises a strong, forwardly
directed crest, reaching as far anteriorly
as does the hypocone. A similar forward
extension of the buccal end of the pos-
teroloph lies along the buccal margin of
the tooth. If the anterior ends of these
two crests should join, it would result in
the formation of a neoloph; we suspect
that this is exactly what happened in ca-
viomorph evolution. Nothing remotely
similar to all this has been reported in
any tooth of any thryonomyoid. Replace-
ment of dm^ did not occur until after M^
came into full usage.

There is a small root in front of dm'*
(Fig. 24B), either that of dm^ or of P^ The
fact that the tooth was associated with
dm"* and was lost before the eruption of
P^ in all known specimens of Branisa-
mys suggests that it was dm^. However,
the tooth produced a wear facet on the
dorsal slope of the anteroloph of dm^, lay
dorsal to the overhang of the anteroloph,
and could not have reached the level of
the occlusal surface of the upper cheek
teeth. In this position, it is difficult to see

444

Bulletin Muacuiti of Comparative Zoology, Vol. 149, No. 7

MT EN

k-AND-^MT EN

PLD

PRD HYD

Figure 26. Tooth of Villarroelomys bolivianus Hartenberger compared with dm4 of Cephalomys arcidens, x5. A. GB
no. 014 redrawn from Hartenberger, 1975: Fig. 2a, and restored according to our interpretation of its probable structure.
B. Rdm4 of Ceptialomys arcidens, ACM no. 3013, redrawn and reversed from Wood and Patterson, 1959: Fig. 20A.

Abbreviations: AND — area derived from anteroconid; EN — entoconid; HYD — hypoconid; MT-
lophid; PRD — protoconid.

-metaconid; PLD — postero-

how this tooth could have erupted except
after dm^, which suggests that it might
have been a P^ that was shed along with
dm^ (Wood, 1969; 1970b: 245-246).

One badly broken dm4 is present in the
Princeton collection (PU no. 21987, Fig.
24G). This preserves no more than the
posterior half of the pattern, which is the
most molarifomi part of the caviomorph
dnii. In view of this, little more can be
said of the specimen. Lavocat (1976: 39;
PL 2, Fig. 5) described a similarly broken
and more worn specimen in a ramus frag-
ment that also includes Mj.

Lavocat (1976: 40-41) and Hoffstetter
(1976: 10, n.) concluded that the type of
Villarroelomys bolivianus (Hartenber-
ger, 1975) was a dm4, not an Mg, and sus-
pected that it may have been a young in-
dividual oi Branisamys, in which Mj had
not yet erupted. Lavocat further stated
that Hartenberger was in essential agree-
ment with him on this point (Lavocat,
1976: 40). We had independently arrived
at the same conclusion. Certainly this
tooth has nothing to do with the hydro-
choerids. Because Lavocat felt that there
was some possible uncertainty as to the
reference to Branisamys, since the tooth
of Villarroelomys bolivianus might rep-
resent an otherwise unknown genus
closely related to Branisamys, he be-
lieved that the name Villarroelomys
should be retained for the time being.

Hartenberger's illustration of the type
oi Villarroelomys bolivianus (1975: Fig.

2A) shows a tooth broken at both ends,
with three crests reaching one margin,
four the other, and with an assemblage of
two or more cuspules near one end. The
side view that he gave indicates, we be-
lieve, that the roots are somewhat diver-
gent. This, together with the number and
distribution of the crests, is convincing
evidence that this tooth is Ldm4 and not
RMg, as Hartenberger originally be-
lieved.

We have restored this tooth, from Har-
tenberger's illustration (Fig. 26A; our fig-
ure differs slightly, in the hypothetical
portion, from the restoration given by
Lavocat, 1976: Fig. 2J). There is suffi-
cient resemblance to dm4 of Branisamys
(Fig. 24G) and to the corresponding tooth
of Cephalomys arcidens (Fig. 26B) so
that we feel considerable confidence in
the general accuracy of the restoration.
There are differences between GB no.
014 and Ldm4 of Cephalomys. The Bo-
livian tooth is considerably larger and not
quite so completely lophodont. Anterior
to the metalophid, the anteroconid area
of Cephalomys consists of an anterolin-
gual cusp behind which there is a curved
crest fomied of an anterior buccal and a
posterior lingual crest, whereas the an-
terior cusp of GB no. 014 is more isolated
and is at least incipiently divided. On the
other hand, the entoconid crest is shifted
forward of the hypoconid-posterolophid
crest as in dm4 of Cephalomys (Fig. 26;
Wood and Patterson, 1959: 340), and the

Oligocexe Rodents of Bolivia • Patterson and Wood 445

central crest, which we interpreted as the
metalophid in Cephalomys, is also as in
that genus.

On the basis of its size, general ap-
pearance, and agreement with what is
preserved in undoubted dm4 of B. luri-
bayensis, we feel that GB no. 014 is dm4
of this species. It certainly cannot be re-
ferred to Sallamys (Fig. 5D), Incamys
(Fig. 19A,B) or Cephalomys bolivianus.

The upper incisor is a robust tooth that
in its course describes a semicircle ex-
tending forward from its point of origin
on the medial wall of the masseteric fossa
above P. The posterior part of the tooth
bulges into the infraorbital foramen, as in
Dinomys (AMNH Mam. no. 46551). The
anteroposterior diameter is greater than
the transverse, the width/length ratio
being between .67 and .75. Near the wear
surface, the anterior face is flat over most
of the width, becoming a little rounded
laterally (Fig. 23A,B). Farther back, the
anterior face is more rounded, and the
pulp cavity becomes very large (Fig.
25B). The tooth is much more like that of
Dinomys than like those of dasyproctids.
The enamel extends for short, nearly
equal distances onto the lateral and me-
dial faces, again as in Dinomys. Near the
wear surface (Fig. 23B), the pulp cavity
is small, and shaped like a distorted fig-
ure 7 rather than like a distorted T as in
Dinomys. The lower incisor was gener-
ally similar (Figs. 23E; 25D,E,G; Lavo-
cat, 1976: 36; Fig. 3E,H). The enlarge-
ment of the pulp cavity beneath the rear
of Ml of PU no. 21987 (Fig. 25G) is re-
lated to the age of the individual (dm4
still present) rather than to the position
in the jaw, as is shown by adult speci-
mens (PU nos. 22172 and 22173, Fig.
25H) that are broken farther to the rear.

Relationships. Hoffstetter and Lavocat
compared Branisamys with Neoreomys,
from which they distinguished it (1970:
172) "notammement par la brachyodon-
tie, et par la reduction de M^, ce dernier
caractere excluant des liens directs."
They were entirely correct in this, and

they were fully justified in comparing
Branisamys with Neoreomys, which, on
the basis of the cheek teeth, it closely re-
sembles. The two forms are definitely re-
lated, although not, we believe, at the
family level. Lavocat (1976) made com-
parisons of Branisamys with other De-
seadan rodents, with Oligocene and
Miocene African rodents, but with no
post-Deseadan South American genera
except Neoreomys, and in this instance
the comparisons were limited to the
cheek teeth. He concluded (1976: 73)
that Branisamys and Incamys should be
placed in the same family and that Bran-
isamys was clearly related to Neoreomys.

The discussion of the cranial charac-
teristics given above and summarized in
Table 7 indicates rather different rela-
tionships for Branisamys from those sug-
gested by Hoffstetter and Lavocat (1970)
and Lavocat (1976). Almost all available
cranial characters point toward affinities
with the Dinomyidae, and suggest rela-
tionships between Neoreomys and the
Dasyproctidae as proposed by Wood and
Patterson (1959: 324-328). It is interest-
ing to note that, in early members of
these two families at least, the cranial
characters give a better insight into fa-
milial relationships than do those of the
cheek teeth. Table 7 demonstrates that,
insofar as the characters available for
Branisamys are concerned, ?Simplimiis
is somewhat aberrant with respect to
known dinomyids, sharing more features
with dasyproctids than with Dinomys.
We give more weight to those in common
with the latter, however, and follow
Fields in his familial assignment. The
family having been in existence since at
least the beginning of the Oligocene, an
early radiation within it may have taken
place, giving rise to the Friasian Simpli-
mus, a side line independent of the later
dinomyid radiation.

Derivation of the cheek teeth of later
dinomyids from those of Branisamys, or
some similar fonn, would not present se-
rious difficulties. The requirements for

446 BuUctiii Museum of Comparative Zoology, Vol. 149, No. 7

a. o

:s:

C/5

tf

3

so

^ r5

Oj

o

O

s

Oj 02

>.

_c

■M

a

•n

c/j

-n

5P ^

<o

S 0)

-o

o

^-T3

O

c

c

. c

C

3

o

C^

>

-a
c

3
O

o
c

"^ -2

O c

t/3

a! iS

o;

o c

Oj*

>,

X, S

c

(A ^

^5

-a

^

0)

. Oj

-o

bc-a

f—

£ =

3

O 0)

3

o o

ll

o =:'

1)

t;

o

t;

o

01

t;

o

0)

c

<

"5 "S

^ . .^
m a;

— E

E-o
ctf n

C TO

o
D2

a;

Of}

t

4-'

3

t«

M

(U

o

O

0)

OJ

>>

-C

4->

X

>.

o

OJ TJ

S a;

a! VI

C cS t"

0)

-0

-a

I I

O

3;

c
o

2

■4^

-a

o

ll

U

a

a

C

o

4^

i^

0;

4-*

'o

a

"s '*^

crt

-t-*

a t-

ol

en

O 3

s 5?-S

pl3 5^ ^

3 "^

o -^

o

c

<L1

o
c

t.

•k^

^ f

Oj

c

_c

Vi

-a

5

M

£

c

3

a

2

o
t4

V3

03

O

3 _C ^

c

C/1

0)

en

V3

>^

03

>,

OJ

c

c<3

S^ o

u

o
03

.Bf o
ll y

;-i

OJ
C

c

a;

3

-Q

>.

C

ca

I-l

^-i

4-*

u

c

-C '^

t«

-0

a>

.5Po

v:

0)

o.

o

*; ^

OJ

01 u bc

01 OJ b

-C > -2

ll

a

a

0)

bc

13 > -2

c/3

0) ■

a

C/5

^ -

M

v:

O

13

C/3

m

V

•^

OJ

>>

>>

a

>>

a

0)

o»

a

OJ
0)

-a

>^ a; o

o
c

0)

a;

o 1,

r- O

S2u

r- O

O 2li

^-^ -4-1

r- O

>

03

dj 4^ -

> = :s

3 c

_Q ,1-" -*-'

01
03

O 2h

TO

-a

1

2^

13

is

t/3 0

bc

ll

Oi

V.

13

3

u

-2

1

J

0

4-»

r-

allary
ind in

>

0

l->

c

3

■4-'

3 S
in 3

c
0

'C

y

J2

c
0

C/5

-^4

a

■S

E

OS

g

W4

"0

03

y

t-^

c

t*4

0

OJ

2"^

^

ij

OS

0

0

a;

c

-2 c«

c
0

E^

0

-5

_c

y
T2

4-»

0

•S 2

■•£

^

is 0

4->

X

•^4

c

1/5

^

0)

cS 0

E ^

0
a

"a

s

E
2

0
0

4->

u

_o

'1^

IH

bC

c

-G
bC
3

2
y

en

03
i-i

2 2

C OJ

2 >

13
4-1

0

2-

a;

OJ

y
4-4

S

03

C

c 2

0) J2

a ■"

>
0
c

2

C/3

0
>

0
V)

C

l-H

Uh

0

>

Cu

U

bn

0

0

Oligocene Rodents of Bolivia • Patterson and Wood 447

this would be progressive elongation of have been contemporaneous either with
the teeth, already foreshadowed in un- the tuff or with the cave. Anthony decid-
wom teeth of Branisamys, further sepa- ed in favor of the latter possibility, and
ration of the lophs and lophids, together considered the species to be Pleistocene
with ever-increasing height of crown, in age. Fields (1957: 324, n. 5) discounted
culminating in the hypselodonty of the this. Hoffstetter (1952: 124-125) has re-
late Tertiary and Quaternary fonns. The ported that on geological grounds the tuff
anteroloph and posteroloph are already, may be of late Tertiary age. It would
in Branisamys, well on the way to inde- therefore appear probable that the spec-
pendent status, and the changes required imen had weathered out of the roof or
to isolate the protoloph as the second of walls of the cave, and that the deposit
the dinomyid upper cheek-tooth crests from which it came is approximately con-
would not seem to have been very pro- temporaneous with part of the Honda
found. Due to the absence of dinomyids Fomiation of Colombia,
from the Colhuehuapian and Santacruz- In Drytomomys, there was a highly
ian of Patagonia, the details of the pro- progressive development of subparallel
cess remain obscure. When the dinomyid diagonal crests in the upper molars, fully
record begins again, we see in Drytomo- separate initially but rapidly uniting with
mys, from the La Venta local fauna of wear along the buccal sides of the teeth.
Colombia, of Friasian age, an early mem- The crests of the premolar seem to have
ber of the spectacular later Tertiary di- been even more independent than those
nomyid radiation. of the molars. This separation of the

Fields (1957: 323) referred his material lophs is associated with a very consider-

of this fomi, correctly we believe, to An- able degree of hypsodonty. The postero-

thony's Drytomomys aequatorialis, but lophid of the lower molars of Branisa;?j(/.s

he synonymized Drytomomys with is already well on the way toward sepa-

Olenopsis. As we have previously indi- ration, with the three anterior lophs

cated (Wood and Patterson, 1959: 361, n. joined buccally, and the two anterior

21), the syntype material of Ameghino's united lingually as well. Drytomomys

type species, the Santacruzian O. unci- aequatorialis shows the same features

nus (1889: 145), is partly lost, partly ref- (Fields, 1957: Fig. 16a), but again with

erable to Neoreomys, and partly clearly somewhat longer independence of crests

different from Fields' material. Until and much greater hypsodonty than in

such time as additional specimens are Branisam(/.s. The various later dinomyids

found in the Santa Cruz, clearly deter- could have been derived, with relatively

minable as O. uncinus and clearly diag- little difficulty, from something very sim-

nostic, we believe that the reference of ilar to Branisamys, by way of something

other species to this genus will serve no like D. aequatorialis.

useful purpose. We therefore retain An- For these reasons, we have no hesita-

thony's generic name, Drytomomys, for tion in referring Branisamys to the Di-

aequatorialis. The Pliocene Olenopsis nomyidae (since this was written, Har-

typicus Scalabrini (Ameghino, 1889: 901) tenberger, 1975, has included the genus,

is not congeneric either with O. uncinus with a query, in this family),

or with D. aequatorialis, and was made The less pronounced, but still very real

the type of Paranamys by Kraglievich resemblances of Bran /sg/jhjs to the Dasy-

(1934: 73). proctidae we interpret as indicating that

The type of D. agqf«afona?i.s was found the two families were closely related in

in Ecuador, apparently on the floor of a Deseadan time. We fomierly (1959: Fig.

cave in tuffaceous rock (Anthony, 1922). 34) showed the Superfamily Chinchil-

On the basis of these scanty data it could loidea as extending back to the Deseadan

448 Bulletin Mtisctini of CoDiixinitivc Zoolomj, ^(>l- 149, No. 7

with tAvo fomilies, the Chinchillidae and
Dasyproctidae, and the Heptaxodonti-
dae, Cuniculidae and Dinomyidae as
arising from the Dasyproctidae at some
unknown (but perhaps mid-Tertiary)
time. This must now be modified since
the Dinomyidae, in the fonn of Brani-
saniys, were already present in the De-
seadan, although not far removed from a
common ancestry with the Dasyprocti-
dae. We now agree with a growing con-
sensus that dinomyids, dasyproctids and
cuniculids should be removed from the
Chinchilloidea and returned to the Ca-
vioidea. The broader affinities of all
groups are discussed below and the im-
plications for caviomorph higher taxono-
my of recent comparative anatomical and
karvological work are considered (pp.
509-512).

Kraglievich (1926, 1930a, 1931, 1932a)
proposed no less than four families, one
of them including four subfamilies, for
the reception of the genera involved in
the later Tertiary radiation of what we
consider to have been dinomyids. Simp-
son (1945: 95-96), who placed these later
Tertiary genera in the Heptaxodontidae,
reduced Kraglievich's swollen array to
three subfamilies. ?Simplimus is a mem-
ber of Kraglievich's Potamarchinae in
this scheme, and Drytomomys would fit
without strain into his Eumegamyinae
( = Dinomyinae), but Branisamys could
not, without knowledge of intennediate
stages, be fitted into any subfamily. In
view of the enlarged dinomyid vistas
now opened up, it would be the part of
wisdom to hold subfamilial divisions in
abeyance for the time being.

Caviomorpha inc. sed.
Luribayomys Lavocat 1976
Luribayomys masticator Lavocat 1976

Luribayomys masticator Lavocat, 1976: 64-68; Fig.
3D; PI. 5, Figs. 1-3.

This taxon is based on a rostrum with
the incisors broken off at the alveoli and
with the roots of both premolars and, in

part, of LM^ For all other Deseadan ro-
dents (except Palmiramys — see below, p.
449) we have some knowledge, often
good, of the cheek teeth, and the taxon-
omy is to a large extent based upon them.
Luribayomys stands outside this circle.
It is about the size of Sallamys, from
which it differs in incisor cross section
and in the possession of a much wider
rostrum. Its rostral structure would ap-
pear to separate it from the other forms —
Platypittaniys, Incamys, Cephalomys,
Branisamys — in which this part of the
anatomy is known. It is much too small
to belong to either Protosteiromys me-
dianus or P. asmodeophihis. The rooted
cheek teeth exclude it from Scotamys.
This still leaves a number of fomis with
which it cannot at present be compared:
Migraveramiis, Deseadomys, Xylechi-
mys, Litodontomys, Dasyproctidae aff.
Neoreomys, Asteromys and Chubuto-
mys, aside from the possibility of its
being a valid genus. It cannot even be
placed as to family. For the present it
must be left indetemiinate, as a clear case
of a nomen vaniim (Simpson, 1945: 27).
We can only join Lavocat in the hope
for a fortimate find that will include both
rostrum and cheek teeth and thus permit
characterization of this genus.

Palmiramys Kraglievich 1932b (by A.E.W.)
Palmiramys walttieri Kraglievich 1932b

Kraglievich (1932b: 314-316) de-
scribed Palmiramys waltheri as a new
genus and species, based on a lower jaw
from the Fray Bentos Fonnation of Uru-
guay. He referred it to the Acaremyinae,
then included in the Erethizontidae, but
gave no figures, and his description was
not very illuminating.

Mones and Ubilla (1978: 154) conclud-
ed that the Fray Bentos was Deseadan
because it contained Propachynicos and
Proborhyaena, both characteristic De-
seadan genera.

Mones and Castiglioni report the dis-
covery of a drawing (1979: 78, Fig. 1),

Oligocene Rodents of Bolivia • Patterson and Wood 449

presumably of the type specimen which One of them, revived in modern fonn

they state may be presumed to have been by Lavocat (1969) and by Hoffstetter and

lost. This drawing accompanied the Lavocat (1970), and subsequently devel-

manuscript of Kraglievich's paper. Mones oped by both authors severally (Hoffstet-

and Castiglioni conclude that Palmira- ter, 1971; 1972; 1975; Lavocat, 1971a;

mijs was probably a dasyproctid. 1973: 254-258; 1974a; 1974b; 1976: 74-

The measurements that Kraglievich 84; 1981), holds that the Suborder Hys-

(1932b: 315) gave for the type of Palmi- tricognathi arose in Africa from waif im-

ramijs waltheri are very close to those of migrant European paramyids, possibly

Migraveramus heatus (Table 1; Fig. 3). related to the European Theridomyidae,

His original identification of Pa/miramjys that had arrived in Africa in late Paleo-

as an acaremyine would also be in accord cene or earliest Eocene time (well before

with its being close to Migraveramus. It the evolutionary origin of the Therido-

would be difficult to visualize, however, myidae). These ancestors then gave rise

how wear of Deseadan octodontid teeth to the Thryonomyoidea, from which

could produce a pattern such as that fig- evolved, in Africa, the Bathyergoidea

ured by Mones and Castiglioni (1979: and, possibly, the Hystricidae (alterna-

Fig. 1). tively, these last arose from thryonomy-

A comparison of this figure with all oids supposed to have reached southern

available illustrations of Deseadan ro- Asia from Africa, as waif immigrants,

dents indicates that the only closely sim- around the middle Eocene). The New

ilar fonn is Incamijs (cf. Figs. 16, 19 and World Infraorder Caviomorpha descend-

Mones and Castiglioni, 1979: Fig. 1). ed from African thryonomyoids that had

Mones and Castiglioni made no specific been rafted, perhaps twice (if the Erethi-

comparisons of Palmiramijs with Inca- zontoidea are as distinct as Bugge has

mijs or any other rodents. Kraglievich's suggested), in middle Eocene time,

original description could fit Incamys, across a then-narrower South Atlantic

but it could equally well refer to some Ocean. The proponents of this view tend

other genus. His measurements (1932b: either to dismiss the awkward fact that

315) are those of an animal half to two- subhystricognathous and hystricognath-

thirds the size of Incamys bolivianus. ous Eocene rodents are known only from

Palmiramys waltheri appears to have North (including Middle) America as an

all the characteristics of a nomen vanum. irrelevancy, or to neglect these rodents

entirely. Hussain et al. have recently

DISCUSSION proposed an extreme version of this hy-

^ ,,. .• -ri_ • r XI- pothesis (see below, pp. 504, 515).
Conflicting Theories of the ^.i , , .r . i 4. u j

r^ ■ ■ t fu r- ■ u Ihe second hvpothesis was sketched

Origin of the Caviomorpha 1 in-m j r ^u i 1 ju

^ '^ by us in 1970 and further developed by

Description and taxonomic assessment Wood (1972, 1973, 1974b, 1975a, 1977b,
of the Bolivian Oligocene rodents having 1980b, 1981); we here carry it further. It
been completed, we now turn to the holds that the Hystricognathi arose in the
broader questions that involve them, late Paleocene of North America as fran-
their ancestors and descendants, and imorph reithroparamyids. A group, or
their relatives in other parts of the world, groups, of these crossed to Asia, via Be-
There are at present two, in large part ringia, early in Eocene time, and there
conflicting, principal hypotheses con- gave rise, directly or indirectly, to the
cerning the origin, interrelationships and Thryonomyoidea and Hystricoidea.
dispersal of hystricognathous rodents. As Another franimorph group, the cylindro-
background for what follows, these op- donts, reached Asia in the late Eocene,
posing concepts must again be compared, giving rise to the Tsaganomyidae in Asia

450

Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

and, later, to the Bathyergidae in Africa.
The thryononiyoids reached Africa, as
waif immigrants, from southwestern Asia
during tlie kite Eocene; the bathyergoids
and liystricoids, as dry land immigrants,
during pre-Burdigalian and late Miocene
times, respectively. The ancestors of the
Caviomorpha reached South America, as
waif immigrants, from southern North
America (i.e., Middle America) in middle
Eocene time. McKenna (1981: 63) has
proposed a possible variant in the meth-
od by which the caviomorphs reached
South America from the north. The pro-
ponents of this view of things consider
the North American Eocene subhystri-
cognathous and hystricognathous rodents
as very relevant indeed.

Earlier, at a time when nothing was
known of North American Eocene hystri-
cognaths and next to nothing of the pa-
leontology of the Old World ones, we
(Wood and Patterson, 1959: 419) conclud-
ed "that the Hystricomorpha and Ca-
viomorpha have derived those charac-
ters, which they hold in common,
independently and subsequent to their
geographic separation." The presumed
ancestors would have been relatively un-
differentiated paramyids. We added:
"This is extreme parallelism," which is
true, but not much, if any, more than has
been reported on numerous occasions
among the rodents (Engesser, 1979: 41-
42; Lavocat, 1951; 1954: 127; 1955: 634;
Vianey-Liaud, 1972; Wood, 1935: 249;
1936b; 1947: 156-161; 1950: 90-97;
1965: 122-124; 1974b: 38-41; 1975a: 76-
77; 1977a: 133-136; 1981: 82; in press a;
Wood and Patterson, 1970: 631). We do
not know how many of the morphologi-
cal characteristics of modern hystrico-
gnaths (Wood, 1975a: Fig. 1) had been
acquired by the Eocene franimorphs;
certainly at least hystricognathy and hys-
tricomorphy had been. And certainly
all features that characterize modern
hystricognaths did not appear simultane-
ously, fully developed (Wood, 1980c:
270), but must have evolved gradually

and, almost certainly, independently and
at different times. Dawson (1977: 197, n.)
commented that the hystricognathy of the
Reithroparamyidae was "'incipient' only
a posteriori." However, as pointed out
elsewhere (Wood, 1980c: 270), unless
hystricognathy arose as a single genetic
mutation of one sort or another, it must
have been a gradual development, its ini-
tial stages barely detectable. Since hys-
tricognathy was the first (and only uni-
versal) hystricognath character to evolve,
the initial changes were presumably the
result of selective pressure for lengthen-
ing the M. pterygoideus internus, as an
adaptation to increase gnawing ability.
This was followed relatively soon (but
with a definite time lag) in many groups
by the evolution of hystricomorphy, re-
sulting from a similar lengthening of M.
masseter medialis, presumably as an
analogous adaptation. This same change
also occurred independently in several
sciurognath lines.

In this theory, the resemblances be-
tween the Old and New World hystrico-
gnaths (and, very probably, among the
Thryonomyoidea, Bathyergoidea and
Hystricidae) other than hystricognathy
(and possibly the correlated deepening of
the pterygoid fossa) and, perhaps, hystri-
comorphy, have all resulted from parallel
evolution during the 50-odd million
years since the ancestors of the Old and
New World hystricognaths separated in
the early Eocene.

The two hypotheses are not mutually
exclusive in their entirety. They are in
agreement that, so far as our present
knowledge goes, all living hystricognath-
ous rodents form a natural suborder, the
Hystricognathi; that Old and New World
hystricognaths are distinct, infraordinal-
ly; and that the thryononiyoids and ca-
viomorphs reached Africa and South
America as waif immigrants, the latter
around the middle of the Eocene. Under
either view, hystricognath evolution in
the Old and New Worlds has gone on in-
dependently for over 40 million years.

Oligocene Rodents of Bolivia • Patterson and Wood 451

Beyond the points just indicated, the
common ground between the two view-
points ends and disagreements begin.
Discussion of the evidence bearing on
these disagreements, as presented be-
low, must range widely, from paleogeog-
raphy to parasitic nematodes, from com-
parisons of faunas to the excruciating
minutiae of molar tooth morphology. As
a beginning, we will now attempt to set
the stage.

First, however, perhaps a third hypoth-
esis, that of Croizat, should be mentioned
for the sake of completeness. As recently
stated by him, he feels it to be a fact that
the ancestors of the Hystricomorpha "ne-
gotiated [the South Atlantic] in the Cre-
taceous or even earlier" and, further, "At
some unspecified time earlier than the
beginning of the Tertiary, certain very
primitive 'proto-rodents' originated in a
then unitarian continent of Gondwana.
When this continent split into a later
South America and Africa, respectively,
these 'proto-rodents,' already on the way
to evolving as taxonomically definable
'hystricomorph rodents,' turned 'Ameri-
can' and 'African' by right of the new con-
tinental geography . . ." (1979: 251). As
pointed out by Wood (1980a), the docu-
mentary evidence for this hypothesis is
nonexistent.

And, finally, as some authors have
pointed out, since there is evidence sup-
porting both the African and the Middle
American origin of the Caviomorpha,
there is justification for adopting a non-
committal position (Simpson, 1980: 140-
141; Hartenberger, 1980: 296).

Eocene Geography

According to plate tectonic reconstruc-
tions, the world in Eocene time was well
on the way toward conditions now exist-
ing. The present configuration of the con-
tinents was at least foreshadowed, al-
though the relations of land and sea and
the spacing of the continents relative to
each other and their interconnections, or

lack of them, were in many cases signif-
icantly different (Fig. 27). South America
and Africa (cum Arabia), which had be-
gun to separate at about 125 m.y.b.p.,
were, by mid-Eocene time, separated by
a South Atlantic Ocean that was some
1300 km wide at its narrowest point. Af-
rica was isolated from Europe and Asia,
and South America from North America.
Middle America, the tropical portion of
the continent, seems to have ended in
what is now Honduras and Nicaragua.
North America had a land connection
with Europe at the beginning of the
Eocene, which was quickly broken
(Wood, 1977b: 100), and another, also of
no great duration, with Asia via Beringia
at the same time; the latter was re-estab-
lished for a while in the latest Eocene or
the early Oligocene. Europe and Asia
were separated by a seaway that lasted
throughout the Eocene. This was the
"Uralian trough" of Patterson (1954:
192), the Turgai Strait of Thenius (1959:
107; McKenna, 1975b: Fig. 1). One result
of its elimination was the "grand cou-
pure, " the long-recognized major change
in Europe between the mammalian fau-
nas of the early and middle Oligocene
(Stehlin, 1909: 502-508). India and
southeastern Asia were in their present
positions by the middle Eocene. The
whereabouts of Australia, New Guinea,
and New Zealand during the Eocene are
less certain — but in any event irrelevant
to the present study; rodents did not
reach Australia until the early Pliocene
(Archer and Bartholomai, 1978: 15;
Wood, in press b).

The evidence provided by Eocene
mammalian faunas, so far as these are
known, is in accord with such continental
relationships. South America and Africa
(to judge from the absence of similarities
between the Eocene faunas of South
America and the Fayum fauna of Africa)
had long been isolated. The faunas of
northern North America and Europe
were closer in early Eocene time than
they were ever to be again. The faunas of

452 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Figure 27. The world in Eocene time showing the distribution of the hystricognath groups and the Paramyidae, and
the evolutionary relationships of hystricognath groups according to Lavocat (dashed lines) and according to us (solid
lines). B — Bathyergoidea; C — Caviomorpha; F — Franimorpha; H — Hystricidae; P— Paramyidae; T— Thryonomyoidea; w —
waif transport. Intercontinental migrations not marked w are assumed to have been via dry land. In our view bathyergids
and hystricids reached Africa in early and late Miocene time, respectively.

Europe and Asia were sufficiently dis-
tinct to show that the Turgai Strait was a
real barrier to interchange, whereas there
were many similarities between the fau-
nas of North America and northern Asia
(Dawson, 1977).

Discoveries in the early Eocene ot
China of such typically Wasatchian mam-
mals as Homogalax and Heptodon, un-
known in Europe (Chow and Li, 1965),
and of a paramyid in the size range of
small species oi Microparamys (Li et ah,
1979) strongly indicate that a connection
between Asia and North America was in
existence at that time. Confirmatory evi-
dence is provided by recent finds of
mammals referable to or allied with
North American genera in the early
Eocene Naran Bulak Fonnation, Mon-
golian People's Republic (Dashzeveg
and McKenna, 1977, and references there
cited). These authors point out that the
presence in the Naran Bulak oi Altonius,
an omomyid of North American anapto-

mori^hine affinities, indicates a climate in
the Bering Strait region sufficiently clem-
ent to support very small primates. The
Eocene, especially its earlier stages, does
seem to have been a rather wami epoch.

As an aside (by B.P.), categorical state-
ments that marsupials were absent in
Asia may prove to have been premature.
Didelphids reached Europe, via the
North Atlantic early Eocene connection,
and they could have reached Asia, via
Beringia, at about the same time. Fur-
ther, it is also conceivable that in late
Cretaceous time they could have inhab-
ited the river valleys of eastern Asia
where they would have encountered en-
vironments similar to those of the west-
ern interior of North America, surround-
ings more congenial to them than the
drier, more upland conditions that seem
to have prevailed in Central Asia at that
time.

Until very recently, all Eocene terres-
trial mammals recorded from India and

Oligocene Rodents of Bolivia • Patterson and Wood 453

Pakistan were of Holarctic affinities, that, in the Eocene, the Pakistan-Indian
There was no trace of any pecuHar en- area was part of the African-Arabian
demic mammals such as one might ex- block, and carried African hystricids to
pect to have evolved on the subcontinent Asia. This idea perhaps was based on the
during its long isolation and postulated then unstudied Eocene rodents of Paki-
northward passage. The large mammals stan (Hussain et al., 1978). However, as
from the middle Eocene of Pakistan, re- indicated below (p. 504), these rodents
cently described by West (1980) include were apparently all sciurognaths and had
a variety of cetaceans, as well as a con- nothing to do with the ancestry of either
siderable number of land mammals, the Thryonomyoidea or the Hystricidae.
These are all, with one exception, refer-
able to Holarctic families, although gen- rM^t,:L,..x- „ ^z Uw^+.-;^^««o+u«.,o
n ^ 1 . ' , ^ ^ Distribution of Hystricognatnous
eraily to endemic genera. Ihe one ex- o^^^^t^ ;„ t^^ c^^^„^
/ . ^, A J.1 1 Rodents in the Eocene
ception IS the genus Antnracobune,

identified by West as a primitive moeri- The Order Rodentia was certainly of

there, which he considered (1980: 530- northern origin. Members of it first ap-

531) might indicate a south Asian origin pear in the latest Paleocene of western

of this group. Sahni and Mishra (1975) North America. These forms, which had

have also reported a probable moerithere already acquired all the basic ordinal

from the middle Eocene of India. No oth- characters, are referable to two very

er possible descendants of the presumed closely related families, the sciurognath-

south Asian Mesozoic or Paleocene mam- ous Paramyidae and the incipiently hys-

mals have been reported. tricognathous Reithroparamyidae, the

Whatever its position may have been latter, in our opinion, the ancestral stock

in the Pennian, the Indian subcontinent of all later members of the Hystrico-

could not possibly have docked at its gnathi.

present position bearing a mammalian Where, in the northern hemisphere,

cargo of the kind revealed by its known the order arose is still uncertain. Europe

Eocene fossils. The striking lack of ende- would appear to have been a most un-

mism in the Indian-Pakistani biota as a likely place. No reithroparamyids are

whole is, in fact, rather difficult to rec- known from there, although Michaux

oncile with plate tectonic theor\'. It has (1964) referred two species to Reithro-

been seriously suggested that the (pre- paramys; he later (1968: 155) made one

sumed) endemic biota was essentially of them the type species oi Meldimys, a

wiped out as a result of climatic stresses member of the sciurognathous Ailuravi-

encountered during the northward pas- nae, and stated (1968: 173-174) that the

sage. This strikes us as fatuous: life is other was indeterminate as to genus;

more adaptable than that, and continents both were reported only from isolated

do not move that rapidly. teeth, so that there was no infomiation as

The presence of what may be primitive to their possible hystricognathy. The

moeritheres in the Eocene of Pakistan paramyids that arrived in Europe at the

and India is the only suggestion of ter- beginning of the Eocene gave rise to a

restrial mammals that show any affinity rodent fauna that evolved in isolation

to those that evolved on Africa while that throughout the epoch, with a consider-

continent was isolated, and it is possible able variety of strictly endemic forms,

that they were the ancestors of the late This leaves us with North America,

Eocene and Oligocene African moeri- where the early Eocene record is fairly

theres. The absence of other mammals of good, or Asia, where it is not. The only

African aspect would seem to refute the early Eocene Asian rodent, so far as we

idea expressed by Hoffstetter (1975: 524) are aware, is that described as Micropar-

454 Bulletin Museum of Comparative Zoolo^ij, Vol. 149, No. 7

amijs Un^chaensis from Hunan (Li et were a number of late Paleocene groups

a/., 1979: 76), which is, we believe, a new independently evolving gnawing inci-

genus and not Microparamys. Dr. Li in- sors, at least sometimes in different geo-

formed one of us (A.E.W., April 1981) graphic areas. With the evolution of the

that he had reached the same conclusion rodents, these competing lines were

and that a new name and a re-assessment quickly replaced except, apparently, in

of the relationships of this form are in restricted environments. One line, the

preparation. In Figure 27 we tentatively lagomorphs, later acquired the capability

suggest a North American origin for the of competing successfully with the ro-

rodents, but stress that an Asian one dents.

would have just about as much evidence Reports of teeth identified as Frani-
in its favor (Hartenberger, 1980: 300). mys? by Sahni and Srivastava (1976:
However, we believe that such an origin 923-924) from the middle Eocene of In-
would not affect our thesis appreciably. dia, and of phiomyids or of "une famille
The group that, in our opinion, is the protophiomyidee" by Lavocat (1973:
most probable ancestral stock for the ro- 165-166) from the middle Eocene of Pa-
dents, the primitive primates (Wood, kistan, are readily interpretable in terms
1962a: 253-254), is better known in of our view that hystricognaths were
North America than in Asia. The Asiatic present in the Eocene of Asia — presum-
Eurymylidae, sometimes cited as having ably these teeth had been housed in hys-
rodent affinities, were close to the ances- tricognathous or subhystricognathous
try of the lagomorphs (Wood, 1942; 1957: jaws. However, it is also possible that all
417-418), but, so far as they are known at of these teeth are referable to the Family
the present time, do not seem to A.E.W. Chapattimyidae of Hussain et al. (1978),
to have had any special relationship to which were members of the sciurognath-
the rodents. Li (1977: 116-117) described ous ctenodactyloids. Since the above was
Heomys orientalis, the most rodent-like written, Hartenberger (1980: 288) has
eurymylid so far reported. Gingerich and adopted this last point of view. Sahni
Gunnell (1979: 151) cited this form as (1980: 26) accepted the ctenodactyloid
"appearing closely related to rodent relationships of these fonns, but consid-
origins." However, a cast of the type ered Chapattimys to be a synonym of
specimen, sent to one of us (A.E.W.) by Metkamys Sahni and Srivastava (1976).
Dr. Li, shows that the molars, particularly There obviously was extensive rodent
M*~^, have a well-developed hypocone, evolution going on in Asia during the
extending linguad of the protocone, Eocene, even though we are still only
much more prominent and quite differ- barely aware of much of it. The large sc-
ent in appearance from anything seen in ries of rodents described by Shevyreva
primitive rodents. Whereas the hypocone (1976) is a beginning of a demonstration
of the rodents arises from the area where of this activity. The rodents that she iden-
the posterior cingulum joins the proto- tified as sciuravids (1968, 1971a, 1971b,
cone, that of Heomys seems to have 1972a, 1972b, 1972c, 1976), following
grown from a posterior cingulum not con- Dawson's (1964: 11) tentative reference
nected with the protocone. Furthermore, of Advenimus to the Sciuravidae rather
the paracone is isolated from the proto- than to the Ctenodact\'lidae, but all since
cone, lying behind the protoloph which transferred to the Ctenodactylidae by
is fomied by the protocone, protoconule Wood (1977a), illustrate once more the
and anterior cingulum, a condition un- development of hystricomorphy among
known among the rodents, but suggestive sciurognaths as an occurrence indepen-
of the situation in early lagomorphs dent of what happened among the hys-
(Wood, 1940: Fig. 114). There apparently tricognaths. Quite a variety of paramyids

Oligocene Rodents of Bolivia • Patterson and Wood 455

has been reported from the Asian Eocene ica, also as waif immigrants, in the mid-

(Chow et al., 1973: 179; Li, 1975; Li et die Eocene (Wood, in press b). They first

al., 1979: 79), although Hartenberger appear in the record in the early Oligo-

(1980: 288) considered Yuomys and Mi- cene Deseadan, by which time they were

croparamys lingchaensis to be ctenodac- numerous and diversified. Had it not

tylids. Even if Hartenberger is correct, been for these successful landfalls on the

and both of these fonns are ctenodactyl- South American and African continents,

ids, they do not seem to me (A.E.W.) to the Hystricognathi would probably now

be anywhere near as distinct from para- be limited to the bathyergids and hystric-

myids as he indicated (1980: Fig. 2). ids, there would be no lively, widely

Cylindrodonts — Ardijnomys and Pseu- ranging hystricognath controversy, and

docylindrodon — had reached Asia by the life on the rodent front would be duller,

late Eocene and middle Oligocene, re- Lavocat (1976: 80) believed that Eocene

spectively (Vinogradov and Gambarian, ocean currents would have facilitated

1952: 14-15; Kowalski, 1974: 155-156; transport of rodents on natural rafts from

Shevyreva, 1976: 41-46, Fig. 16). As Africa to South America across an ocean

pointed out elsewhere (Wood, 1980b: 6; at least 1500 km wide. He stated (1976:

1981: 85-87; in press a: Fig. 1), the 61) that "diverses observations de navi-

cylindrodonts were incipiently hystrico- gateurs ne permettent pas de considerer

gnathous (subhystricognathous). The pres- comme infranchissable a des radeaux na-

ence of the specialized, fully hystrico- turels" such a distance. Unfortunately, he

gnathous Tsaganomys and its relatives in gave no documentation for such "diver-

the middle Oligocene of Central Asia ses observations" (Wood, 1977b: 104, n.).

(Shevyreva, 1976: Fig. 16), which we More recently, he has gone even farther,

recognize as a family of the Bathyer- stating "mariners are said to have found

goidea (see below, pp. 515-516), and living rodents on trees rafted as far as

which are believed by at least one of us 2,000 km from the shores (Chaline, pers.

(A.E.W.) to have been derivatives of the comm.)" (Lavocat, 1978: 86). This is still

subhystricognathous cylindrodonts, is an anecdotal evidence, and not documenta-

indication that much more was going on tion. He and Hoffstetter (1975: 526) also

in this area. believed that currents sweeping west-

Thryonomyoids reached Africa as waif ward through the gap between North and

immigrants late in the Eocene (Wood, in South America would have been strong

press b), and, we believe, unquestion- enough to have prevented rafting from

ably from southwestern Asia. They first the fomier to the latter. We know of no

appear in the record in the early Oligo- evidence as to the velocity of such sup-

cene Jebel el Qatrani Formation, where posed currents (see also Simpson, 1978:

they are rare faunal elements (Table 11) 326). There should, logically, be similar

that had just begun their radiation. Sav- currents now sweeping westward be-

age (1969: 69; 1971: 220) reported ro- tween Australia and New Guinea, which

dents as being present in the late Eocene do not seem to have prevented the east-

of Libya. Later he concluded (Wood, ward spread of Asiatic mammals, espe-

1974b: 33) that these deposits were early cially the Muridae, through the East In-

Oligocene, contemporaneous with the dies to New Guinea and thence to

lower level of the Jebel el Qatrani For- Australia.

mation of Egypt. There is no certain evi- Moreover, Lavocat's beliefs are based

dence that any non-thryonomyoid ro- on the assumption that all was quiet

dents were present in Africa before early around the Caribbean at the time, which

Miocene time (see below, p. 467). was far from the case. Reporting on the

The caviomorphs reached South Amer- geology of northwestern Colombia, Haf-

456 Bulletin Muscudi of Comparative Zoology, Vol. 149, No. 7

fer, whose opinions cam' weight due to
his N'ears of work there as a petroleum
geologist, found (1970: 615-618, Fig. 6)
clear evidence of two Tertiary episodes
of orogenic uplift, each followed by sub-
sidence, prior to the late Pliocene ele-
vation of the Isthmus of Panama. It
should be noted that Haffer's work was
carried out in complete independence of
any consideration of the South American
Tertiary mammalian record, to which he
did not refer. The first orogenic uplift oc-
curred toward the end of the middle
Eocene, the second toward the end of the
middle Miocene. We regard it as no mere
coincidence that these uplifts coincide
with the presumed times of arrival of,
first, rodents and primates in South
America and, second, of ground sloths in
North America and an extinct group of
procyonids in South America.

To digress for a moment, infatuation
with Africa as a source for South Ameri-
can Tertiary mammals has not yet em-
braced these raccoons, perhaps because
no member of the famih' has ever been
recorded from Africa. It has, however,
extended to the hesperomyine cricetids,
whose fossil record in South America
dates from the end of the Pliocene.
Hershkovitz (1972: 325) has suggested
that the ancestors of these rodents may
have been rafted across the Atlantic from
Africa in the early Teriary. Satisfactory
evidence of the presence of hespero-
myines in North America since the late
Miocene (Baskin, 1978: 131-134) dispos-
es of this possibility, which in any event
ran directly counter to the complete ab-
sence of possible hesperomyine ances-
tors in Africa or of any non-caviomorph
rodents in the South American record
prior to latest Tertiary time.

The orogenic episodes reported by
Haffer are by no means all that was hap-
pening in this area. During later Creta-
ceous time the Caribbean-East Pacific
Plate with its island arc advance guard
was moving eastward between Central
and South America (Malfait and Dinkel-

man, 1972). A partial connection through
such an island arc is what would be re-
(juired to account for the presence during
the early Tertiary of, e.g., notoungulates
in both Americas and of condylarths in
South America, which last, had they been
found in the north, would without hesi-
tation have been placed in Holarctic fam-
ilies. Malfait and Dinkelman further con-
cluded that, prior to the early Oligocene,
the Caribbean Plate had decoupled from
the East Pacific one, and that the Middle
America Trench had extended south-
southeastward and had become bordered
along its northern flank by a chain of vol-
canic islands. These, following a period
of quiescence, had combined by middle
Miocene time to fomi the western por-
tion of the present Isthmus of Panama.
That these plate events bore some rela-
tion to the episodes of orogenic uplift and
of subsidence reported by Haffer (whose
work was evidently unknown to Malfait
and Dinkelman) in extreme northwestern
South America is a reasonable assump-
tion. In other words, waif transport be-
tween the Americas would have been
possible at the times that the fossil record
indicates that it occurred. Ladd's views
(1976) on the relation between the Amer-
icas and the Caribbean Plate do not con-
tradict this possibility.

More recent work seems to give even
stronger support to the probability that
there was a Central American-Antillean
land area that could well have been in-
volved in the caviomorph migration to
South America. Perfit and Heezen (1978)
indicate that there was extensive uplift of
Middle America from southern Mexico to
include Nicaragua, as well as of the Cay-
man Ridge and the Nicaraguan Plateau
now beneath the western Caribbean. As
a result, there was deposition of a thick
series of continental sediments all through
this area and the Greater Antilles during
the latter part of the Eocene {op. cit.:
1155, abstract; 1169). "Each clastic rock
type appears to have a red-bed equiva-
lent suggestive of subaerial and/or deltaic

Oligocene Rodents of Bolivia • Patterson and Wood 457

depositional environments .... The
presence of plutonic clasts suggests a
great deal of erosion and uplift in order
to expose the [Cretaceous] intrusives by
the Eocene" (op. cit.: 1169). Their Fig-
ure 10 indicates that there was a pre-Oli-
gocene Cayman Ridge and Nicaraguan
Plateau land mass that was continuous or
nearly so from Nicaragua through Jamai-
ca, Oriente Province of Cuba and His-
paniola. Here the rodents would have
made enough easting so that the\ could
have achieved a landfall in northern
South America by island-hopping through
the Lesser Antilles (either the present is-
lands or their Eocene predecessors as
suggested by Malfait and Dinkelman,
1972), crossing Lavocat's hypothetical
east-to-west current from northeast to
southwest (Wood, 1977b: 104; 1981: 87-
88).

There seems to be a vague impression
afloat that plate tectonic theory is some-
how more favorable toward latest Creta-
ceous and earlier Cenozoic raft transport
from Africa to South America than toward
such transport between the Americas. If
anything, as indicated above, the evi-
dence shows that it was the other way
around. Webb (1978: 395) reached con-
clusions similar to ours as to the distances
separating South America from Africa
and North (or Middle) America during
the Oligocene.

If the caviomorphs, or their immediate
ancestors, reached South America from
the north, the place of departure of the
waif emigrants was southern, or tropical.
North America. Southward from the bor-
der between the United States and Mex-
ico this portion of the continent is prac-
tically a terra incognita as regards its
earlier Tertiary mammalian faunas. It is
a large region, whose size is perhaps not
always appreciated. At the beginning of
the Cenozoic, its area was about half that
of present-day Europe west of the USSR,
and it spanned some 20° of latitude, a dis-
tance approximately equal to that be-
tween the Straits of Gibraltar and north-

ern Scodand (Wood, 1981: 87). This area
is all part of the region defined as, bio-
geographically. Middle America by Fer-
rusquia (1978: 198). Middle America
"must long have been a center of adap-
tation and local radiation of faunal ele-
ments specifically adapted to its special
conditions . . ." (Simpson, 1950: 389).

Reviewing the Recent mammalian fau-
na and what evidence of past ones was
then available, Patterson and Pascual
(1972: 255-257, Table 3) concluded that
"the North American tropical peninsula
was evidently a fact of ancient geog-
raphy, and there is convincing Recent
and suggestive Tertiary evidence that its
mammals are and were distinctive to
some degree within the continental fauna
as a whole." Such suggestive evidence is
increasing. Jacobs (1977: 518) and Baskin
(1978: 133) see indications of southern
elements within late Miocene rodent fau-
nas in Arizona. Wood (1973: 32; 1974a:
102-104; 1975a: 78; 1981: 87) has con-
cluded that rodent faunules from the
middle Eocene to the early Oligocene of
southwestern Texas and from the middle
to late Eocene of southern California, the
latter now some 200 km north of their
original position due to displacement
along the San Andreas rift (Merriam,
1972), were the northern representatives
of a very diverse southern (Middle Amer-
ican) rodent fauna, quite different from
the better known faunas from farther
north in the western United States. All
the above localities were on or near the
northern fringe of the North American
tropics, and have been included in Mid-
dle America by Fermsquia (1978: 198)
who placed the northern boundary of
Middle America at the 35th parallel.

The only early Tertiary (probably
Eocene) rodents so far found deep within
Middle America, in Guanajuato, Mexico,
are not like anything known elsewhere in
the continent, and thus add to the grow-
ing evidence. One, Floresomys, may be
a sciuravid, but, if so, "is a divergent
member of the family" (Black and Ste-

458

Bulletin Mu.seuni of Comparative Zoology, Vol. 149, No. 7

phens, 1973: 9); the other, Guanajuato-
niijs, is a hystricognathous form that can-
not at present be allocated to family
(Black and Stephens, 1973: 5). Neither
Guanajuatomiis nor the likewise hystri-
cognathous but not closely related Pro-
lapsus from the middle Eocene of south-
western Texas could have been on or
near the ancestry of the Caviomorpha, to
judge from their cheek-tooth structure.

The Uintan Protoptychus, known from
adjoining areas of northeastern Utah and
southwestern Wyoming, was hystrico-
morphous and probably hystricognath-
ous. A partial skeleton of this animal, con-
tained in a block of hard matrix, is in the
Field Museum of Natural History. The
mandible, the only one known, is at pres-
ent visible only in lateral view, which
does not allow determination of the
amount of hystricognathy that it had at-
tained. Protoptychus had a tooth pattern
that would have been ideally suited to
have given rise to what we believe to be
the primitive caviomorph type — the four-
crested pattern seen in Deseadan echi-
myids and octodontids. We agree with
Wahlert's observation (1973: 12) that the
"rarity of Protoptychus in fossil collec-
tions supports the possibility that it, too,
is based in a stock evolving elsewhere
than in the western United States,"
which we suspect to have been in Middle
America.

The earliest rodent faunas of southern
California are, in general, rather similar
to their contemporaries of the Rocky
Mountain area, suggesting relatively lit-
tle regional differentiation at that time
(Golz and Lillegraven, 1977: 61 and Ta-
bles 2-3). Endemism in southern Cali-
fornia seems to have increased fairly rap-
idly in the latter half of the Eocene,
suggesting (Wilson, 1980: 276) that there
was increased climatic differentiation at
this time that led to the separation of the
distinct Middle American faunas dis-
cussed above.

Berggren et al. (1978) have suggested
that the Uintan be regarded as middle

Eocene in age. The Ukelihood of such an
age was first indicated by the discovery
of Notiotitanops, a brontotheriid of Uin-
tan if not Duchesnean evolutionary
grade, in the marine middle Eocene Lis-
bon Formation in Mississippi (Gazin and
Sullivan, 1942), and reinforced by Ga-
zin's somewhat guarded observation
(1955: 15) that Uintan artiodactyls were
behind those of the late Eocene of Eu-
rope in their degree of advancement.

The intercalated marine and continen-
tal faunas of the San Diego area have
shown that earlier Uintan continental
beds are equivalent to marine middle
Eocene, but that the later Uintan (equiv-
alents of Myton or Uinta C) is upper
Eocene (Golz and Lillegraven, 1977: 43
and Fig. 6). If this middle Eocene age of
the earlier Uintan (Wagonhound, or Uin-
ta A + B) be accepted, then the late early
Uintan (upper Uinta B) Protoptychus
would be middle Eocene and would
have lived at approximately the time
when the first caviomorphs reached
South America. Some southern member
or members of the Protoptychidae could
have been involved in the crossing.

The recent discovery that the North
American Oligocene Cylindrodon was at
least subhystricognathous (Wood, 1980b:
6; 1981: 85-87) shows that there was
another stock in the middle Eocene of
Middle America (Wood, 1974a: 102-104)
that may have been morphologically ca-
pable of having given rise to the Cavio-
morpha, although at least the Oligocene
members of the family were specialized
in a different direction by having ac-
quired uniserial incisor enamel. This
does not, however, seem to have been
true of the middle Eocene Mysops,
which still retained pauciserial enamel
(Wahlert, 1968: 15). The middle Ameri-
can fomis of Mysops, in particular, have
tooth patterns that could have given rise
to those of the Caviomorpha (Wood,
1973: Figs. 3^). Unfortunately (so far as
A.E.W. is aware), no specimens of My-
sops are on record that preserve enough

Oligocene Rodents of Bolivia • Patterson and Wood 459

of the angular process to indicate wheth-
er or not this genus showed any tenden-
cies toward hystricognathy.

There would thus seem to have been
a considerable variety of hystricognath-
ous, and, at least in part, hystricomor-
phous rodents inhabiting Middle Ameri-
ca during the latter half of the Eocene.
On the basis of presently available infor-
mation, this area would seem to have
been the center of evolution of the Fran-
imorpha.

It is, of course, impossible for anyone
to know what were the dental character-
istics of the ancestors of the Deseadan
rodents, until such forms have been
found and tied in with their Deseadan
descendants by a continuous series of in-
termediate stages. It is easy to assume
that one knows, as we have done previ-
ously, and as Hoffstetter and Lavocat
have also done. Since a variety of rodents
is known from the Deseadan, some must
be considered more primitive than oth-
ers, and each author will choose those
that come closest to fitting in with his
opinions as to the source of the fauna.
There are genera with brachyodont
cheek teeth and ones with hypsodont
ones; ones in which the individual cusps
are clearly distinguishable from the
crests, and ones in which they are not; in
each case we consider the former the
more primitive, but in each case we be-
lieve that Lavocat and Hoffstetter have
selected the latter. The Deseadan Platij-
pittamys had cheek teeth that are among
the most brachyodont of any of the
known Deseadan genera, and they clear-
ly show the presence of the individual
cusps. There are only four crests, with no
trace of a fifth. This is the only Deseadan
rodent, whose skull is known, in which
the infraorbital foramen is relatively
small and the masseter either did not
penetrate the foramen or did so to a ver>'
limited extent. Landry (1977: 460) stated
that the infraorbital foramen of Platypit-
tamys was "about the same size as that
of a modem Proechitnys, and not signif-

icantly reduced as" Wood believed. A
comparison of the figures of the skull of
Proechitnys, as given by Ellemian (1940:
Fig. 11) and Woods (1972: Fig. lA) shows
the foramen to be considerably larger
than that of Platypittamys (Wood, 1949:
Fig. 2A). Incidentally, Wood has never
believed that the infraorbital foramen of
Platypittamys "was significantly re-
duced," but rather that it was primitive-
ly small. We feel that the cumulative evi-
dence is that Platypittamys and the
related Migraveramus are the closest
approaches, now available to us, to con-
ditions in the earliest South American ca-
viomorphs.

If we are correct as to the significance
of Platypittamys and Migraveramus, the
pre-South American ancestry of the Ca-
viomorpha should be sought among four-
crested hystricognathous, but perhaps
not yet hystricomorphous, rodents.

Summarizing to this point, Eocene
subhystricognathous and hystricognath-
ous rodents, one or more of which are
rather good structural ancestors for the
Caviomorpha, inhabited North (and par-
ticularly Middle) America, the onh; area
where such rodents have so far been def-
initely recorded. Members of this group
could have reached Asia in the early
Eocene. There was an Eocene hystrico-
gnathous radiation in Middle America.
There is evidence that this area has been
a faunal subregion of the continent
throughout the Cenozoic (Ferrusquia,
1978). Eocene geography would appear
to have been such as to have permitted
rafting from North to South America in
middle Eocene time.

We regard such rafting of the ances-
tral caviomorphs (and, presumably, plat-
yrrhines) as being distinctly easier of
accomplishment than the transatlantic
crossings of over 1300 km of open water
postulated by Hoffstetter and Lavocat (it
would seem unlikely that winds and cur-
rents would have combined to propel
rafts across the shortest possible distance
between Africa and South America).

460 Bulletin Museum of Coinpurutive Zoology, Vol. 149, No. 7

Nevertheless, their views cannot be re- 8 or 9 families (Wood, 1980b: Table 1), in

jected out of hand. No rafting is easy of contrast to the 5 genera belonging to at

accomplishment, which is why there is most 3 closely related families from the

evidence of so very few successful ones, entire Oligocene of Africa. Most recently,

A Caribbean crossing may have been Lavocat (1978: 84) suggested that the

somewhat less hazardous than a transat- thryonomyoids, bathyergoids, anomalu-

lantic one, but this does not prove that rids and pedetids were all members of

the latter could not have occurred. Nor the African Eocene faunas, and proposed

can the possibility, envisioned by Lavo- that all of these groups shared a common

cat (1973: 166), of a latest Paleocene or stem with the Theridomyoidea. This last

earliest Eocene rafting of param\ ids from idea seems extraordinarily unlikely to us;

Europe to Africa be summarily ruled out. the theridomyoids are a sciurognathous

More recently, Lavocat (1978: 76) stated but hystricomorphous middle Eocene to

categorically that "... the rodents as well Oligocene European group that (unless

as at least most of the other elements of they were related to the Anomaluridae)

the Oligocene fauna, are probably related seem to us never to have spread out of

to animals that reached Africa much ear- Europe. Lavocat's suggestion of a rela-

lier, perhaps very soon after the end of tionship of this group of families and su-

the Paleocene, and probably before the perfamilies appears to us to be a retreat

Lutetian." He also said (1978: 84), refer- from his belief in the fundamental im-

ring to an early Eocene invasion of Africa portance of hystricognathy in rodent clas-

by rodents, that "such an invasion nee- sification and a return to the acceptance

essarily did occur. ... At least as early as of the basic importance of hystricomor-

the Lutetian this . . . invasion had given phy (see below, p. 513).

rise to the first Hystricognathi with all the Were caviomorphs the descendants of

essential characters of that group already one or two successful middle Eocene

developing." It should be stressed that raftings of African thryonomyoids to

there is no direct evidence for such an South America, one would hardly expect

early Eocene or Paleocene rodent inva- that, by early Oligocene time, they would

sion of Africa, or for the presence in the have become more diversified, more nu-

middle Eocene of Africa of fully devel- merous, and in a number of cases more

oped hystricognaths. specialized than their contemporary Af-

Successful raftings of rodents, of the rican relatives, the descendants of the

kinds and at the times postulated by La- stock from which they supposedly came,
vocat and Hoffstetter, either from Europe

to Africa or from Africa to South America, j^^ Composition of the Deseadan and

must have had faunal and evolutionary p Rodent Faunas
consequences, and this tact pemiits the

testing of their hypotheses against the Twenty-four species of Deseadan ro-

fossil record. dents, not all fomially described, repre-

Had paramyids reached Africa by the senting perhaps 16 genera, have been re-
beginning of the Eocene, they would cer- corded to date (Table 8). They are
tainly have radiated rapidly in that vast assignable to at least 7 families, and thus
continent that contained no potential are almost as diverse as their North
competition, and would have given rise American contemporaries. Rodents form
by early Oligocene time to a rodent fauna large percentages of the total mammalian
at least as diversified as that of contem- assemblages in the better known local
porary North America. The known ro- faunas. As noted in the Introduction (p.
dents of the early Chadronian of North 374), each of these contains one rodent
America include 23 genera belonging to species that outnumbers all the rest, one

Oligocene Rodents of Bolivia • Patterson and Wood 461

Table 8. Rodents reported from various Deseadan localities.

Patagonia

Bol

livia

Rinconado

Laguna

de los

La

Cabeza

de los

Lopez

Flecha

Blanca

Machos

Salla

Lacayani

Platypittamys brachyodon

2

Migraveramus beatus

1

Sallamys pascuali

12

Xylechimys obliquus

1

Deseadomys arambourgi

1

1

Deseudo m i/.s loo m is i

1

Scotamys antiquus

ca. 45

1

Scotamys sp.

*

Cephalomys arcidens

12

100+

Cephalomys plexus

1

32+

Cephalomys sp^

1

Cephalomys bolivianus

1

Litodontomys chubutensis

1

Incamys bolivianus

1

79+

Dasyproctidae aff. Neoreomys

1

Branisamys luribayensis

19

Asteromys punctus

1

1

Asteromys sp.

*

Chubutomys simpsoni

1

Eocardiidae indet.

2

*

Protosteiromys medianus

7

1

Protosteiromys asmodeophilus

2

Caviomorpha indet.

1

1

Octodontoidea gen. et sp. nov. 1

?

Octodontoidea gen. et sp. nov. 2

?

Luribayomys masticator

1

Number of species

1

7

13

2

8-9

3

Number of specimens

2

69

144 +

2

115+

*

In addition, one specimen oi Palmiramys waltheri from Uruguay.

* Species listed by Hartenberger, 1975. Number of specimens unknown.

or two others that are moderately abun-
dant, with the remainder represented by
one or two specimens; the predominant
species is different in each local fauna.
At any one Deseadan locality, so far as is
known, only one fossiliferous zone is rep-
resented, in contrast to the Fayum situa-
tion (see below, p. 464).

Professor E. E. Williams infonned us
that he had encountered rather compa-
rable conditions in collecting Neotropi-
cal reptiles. We therefore looked for data
on numbers of Recent Neotropical ro-
dents at specific localities. These were
found in the records of a plague survey

carried out in northeastern Brazil from
1952-55 by the then Servigo Nacional de
Peste (now Superintendencia da Cam-
panhas de Saiide Publica), in the course
of which nearly 40,000 specimens of ro-
dents were collected. Ectoparasites ob-
tained have been published upon (Gui-
maraes, 1972 and references therein), but
almost nothing, other than the essential
preliminary task of identification (by
Moojen), has been done on the hosts. Ar-
coverde de Freitas (1957, Table 1) has
given the grand totals of the fomis col-
lected in the Noreste (an unofficial name,
comparable to Patagonia or New En-

462 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Table 9. Rodents collected during the plague survey of 1952-55 in northeastern Brazil

(excluding State of Bahl\).

Zijgodontointjs lasiurus pixuna
Oryzotnys subfluvus
Cercomys ctinicularius
Wicdoinys pyrrliorhinos
Akodon a. arviculoides
Galea spixii wellsi
Oryzotnys tiigripes
Holochilus b. brasiliensis
Oxymycterus angularis
Kerodon rupestris
Nectomys stpiamipes
RhipiditJiys masticalis cearanus

Data from Arcoverde de Freitas, 1957, Table 1.

3,486

Echimys dasythrix lamarum

133

7056

Oryzoniys lamia

59

5700

Caloinys sp.

50

2643

Calomys laucha tener

36

2213

Oryzomys sp.

18

2137

Proechitnys guyannensis arescens

16

1944

Dasyprocta prymnolopha

12

967

Coendou p. prehensilis

11

948

Calomys callosus expulsus

4

523

Sciunis (Guerlingtietus) sp.

4

493

Oryzomys capito laticeps

2

439

Echimys sp.

1

Total

38,990

gland, bestowed upon the northeastern
coastal states from Maranhao to Sergipe;
Bahia is not part of the Noreste), and the
subtotals by Distritos, large areas that
cover three to thirteen Municipios (coun-
ties). His grand totals are given in Table
9. Seventeen genera, over a third of the
total recorded in the entire country
(Viera, 1955) and 24 species, are repre-
sented in this comparatively small and,
for the most part, far from lush part of
Brazil. It is interesting to note that the 7
commonest species make up 90% of the
whole, and that each of 12 species con-
stitutes less than 1%. Arcoverde de Frei-
tas' breakdown shows, furthennore, that
relative abundance may vary widely from
one Distrito to another.

A Distrito, however, is too large an area
for the kind of comparison needed. For
this, the lists of parasitized specimens
sent by the Servigo to Dr. Guimaraes
have been utilized, and from them the
data presented in Table 10 have been
compiled. The localities (Fig. 28) have
been chosen with an eye both to geo-
graphic proximity or spacing, and to the
sampling of different environments.
Three of the localities, G, H, and I, are
single collecting stations; the rest are the
combined totals for collecting stations
within a Municipio.

These environments may be briefly de-

fined (Dr. P. E. Vanzolini, personal com-
munication) as follows:

Caatinga: rainfall 300-800 mm per an-
num, vegetation highly xerophytic with
Cactaceae very conspicuous, soils very
shallow; a harsh environment.

Agreste: rainfall 1000-1100 mm, vege-
tation xerophytic but Cactaceae less con-
spicuous, soils somewhat better; a miti-
gated Caatinga.

Brejo: an island of forest within Caatin-
ga or Agreste resulting from higher ele-
vation and increased, although locally
variable, rainfall; vegetation not summer
deciduous.

Atlantic Forest: rain forest with highly
variable rainfall, 1200-4000 mm.

Table 10 shows that not only is the pre-
dominant species frequently different
but also that relative abundances vary
widely from locality to locality, even be-
tween those in close proximity, e.g., H
and I (Fig. 28), or in similar general en-
vironments, e.g., E and G. Two forms,
Oryzomys suhflavus and Zygodontomys
lasiurus pixuna, are well represented at
most localities, while almost half (11 out
of 23) of the species occur at only one
locality. In some cases, such as B and I,
two species are about equally predomi-
nant numerically. As a rule, however, one
species is predominant, although not in
all instances (as at A and D) overwhelm-

Oligocene Rodents of Bolivia • Patterson and Wood 463

lOOOKM

Figure 28. Map of northeast Brazil showing localities where rodents listed in Table 10 were collected (modified from
Guimaraes, 1972: Map 1). A-J: Collection localities listed in Table 10. States: 1— Piauhy; 2— Ceara; 3— Rio Grande del
Norte; 4 — Parahyba; 5 — Pernambuco; 6 — Alagoas; 7 — Sergipe; 8 — Bahia. Scale as shown.

ingly so. The situation at C, where Cer-
comys cunicularius is so remarkably nu-
merous, might lead to the suspicion that
a ratada, or population irruption, had
been tapped by the collectors. This does
not appear to be the case, since the rec-
ords show peaks of abundance following
the rainy season during each of the years
1952-54, which is not the classical ratada
pattern. Cercomys cunicularius is an in-
habitant of rocky places. Presumably lo-
cality C is such an area.

A general resemblance to the relative
abundances encountered in Deseadan lo-
cal faunas is apparent. The frequency dis-
tributions of La Flecha and Salla resem-
ble those of localities E and G rather
closely, and Cabeza Blanca shows some
similarity to F. A collecting station for liv-
ing fonns, or a group of such stations, is
of course not quite the same thing as an

area in which sediments and skeletal re-
mains are deposited. Sustained trapping
will in time reveal the presence of even
the rarest species and provide accurate
estimates of relative abundance. The vi-
cissitudes of taphonomy attendant upon
the accumulation of a fossil assemblage,
coupled with the accidents of collecting,
can hardly be expected to yield precisely
comparable results. Allowing for this,
however, we believe that relative abun-
dances of rodents at Deseadan localities
were much like those now obtaining in
Brazil (cf Tables 8 and 10).

Deseadan caviomoq^hs were a rather
diversified lot, ranging in size from small
to fairly large (for rodents) and in crown
height from brachyodont to hypselodont.
Given the large numbers of the predom-
inant forms — Cephalomys is the com-
monest Patagonian Deseadan mammal,

464 BuUctiu Museum of Conifxiratiic Zoology, Vol. 149, No. 7

Table 10. Rodents collected at various localities in northeastern Brazil.

B

D

J

Totals

Cercomys cunicularius'
Onjzomys siibflavu.s
Zygodontoniys lasiurus pixuiia
Akodon a. arviculoides
Onjzomys nigripcs
Oxymycterus angularis
Holochilus b. brasiliensis
Calomys sp.

Rhipidomys masticalis cearanus
Galea spixii wellsi
Nectomys squaniipes
Onjzomys lamia
Proechimys iheringi denigratus
Calomys laucha tener
Proechimys guyannensis arescens
Kerodon rupestris^
Wiedomys pyrrhorhinos
Proechimys a. albispinus
Coendou p. prehensilis
Echimys armatus
Calomys callosus expulsus
Onjzomys capito laticeps
Dasyprocta prymnolopha

Total

4

95

144

1
19
32
32

100

20

5

15

2

3

413
460
162

12
1

2
13

1424

1
86
32

51

112

33
245

82

3
1

7

1
101

6

36

43
33
173
11
19
1

31
5
80
14
12

22
59
12
67

1

56

2
262
121
12
17
91
69

39

9
1

19
13

5
9

7
4

1878

1345

644

278

230

156

115

112

111

80

39

20

19

18

15

15

11

9

7

4

2

1

I

5110

Localities, here and on Fig. 28, are: A. Ceara, Municipio de Sao Benedito, Brejo; B. Pernambuco,
Municipio de Triunfo, Brejo; C. Pernambuco, Municipio de Bodocb, Caatinga; D. Bahia, Municipio de
Vitoria da Conquista, transitional between Caatinga and Atlantic Forest; E. Alagoas, Municipio de Pal-
meira dos Indios, Agreste; F. Alagoas, Municipio de Limoeiro de Anadia, Agreste; G. Pernambuco, Mun-
icipio de Caruaru, Fazenda Caruaru, Agreste; H. Pernambuco, Municipio de Garanhuns, Sitio Barrocao
3°, Atlantic Forest with areas of Agreste; I. Pernambuco, Municipio de Garanhuns, Fazenda Cristovao,
Atlantic Forest with areas of Agreste; J. Alagoas, Municipio de Vigosa, Atlantic Forest.

' C. c. inermis at D; C. c. laurentius at the other localities.

^ An inhabitant of rocky places, this species is perhaps under-represented.
Rattus and Mus, some of which were taken, have been excluded.
Determinations by J. Moojen. Certain nomenclatural emendations were made by B.P.

Incamijs apparently the commonest Bo-
livian one — we suspect that by early Oli-
gocene time the caviomorphs were well
on the way toward achieving the numer-
ical position and ecological diversity in
the South American mammalian fauna
that is held by rodents today. In the rich
Santacruzian fauna of earlv Miocene age
(21.7-18.5 m.y.b.p.; Marshall et al, 1977)
there are only 19 genera of rodents. Scott
(1905), largely following Ameghino, a
splitter of the first order, listed nearly 70
species, surely too high a number. Revi-
sion at the specific level has yet to be
done. Data sufficient for any attempts at

numerical analyses of the fauna are not
available in the literature.

The Fayum thryonomyoids present
quite a different picture. Three levels in
the Jebel el Qatrani Formation yield
mammalian remains: the lower fossil
wood zone, the Yale Quarrv G level, and
the upper fossil wood zone; the lower
and middle horizons are separated by an
interval of 140 feet, the middle and upper
by one of 150 feet (Simons, 1968: 21).
The time involved in the deposition of
these deposits apparently spanned most
of the Oligocene (Simons, 1968: 4).

The rodents from this sequence com-

Oligocene Rodents of Bolivia • Patterson and Wood 465

Table 11. The Fayum rodent faunas.

Lower lossil

Upper
fossil wood

wood zone

(Quarries

Quarry

zone

A, B, E)

G

(Quarr> I)

Phiomys andrewsi

30

2

Phiomys lavocati

9

Ph io m ys paraphiomyo ides

6

Phiomys aff. paraphiomy aides

4

Phiomys or Metaphiomys sp.

2

6+

Metaphiomys schaubi

6

59

Metaphiomys headnelli

2

a. Metaphiom.ys sp.

1

1

Phiocricetomys minutiis

1

Gaudeamus aegyptius

11

Neosciuromys simonsi

1

Unidentified

113+

9

Number of species

6

4

5

Number of specimens

59

186+

18

Data essentially from Wood (1968). The unidentified specimens have been recorded by Conroy (1976;
20) for Quarry 1 and by Simons (personal communication) for Quarry G; all were collected since 1968.
Since this table was prepared, a collection in the MCZ of about 120 isolated cheek teeth and 1 lower jaw
fragment, sorted from matri.x from Quarry- G, has become axailable. The jaw and 53 cheek teeth are
referable to Phiomys, the remainder to Metaphiomys.

prise 8 to 11 species (3 possibly distinct Table 12 compares the numbers of
fonns were not named) representing 5 mammalian species collected by Loomis'
genera (Table 11) and 3 families. Wood expedition to Cabeza Blanca, Patagonia,
(1968) placed all forms in one family, the with the numbers collected in one quarry
Phiomyidae, a step justified on the basis (Quarry I) in the upper fossil wood zone
of their morphological similarity. Lavocat of the Fayum during several field sea-
(1973: 20), in agreement, pointed out that sons. The difference between the two in
familial separation is only made possible the percentage of rodents is striking. At
by better knowledge of their Miocene neither locality are the sediments uni-
descendants. We have previously stated fonn, and as regards the total number of
(1959: 389) that Deseadan caviomorphs specimens collected the two are not ver\
fomied a relatively homogeneous group, different. Cabeza Blanca is a small hill.
This they did, but they were very much about 240 m long and averaging about 60
more diversified morphologically than m in width, capped over much of its ex-
their Fayum contemporaries. The much tent by the Miocene Monte Leon ("Pa-
greater homogeneity of the Fayum than tagonian") Fonnation; the Deseadan sed-
of the Deseadan rodents, and the fact that iments there are 12.7 m thick (Loomis,
they were, individually, of smaller size, 1914: 6, 21). Quarr>' I measures roughly

makes it difficult to understand the state-
ment by Lavocat (1969: 1497) that the
South Atlantic "a joue un role de filtre
important, la faune de Rongeurs qui a
reussi a passer en Amerique du Sud etant

100 by 50 m and the sediments are about
1 m thick (E. L. Simons, personal com-
munication). All matrix in Quarr\- I was
meticulously examined, whereas no
quarrying was carried out at Cabeza

beaucoup moins variee que la faune au- Blanca, all specimens collected having

tochthone d'Afrique, et constitute sem- been found on the surface or in the pro-

ble-t-il essentiellement par de petites cess of weathering out. To compare the

formes arboricoles." two seems fair, and it is clear that rodents

466 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Table 12. Comparison of Patagonian and Fayum local faunas.

Cabeza Blanca

Jebel

el Qatrani, Quarry I

N

%

N

%

Notoungulata
Rodentia

125
102

44.3
36.1

Artiodactyla

Primates

Hyracoidea

Creodonta

Rodentia

147
105

72
22
18

38.3
27.2

Edentata

Litopterna

Marsupialia

Astrapotheria

Condylarthra

Total

23

15

11

6

1

283

8.5
5.3
3.9
2.1
0.4

18.6
5.2

4.7

Proboscidea
Chiroptera
Insectivora s.l.
Sirenia
Embrithopoda

Total

8
6
3
3
2

386

2.1
1.1

0.8
0.8
0.5

Numerical data for Cabeza Blanca from Loomis (1914: 19-20, 216); for Quarr>' I from Conroy (1976: 20,
22). All specimens identifiable to order — cranial, postcranial and dental — are counted. Of the Cabeza
Blanca notoungulates, 31 (24% of the members of that order) are comparatively small, rather rodent-
like members of the suborders Typotheria and Hegetotheria.

really are far less abundant at Quarry I
than at Cabeza Blanca; it may be added
that five other Yale quarries in the upper
fossil wood zone contained no rodent re-
mains at all. Had the large forms, arsi-
noitheres and gomphotheres, been better
represented in Quarr>' I, the percentage
of rodents would have been still lower.
The Cabeza Blanca faunule is more near-
ly a representative sample of the ambient
mammalian fauna; in Loomis' collection
the smaller mammals constitute about
59% of the whole, medium-sized ones
about 16%, and large ones 25%.

As regards the other Fayum levels, the
lower fossil wood zone has yielded 59
specimens of rodents. This is the longest
worked of the Jebel el Qatrani horizons,
and Professor Simons is willing (personal
communication) to estimate that about
2000 specimens of mammals have been
collected from it. Taking this estimate,
rodents constitute roughly 3% of the
whole, despite the concentration of the
Yale parties on finding the smaller faunal
elements, which resulted in their in-
creasing the number of known rodent
specimens from the zone nearly fivefold.

In the middle Jebel el Qatrani, at Quar-

ry G, the fossiliferous bed is homogene-
ous, 14 m long, 1 m wide and 4 cm thick
at the center, tapering to feather edges at
the sides, evidently what remained of the
bed of a small stream. Except for one
tooth of Moeritherium and one of an an-
thracotherid, the mammals are all small
(E. L. Simons, personal communication).
Here, rodents constitute about 58%, pri-
mates 27%, hyracoids 10%, creodonts 3%
and "insectivores," s.l., 1% of the total.
The stream evidently ran through or near
a habitat congenial to Metaphiomys
schaubi, by far the commonest species in
the deposit. Apart from a maxillary frag-
ment of F/iio//i{/5 paraphiomyoides found
about a meter above the quarry, no other
fossils were found near Quarry G (G.
Meyer, personal communication). Quarry
G is obviously a special case.

The available data indicate quite clear-
ly that the rodents of the Jebel el Qatrani
were on the whole less numerous rela-
tive to their contemporaries than were
those of the Deseadan.

Passing on to taxonomic diversity,
there is a paucity of rodent species in
each of the three levels — 6 in the lower
fossil wood zone, 4 in Quarry G, and 5 in

Oligocene Rodents of Bolivia • Patterson and Wood 467

the upper fossil wood zone, compared
with 13 recorded from one level at Ca-
beza Blanca. The difference in total taxo-
nomic diversity is striking — 7 families,
perhaps 16 genera and 24 species in the
Deseadan as against 3 families, 5 genera
and 8 to 11 species in the Fayum, a point
touched on by Hartenberger (1975: 429).
This indicates to us that caviomorphs had
been evolving considerably longer in
South America than had thr>'onomyoids
in Africa.

The Fayum and Deseadan rodent fau-
nas are far apart as regards structural di-
versity. All Fayum rodents are small, ex-
cept Neosciuromys ("Paraphiomys")
simonsi (see below, p. 520), which was
only as large as Ondatra or Cynomys.
The Fayum rodents, in their entirety, are
more specialized than any Oligocene ro-
dents from any other part of the world in
their marked degree of retention of dnV,
a character not met with in any known
Deseadan rodent (this and other dental
differences are discussed at length be-
low, pp. 491-503). On the other hand, the
Fayum thryonomyoids are far behind many
Deseadan caviomorphs in cheek-tooth
height. Nearly all of them are brachy-
odont, only Neosciuromys and Gaudea-
mus being incipiently mesodont. In the
Deseadan, Protosteiromys is brachy-
odont; Platypittamys and Migraveramus
are essentially brachyodont; Sallamys
and Branisamys are brachyodont with in-
cipient unilateral hypsodonty; Deseado-
mys, Xylechimys and Asteromys are me-
sodont, the first with some unilateral
hypsodonty in the uppers (the other two
are known only from parts of mandibular
rami); Cephalomys and Incamys show
marked unilateral hypsodonty, especially
in the uppers; Litodontomys and Chii-
butomys (both known only from parts of
mandibular rami) are hypsodont; and
Scotamys is hypselodont. In cheek-tooth
structure the Deseadan caviomorphs are
much more varied than the Fayum
thryonomyoids, and several of them are
considerably larger. All this likewise sug-

gests that caviomorphs had been resident
in South America longer than had thryon-
omyoids in Africa.

There is one outstanding fact concern-
ing the Fayum faunas: of all the rodents
in the Yale and other museum collec-
tions, not one is attributable to any group
other than the Thryonomyoidea. Lavocat
(1973: 196, 252) reported that Simons had
shown him an ulnar fragment from the
upper levels of the Jebel el Qatrani For-
mation that was unquestionably that of an
anomalurid. Search of the Yale collec-
tions has uncovered no trace of such a
fragment. Simons infomis us (personal
communication to B.P.) that he has no
recollection of ever having seen such a
specimen, of ever having shown it to La-
vocat, or of having Lavocat tell him that
it was an anomalurid. Lavocat (1978: 80)
stated, of anomalurids, that "their Oligo-
cene record is extremely poor except for
the Fayum ulna." This statement sug-
gests that there is more basis for assum-
ing that there were Oligocene anomalu-
rids in Africa than is supported by the
evidence, since the Oligocene record of
anomalurids (except for the supposed Fa-
yum ulna) is nonexistent.

If we contrast the Fayum assemblage
with that present in one North American
early Oligocene local fauna, the Porvenir
Local Fauna in the Chambers Tuff For-
mation of southwestern Texas, we note
that the latter includes 6 families of ro-
dents, one represented by 3 and two oth-
ers by 2 subfamilies; perhaps 13 genera;
and some 15 species — and all this diver-
sity is revealed by a total of only 44 spec-
imens (Wood, 1974a: Table 1). To us it is
inconceivable that, if primitive rodents
had made a landfall on Africa in latest
Paleocene or earliest Eocene, as pro-
posed by Lavocat (1973: 166), so few ro-
dent species, all of them members of one
closely knit group, would occur in three
levels of a fonnation laid down, during
most of the Oligocene, at least 16 m.y.
after such a landfall. Rodents do not
evolve so slowly, especially in the com-

468 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Table 13. Rodents from the early Miocene of East Africa.

old natives (Thnoiioniyoidea)

Recent immigrants (all others)

Diatiumtotnys luederitzi

Paraphioimjs pigotti
Paraphioniys s. stromeri
SiDioniniijs gciiovcfae
Phionu/s f{. andrewsi
Paraphioniys stromeri hopwoodi
Myophiomys araiuhourgi
Epiphiomys coryndoni
Thryonomyoidea indet.
Kenyamys marine
Elmerimys woodi
Andrewsimys parvus

Total counted specimens
Percent of grand total

9 genera, 11 species or subspecies

2180 +

Paranomalurus soniae

70

abundant

Bathycrgoides neotertiarius

62 + "4 boites

at 3 levels

Afrocricetodon s. songhori

65

of Napak

Megapedetes pentadactylus

53 + incisor

1734

fragments

110

Paranomalurus hishopi

51

46

Vulcanisciurus africanus

32

23

Notocricetodon petteri

21

20

Paranomalurus walkeri

20

19

Proheliophohius leakeyi

14

14

Protarsomys macinnesi

9

12

Afrocricetodon songhori korui

5

9

Afrocricetodon sp. 2

4

3

Sciuridae gen. et sp. indet.

3

2

Megapedetes sp.

3

Zenkerella ivintoni

2

Afrocricetodon sp. 1

Total counted specimens

1

4172

415

90.95%

Percent of grand total

9.05%

10 genera, perhaps 16 species or subspecies

Numbers based on the maximum figures given by Lavocat (1973). Specimens without precise locality,
cited merely as "Kenya" are included, but those from Kirimun and "Lodwar' are excluded. The "4 boites"
of Bathycrgoides neotertiarius are of necessity not included.

plete absence of any competition. We
must conclude, therefore, on the avail-
able evidence, that Lavocat's hypothesis
of a very early date of arrival of rodents
in Africa must be rejected.

To what extent the Fayum rodents
were representative of the early Oligo-
cene rodent fauna of the African conti-
nent is a question that must now be con-
sidered. Conditions prevailing during
the deposition of the Jebel el Qatrani
Fonnation — rivers, streams, lakes, ponds,
forests and, possibly, savannah (Simons,
1968: 15-16) — would certainly seem to
have been suited to the support of a ro-
dent fauna at least as varied as that of
Cabeza Blanca (cf. Tables 8 and 10), had
such been in existence in Africa at the
time.

Comparisons of the Fayum faunas with
those of the late Burdigalian deposits of
East Africa are instructive in this connec-
tion. Table 13 lists numbers of specimens
of all East African rodent taxa, and the
percentages that belong to the Thryon-

omyoidea and to the other rodent fami-
lies and superfamilies that had, in the
early Miocene, only recently invaded the
African continent.

In the compilation of Tables 13 and 14
we have utilized the numbers of speci-
mens enumerated by Lavocat (1973) un-
der the taxonomic headings in the main
body of his text, and those listed in his
summary table on page 245 of that work.
These two sources often disagree; in fact,
they conflict over half the time. Most of
the discrepancies are minor, but some are
glaring. For example, 70 specimens of
Paraphiomys pigotti and 48 of Diaman-
tomys luederitzi from Rusinga are enu-
merated in the text (Lavocat, 1973: 22-
23, 52), whereas 487 and 1647, respec-
tively, are listed in the table; for D. lued-
eritzi, 74 are enumerated from Songhor,
but 1104 are listed in the table, although
the text also specifies "nombreux autres
echantillons" of upper teeth of D. lued-
eritzi from Rusinga and Songhor (Lavo-
cat, 1973: 52). Since specimens of upper

Oligocene Rodents of Bolivia • Patterson and Wood 469

Table 14. Numbers, percentages of total rodents, and numbers of taxa of recent immigrant

RODENTS IN THE EARLY MiOCENE OF EaST AFRICA.

Number of
specimens

Percent
of total
rodents

Genera

Species &
subspecies

Bathyergoidea

Anomaluridae
Pedetidae
Cricetidae
Sciuridae

Total

76 +

1.66

2

2

boites"

143

3.12

2

4

56

1.22

1

2

105

2.29

3

6

35

0.76

2

2

415

9.05

10

16

The "4 boites" of Bathyergoidea have of necessity been omitted from the calculations.

teeth of fossil rodents are usually dis-
tinctly rarer than those of the lowers, the
fact that Lavocat lists only 28 specimens
of lower teeth of this form from Songhor
and 22 from Rusinga strongly suggests
that the "nombreux autres echantillons"
totalled significantly less than the 50
specimens of lower teeth. Lavocat's table
(1973: 245) states that Diamantomijs
luederitzi is "Abondant" at three of the
four levels of Napak, whereas, on page
52, he enumerates only 31 specimens
from that locality. In the list of referred
specimens of Bathijergoides neotertiar-
ius (Lavocat, 1973: 110), there are includ-
ed (in addition to 33 listed specimens)
"4 boites" from Songhor, that we have
perforce neglected in our calculations.
Lavocat's table (1973: 245) merely lists
this species as being "Abondant" at Son-
ghor. In all cases of conflict we have used
the higher of the numbers given by Lavo-
cat. Even so, our Table 13 is clearly not
complete. Lavocat enumerated, in his
text, a total of 76+ rodent specimens from
Napak, whereas Bishop (1967: 48) re-
corded 716 as having been collected
there between 1958 and 1965. We must
assume that these presumably unstudied
specimens would not materially change
the relative abundances shown by our
Tables 13 and 14.

Numerically, the old natives (Thryon-
omyoidea) greatly outnumber the invad-
ers (4172 to 415), the latter constituting

only 9% of the total (Table 13). As regards
taxonomic diversity, the invaders have an
edge, 10 genera and 16 species or sub-
species, as against 9 genera and 11
species or subspecies of thryonomyoids.
This diversity results from the then-re-
cent establishment of a land connection
with Eurasia. In the early Miocene of
East Africa we are witnessing the irrup-
tion of already differentiated southwest
Asian faunal elements, that had not yet
succeeded in dominating the autochtho-
nous African rodent fauna. Lavocat (1973:
247) suspected that the invading cricetids
had differentiated prior to their arrival in
Africa; we go farther, suspecting that this
was the case for all the non-thryonomy-
oid families, each of which was contrib-
uting an approximately equal amount to
the total rodent fauna (Table 14). Dia-
mantomijs luederitzi and Paraphiomys
pigotti constitute 85% of the total speci-
mens; the eight commonest fomis make
up well over 90% (almost 95%), and the
remainder, 17 in all, each contribute less
than 1%.

Here in the early Miocene we see ro-
dent faunas in full flower, as in the De-
seadan of South America. The resem-
blance to the Recent faunas in the
Brazilian Noreste (Tables 9, 10) is rather
close, and the contrast to the Fayum ro-
dent faunas is striking.

Lavocat (1973: 246), on the basis of dif-
ferences in relative abundance of certain

470 Bulletin Museum oj Coiupanilivc Zoology, Vol. 149, No. 7

Table 15. Ordinal percentages in early
Miocene faunas of East Africa.

Mtwan-

Soiighor

Koni

gaiio

Napak

Rodentia

82.5

86.5

68.6

64.0

Pmboscidea

—

—

0.3

10.9

Artiodactv la

10.8

3.9

16.5

9.3

Primates

2.5

4.3

3.0

6.0

Others

4.1

5.3

11.6

9.8

Data from Bishop (1967: 48) tor Napak (four
levels lumped); from Bishop and Whyte (1962:
1286) for the other three. Cranial and dental spec-
imens onlv.

rodent species, has tentatively suggested
that the East African early Miocene fauna
of Rusinga might be a little older than
that of Songhor. As the Brazilian data
show, this conclusion does not follow.

The early Miocene rodents of East Af-
rica had achieved the numerical predom-
inance (Table 15) characteristic of the or-
der in areas where it has become well
established, due primarily to the success-
ful diversification of the thryonomyoids
during the 10 m.y. of the Oligocene. That
the Fayum thryonomyoids were far from
having achieved such a position would
seem evident (cf. Tables 12 and 15).

Under the circumstances, we are
forced to conclude, for all the reasons dis-
cussed above, that the Fayum rodents
were indeed representative of the early
Oligocene rodent fauna of Africa.

The degree of diversity and numerical
abundance attained by the Fayum
thr>'onomyoids, whether placed in one
family as was done by Wood (1968), in
three closely related families as we do
here, or in four as was done by Lavocat
(1973: 160), was certainly no greater and
probably less than that of the Recent Aus-
tralian murids. These rodents belong to
three subfamilies, the Murinae of Asian
origin, the Hydromyinae and the Pseu-
domyinae. The Hydromyinae originated
in New Guinea, the Pseudomvinae in
Australia (Tate, 1951: 217-221),' the two
areas having been separated a significant
portion of the later Tertiary. Intermigra-

tions occurred during Pleistocene pe-
riods of low ocean levels. A single rodent
incisor has been reported from the Plio-
cene Otibanda Formation of New Guinea,
dated at 5.7-7.6 m.y.b.p. (Plane, 1967:
56). In Australia, the "Pseudomyinae dif-
ferentiated, spreading into all the usual
rodent niches, including burrowing
(Mastacomys), saltatorial (Notomys and,
partially, Coniluru.s) and arboreal {Me-
sembriomy.s and, partially, Coniluru.s), as
well as the more typical scampering
niche . . ." (Wood, in press, b).

Isolated pseudomyine molars have
been reported from the Pliocene Chin-
chilla local fauna of southeastern
Queensland, dating from about 4 m.y.b.p.,
and a single rodent incisor was found in
the slightly older (4.5 m.y.b.p.) Bluff
Downs Formation of northern Queens-
land (Archer and Bartholomai, 1978: 15-
16), but no rodents are included in the
extensive microfauna of the Hamilton
Formation of Victoria, dated at 4.3
m.y.b.p., in spite of extensive sieving
(Turnbull and Lundelius, 1970; Archer
and Bartholomai, 1978: 15). This strongly
suggests that the rodents had only just
reached northern Australia in Bluff
Downs time, and that their apparent ab-
sence from the Hamilton fauna repre-
sents a real absence from the southern
border of the continent at that time (Ar-
cher and Bartholomai, 1978: 15; Turn-
bull, personal communication; Wood, in
press b). The success of introduced rab-
bits in Australia should be a clue as to
how rapidly rodents would have spread
when they reached Australia (or the
equally island continents of Eocene
South America and Africa).

"That is, the present diversity of the
Pseudomyinae represents the maximum
differentiation that members of this rap-
idly reproducing and rapidly evolving
order could achieve, in an area of conti-
nental size, with no real competition of
any kind, in about 4.5 million years"
(Wood, in press b).

The comparison of the diversification

Oligocene Rodents of Boli\ l\ • Patterson and Wood 471

of the Fayum thryononiyoids with that of
the Australian pseudoniyines strongly
implies that rodents had been in Africa
no more (and probably significantly less)
than 4.5 m.y. before the deposition of the
lower fossil wood zone. This would mean
that rodents first reached Africa no earlier
than late in the Eocene. The much great-
er diversity and abundance of the Desea-
dan caviomorphs is in accord with their
having been in South America two or
three times as long, probably having
reached South America 10-12 m.y. be-
fore the Deseadan, presumably immedi-
ately after Musters time (Wood, in press
b). Therefore, derivation of the Cavio-
morpha from the thryononiyoids would
appear to us to be an untenable hypoth-
esis.

As we have pointed out before (1959:
397, n. 32), if transatlantic rafting of ro-
dents had occurred, on faunal grounds it
would be more likely to have been from
South America to Africa. Now as then, we
do not take this possibility seriously.
Platyrrhine primates also make their first
appearance in the Deseadan, and it is not
surprising that a nearly identical gamut
of opinions regarding their dispersal has
been run. Hoffstetter (e.g., 1972) has ad-
vocated rafting from Africa, while Szalay
(1976: 426) has suggested rafting in the
reverse direction. In answer, Hoffstetter
(1977) cited the much greater abundance
and diversity of the Fayum primate fauna
as a telling argument against such a view.
He is quite right — and precisely this
same argument is equally valid evidence
against the possibility of a rafting of
thryononiyoids to South America. Rela-
tionships and place of origin of the platyr-
rhines are open and lively questions.
Transatlantic rafting of the ancestral
stock is perhaps a little less unlikely than
rafting of ancestral caviomorphs (there is
not so much evidence against it), and so
cannot be ruled out entirely. Neverthe-
less, we adhere for the present (Wood,
1981) to the third and older alternative,
that catarrhines and platyrrhines have de-

scended from a common ancestral stock
that lived in the southern parts of the
northern hemisphere, and which reached
Africa and South America independently.
For further discussion of the difficulties
involved in transatlantic rafting, see Si-
mons (1976), who has infomied us (per-
sonal communication) that in his opinion
such resemblances as exist between the
Deseadan primate Branisella and the Fa-
yum primates are of a general nature and
indicate no special affinities.

With that, we may proceed to a consid-
eration of certain other items of evidence
that have been cited to support a thryon-
omyoid ancestry for the caviomorphs.

The Evidence from Parasitic
Nematodes

Hoffstetter (e.g., 1975: 522-523) and
Lavocat (1973: 256) have indicated that
evidence from certain parasitic nema-
todes favors an African origin for the Ca-
viomorpha. Use of the data from these
parasites is complicated by the complete
absence of a fossil record to provide a
time scale for nematode evolution, by the
fact that there are probably very large
gaps in our knowledge of present-day
nematodes and their distribution, and by
the fact that it is as difficult for paleon-
tologists to interpret the data from para-
sites as for parasitologists to interpret
those from fossils. Nevertheless, since
nematodes have been cited to support
thryonomyoid-caviomorph relationships,
we have ventured into this gray area of
interpretation, feeling that, at least, we
can speak on the subject with as much
authority as do Hoffstetter and Lavocat.

The nematode groups in question are
the Heligmosomidae (Durette-Desset,
1971) and the Oxyurinae of the Family
Oxvuridae (Quentin, 1973a, 1973b,
1973c).

Durette-Desset's concept of the Helig-
mosomidae would seem to be that of a
grade rather than a clade, since she evi-
dendy believed (1971: Figs. 95-96) that

472 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

tlie tnolutionan- "lignees" (all but one
consisting ot separate suhianiilies) arose
independently from different genera of
the ancestral Family Trichostrongylidae,
and developed their heligmosomid fea-
tures as parallelisms. The major host
group is the Rodentia, heligmosomids
being reported in 17 of the living fami-
lies.

Among Old World hystricognaths,
Thryoiiomys is infected by HeligmoneUa
(Heligmonellinae), the hystricids by He-
ligmotiella and by Paraheligmonina
(Brevistriatinae) in Africa and by Cordi-
cauda (Brevistriatinae) in the Oriental
Region, and the bathyergids by Ort-
leppstrongylus (Impalaiinae). There are
no records for Petromus.

Within the caviomorphs, there is a
sharp cleavage. Cuniculidae, Caviidae,
Hydrochoeridae and Chinchillidae har-
bor Vianella (Vianniinae) whose other
hosts are didelphids; presumably a trans-
fer from the latter to the fonner was in-
volved here. Echimyidae, Dasyproctidae
and Erethizontidae are afflicted by mem-
bers of the Subfamily Pudicinae, the
echimyids by Pudica, Heligmostrongy-
lus and Pseudoheligmosomum (Capro-
myinae, only, for the last), the dasyproc-
tids by Pudica and Heligmostrongylus,
and the erethizontids by Heligmostron-
gylus. In addition, Paraheligmonella (ac-
cording to Durette-Desset, 1971: 50, the
reference of this genus to the Heligmo-
nellinae is questionable) has been re-
corded from the echimyids Cercomys
and Capromys, and a species of Stile-
strongylus (Nippostrongylinae), a genus
otherwise confined to Neotropical crice-
tids, occurs in an echimyid, presumably
another case of transfer. Heligmosomids
have not been reported from the Octo-
dontidae, Abrocomidae or Dinomyidae.

Durette-Desset doubted that host spec-
ificity is of primary importance in the
Heligmosomidae, believing, rather, that
diversification of the family came about
mainly by various evolutionary "bursts"
that accompanied expansion of suitable

host groups. She postulated that helig-
mosomids were not in existence in the
Paleocene (1971: Fig. 86), the various
"lignees" having arisen during the
Eocene. The "lignee" of interest in the
present connection is that of Heligmo-
neUa. From the Subfamily Heligmonel-
linae there arose, in the late Eocene or
early Oligocene according to Durette-
Desset, the Pudicinae in South America
and the Brevistriatinae in the Old World,
and, in the later Tertiary, coinciding with
the rise of the Muridae, the Nippostron-
gylinae. With the exception oi Parahelig-
monella in certain echimyids, the helig-
monellines are Old World in distribution
(Palaearctic talpids, Palaearctic and Ori-
ental lagomorphs, and an Oriental sciurid
in addition to Thryonomys and African
hystricids).

However, Durette-Desset (1971: 50)
apparently considered it uncertain as to
whether Paraheligmonella was actually
a heligmonelline. "Le fait que nous clas-
sions ce genre dans la meme sous-famille
qn HeligmoneUa (mais en le considerant
comme un pen plus evolue) implique
une invasion de I'Amerique du Sud par
les Rongeurs africains. Nous devons in-
sister sur la fragilite de notre hypothese,
puisqu'elle ne repose que sur la structure
du synloph. D'autres elements, an con-
traire, sont deja assez characteristiques
de la faune sud-americaine (division de
la dorsale, aretes discontinues)." It is this
very, very tentative suggestion, that the
subfamily crossed the South Atlantic in-
side rafted thryonomyoids, that Lavocat
and Hoffstetter have emphasized.

Other interpretations would seem pos-
sible. In the first place, as we understand
Durette-Desset's reservations, it seems
entirely possible that her very tentative
reference o{ Paraheligmonella to the He-
ligmonellinae may be incorrect, and that
the features uniting Paraheligmonella
and the Heligmonellinae evolved as par-
allelisms, an amount of parallelism ap-
parently no greater than that which she
accepts in the parallel evolution of the

Oligocene Rodents of Bolivia • Patterson and Wood 473

subfamilies from the ancestral Tricho-
strongylidae. If so, heligmosomid sup-
port for transatlantic rafting collapses at
the outset. Secondly, we do not see that
there is any basis for assuming that the
heligmosomids were not in existence
during the Paleocene, especially since
didelphid marsupials and insectivores,
two other groups parasitized by them,
were then present, and remembering the
lack of host specificity that Durette-Des-
set stresses. In discussing the relation-
ships between heligmosomid evolution
and rapid radiations of host groups, Dur-
ette-Desset has omitted the most impor-
tant rodent evolutionar\' "burst" of all,
the protrogomorph-franimorph radiation
of latest Paleocene and early Eocene
time. Rodents being the major host group
(approximately 80% of the species listed
by her infest members of this order), it
seems unlikely to us that heligmosomids
would not have profited from this radia-
tion. We suggest, therefore, as an alter-
native possibility, that heligmonellines
arose in Paleocene time in Holarctica,
parasitized the ancestral hystricognath
infraorder Franimorpha and their de-
scendants, and subsequently gave rise to
the Pudicinae in South America and to
the Brevistriatinae in the Old World.
Parahelig77ionella, if correctly referred to
the Heligmonellinae, would be, in this
view, a surviving New World heligmo-
nelline, not a transatlantic immigrant
from the Old World.

In passing, it may be stressed that
bathyergids are parasitized by members
of the Subfamily Impalaiinae, which
does not infest any other Old World hys-
tricognaths. We believe that this was con-
sidered by Durette-Desset to be the most
primitive of the "lignees"; she indicated
that the subfamily originated in the mid-
dle Eocene (1971: 98), and in her phy-
logenetic chart (1971: Fig. 95) she shows
this as the left hand "lignee." If Lavocat
was correct (1973: 256) in his opinion that
the distribution of nematodes is a key to
host interrelationships, we feel that these

data should have given him pause in at-
tempting to derive the Bathyergoidea
from the Thryonomyoidea (1973: 150-
153).

In the case of the Oxyurinae, the major
host group is again the Rodentia; mem-
bers of 13 families are at present known
to be infested by oxyurines. Other hosts,
so far as recorded, are: Marsupialia (Aus-
tralia only), Dermoptera, Lemuroidea,
Platyrrhini, Leporidae and selenodont
Artiodactyla. This is a very mixed bag,
which suggests that knowledge of the
distribution of the subfamily may be far
from complete. No oxyurines have so far
been recorded for either Thnjonomys or
Petronius. Four genera — Hilgertia, Het-
eromijoxyuris, Helminthoxys and Evagi-
nuris — are of interest in the present con-
nection.

Hilgertia, which is not very closely re-
lated to the others (Quentin, 1973c: Figs.
21, 23-24), is known from Ctenodactylus
and Heterocephalus. Ctenodactylids were
certainly of Asian origin (Wood, 1977a)
and we believe bathyergids were, as well
(see below, p. 516).

The other three genera, which form a
closely related group, are, with the pos-
sible exception of two species of one of
them. New World fomis. Heteromyoxy-
uris infests heteromyids, a North Ameri-
can family which does not, as Quentin
stated (1973c: 1092), date from the early
Eocene but from the early Oligocene.

The various species of Helminthoxys
are divided by Quentin (1973c: 1081)
into four "lots." The first of these infests
Lagidium, the second Capromys, the
third Microcavia C'Caviella"), and the
fourth Dasyprocta and Cercomys. Rely-
ing on our diagram (1959: Fig. 34) of ca-
viomorph relationships that hypothe-
sized early Miocene dates of origin for
the Caviidae and the Capromyidae (in
which, following a consensus, we then
included the myocastorines), Quentin
suggested (1973c: 1092) that infestation
of these by two of the principal "lots" of
Helminthoxys "indiquerait effective-

474

Bulletin Museum oj Comparative Zoology, Vol. 149, No. 7

nient une evolution plus recente dv ce
genre. " This may have been the case, but
we know of no reason why the caviids
could not have inherited these parasites
from the extinct, broadly ancestral Eocar-
diidae. The Capromv inae, as we now re-
strict them (p. 393), are unknown prior
to the Antillean late Quaternary. The oth-
er two families harboring species of this
genus, like the Eocardiidae, go back to
the Deseadan, so there seems to be no
reason why Hehuitithoxys could not date
at least from the early Oligocene.

Species of Evaginuris have been re-
corded, in the New World, from EretJii-
zon, Coendou, and Ditioinys. The type
species, the North American Evcuiginuri.s
compar, has been found in the domestic
cat (the source of the type material) and in
Odocoileus, as well as in Erethizon dor-
satum and E. epixanthum. With this dis-
tribution, one wonders where else the
species may occur. Interest here, how-
ever, centers on the two Old World re-
ports of species of this genus.

Evaginuris evoluta (von Linstow) was
originally — and summarily — described as
a new species of Oxyuris (von Linstow,
1899: 20). When Quentin examined the
type material, he found it to be "en mau-
vais etat," "3 femelles dessechees, en
fragments .... Ce materiel a pu etre re-
hydrate en partie . . . par le procede du
phosphate monosodique" (1973c: 1068).
How much of the original structure is re-
capturable by this technique is not for
non-specialists to say. However, there is
some uncertainty as to the provenance of
these fragments, although Quentin
(1973a: 2015) cited E. evoluta as "para-
site d'un Hystrix hrachyura originaire de
Malaisie." Von Linstow, however, mere-
ly listed '^Hy.strix hrachyura (?)*" with-
out further details, although for other
species of nematodes that he described
in the same paper he was meticulous in
recording localities and collectors. He
added, in the footnote: "*Manuscript-
name, dessen Identitat nicht mehr fest-
zustellen ist." We are thus left with un-

answered questions. Was von Linstow in
doubt as to whether the manuscript
name, i.e., the host datum, was correctly
associated with the caecum in which he
found the oxyurine specimens, or was it
the identification of the host that he ques-
tioned? If the host datum was essentially
correct, could the animal have been a zoo
inmate, housed in a cage adjacent to that
of, perhaps even shared with, some er-
ethizontid?

A second species of Evaginuris, E.
stossichi (also originally referred to Oxy-
uris), has been reported from Hystrix
cristata (Quentin, 1973a: 2015). Again,
there seem to be problems with this
species, which was originally given a
very incomplete description (Quentin,
1973b: 1403). In 1956 "Vuylsteke . . .
identifie a nouveau I'espece chez un
Hystrix cristata du Jardin Zoologique
d'Anvers . . ."; these specimens were
studied by Quentin and were the basis
for his transferral of the species to Evagi-
nuris. Since this specimen of Hystrix
cristata was definitely a zoo inmate, the
question raised above about the source of
the Evaginuris evoluta infestation is
equally applicable here.

We cannot answer the questions we
have raised, but it seems obvious to us
that the records are somewhat suspect,
and that they should be confinned from
field collected specimens before they can
be accepted at face value.

Even should the records be confirmed,
there would be no necessity to invoke
transatlantic rafting to account for the oc-
currence of Evaginuris in New and Old
World porcupines. Quentin believed that
Evaginuris was descended from Hcter-
omyoxyuris or something very similar.
The presence of this last genus in the
wholly American (and, in the Oligocene,
exclusively North American) Heteromyi-
dae argues strongly that members of this
group of oxyurine genera were present in
the North American Eocene. They could
not have been parasites on the Hetero-
myidae, nonexistent at that time. Some

Oligocene Rodents of Bolivia • Patterson and Wood 475

extinct family or families ol rodents were
presumably involved. If these had a Hol-
arctic distribution, Heteromyoxyuris and,
very possibly, Evaginuhs, also had a sim-
ilar distribution, and could have inde-
pendently transferred to Asian hystricid
descendants of the ancestral rodent hosts,
as well as to Middle America, where they
would have been ready to migrate to
South America with the ancestral cavio-
morphs. It seems to us that those who are
not willing to accept this view must face
the difficult paleobiogeographic prob-
lem of explaining how the ancestral
Evaginuris was able to reach the Old
World tropics after having evolved from
the North American Heteromyoxyuris.

Furthennore, if it is proposed that the
distribution of Evaginuris be considered
a support for the hypothesis of the deri-
vation of the Caviomorpha from the
Thryonomyoidea, it seems most surpris-
ing to us that the Old World members of
Evaginuris should be confined to hystri-
cids, absent from Africa, so far as anyone
knows, until late Tertiar>% and be un-
known in the thryonomyoids, the group
some of whose members supposedly
crossed the South Atlantic carrying Evag-
inuris.

In summary, the evidence from nema-
todes cited as favoring transatlantic raft-
ing is shaky indeed, and the support it
offers for an African origin of the Cavio-
morpha has been grossly exaggerated.
Further, even if the evidence for close
relationships of the nematodes of cavio-
morphs and Old World thryonomyoids
were accepted as being as clear-cut as in-
dicated by Hoffstetter (1975: 522-523)
and Lavocat (1973: 256), the data are ca-
pable of other, at least equally plausible,
interpretations.

Comparisons of Caviomorpha and Old
World Hystricognathi

We may now pass to a consideration of
the hystricognathous rodents themselves,
investigating the resemblances between

the New and Old World forms. The
points of similarity are numerous, which
of course accounts for the controversies
that have raged about the degree of re-
lationship. There are three questions.
First, which of these similarities are con-
fined to the Caviomorpha and the Old
World hystricognaths, and which are
found as well in one or more of a variety
of sciurognathous rodents? Secondly,
which features are found in the North
American Eocene hystricognaths, and
which are limited to the Oligocene and
later caviomorphs and Old World hystri-
cognaths? Thirdly, are the resemblances
that are confined to the Oligocene and
later hystricognaths of so overwhelming
and so exclusive a nature as to demand
direct descent of the caviomorphs from
hypothetical middle Eocene African
thryonomyoid ancestors, or could they be
the results of independent descent from
a common, northern hemisphere frani-
morph ancestry combined with parallel
evolution occurring subsequent to geo-
graphic separation? We shall discuss the
answers to the three questions together.

(1) General. Lavocat (1973: 168) has
listed, without discussion, 16 "principal
common characters" of the Old World
hystricognaths (his Phiomorpha) and the
Caviomorpha. We believe that he omit-
ted discussion because he had recently
covered some or all of these points in de-
tail (Lavocat, 1971a), although he did not
cite this paper in this connection. These
characters — some of them emphasized as
early as 1839 by Waterhouse — and others
have been reviewed repeatedly (Alston,
1876; Winge, 1887; Tullberg, 1899; Lan-
dry, 1957; Lavocat, 1974b, 1976; Wood,
1974a, 1975a; Hoffstetter, 197.5— to give
only a partial list). We regret the neces-
sity of again going over this ground.

Lavocat's 16 characters are discussed
below, in a somewhat different order
than that in which he listed them; the
italicized statements heading each sec-
tion are his identifications of the charac-
ters, in translation. We have added a few

476 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 7

other characters that he did not mention, parallehsni (Wood, 1975a: Fig. 1). Fur-
In the course ofour comparisons we have thermore, if hystricognathy can be as-
examined approximately 1000 skulls rep- sumed to have developed more than once
resenting nearly all Recent hystricognath among rodents as parallelisms, we be-
genera, mostly from the Museum ot Com- lieve that there is no justification what-
parative Zoology. ever for assuming special relationship

Lavocat (1971a) went into some detail between the Caviomorpha and the Old

as to structural similarities between the World hystricognaths. We prefer to ac-

caviomorphs and the Old World hystri- cept, as a working hypothesis, the pos-

cognaths, which he felt were too great tulate that hystricognathy originated only

(op. cit.: 521) to have developed as par- once, in the late Paleocene, probably of

allelisms. He stated that the occurrence North America, and that it began as the

of the same features in sciurognathous condition that Wood has variously re-

fonns does not prove relationship. This ferred to as "incipiently hystricognath"

is the problem that faces the student of (1962a: 117, 122; 1974b: 31, 42-43;

rodents — when are identical structures 1977b: 104; 1981: 81, 82), "hystricognath

proof of close relationship, and when condition" (1962a: 142), or "subhystri-

does identity merely indicate parallel- cognathous" (1975a: 79; 1981: 86), rather

ism? We have tried to find data support- than springing into existence full-blown,

ing or contradicting Lavocat's statements, like Athena (Wood, 1980c: 270). We

The necessity of providing documenta- therefore include in the suborder all

tion for our disagreement with some of forms that are subhystricognathous or

his conclusions has considerably length- hystricognathous, believing them to have

ened this section of our paper. been derived from late Paleocene frani-

(2) Hystricognath mandible. This is morphs such as Franimys. Since some
the basic character of the Suborder Hys- franimorphs were fully hystricognathous
tricognathi, and the only feature known (Wood, 1972: Fig. 1; 1974b: Fig. 5), in-
to us that is universally present in hys- heritance from a common ancestor is, at
tricognaths and never present in sciuro- the very least, as plausible a hypothesis
gnaths. Hoffstetter (1973: 159, 161) to explain the possession of such mandi-
seemed willing to accept the multiple bles by the Caviomorpha and the Old
origin of hystricognathy, not believing World hystricognaths as is one postulat-
that Prolapsus or Tsaganomys had any- ing a direct thryonomyoid-caviomorph
thing to do with the Hystricognathi. He relationship.

eliminated Tsaganomys from the Hystri- One of the major changes that devel-

cognathi because it was not hystricomor- oped among rodents in general during

phous, but apparently was willing to ac- the Eocene was the increase in gnawing

cept the equally non-hystricomorphous ability. This involved a number of struc-

bathyergids as members of the Hystrico- tural changes: the forward growth of parts

gnathi. It is possible that hystricognathy of the masseter, increasing the length of

evolved more than once in the course of that muscle for more efficient gnawing

rodent evolution, but we believe that ac- (Wood, 1965: 116-119; Lavocat, 1971a:

ceptance of such an occurrence would 519); in some fomis, the lengthening of

mean that there are no morphological fea- the M. pterygoideus internus, which

tures (and probably no non-morphologi- served the same function as the length-

cal ones) that can be used in establishing ening of the masseter; and the modifica-

subordinal divisions of the Order Roden- tion of the histology of the incisor enam-

tia, since all other structures have already el, increasing its strength (Wilson, 1972:

been shown to have evolved indepen- 220-222). One very important point to

dently, in various groups of rodents, by note here is that these changes did not

Oligocene Rodents of Bolivia • Patterson and Wood 477

occur instantaneously as the result of a tence — is dependent on the size, and
single mutation, but built up gradually hence importance in the chewing func-
over considerable periods of time. Nor tion, of two masticatory muscles. These
were they universally correlated with are M. masseter lateralis profundus, pars
each other, but rather each feature posterior, deep division, whose fibers in-
evolved independently. Any slight shift sert on the lateral surface of the ascend-
in any one of these features would have ing ramus and on the posterior margin of
been selectively advantageous. All of the post-condyloid process; and M. pter-
these changes may be observed taking ygoideus externus, which inserts below
place among North American Eocene the condyle and on the medial face of the
franimorphs. process (Woods, 1972: 127-128, 130-

We do not know whether the Cavio- 131). Dependence of the process on the
morpha, the Thr>'onomyoidea, the Bathy- muscles is neatly demonstrated by
ergoidea and the Hystricidae developed Ctenomys, "in which the pterygoideus
full hystricognathy independently from externus is somewhat small, the expan-
subhystricognathous franimorph ances- sion has moved laterally and exists as a
tors, inherited it directly from hystrico- tubercle" (Woods, 1972: 131) below and
gnathous franimorphs, or whether the lateral to the condyle (Ellemian, 1940:
Old World fomis inherited it from a com- Fig. 38) for insertion of the deep division
mon ancestor later than their common oi \he M. masseter lateralis. The \ery ire-
ancestor with the caviomorphs. There is quent association of the process with hys-
at present no way of choosing among tricognathy is, we suspect, because the
these choices other than following one's development of the process has permit-
personal preferences. ted the anteroposterior lengthening of

A mandibular feature, not mentioned the M. masseter lateralis profundus,

by Lavocat, that very frequently accom- pars posterior, deep division, and of the

panics hystricognathy, although neither M. pterygoideus externus, serving the

universally nor exclusively, is a projec- same function as the similar lengthening

tion extending posteriorly behind and of M. pterygoideus internus, that pro-

below the condyle. This, the post-con- duced hystricognathy. We suspect that

dyloid process of Woods (1972: Fig. lA), the process arose independently in the

varies in degree of prominence from ab- Old and New Worlds, although it is pos-

sent (or at least scarcely defined) in, e.g., sibly present in some franimorphs other

bathyergids, Octodon, Echimys and Tri- than Reithroparaniys huerfanensis, and

chys, to large in, e.g., Kannabateomys, may or may not have been present in the

Dasyprocta, Dinomys and Hydrochoe- numerous franimorphs in which this area

rus, among living fomis. It is slightly to has not been preserved,

moderately developed in Santacruzian Although not strictly relevant to the

caviomorphs and large in the East Afri- question at issue, another direct conse-

can Paraphiomys, but small in Petronius quence of hystricognathy or sciurognathy

and Thryonomys. Platypittamys pos- deserves mention. This is the course tak-

sesses a very small post-condyloid pro- en by the lower incisor in those forms in

cess, Incamys one of moderate size, which the tooth has lengthened poste-

Within the Franimorpha, this area has so riorly. Since the axis in the posterior por-

far been described only in two species of tion of this tooth in rodents is aligned

Reithroparaniys. The process is lacking dorsolaterally, the elongating base in

in R. delicatissimus, incipient in R. huer- hystricognaths can only terminate be-

fanensis (Wood, 1962a: Figs. 41D, 46F). neath the condyle, which is situated

The degree of prominence of the post- above it. In those sciurognathous forms

condyloid process — indeed its very exis- in which elongation occurs, the base of

478 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

the incisor can only extend upward lat- gatiouiys. The skull is not known for any

eral to the condyle, which is situated me- Fayuni thryononiyoid, but we suspect

dial to the axis. The result is that, de- that their pterygoid fossa opened into the

pending on the extent of elongation orbito-temporal fossa,
achieved, the base fonns a protuberance Among the few known Deseadan ca-

on the outer side of the ascending ramus, viomorph skulls, only one specimen is

ranging from small and low to so large as sufficiently complete and adequately

to reach the level of the condyle. enough prepared to reveal details of the

(3) Pterygoid fossa opens anteriorly pterygoid region. This is PU no. 21945,

into the orbitotemporal fossa. The pter- Incamys holivianus (p. 409). Here, con-

ygoid fossa of hystricognaths is bounded nection with the orbit is definitely pres-

by the alisphenoid laterally, the palatine ent. The character was probably present

ventrally, anteriorly, and medially, and in all Deseadan caviomorphs.
by the pterygoid posterolaterally; de- In the three described early frani-

grees of participation of these elements morph skulls, those of Franimys, Rei-

are highly variable, as are the size and tJiroparamys and Protoptychus, this area

shape of the fossa. "The main trends ... is either damaged or missing. Wahlert

are toward elongation of the internal {1973: 8) remRrked, oi Protoptychus, that

pterygoid and horizontal positioning of the fossa "... is very deep, and inade-

the muscle. These trends [result from] quately preserved for full description."

the movement of the origin of the muscle The breakage in all of these skulls may

deep into the pterygoid fossa, and the have occurred because the area was

movement of the insertion to the tip of weakened by perforation of the fossa, but

the extended and laterally displaced an- there can be no certainty on this point at

gle" (Woods, 1972: 131). present.

The opening into the orbito-temporal With the opening of the dorsal side of
fossa is present in all Recent hystrico- the fossa, the origin of the M. pterygoi-
gnaths except Hydrochoerus, in which it dens internus passes forward mediad of
has been secondarily blocked by the rear- a bridge of bone of varying length, the
ward extension of M^. The opening is a alisphenoid bridge of Woods (1972: 130).
feature much easier to detect in living The bridge is formed in large part by the
than in fossil specimens. It is present in alisphenoid, dorsally, and to a much less-
all early Miocene Santacruzian cavio- er extent by the pterygoid, ventrally.
morph skulls in which the area is ade- Woods reported the bridge as lacking in
quately preserved; Scott (1905: 394) Erethizon. However, a large series of
doubted its occurrence in Neoreoniys, specimens in the MCZ reveals the bridge
but we have found it to be present at the to be highly variable, ranging from pres-
bottom of a deep and very narrow ptery- ent, but small and delicate, to interrupted
goid fossa. An opening definitely occurs in the center, to absent with no separate
in the East African early Miocene Para- alisphenoid canal. Reduction is certainly
phioniys pigotti and Bathyergoides neo- secondary. A specimen of Echinoprocta
tertiarius (Lavocat, 1973: PI. 24, Figs. 3, shows a well-developed bridge and ca-
4; PI. 7, Figs. 1, 5), and we assume one nal. Within a series of Coenrfow, variation
to have been present in other contem- extends from a well-developed bridge
porary African hystricognaths, although with a canal, to a delicate one without a
Lavocat's figures do not show the open- canal.

ing in Paraphiomys stromeri (Lavocat, The area of origin of M. pterygoideus

1973: PI. 24, Fig. 1) or Diamantomys internus extends forward to the vicinity

luederitzi (Lavocat, 1973: PI. 24, Fig. 2). of the sphenopalatine foramen, in various

The character is fully developed in Tsa- fomis lying in a conspicuous channel in

Oligocene Rodents of Bolivia • Patterson and Wood 479

the adjacent bone. As the muscle en-
larged, its passage encroached upon the
alisphenoid canal. In some forms passage
and canal are separated by a delicate bar
of bone within the passage; in others a
small projection remains to mark the for-
mer position of the separation; in yet oth-
ers all trace of a division has been lost.
In extreme cases, as in bathyergids, the
origin of the muscle extends, via an open-
ing, into the brain case as well. Where
this occurs, the basisphenoid and the
posterior portion of the presphenoid rise
steeply to their dorsal midline, and fonn
a narrow ridge flanked by depressions
for the origin of the muscles.

Deepening and, ultimately, perforation
of the fossa is a character definitely, al-
though not exclusively, associated with
hystricognathy. As is well known, it oc-
curs elsewhere in the order among geo-
myoids and in Spalax; it is also present
in Aplodontia, in which the perforation
is large. Anterior to it and lateral to the
sphenopalatine foramen in Aplodontia is
a shallow but clearly defined channel as
in various hystricognaths. We have not
had the opportunity to dissect preserved
material, but in a series of skulls we have
detected dried muscle fibers within the
passage and attaching to the channel
leading forward from it, presumably in-
dicating that it carried M. pterygoideus
internus. The perforation in Aplodontia
has usually been considered, as by Wah-
lert (1974b: 404), to be a very large ali-
sphenoid canal. The internal maxillary
artery and vein no doubt pass through to-
gether with the muscle, but, again as in
various hystricognaths, no trace of a bony
separation remains.

In all forms with a perforated pterygoid
fossa, the angular region of the mandible,
either the whole, as in hystricognaths, or
the posterior part, as in the sciurognath-
ous fonns, is everted. The presence of a
perforated pterygoid fossa in some sci-
urognaths adecjuately demonstrates, we
believe, that this structure was capable of
evolving independently, as parallelisms.

among rodents. Furthemiore, there is no
evidence as to whether the perforation of
the fossa arose among Eocene frani-
moqohs, or not until the establishment of
the modern infraorders.

We believe that hystricognathy began,
in the early Reithroparamyidae, as a
means of increasing the length of M.
pterygoideus internus, because of its im-
portance to the functioning of the lower
incisors. At this stage, even ver>' slight
and incipient hystricognathy would have
been of selective value and hence pre-
served. Selection for increasing incisor
usage would have resulted in lengthen-
ing the internal pterygoid, which could
have taken place either by moving the
origin of the muscle (deepening the fos-
sa), or moving its insertion (developing
hystricognathy), or both. We believe that
these two features (hystricognathy and
deepening of the pterygoid fossa) evolved
together, as secondary results of the mus-
cle elongation (Lavocat, 1971a: 519;
Woods, 1972: 131; Wood, 1980c: 269-
270).

(4) Inferior border of the zygomatic
arch very often bearing an elevated ver-
tical blade for the origin of the masseter.
What Lavocat refers to here, at least in
part, is evidently the zygomatic masse-
teric fossa of Landry (1957: Fig. 11), the
jugal fossa of Woods (1972: Fig. 1). This
depression, when present, serves, togeth-
er with the ventral margin of the poste-
rior part of the malar, as the area of origin
of M. masseter lateralis profundus, pars
posterior, deep division (Woods, 1972:
127-128).

Among Eocene rodents, there is a well-
developed jugal fossa in Reithroparamys
delicatissimus, the only franimorph for
which the zygomatic arch has been de-
scribed (Wood, 1962a: Fig. 41C). A fossa
is also present in Paramys delicatus and
P. copei (Wood, 1962a: Figs. 2B, 13B),
Ischyrotomus horrihilis and /. oweni
{op. cit.: Figs. 68C, 71C), and Sciuravus
nitidus (Dawson, 1961: PI. 5). The jugal
fossa is poorly developed to nonexistent

480

Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

in Thisbemys corrugatus (Wood, 1962a:
Fig. 36A), the only other Eocene rodent
in which this area has been figured. It
would thus seem that presence of a jugal
fossa is a primitive rodent character.

In hystricognaths, the character is an
extremely variable one. Among living
Old World hystricognaths, the fossa is ab-
sent in the Bathyergidae and weakly, if
at all, developed in the Hystricidae.
Woods (1972: 127) stated that the fossa is
absent in Thryonomys but present in
Fetromus, although no part of the muscle
takes origin from it. We were unable to
identify a jugal fossa in Petromus, but
found that both genera have a more an-
teriorly situated depression in the lateral
side of the zygomatic arch. This lies
mainly in the maxilla, which in these
fomis extends farther back in the ventral
portion of the arch than is generally the
case in other hvstricognaths (Ellerman,
1940: Figs. 26-27, 29-30). This anterior
depression is larger in Thryonomys, but
other than that we can detect no differ-
ence between the two fomis, and consid-
er both as lacking a jugal fossa in the
strict sense. Lavocat (1973: 168) and
Woods (1972: 127) apparently both
equated the anterior depression of Petro-
mus with the true jugal fossa, thus lump-
ing two different bone-muscle relation-
ships as one supposed character.

Among caviomorphs, the jugal fossa is
absent or virtually so in the Erethizonti-
dae, Chinchilloidea and Cavioidea, being
essentially limited to the Octodontoidea.
Even within this superfamily it varies
from absent to large in the Octodontidae
and from small to large in the Echimyi-
dae.

The East African Miocene Paraphio-
m,ys and Diamantomys in general resem-
ble Petromus and Thryonomys, with the
exception that the more anterior (not the
jugal) fossa presents more ventrally. Such
Santacruzian erethizontid, cavioid and
chinchillid specimens as are sufficiently
well preserved to show this region lack
the jugal fossa, as does the Deseadan ca-
vioid Incamys. Of fossil octodontoids,

the Santacruzian echimyid Acarechimys
minutus (see Appendix 2) has a small fos-
sa (Scott, 1905: PI. 67, Fig. 10b) while the
Deseadan octodontid Platypittamys
shows no trace of one.

Since the jugal fossa serves as part of
the area of origin of the deep division of
M. masseter lateralis, pars posterior, and
the post-condyloid process of the man-
dible as part of the area of its insertion,
one might expect a fair degree of corre-
lation between the two. This is definitely
not the case, however. The process varies
from small to large in those caviomorph
superfamilies in which the fossa is essen-
tially lacking, while within the Octodon-
toidea every change is rung from absence
of both, to small size of one and large size
of the other, to large size of both.

Presence of a jugal fossa is, therefore,
not of any significance at higher taxonom-
ic levels, and is, in any case, apparently
a primitive rodent characteristic. The
myological character associated with it,
the high degree of differentiation of the
deep from the superficial division of M.
masseter lateralis, pars posterior, is sig-
nificant. It occurs in all living hystrico-
gnaths that have been investigated in suf-
ficient detail to reveal it. There is some
evidence that the division of this muscle
may be associated not only with hystri-
cognathy, but also with hystricomorphy,
since the hystricomorphous but sciuro-
gnathous ctenodactylids possess a well-
defined jugal fossa. Whether or not this
is accompanied by any marked degree of
differentiation of the deep from the su-
perficial division of M. masseter later-
alis, pars posterior, apparently is not
known.

These data indicate that the presence
of a jugal fossa is not a hystricognath char-
acter; it could have been independently
inherited from Eocene franimorph ances-
tors in those fomis where it is present; or
it could have been secondarily reac-
quired independently in the octodon-
toids and hystricids.

It should be pointed out that modifi-
cation of the masseter complex began

Oligocene Rodents of Bolivia • Patterson and Wood 481

very early in rodent histor>', as indicated uosity of the suture does not necessarily
not only by the hystricomorphous frani- accompany heaviness of the malar, and
morphs but also by the large jugal fossa both features are extremely varied among
borne on a descending flange, which ex- the Hystricognathi. The bathyergids dis-
tends over the entire length of the malar play neither. In hystricids, the suture var-
in the Bridgeran manitshine paramyid ies from simple to slightly sinuous, and
IscJnjrotomus Jiorribilis (Wood, 1962a: the malar may be either heavy {Atheru-
Fig. 68C). rus, Hystrix), moderately heavy {Thecu-

(5) In both, the infraorbital foratnen is rus), or comparatively slender (Trichys).
of the hystricomorphous type. This is in- Thryonomys has a very heavy malar and
deed a resemblance between the New a moderately to highly sinuous suture,
and some of the Old World members of Petromus one that is but little expanded
the suborder, involving, as it does, the with only a slightly sinuous suture,
same structural modifications and for- Among living caviomorphs, heaviness
ward migration of the origin of M. mas- ranges from extreme (e.g., Chaetomys) to
seter medialis, but it is not universal nonexistent (e.g., Echimys), and sinuos-
within the suborder — the protrogornor- ity of the suture, as, e.g., in Dinomys,
phous Tsaganomyidae and Bathyergidae Myocastor and Spalacopus (which has a
(p. 517) lack it, and the foramen is very slender malar) is the exception, not the
small for a hystricognath in Platypitta- rule.

mys (Wood, 1949: Fig. 2; 1974b: Fig. 3A). The East African Paraphiomys (Lavo-

Nor, as is very well known, is this con- cat, 1973: Figs. 2A, 3A) and Diamanto-

dition confined to the Hystricognathi. It mys (Lavocat, 1973: Fig. 4A,B) possess

occurs in the European Eocene to Oli- malars ranging from moderately slender,

gocene Theridomyoidea; the Asian, Af- anteriorly, in P. stromeri and P. pigotti,

rican and European Ctenodactylidae; the to heavy, and the suture resembles that

African Anomaluridae and Pedetidae; the of Petromus. It is not preserved in any

Eurasian, North American and African other described Oligocene or Miocene

Dipodoidea; and the African glirid Gra- thryonomyoid.

phiurus — all sciurognathous fonns. Pro- The Santacruzian caviomorphs, like

toptychus. Prolapsus and, very possibly, their living relatives, display variability

Rapamys (Wood, 1962a: 148, Fig. 52A) as regards both features. Only in Inca-

show that hystricomorphy was early ac- mys among early Oligocene hystrico-

quired by North American franimorphs. gnaths is any part of this region preserved

We suspect that it may have developed (Fig. 12), and here the suture, although

independently among Asian hystrico- irregular, could hardly be described as

gnaths descended from protrogomor- sinuous. The drawing of the skull of Sa/-

phous franimorphs, and that the fossorial lamys given by Lavocat (1976: Fig. 1)

bathyergoids are descended from frani- shows the zygoma fairly well preserved,

morphs in which hystricomorphy had not but with a straight suture. His stereopho-

become established. As has been dem- tograph (1976: PI. 1, Fig. 4) seems to

onstrated several times (Wood, 1974b: show about the same thing, except that

38-41; 1975a: 75-77; 1977a: 133-136; ^^-,1^ ^^e dorsal and ventral roots of the

1981: 82) hystricomorphy evolved inde- zygoma are present. The onlv franimorph

pendently a considerable number of ,^hose arch area is known is Ke/^/iropara-

times — perhaps as many as ten or i j. ,. . ■ i • i ^i ^ ■

.1 ] 4.U •] A V mys delicatissimus, in which the malar is

twelve — and cannot be considered prooi •'

ofrelationship, except insofar as parallel- slender and the suture is not sinuous

ism indicates some relationship. (Wood, 1962a: Fig. 41C). There is thus

(6) The malar heavy in front, with a no evidence for (but only a little against)
sinuous suture with the maxillary. Sin- a sinuous suture in the early members of

482 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

either the Caviomorpha or the Old World that of the Hystricidae "de fagon inatten-

hystricognaths. due" (Parent, 1976: 244). We ean see no

Among those rodents that are hystri- evidenee in this that caviomorphs were
comorphous sciurognaths, variation in descended from any Old World group; in
anterior heaviness of the malar is also en- fact, the situation in erethizontids, as re-
countered, to an extent equal to that in ported by Parent (1976: 244), can be in-
hystricognaths. The heaviness ranges terpreted as ruling out such a relation-
from nonexistent, as in anomalurids, ship. The degrees of differentiation seen
through moderate, as in such zapodids as within the Old and New World groups
Sicista, to very great, as in Pedetes. This are what would be expected, given a rath-
heaviness is almost never present in sci- er distant common ancestry,
uromorphous or myomorphous rodents. Knowledge of the middle ear cavity in
Castor being a notable exception. It hystricognaths does not extend farther
seems obvious to us that this is a char- back in time than the early Miocene
acter associated with evolving structural (Lavocat, 1973, for Paraphiomys; Vuce-
stresses, whose acquisition has been su- tich, 1975, for Perimys). The bulla is un-
perimposed on a pre-existing hystrico- known in Fayum thryonomyoids, and its
morphous base. Sinuosit>' of the suture dissection has not been practicable in In-
occurs in Paramys delicatus (Wood, caiiiys. Among the known Franimorpha,
1962a: Fig. 2B), but not in other ade- the tympanic is free of the skull only in
quately known Eocene rodents. the very early Franimys. Here, the ven-

Anterior heaviness of the malar and tral aspect of the periotic, although "dif-

sinuosity of its anterior suture seem to us ficult to interpret" (Wood, 1962a: 142) is

to be worthless as indications of cavio- in general similar to that of Paramys, as

morph-thryonomyoid relationships. is to be expected. Comparison with most

(7) Middle ear, malleus and incus of later fomis, in which the tympanic sur-
the same type. Malleus and incus fused, rounds the middle ear cavity, is difficult.
The statement regarding the middle ear being limited in large part to the pro-
is apparently based largely on the inves- montorium. There is a decided differ-
tigations of Parent, whose work is avail- ence in the size and shape of this latter
able to us only in abstract fomi (Parent, structure between the primitive condi-
1976; Lavocat and Parent, 1971). Parent tion exhibited by the early Eocene forms
found that hystricognaths possess various and the more advanced one shown by the
middle ear characters in common, which Miocene genera that lived some 30 m.y.
is not surprising, and that thryonomyids, later. An assessment of this difference
petromurids, hystricids and bathyergids must await some knowledge of interme-
can be distinguished from each other in diate stages, but we suspect that, under
this region, which is somewhat more sur- either hypothesis for the origin of the
prising. Among the Caviomorpha, he rec- Caviomorpha, the common ancestors of
ognized three groups: erethizontids (in the living caviomorphs and the Old
which the subordinal characters are "peu World hystricognaths were not greatly
marques"), octodontoids, and cavioids advanced beyond the paramyid-reithro-
plus chinchillids. Vucetich, however paramyid level.

(1975: 486), found features of the middle Malleus and incus of the same type. As

ear that separate the Chinchillidae from may be seen from the figures and descrip-

the Cavioidea. Not surprisingly, parallel- tions given by Doran (1879: 413-419, PI.

ism crops up in middle ear structure. Cer- 60), Tullberg (1899: PI. 24), Cockerell et

tain hystricognath features are met with al. (1914: Figs. 117-119) and others,

in Anomalurus and Pedetes, and Agouti these ossicles are heavily built and more

("Cuniculus") has a bulla that resembles or less in line dorsally. The head of the

Oligocene Rodents of Bolivia • Patterson and Wood 483

malleus is large and elongate, vatying
from moderately (Bathyergidae) to ex-
tremely so (Chinchilla); the neck is short,
and the lamina small.

Ossicles are frequently absent from
within the bullae of fossil mammals. This
is not surprising in view of field obser-
vations by one of us (B.P.) that fly larvae
in the course of their development can
pullulate through even the smallest
openings in decomposing bodies and are
fully capable of dislodging and moving
out such small objects as ear ossicles.

Knowledge of the ossicula auditus of
hystricognaths does not extend farther
back than the Miocene. Among fossil hys-
tricognaths only three examples of the
malleus and incus have been reported:
Drytomomys aequatorialis (Fig. 23A-C
and Fields, 1957: 343, 347-349; Fig.
23E,F); Perimys sp. (Vucetich, 1975: 484,
488); and Faraphiomys pigotti (Lavocat,
1973: 34; PI. 13, Fig. 4; PI. 40, Fig. 1). All
three sets, as might be expected, are of
hystricognath type. Drytomomys is sim-
ilar to Dinomys, but Perimys, a chinchil-
loid, is closer to Cavia, especially in the
relative shortness of the head of the mal-
leus, than to the living members of its
superfamily. Paraphiomys seems to us to
be close to Petromus; it is certainly very
different from the bathyergids, and the
head of the malleus is more slender than
in Thryonomys. There is, as Lavocat not-
ed (1973: 34), a resemblance to that of
Dasyprocta.

In assessing the ossicular evidence, the
same difficulty is encountered as in the
case of middle ear structures, namely
lack of infonnation, in this case total, re-
garding pre-Miocene evolutionary stages.

Malleus and incus fused. Fusion of the
malleus and incus is not confined to the
Hystricognathi, since the Ctenodactyli-
dae show it also, and Doran (1879) re-
ported it as being variably present in Di-
pus. It would seem to be universally
present in Old World hystricognaths. It
is not a universal feature, however,
among the Caviomorpha. Vucetich in-

fonns us (personal communication) that,
in perhaps 1% of adult Cavia whose
ossicles she has seen, she was able to
separate the malleus and incus without
breakage. We have recorded (1959: 292-
293) a lack of fusion in some adult spec-
imens of Octodon, Aconaemys, Spalaco-
pus and Echimys; and such a lack had
earlier been reported in Proechimys
(Cockerell et al., 1914: 372), and, earlier
still, in Myocastor (Hyrtl, 1845). There is
therefore no justification for the unqual-
ified statement that fusion is universal in
the Hystricognathi.

To the above list of non-fused ossicles,
we now add a young adult of the late
Miocene dinomyid Drytomomys aequa-
torialis. Fields' figures (1957: Fig. 23E,F),
while suggesting fusion, showed a very
definite groove between the bones. Sub-
sequent to his work, the specimen was
badly damaged with accompanying loss
of the manubrium mallei and of the in-
cus, the two ossicles parting at the
groove. A thin film of matrix extended
over the entire articular surface of the
malleus separating it from the incus; fu-
sion had not occurred (Fig. 29).

Lavocat reported that the malleus and
incus oi Paraphiomys pigotti were fused,
presumably because they were lying
within the bulla in an articulated posi-
tion. His figure (1973: PI. 13, Fig. 4) sug-
gests that there is a matrix-filled groove
at their junction. This raises the possibil-
ity that, as in the case of Drytomomys,
they may have been separable in life. In
fact, it would seem to us that, although
lack of fusion can be demonstrated in fos-
sil rodents if the ossicles are preserved
separately, fusion can only be proven by
serial sections or when a continuous
groove between the ossicles is no longer
present.

Vucetich (1975: 484), although not
mentioning fusion, remarked of Perimys
sp. that the suture between the two os-
sicles was simple on both sides.

In some, but not all. Old World hystri-
cognaths the area of fusion extends to the

484 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

-^^A B

Figure 29. Malleus of Drytomomys aequatorialis, UCMP no. 41636, xi5. A. Dorsal view. B. Ventral view. C. Surface for
articulation with the incus.

Abbreviations: D — dorsal side; M — broken manubrium.

cms longus of the incus, moderately so in heavy ossicles such as characterize most

Thnjonomys (Tullberg, 1899: PI. 24, Fig. of the Hystricognathi.

14; Wood and Patterson, 1959: 293, n. 4), If the Caviomorpha were descended

although Doran's figure does not show from rafted thryonomyoids, it seems most

this, and almost as far as the incudal ar- amazing to us that the more primitive

ticulation in bathyergids (Hyrtl, 1845; condition (occasional absence of fusion)

Doran, 1879: 413; Tullberg, 1899: PI. 24, should be rather widespread among the

Figs. 1-2). Hystrix (Doran; Tullberg), Caviomorpha, presumably descended

Petromus (Tullberg and personal obser- from a small founder population, and un-

vation), and all caviomorphs for which known in the Old World fonns. We know

ossicles are known display no such ex- of no basis for assuming that there has

tension. We do not, of course, feel that been a selective advantage for a neotenic

the similar degree of fusion of the mal- retention of embryonic lack of fusion

leus with the crus longus of the incus in- among the Caviomorpha that did not act

dicates closer relationships of Thryono- on the Old World fomis. Therefore, we

mys to the bathyergids than to Petromus. conclude that the status of fusion of the

Among those hystricognaths — and they malleus and incus supports one of two
constitute the great majority — in which possibilities. Either the Old World hys-
fusion occurs, the time of onset of the fu- tricognaths were descended from cavio-
sion during ontogeny is variable. Avail- morph ancestors; or (much more proba-
able data are few but sufficient to dem- bly, in our opinion), fusion was either
onstrate this. Doran has recorded fusion incipiently present among the Franimor-
in newborn Cavia, "even in very young" pha or was acquired independently, by
Dasyprocta, and in "quite young" Hys- parallelism, by the Old World and New
trix, whereas in Hydrochoerus the ossi- World hystricognaths, as well as in such
cles are "often free in young" specimens; unrelated rodents as the sciurognathous
we have recorded fusion in juvenile ctenodactylids (Wood, 1975a: Fig. 1).
Thnjonomys and in adult, but not in ju- (8) Similar arrangement of openings at
venile, Ahrocoma. Thus the available the base of the skull. We are not sure to
data suggest that the ontogenetic onset of what this refers, unless to the foramina
fusion is earlier in Old World than in for the cranial arteries. In view of the re-
New World forms, and all stages from in ported similarities in cephalic arterial cir-
utero fusion to complete lack of fusion culation between the Caviomorpha and
may occur in the latter. the Old World hystricognaths (Guthrie,

There has been no documented selec- 1963: 472; Bugge, 1971: 519-522; 1974a:

tive basis for the fusion of these ossicles. 70-71; 1974b: 65-69), a similarity of the

However, Dr. E. C. Oaks has suggested foramina through which the arteries pass

to us {in lit.) that it may be an adaptation is not surprising. Unfortunately, Old

against possible dislocation of unusually World Oligocene material (other than of

Oligocene Rodents of Bolivia • Patterson and Wood 485

tsaganoniyids) is unavailable to demon- tween Coendon and the other hystrico-
strate either similarity or dissimilarity to gnaths is so striking that Bugge (1974a:
the Deseadan caviomorphs. 56) felt that the erection of a new Sub-
Recent studies by Bugge (1971, 1974a, order Erethizontomorpha was needed
1974b), Guthrie (1963) and Wahlert (Bugge did not use the Suborders Sciuro-
(1974a) have provided enough new infor- gnathi and Hystricognathi that we are us-
mation on the cranial circulation of ro- ing; his suborders are therefore equiva-
dents so that the arrangement of these lent to our infraorders).
blood vessels (or of the foramina pre- The cranial circulation of the Chin-
sumed to have been occupied by them) chillidae, Octodontoidea, Thryonomys
can be used as an important adjunct to and the Bathyergidae is essentially iden-
rodent classification. Bugge (1974a, tical, with the internal carotid, proximal
1974b) has developed a classification of part of the stapedial and varying amounts
the rodents, using the losses, during on- of the central part of the supraorbital ar-
togeny, of arteries (especially the stape- teries having been lost (Bugge, 1974a:
dial and internal carotid) from the prim- Figs. 14D, 15A). The Hystricidae differ
itive mammalian pattern. He reported by developing a transient anastomosis
(1974b: 78) that the "stapedial artery (a6) between the distal part of the inter-
obliterates when the bulla and pars pe- nal carotid and the infraorbital arteries
trosa are nearly connected and when the (Bugge, 1974a: Fig. 15C). Slightly differ-
chewing musculature is very strongly de- ent anastomoses occur in the cavioids,
veloped, and I think that there is some with a transient connection (a4) from the
mechanical explanation for obliteration distal internal carotid to the supraorbital
of the stapedial artery. But I do not know in Dolichotis (op. cit.: Fig. 14B); a tran-
why the internal carotid artery is oblit- sient one (a5") from the same source to
erated in some cases and not in others." the infraorbital in Cavia (op. cit.: Fig.
It would seem possible that the same 14A); and a pennanent connection in ap-
types of processes would have affected proximately the same area (a6) in Galea
the internal carotid as the stapedial ar- and Dasyprocta {op. cit.: Fig. 14C).
tery. If Bugge was correct, that the fusion Problems arise in connection with re-
of the bulla with the skull was responsi- liance on the pattern of cranial circula-
ble for the loss of the stapedial artery, this tion as a guide to rodent relationships,
loss should have occurred among Eocene We do not know of the selective pres-
franimorphs, in some of which the bulla sures that led to the changes, other than
is fully developed and fused with the Bugge's suggestion that the loss of the
skull (Wood, 1962a: Fig. 41B,C; Wahlert, stapedial artery was related to the fusion
1973: Fig. 2; 1974a: Fig. 5). of the bulla with the skull. As a result, it
As pointed out by Bugge (1974a: 56, is difficult to assess the weight to be giv-
Fig. 15B), the arterial pattern oi Coendou en such changes. How can one be sure,
prehensilis is quite different from that of in dealing with fossils, that an opening in
other hystricognaths; it is apparently the skull carried the blood vessels that
most similar, among living rodents (other one would anticipate, on the basis of the
than £r£>f/iizo?j, in which, apparently, the study of Recent forms? This point is
cephalic arteries have not been studied, raised by the observation by Wahlert
although a carotid canal was reported by (1974a: 373) that the carotid canal of Mar-
Hill in 1935), to that of Castor fiber mota monax carries the inferior petrosal
(Bugge, 1974a: Fig. IOC), and seems to sinus rather than the internal carotid ar-
be almost identical to what Bugge ter>% as is usually the case among rodents.
(1974a: Fig. 15D) described i )r Orycto- Bugge (1974a: Fig. IIB) indicates the in-
lagus and Lepus. The difference be- ternal carotid artery to be absent in Mar-

486 Bullet ill Mtiscuni of Comparative Zoologij, Vol. 149, No. 7

mota marmota as well. Finally, how does all the caviomon^hs (including the Ere-

one distinguish parallelism or conver- thizontidae) are descended from thryon-

gence in the gain or loss of blood vessels omyoids, it is surprising that no trace of

from the inheritance of a pattern from a the internal carotid has been preserved

common ancestral source? in any of the Old World hystricognaths

The similarities cited above between that Bugge dissected. However, and this
Coendou, Castor, and leporids, and the we consider to be an important point, if
likewise great similarities between Miis- the caviomoq^hs (including the Erethi-
cardiniis and Aplodontia (Bugge, 1974a: zontidae) are descended from thryon-
Fig. 13C,D), are certainly not indications omyoids, it must follow either that the
of close relationships. The first three erethizontids have reacfjuired the inter-
have interrupted the stapedial, supraor- nal carotid artery or that the loss of the
bital and internal ophthalmic arteries, internal carotid has occurred indepen-
and have acquired anastomoses al, a2, dently, as a parallelism, in the Old and
and a3' in the adult as well as a3 tran- New Worlds.

siently during ontogeny (Bugge, 1974a: There is some evidence as to the situ-

Figs. IOC; 15B,D). Muscardintis and Ap- ation in the fossils. Unfortunately, no cra-

lodontia have lost the internal carotid, nial material is available from Oligocene

stapedial and supraorbital arteries and thryonomyoids. Bugge (1974a: 63)

have developed the a3 anastomosis; Ap- thought that ". . . it is a reasonable as-

lodontia has also acquired an a3" anasto- sumption that both the internal carotid

mosis. Since none of these fonns are at artery and the stapedial artery were al-

all closely related, the losses and gains ready obliterated in the Oligocene-Mio-

must be presumed to have occurred as cene phiomyids . . . ." Lavocat (1973: 63),

convergent adaptations. If these similar- however, reported that, in Diamantomys

ities are not indicative of relationship, we hiederitzi, "Le foramen carotidien s'ouvre

see no real reason for believing that the en avant de la gouttiere . . . qui se creuse

similarly close resemblances between entre la region posterieure du basi-occi-

the chinchillids, octodontoids, Thryono- pital et la bulle. C'est un trou de toute la

mys and the bathyergids (Bugge, 1974a: hauteur de la gouttiere." As indicated

Figs. 14D, ISA) could not either be in- above (p. 411), there is possibly a small

heritances from a common Eocene fran- carotid canal in Incajyujs. These obser-

imorph ancestry or the result of parallel- vations suggest that the internal carotid

ism. The fact that three different types of artery was still present in Diamantomys

anastomosis occur in the caviids (Bugge, and possibly in Incamys, again indicat-

1974a: Fig. 14A-C) suggests that similar- ing that the artery has been lost indepen-

ities and differences in the cephalic cir- dently in the New and Old Worlds. In-

culation need to be treated with caution, cidentally, Wahlert (1974a: Fig. 10)

The pattern of circulation described in identified a carotid canal in the North

Coendou by Bugge (1974a: 56) is certain- American early Oligocene franimorph cy-

ly much more primitive than that in any lindrodont Ardynomys occidentalis.

other hystricognath that he dissected. As We feel, in summar\', that the cranial

he has indicated (Bugge, 1974b: 78), circulation very probably indicates that

"these arteries pass through canals in the the Caviomorpha and Old World hystri-

cephalic bones and when these arteries cognaths are related, but that almost cer-

are [lost, thel canals are also obliterated, tainly one of their similarities, the loss of

Therefore 1 would think it most unlikely the internal carotid artery, occurred in-

that such osseous canals should re-open dependently in the two hemispheres,

for the internal carotid arter>' or the sta- How many of the other circulatory simi-

pedial artery." If Lavocat is correct that larities are the result of parallelism and

Oligocene Rodents of Bolivia • Patterson and Wood 487

how many were inherited from common
franimorph ancestors, we have no way of
knowing at present.

(9) Alisphenoid remaining in a low po-
sition. This is an advanced character, not
a primitive one as Lavocat's wording
might suggest. In most Eocene rodents,
as may be seen in the figures of Wood
(1962a; 1974b: Fig. 4B) and Wahlert
(1974b), the ahsphenoid extends dorsally
to or above the posterior root of the zygo-
ma, to meet a descending portion of the
parietal, the frontal and squamosal being
separated. In nearly all later rodents, a
reshuffling of relationships between
these bones has taken place. The de-
scending portion of the parietal was
quickly lost and a squamoso-frontal con-
tact was established, its extent largely
depending on the degrees of reduction of
the anterolateral portion of the parietal
and of the dorsal portion of the alisphe-
noid. The latter may retain the size met
with in the primitive rodents (e.g., Aplo-
dontia, many sciurids, Pedetes) or may
increase dorsally until it nearly (e.g., Ta-
niias, Pteroniys) or quite reaches the pari-
etal (e.g., various sciurids, Castor).

In the other direction, the dorsal por-
tion of the alisphenoid may be reduced
to varying degrees until it comes to lie
ventral to the level of the root of the zygo-
ma or to that of the glenoid cavity. This
occurs in many muroids as well as in
the Hystricognathi. Within the latter, as
Landry (1957, Table 2) has noted, the
bathyergids do not display quite the same
degree of reduction as do the others.
Incamys, the only early Oligocene hys-
tricognath in which this portion of the
skull is known, has a low alisphenoid that
extends no higher than the root of the zy-
goma (Figs. 12, 15A). In the Eocene Pro-
toptychus the critical area is obscured,
but from what is visible it would seem
probable that squamosal-frontal contact
had become established (Wahlert, 1973:
Fig. 2). There is at present no real evi-
dence as to whether alisphenoid reduc-
tion took place independently in the sur-

viving hystricognath infraorders or
whether the reduction represents inher-
itance from a common franimorph ances-
tor. In any event, on the basis of our pres-
ent knowledge, this character can be
considered no more a support for deri-
vation of the Caviomorpha from the Thry-
onomyoidea than for independent deri-
vation of the two from the Franimorpha.

(10) Same type of squamosal-mastoid-
parietal-occipital junction. We believe
that this statement refers to the presence
of a lateral strip of the supraoccipital that
extends downward, separating the squa-
mosal from the mastoid process (Lavocat,
1971a: 520). It is present, but does not
project freely, in all living Old World
hystricognaths, being very large in bathy-
ergids; a small blunt tenninal enlarge-
ment in an old individual oi TJiryonomys
is the closest approach we have seen in
thryonomyoids to the projection that oc-
curs in some caviomorphs.

In living caviomorphs, the strip is lack-
ing entirely in the Hydrochoeridae, Das-
yproctidae, Dinomyidae and Cuniculi-
dae, and is variable in the Caviidae,
being absent in Cavia, Galea and Doli-
chotis, present in Microcavia, and inter-
mediate in Kerodon. In most echimyids
it projects as a small blunt nubbin, as a
blunt projection in capromyines, and cul-
minates in the very long, independent
processes of Myocastor and Elasmodon-
tomys. Among the remaining cavio-
morphs, it is limited to an enclosed strip
of bone. Landry pointed out that the lat-
eral process also occurs in various sciuro-
gnathous rodent groups, "and seems . . .
to be a more or less random feature
among rodents" (1957: 8). It was not pres-
ent in Reithroparamys or Protoptychus.
Platypittamys appears to lack it, and it is
small in Incamys (Fig. 12). The area is
preserved in no other Oligocene hystri-
cognaths.

It seems highly probable to us that this
feature evolved independently in the
Old and New World hystricognaths; cer-

488 Bulletin Museum oj Comparative Zoology, Vol. 149, No. 7

tainly there is no present evidence that area, and he specifically stated (1962a:

it did not. 14) that "The maxillo-palatine and max-

(11) Frequently, elongate cleft be- illo-orbitosphenoid sutures run behind
tween auditory bulla and squamosal, the rather large sphenopalatine foramen
This character is highly variable. The va- . . . ." Later on the same page, he stated
cuity seems usually to develop from the "Within the orbit, all specimens, except
vicinity of what Wahlert (1974b: 374) the type of P. copei, are broken so that
called the post-alar fissure, a foramen or the suture between the palatine and the
foramina on or near the squamoso-tym- orbitosphenoid cannot be located. In
panic junction anteriorly; he believed [that] fomi, however, it can be seen to run
that it transmits a vein tributary to the from the posterior tip of the maxillary
jugular sinus system. All stages from ab- posterodorsally into the foramen lacerum
sence, through a small foramen, to a com- anterius." In Ischyrotomus, the palatine
plete cleft can be seen among living Old is fused with the orbitosphenoid, and is
World hystricognaths, and the cavio- clearly separated from the maxillary
morphs as a whole are just as varied. The (which includes the sphenopalatine fo-
cleft does not occur in Incamys, Reithro- ramen) by a suture (Wood, 1962a: 189,
paramys or Protoptychus. Among other Fig. 71C). The description that Lavocat
rodents, it seems to be frequent among gives (1971b: 118) for the situation in
the Muroidea. We suspect that the vary- Thryonomys seems to be almost exactly
ing degrees of development are related the same as that in the paramyids.

to differing stresses in the skull bones. It would seem, therefore, that what

and that this feature has no significance Lavocat was trying to say was that the

at higher taxonomic levels. hystricognaths have retained the primi-

(12) Maxillary reaches the rear of the tive rodent condition. This does not seem
floor of the orbito-teniporal fossa. We to be valid evidence of a thryonomyoid
find this a rather vague characterization, ancestry for the caviomorphs.

and are not sure what Lavocat was trying (13) Habitual presence of an interpa-

to say. We believe that he was referring rietal. This bone is no doubt present em-

to this feature earlier on the same page bryologically in all members of the order.

(1973: 168) when he stated that "Nous Thereafter it may remain distinct, fuse

avons recemment montre (Lavocat, 1972) with the parietals, the supraoccipital or

que les relations du maxillaire et du pala- both, again, we suspect, in relation to

tin dans la region du plancher de la cavi- skull stresses. All stages of fusion, or lack

te orbitaire sont comparables." Unfortu- of it, are met with in both caviomorphs

nately, he does not include the 1972 and Old World hystricognaths, as well as

paper in his list of references, and we in all other rodents. A separate bone is

have been unable to locate it. We suspect present in the adult, among the pertinent

that this was a reference to Lavocat, fossils, in Reithroparamys (Wood, 1962a:

1971b, which he likewise does not cite, Fig. 41A), Protoptychus (Wahlert, 1973:

but which does discuss this area of Para- Fig. 2), Paraphioniys and Diamantomys

mys delicatus. It is difficult to be sure just (Lavocat, 1973: Figs. 3C, 4B). There is no

what he meant, especially when he stat- trace of a separate element in Incamys

ed (Lavocat, 1971b: 119) that Wood (Figs. 7, 8) or in any of the Santacruzian

(1962a) "ne montre en effet aucune su- genera figured by Scott (1905), but it was

ture separant le palatin du maxillaire au clearly distinct in Platypittamys (Wood,

niveau du trou sphenopalatin." It would 1949: Fig. 1). "Habitual presence of an

seem quite clear to us that Wood (1962a: interparietal" is more descriptive of the

Fig. 3A) does show a suture separating Muroidea than of hystricognaths. The

the palatine from the maxillary in this character has no significance at higher

Oligocene Rodents of Bolivia • Patterson and Wood

489

taxonomic levels. Lavocat (1971a: 520)
and Hoffstetter (1975: 512) apparently
agreed with us on this point.

(14) Lacrimal foramen within the orbit
and not on the muzzle in front of the zy-
gomatic arch. Lavocat was correct in his
observation: the Old World hystrico-
gnaths and caviomorphs share this char-
acter, but they also share it with most oth-
er rodents, including the Param\idae, all
other protrogomorphs, all the Franimor-
pha, Castor, Pedetes, anomalurids, sci-
uroids, etc. A forward displacement of
the foramen occurs in muroids, glirids
and dipodids. The intraorbital position
was certainh' a primitive rodent character
retained by, among many others, the
Bathyergoidea, Thryonomyoidea, Hystri-
cidae and Caviomorpha.

(15) Skeleton: same type of tibia in
Platypittamys and Paraphiomys. Lavocat
described the tibia of Paraphiomys
(1973: 91-92), but made no comparisons
whatsoever with that of Platypittamys.
We could find no indication in his 1973
paper as to what he meant, but, earlier
(Lavocat, 1971a: 521) he had stated that
the arching ("cambrure") of the tibia in
Platypittamys was similar to that in Old
World fomis. This does not seem to us to
be ver>' significant, in which conclusion
Hoffstetter (1975: 513) appears to agree
with us.

(16) Same type of brain. Lavocat (1973)
gave no documentation for this state-
ment, and again we are not sure what he
intended to imply. Earlier (1971a: 521)
he stated that, while most rodents have
smooth brains, a large number of cavio-
morphs have suggestions of fissuration,
which are also present in Thryonomys
and Paraphiomys. In his monograph on
the East African Miocene rodents, Lavo-
cat described (1973: 36-.37) endocasts of
Paraphiomys pigotti, P. stromeri and
Diamantomys luederitzi, only the first of
these being illustrated (1973: PL 1, Fig.
3; PI. 35, Figs. 1, 3, 5, 6), and he made
comparisons, among living forms, only
with Thryonomys, for which he figured

a plastic endocast (1973: PI. 35, Fig. 2).
No comparisons were made with the
brains or endocasts of any caviomorphs,
and there was no discussion of any of the
literature. Unfortunately, we found that
most of the details that Lavocat described
in the text were not recognizable on his
plates.

Hoffstetter (1975: 522) thought that
there was a particularly marked tendency
toward gyrencephaly in the hystrico-
gnaths, and that perhaps there was "una
tendencia precoz, posiblemente hereda-
da del antecesor comun de los Phiomor-
pha y Caviomorpha." This last statement
is ver\' close to our opinion, although per-
haps we differ from Hoffstetter as to what
was the common ancestor.

A series of papers by Pilleri (1959a-
1959J; 1960a-1960f; 1961a, 1961b; 1962)
indicate that, in general, brains of sciuro-
morphous rodents are more advanced in
gyrencephaly than those of myomor-
phous ones and less advanced than those
of hystricognaths. The distinctions among
them seem to be largely ones of degree.
These degrees of complexity are, again in
general, in agreement with the individ-
ual size attained by members of these
groups, sciuromorphous rodents being on
the average larger than myomorphous
ones and smaller than hystricognaths.
The Muroidea, currently the most suc-
cessful and widely distributed of the ma-
jor rodent groups, have brains, so far as
known, with smooth neopallia. Stud\' of
rodent brains began with those of murids
and sciurids, and the impression that ro-
dent brains were all of simple type be-
came established early. Hydrochoerus
was noted as an exception to the general
rule by Dareste (1855). Subsequently,
Beddard (1892) described or mentioned
the brains of a number of larger rodents
in which there were cerebral convolu-
tions.

Landr\' (1957: 20-22) considered that,
among rodents, cerebral convolutions are
positively associated with increased size,
but are probably more often present in

490 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

hystricognathous forms than in sciuro-
gnaths of corresponding size. Pilleri
(1959b: 77) pointed out that with de-
crease in relative brain size (approxi-
mately the equivalent of increase in body
size) the gyrencephaly progressively in-
creases, but that the cane rat Choeromys
(^Thnjonomys) harrisoni is an excep-
tion, being less convoluted than would
be expected on the basis of its relative
brain size.

Pilleri's illustrations show convolu-
tions to be present in: Marmota (1959a:
Fig. 8), Castor (1959a: Fig. 10; 1959e;
1959i; 1959J), to a slight degree in Aplo-
dontia (1960a: Fig. 16), Sciurus niger
(1960b: Fig. 1), slightly in Citellus
(1960b: Fig. 5), and in all hystricognaths
that he studied (except Heterocephalus
and Petromus), namely Hystrix, Atheru-
rus, Cavia, Dolichotis, Lagostomus,
Chinchilla, Myocastor and Choeromys
{=Thryonomys) (1959b, 1959d), My-
oprocta, Dasyprocta (1959f, 1959g,
1959h), and Paradolichotis, although that
fonn, included by some in Dolichotis, is
less convoluted than Dolichotis (1959g).
The only possible exception that he
found was Erethizon (1959b: 53), al-
though the brains of this fonn that he il-
lustrated the following year (1960e: Fig.
1) look convoluted. He stated, however,
that the brain was smooth (1960e: 370).
The brain of Heterocephalus, in contrast
to those of most genera just discussed, is
smooth; resemblances to the brain of
Geomys were pointed out by Hill et al.
(1957: 503) and to that oi Aplodontia by
Pilleri (1960a: 30-31). The Heteroceph-
alus brain (the only bathyergid described
by Pilleri) seems quite different from
those of all other hystricognaths so far
studied, and impressed Pilleri (19601) as
being very primitive. Later, Pilleri (1962:
494) indicated that the brain oi Petromus
was much more primitive than the brains
of the caviomorphs that he had studied,
and that Petromus and Thryonomys
"haben sich friihzeitig von den Ischyro-
myiden getrennt und sind heute unter
Bewahrung dieser primitiven hirnmor-

phologischen Merkmalen nur auf den af-
rikanischen Kontinent beschriinkt."

Pilleri (1962: 494) also stated that there
was a striking resemblance between the
asulcate brain of Petromus typicus and
that of the ctenodactylid Pectinator
specki (1962: 494), that "Pectinator und
Petromus stammesgeschichtlich gemein-
same Wurzeln haben" (1962: 495), and
that "the brain of Pedetes caffer . . .
shows unmistakable relationships to the
Hystricomorpha" (1960d: 382). Since
both Pectinator and Pedetes are sciuro-
gnaths, either brain structure is not al-
ways a good guide to rodent relation-
ships, or the structure of the angular
process of the mandible is not. Draseke
(1929: Fig. 3), however, illustrated a
brain of Pedetes with an incomplete sul-
cus lateralis on each hemisphere. It looks
as much like that of Castor as like most
of the hystricognath brains figured by Pil-
leri.

An interesting point is suggested by
the data so far available. No clear evi-
dence of neopallial fissuration is met
with in rodents with a head and body
length (hbl) of less than about 350 mm,
at which size a rudiment of the parasag-
ittal (or longitudinal) fissure may appear.
Very few sciuromorphous and even few-
er myomorphous rodents attain a hbl of
350 mm or more. The largest sciurid
whose brain is known, Marmota flavi-
ventris (hbl 470-700), has a depression
(Griibchen) running lengthwise that is
the homolog of the sulcus lateralis (Pil-
leri, 1960b: 60). In Castor (hbl 735-
1300), there is a sulcus lateralis that may
be either a simple groove 15-22 mm long
(about 30^5% the length of the cere-
brum), or it may be subdivided into two
or three parts by short breaks in its course
(Pilleri, 1959c: 103, 1959e: Fig. 2). Like-
wise, the largest myomorphous rodent
whose brain is known, Hypogeomys (hbl
to 350), has a short sulcus lateralis and is,
as a whole, "rather highly differentiated
for a member of the . . . Myomorpha"
(Pilleri, 1961a: 430).

Greater fissural complication frequent-

Oligocene Rodents of Bolivia • Patterson and Wood 491

ly accompanies increase in size. The only gests a phylogenetic relationship (Pilleri,

possible exception known to us may be 1960a: 34 — summary), a suggestion that

Octodon (hbl 125-195). Beddard (1892: will not appeal to many students of the

609) stated that the brain of this animal order.

possesses a parasagittal (longitudinal) fis- (17) Same type of upper and lower

sure, but gave neither a figure nor further molars in the Deseadan Eosteiromys and

description. Since other parts of his work the Phiomyidae. Same type of primitive

do not agree entirely with reports of later premolar in certain Fayum phiomyids

investigators — Draseke (1942) for Cap- and certain Deseadan caviomorphs. We

romys and Coendou and Pilleri (1959b: assume that Lavocat's use of the Colhue-

60; 1959f; 1959g; 1959h) for Dasyprocta huapian Eosteiromys was a lapsus cala-

and Cuniculus, who found certain neo- mi ior the Deseadan Protosteiromys. The

pallia to be simpler than indicated by following discussion, however, applies

Beddard — we believe that Beddard's re- whichever genus he meant. For the ter-

port on Octodon requires confirmation. minology that we use for the parts of the

Endocasts, so far as we are aware, have cheek teeth, see Figs. 1-2, pp. 376-377).
never been described for fossil cavio- General. The most detailed argument
morphs, so fossil and Recent Old World for close dental similarity between the
hystricognath brains can only be com- Caviomorpha and Old World hystrico-
pared with those of Recent New World gnaths was made by Stehlin and Schaub
genera. The endocast of Paraphiomys (1951), who based their entire philoso-
(Lavocat, 1973: 36-^37; PI. 1, Fig. 3; PI. phy of rodent interrelationships, as
35, Figs. 1, 3, 5, 6) does not look much spelled out later by Schaub (1953a,
like that which he figured for Thryono- 1953b, 1958) on the supposedly funda-
mys (Lavocat, 1973: PI. 35, Fig. 2), but is mental importance of a tooth pattern con-
closely similar to the actual brains of sisting of five transverse crests in the
Choeromys {=Thryonomys) harrisoni upper molars (the "Theridotnys-Trecho-
(Pilleri, 1959b: Fig. 12) and of A7ioma- mysplan"), the fifth crest (the mesoloph,
lurus pusillus (Pilleri, 1959a: Fig. 14). the last of the five to evolve) occurring in
The endocast of Parap/iiom(ys is also sim- the center of the tooth, as in the upper
ilar to that oi Pseudocylindrodon texaniis teeth oi Phiomys and Metaphiomys (Fig.
(Wood, 1974a: Fig. 16), an Oligocene 30C,D,F). This pattern is exceedingly
franimorph. common among rodents and is a dental

In general, we conclude that: (1) the adaptation that occurred independently
brains of large rodents tend to be more many times throughout the order (e.g.,
convoluted that those of small ones; (2) Theridomyoidea, Eomyidae, Cricetidae
the brains of hystricognaths are apt to be if not derived from eomyids, Dipodoidea,
somewhat more convoluted than those of Castorimorpha, Anomaluridae, and what-
sciurognaths of the same size; and (3) at ever number of hystricognaths devel-
the present time our knowledge of brain oped it independently). This pattern, we
fonn in rodents and its taxonomic value believe, is an adaptation to the need for
leaves much to be desired. As a caveat, an increased chewing area in animals
we note that Pilleri (1960e: 375) reported that had reduced the cheek teeth at least
that study of six brains of Eref/iizoji c?or- to Pi^~', Mg^ before a shift to the con-
satum dorsatum from two localities in sumption of harsh food led to strong se-
Michigan showed great variability in the lective pressures for more efficient grind-
form of the cerebral hemispheres. He ing surfaces. If this is correct, it should
also considered that the similarity of the come as no surprise to anyone familiar
brains of Heterocephalus and Aplodon- with rodent evolution that a superficially
tia, together with the presence in each similar and functionally identical tooth
genus of two superior venae cavae, sug- pattern should have evolved in both the

492 Bulletin Museum of Coinpuriiticc Zoology, Vol. 149, No. 7

Cavioniorpha and Old World hystrico- ences between the thryonomyoids and

gnaths as a reaction to the same selective the theridomyoids (such as that the latter

pressures. This, in itself, cannot be taken were sciurognathous) for anyone to admit

as evidence of relationship below the or- a close relationship between the two.

dinal level. However, later (Lavocat, 1976: 84-86) he

We have argued in detail (1959: 407- has returned at least part way toward the

412) against the postulate that the ''Ther- Schlosserian view point in his advoca-

idomys-Trechomysplan" had the basic cy — mistaken in our opinion — of a rela-

importance that Schaub believed, and we tionship at some stage between the

see no need to repeat all our arguments, Thryonomyoidea and the sciurognathous

since we feel that no evidence has been Theridomyoidea. Lavocat, however, ap-

brought to bear against them. Lavocat's parently does not follow Schlosser (1884b:

argument for a dental identity of early 102-105) in deriving the Castoridae from

caviomorphs and thryonomyoids arises, the Theridomyoidea.

we believe, from an acceptance of a com- In contrast to Schlosser and Lavocat,

mon ancestral five-crested pattern as the Stromer, the first student of significant

basis for much of rodent dental evolution, quantities of Miocene African rodents,

Such a point of view (due in our opinion questioned relationships between

to a mistaken belief that the pentalopho- thryonomyoids and caviomorphs. "Die

dont Theridomyoidea, so important in mir vorliegenden Reste lassen sich, so-

the European Eocene and Oligocene, weit vorlaufig bestimmbar, fast alle mit

could not have died out without issue) Hystricomorpha im weitesten Sinne am

has often been implicit in work on rodent besten vergleichen, scheinen jedoch be-

evolution by European students since merkenswerter Weise keiner siidameri-

the days of Schlosser, who considered kanischen rezenten oder fossilen Form

(1884b: 102-105) that the theridomyoids besonders nahe zu stehen, sondern viel-

were not only ancestral to the Hystrici- mehr Fomien aus dem Oligozan Agyp-

dae and Caviomorpha, but also (likewise tens und Europas, die leider wie sie fast

on the basis of equal similarities of the samtlich ungeniigend bekannt sind"

tooth structure) to the Castoridae. Lavo- (1926: 147-148).

cat, not citing Schlosser's earlier work, For our part, we feel that there is not
stated (1976: 84) that "Frappe de son cote the slightest possibility of any special re-
des resemblances dentaires egalement lationship between the thryonomyoids
tres etroits, Schlosser (1911) avait pro- and theridomyoids. In this conclusion,
pose I'hypothese d'une origine Therido- we are merely reinforcing those formerly
myidae du genre Phiomys de I'Oligocene reached by Lavocat (1955). If such a re-
du Fayoum" (as indicated by Wood, lationship were to be established, we be-
1968: 34, the specimens figured by lieve that it would demonstrate that the
Schlosser, that he referred to Phiomys, Suborder Hystricognathi was a wastebas-
were actually Metaphiomys schauhi and ket, because the Pseudosciuridae, the
Gaiideamus aegyptius). most primitive theridomyoids, had them-

Schaub's Pentalophodonta (1953a) has selves only just originated in the middle

much in common with a phylogenetic Eocene, at a time when there were al-

tree offered more than half a century ear- ready definite hystricognaths in North

Her by Schlosser (1884a: 327); the 'T/kt- America, and nothing is clearer than that

idomys-Trechomysplan^ of Stehlin and theridomyoids never reached the New

Schaub (1951) could be said to be merely World and that Europe was isolated from

an expansion and up-date of this diagram, the rest of the world during the period

As recently as 1971, Lavocat (1971a: 518) from quite early in the Eocene until early

believed that there were too many differ- or middle Oligocene (Wood, 1977b: 100).

Oligocene Rodents of Bolivia • Patterson and Wood 493

Although we think it is improbable that third crest from the front, the metaloph,
the theridomyoids were ever able to es- even when interrupted in its course, is as
tablish themselves outside Europe, we high and as extended buccally as the pro-
are not ready to abandon completely the toloph, whereas, in five-crested Fayum
possibility of a theridomyoid-anomalurid genera (Fig. 30C,D,F), the third crest,
relationship. At least, the dental similar- which is certainly the mesoloph, is lower
ity between these two groups seems to and less extended buccally than the pro-
us to be much greater than that between toloph. The fourth crest from the front in
the theridomyoids and thr>'onomyoids or Fayum fonns, the metaloph, is as high
between the Thryonomyoidea and the and as buccally situated as is the proto-
Caviomorpha. In addition, both groups loph, just as is the third crest, the meta-
are hystricomorphous sciurognaths. La- loph, in the caviomorphs.
vocat (1967: 498) has pointed out differ- Among known early Miocene thryon-
ences in middle ear structure between omyoids there is considerable variation
theridomyids and anomalurids, but the in mesoloph development, the crest rang-
comparison was perforce between Quer- ing from low and little extended buccally
cy Phosphorites forms on the one hand (Myophiomys, Lavocat, 1973: PI. 42, Fig.
and Recent ones on the other. Before 5; Paraphiomys, op. cit.: PI. 26, Figs. 2-
coming to any final conclusion as to ther- 5; PI. 44, Fig. 7) through low and mod-
idomyoid-anomalurid relationships, we erately extended {Andre wsimys, op. cit.:
would like some evidence from inter- PI. 42, Fig. 4), low and well extended
mediate stages. (Phiomys, op. cit.: PI. 41, Fig. 6) to high

Upper Molars. We currently recognize and well extended {Kenyamys, op. cit.:
thirteen identifiable named genera of PI. 27, Fig. 5; Simonimys, op. cit.: PI. 42,
Deseadan rodents (Table 8; Luribayo- Fig. 3; Epiphiomys, op. cit.: F\. 42, Figs,
niys and Palmiramys we regard as no- 1-2; and Diamantomys, op. cit.: PI. 25,
niina vana), five of which {Asteromys, Fig. 7). This variability is what one
CJiubutoniys, Litodontoniys, Migravera- would expect given the structural ances-
miis and Xylechimys) are known only try represented by the Fayum thryon-
from lower jaws. Of the other genera, the omyoids. The mesoloph of five-crested
upper molars are four-crested in Platy- thr\'onomyoids is an evolving, not degen-
pittamys (Wood, 1949: Fig. 3A; 1974b: crating, crest; it is a specialized charac-
Fig. 2A), Deseadomys (Wood and Patter- ter, not a primitive one.
son, 1959: Fig. 4), Incamys (Fig. 17A,B) Lavocat's charge (1973: 169) that our
and Sallamys (Fig. 5A). All the known hypothesis that the four-crested upper
teeth of Scotamys are heavily worn, but molars of Platypittamys and other four-
it seems to us that the tooth pattern was crested Deseadan genera were primitive
likewise a four-crested one (Wood and skirts the possibility that the four-crested
Patterson, 1959: 315). In Branisamys, conditions could have been derived by
Cephalomys, Protosteiromys and the un- simplification from a five-crested one is,
described dasyproctid (p. 430) there is of course, correct, but it would seem to
evidence of a fifth crest in the upper mo- us highh' surprising if most genera that
lars. This crest, however, was not a me- possessed brachyodont, primitive-look-
soloph, but originated from the anterior ing cheek teeth had indulged in such a
slope of the posteroloph, being what we simplification. We believe that there is
call the neoloph. Protosteiromys (Wood considerable evidence in support of our
and Patterson, 1959: Fig. 30) and Brani- position. Even if there were no such evi-
samys (Fig. 24B), in particular, seem to dence, Occam's Razor would argue
us to demonstrate this quite plainly. In against acceptance of Lavocat's conten-
five-crested Deseadan caviomorphs, the tion without some supporting documen-

494 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

00 Q

■o Ll
o

O co'

■M
O
00

^^

— c

>- D

o .-.

o ',

t E
o a,"

CO (0

5 "2

3 O

00

CD

ii 05 o

CO

05

■Q

C
CD

■~ CO
-C O
Q. 00

o

c

iri ^

V ^

^ Q.
CO >

0) ^-
O DC

O M

O 3

> CD

• - O

O C

T— »*—

CO CD

^ 5L

6 E

^ S

CL O

-^ o

TO
CD ^

E

o cr\j
2c2

CD <M
CJ)

CO T-

(D

0) o
CO C

^^
C C
O 3

CD ^

0 1

CD OC
C\J .
C35 Q)
'- Q.
. >-
O *"
C w.'

1 I

^55

. . N

(3 V~

CD

DC 01

3

■r- CO 3 Q) CO

X S TO d) E ^

"- TD CO .*=■!=

■DC . CO C/3 C

C, E an
C. P. a
. E. M
od, 196i
dubius
. Diama

ds. A,
ID.
g. 6A
; Wo
mys
It. J

■p o>^^ ?i

me thryonomy
ood, 1968, Fii
; Wood, 1968,
YPM no. 213
2). H. Pomoi
areas are cem

Mioce
71; W
21309

973:2
otted

X

CO

CM

and
. 180
1 no.
m-' \
cat, 1
23. D

o>

6

S O ^ -D o X

c

C C Q_ _, ^CD
Q) !i CO CM
O 2 > ._- _I 05

o

Oligc
,. YPI
M' •\
tiaub
fide
no. 1

2

p 1 T " t: o

o □: E 5 2,'^
co^ES.gi

O ._- . - LL

DC
Q.

0^ S TO-:rEt

<U i- x: cu O '-r

0)

Cheel<t
. P. ana
omys sc,
hat papi
Paraphi
pe, Rdm

E
o

1-

CO

m ■■= •" II >.

.3

re 30.
2A. 1
etaph
it " in
x13 (
Tier, t

TO

.l.^MsS

2

LL LL C3 ; 1- CO

c

Oligocene Rodents of Bolivia • Patterson cnul Wood 495

tation, which we beheve is entirely ab-
sent. We beheve that the only basis tor
assuming that the caviomorph tooth pat-
tern was derived from that of the thryon-
omyoids is the prior assumption that
thryonomyoids were the ancestors of the
caviomorphs.

If, as we believe, the Caviomorpha
were descendants of franimori^h reithro-
paramyids, their upper cheek teeth must
ultimately have been derived from a four-
crested ancestral stage of the type seen
in Protoptychus (Wilson, 1937: Fig. 1) or
cylindrodonts (Wood, 1973: Fig. 3), and
we consider it supererogatory to assume
an intermediate five-crested stage be-
tween ancestral and descendant four-
crested ones until there is real evidence
that such a stage actually existed.

Upper Premolars. Lavocat (1973) has
greatly confused the discussion of pre-
molar resemblances by referring to the
teeth, in front of the pennanent molars,
indiscriminately as D ( = deciduous mo-
lar) and P ( = permanent premolar). In
particular, he discussed the deciduous
teeth of Metaphiomijs and the probable
pennanent premolar of Gaudeamus to-
gether (1973: 170), with no indication
that they were probably not homologous
teeth. As a result, we do not know wheth-
er his statement that there was the
"meme type de premolaire primitive
dans certains Phiomyidae du Fayoum et
certaines formes du Deseado" (1973:
168) was based on the comparison of per-
manent premolars, deciduous molars, or
the deciduous molars of the African forms
with the pennanent premolars of the Ca-
viomorpha. Unfortunately, although one
would not get this impression from Lavo-
cat (1973), the single tooth oi Gaudeamus
(if it is P* and not dm^) is all that there is
to go on as to what the upper premolars
of the thryonomyoids were like, and that
one tooth comes from one of the most
specialized of the Fayum genera. In a lat-
er paper (Lavocat, 1978), he says very lit-
tle about the antemolar cheek teeth of
thryonomyoids, although he does state

that P or dP^ are present in some forms
(1978: 71, 78), and he included in his def-
inition of the Family Phiomyidae {op.
cit.: 71) the statement "Milk teeth with
delayed replacement in some fonns, per-
sistent in others . . . ." He does not men-
tion that replacement of deciduous teeth
is unknown in all other thryonomyoids.

Upper premolars are known in seven
of the Deseadan genera. The tooth is in-
cipiently five-crested in Branisamijs
(Fig. 23C,F); four-crested in Incami/s
(Figs. 17A, 18C) and Sallamys (Fig. 4A);
three-crested in Deseadomys (Wood and
Patterson, 1959: Fig. 4A), Cephalomys
(Wood and Patterson, 1959: Figs. 13A,
14A,B) and Platypittamys (Wood, 1949:
Fig. 3A,B); and probably three-crested in
Scotamys. The teeth of the three-crested
fonns look primitive to us, and we be-
lieve that these teeth are close to the an-
cestral caviomorph condition.

Upper premolars of thryonomyoids are
unknown except for the one specimen of
Gaudeamus aegyptitis that is more prob-
ably P* than dm^ (Wood, 1968: Fig. 14A
and Table 2). All other antemolar upper
cheek teeth (except for the peg-like dm^
or P^) known among thryonomyoids from
the early Oligocene to the present are
dm"*. The premolar of Gaudeamus is a
three-crested tooth and, as Lavocat has
stated (1973: 170), it shows some similar-
ity to that of Platypittamys (Wood, 1949:
Fig. 3A,B), although the two are far from
identical and, if one were to have been
derived from the other, that of Platypit-
tamys seems obviously to be the more
primitive in that it is much more brachy-
odont, the cusps are more distinct, and
the crests are less fully developed. The
pennanent premolars of both Platypit-
tamys (Wood, 1949: Fig. 3A,B; 1973b:
Fig. 2A,B) and Gaudeamus (Wood, 1968:
Fig. 14A) could have been derived from
one like that of the hystricomorphous
franimorph Protoptychus (Wilson, 1937:
Fig. 1).

Lower Molars. The lower molars of
Phiomys, the most primitive known

496

Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

thryonomyoid, were shown by Wood
(1968: 32-^3) to var\' widely as to crest
number. Within this primitive genus can
be found "a complete sequence from
five-crested specimens of Ph. andrewsi
through four-crested ones of the same
species, to the nearly three-crested ones
of Ph. paraphiotiiyoides and ol Ph. la-
vocati. Since the first and last of these
species are contemporaneous animals, it
is possible to assume either that evolu-
tion is proceeding from five-crested to
three-crested, or the reverse sequence, or
from four-crested in both directions."
Wood's preference was for the last of
these choices: "the four-crested would
be the most primitive, and evolution
would be in both directions, toward sim-
plification or complication." We adhere
to this interpretation. If Wood was correct
on this point, this alone would rule out
any possible relationship between the
thryonomyoids and the theridomyoids. It
would also support the idea of parallel
evolution of the Thryonomyoidea and
Caviomorpha from northern hemisphere
franimorphs that had four-crested molars.
Among the early Miocene thryonomy-
oids, the number of crests is either four
or three. A point that must be stressed
here is that in those early thryonomyoids
that were "experimenting" with five-
crested lower molars, the fifth crest (the
third from the front) is certainly a meso-
lophid; in some individuals traces of a
mesoconid and (rarely) of a metastylid
can be seen.

The lower molars of Deseadan rodents
are generally four-crested (Asteromys,
Branisamijs, Cephalomys, Chubutomy.s,
Incamys, Migraveramus, Platypittamys
and Protosteiromys). In Xylechimys
(Patterson and Pascual, 1968: Fig. 2A)
and Deseadomys arambourgi (Wood and
Patterson, 1959: Fig. 5) the teeth are in-
tennediate between being four-crested
and three-crested, and D. loomisi is
three-crested, one of the crests, the meta-
lophid, of the four-crested fonns having
been lost (Wood and Patterson, 1959:

311, Fig. 7). The teeth of Sallamys (Fig.
5D-F) seem intermediate in this respect
between those of the two species of De-
seadomys. The lower molars of the
known specimens of Litodontomys and
Scotamys are too worn to pennit certain
detennination of the number of crests,
although the teeth are compatible with
there having been four. There is, there-
fore, no evidence that early caviomorph
lower molars were ever five-crested; no
known lower molar of any Deseadan ro-
dent shows any indication of a mesocon-
id, a mesostylid or a mesolophid. When,
as in some specimens of post-Deseadan
erethizontids, an incipient fifth crest de-
velops on the lower molars, it arises from
the posterior face of the anterolophid
(Wood and Patterson, 1959: 381), not
from the ectolophid, and hence is com-
pletely unrelated to the fifth crest of
thryonomyoids.

A striking feature of the lower molars
of nearly all thryonomyoids is the pres-
ence of an anteroextenial basal cingulum.
Among the Fayum forms it is present in
Phiomys (Fig. 30A,B; Wood, 1968: Figs.
1; 2A,B; 3A,B; 4A) although very small or
absent in P. lavocati (Wood, 1968: 47;
Fig. 5A,C,D), small in Paraphiomys si-
monsi (Wood, 1968: Fig. 5F), transferred
below (p. 520) to Neosciuromys, but
large in Metaphiomys (Fig. 29E; Wood,
1968: Figs. 8-10, 12H, 13A) and Phiocri-
cetomys (Wood, 1968: Fig. 16A,D). This
last genus is peculiar among thryonomy-
oids in having but three lower cheek
teeth, either dm4-M2 or Mj.g (see p. 521).
The second tooth has a very heavy ante-
rior cingulum that extends back along the
buccal face of the tooth and the buccal
half of the posterior face; cuspules occur
here and there along its length. In the
last tooth, the cingulum, very heavy and
with cuspules, reaches posteriorly to the
hypoconid, and in the first is limited to
the base of that cusp. Only in Gaiidea-
miis, a specialized fomi, is the anterior
cingulum entirely lacking (Wood, 1968:
Figs. 14E,G,H; 15B-E,G).

Oligocene Rodents of Bolivia • Patterson and Wood 497

An anterior cingulum, like that of P/iio- pennanent successor was Phiomys an-

mys and Mctaphiomys, is very promi- f/rfit;^/, the most primitive known thryon-

nent among the Miocene thryonomyoids omyoid, and in this fomi replacement

(Fig. 30G-J; Stromer, 1926: PI. 41, Figs, was delayed long after the normal time

24, 25C, 32C; PI. 42, Fig. 28; Lavocat, of replacement. The single pennanent

1973: PI. 25, Fig. 6; PI. 26, Figs. 7-10; PL lower premolar of Gaudeamus was un-

27, Figs. 1, 3, 4, 9; PI. 28, Figs. 1, 2, 4, covered, deep within the jaw, when the

1 1-13). These cingula are prominent and deciduous tooth was removed from above

elevated in Petromus (Wood, 1962b: Fig. it (Wood, 1968: 72); whether or not it

1E,F), but are absent in Thryonoinys would ever have erupted is unknown;

(Wood, 1962b: Fig. 2A-C) as in its prob- certainly eruption would not have oc-

able ancestor Gaudeamus. curred until rather late in life. All three

There is no trace of such a basal ante- known Fayum specimens of P4 are from

rior cingulum in any bathyergoid (Lavo- the lower fossil wood zone and all have

cat, 1973: PI. 29, Figs. 5-6; PI. 30, Figs, essentially the same pattern. There is no

1-6), hystricid or caviomorph, so far as suggestion of an anteroconid on any of

we are aware. This feature, by itself, it them. The metaconid is the highest ele-

seems to us, rules out all known thryon- ment of the anterior part of the tooth, and

omyoids (except Gaudeamus, which, of is continued anteriorly by a cingular arm

course, is much too advanced in the crest- that crosses in front of the protoconid in

ing of its cheek teeth to be a possible ca- one specimen of Phiomys andrewsi

viomorph ancestor) from possible ances- (Wood, 1968: Fig. ID) and in Gaudea-

try to any caviomorph, hystricid or mus {op. cit.: Fig. 14E). There may (P.

bathyergid. The presence of this cingu- andrewsi) or may not (Gaudeamus) be a

lum on the molars of thryonomyoids mesoconid. There is a complete ectolo-

means that any thrvonomyoid that could, phid in P. andrewsi, but none in Gau-

morphologically, have been ancestral to deamus. The entoconid is completely

the caviomorphs must be a hypothetical isolated in all three teeth, except from the

Eocene fomi with an assumed set of char- posterolophid through which it is con-

acters quite different from those of any nected with the hypoconid.
known member of the superfamily. We In the Deseadan Caviomorpha, P4

consider it highly doubtful that such a primitively had a complete ectolophid

hypothetical fomi could be included in (Asteromys, Cephalomys, Deseadomys,

the Thryonomyoidea. We believe it Migraveramus, Platypittamys and Sa/-

would have been a franimorph. The Cha- lamys), but progressively it became in-

pattimyidae (Hussain et ah, 1978) do not terrupted (Incamys, Fig. 19C-E). There

fill the requirements for such an ancestral is no anteroconid in Asteromys (Wood

fomi, nor do their lower teeth possess an and Patterson, 1959: Fig. 26A), Desea-

anterior cingulum of the sort found in the domys (op. cit.: Fig. 5), Platypittamys

Thryonomyoidea. (Wood, 1949: Fig. 3C-D; 1974a: Fig. 2C-

Lower Premolars. The lower premolars D) or Sallamys (Fig. 5C). There is no an-

of thryonomyoids are unknown except for terior arm of the metaconid nor a meso-

those described by Wood (1968: Figs, conid in any of these genera. The ento-

1B,D; 14E) from the Jebel el Qatrani For- conid may be connected directly with the

mation (Fig. 30B). He was able to find hypoconid with no posterolophid (Platy-

only 3 pennanent lower premolars (P4) as pittamys); through the hypolophid with

compared with 57 deciduous ones (dm4) a distinct, isolated, posterolophid (Ceph-

in the material at his disposal. The only alomys and Sallamys); only through

species in which there certainly was re- the posterolophid (Deseadomys); through

placement of the deciduous tooth by its both the hypolophid and posterolophid

498 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

(Asteronufs and Migraveramus); or there the much greater difference of pattern of

may be no hypoconid-entoconid connec- the molars of the same Deseadan genera

tion at all until after very considerable is completely unjustified,
wear (Incainijs). Deciduous Teeth. The best opportuni-

Of all the Deseadan lower premolars, ty for comparing thryonomyoid and ca-

the closest to those of Phiomys are those viomorph antemolar teeth lies in the de-

oi Platypittauiys. We believe that this is ciduous dentition, dm/. Lavocat (1976)

because the former is close to the primi- had available both upper and lower de-

tive thryonomyoid pattern and the latter ciduous teeth of Incamys and Branisa-

close to the primitive caviomorph one, mys, although those of the latter were

and that both the thryonomyoids and ca- broken. Although deciduous teeth are al-

viomorphs were descended from primi- most universally present in thryono-

tive (and hence similar) franimorph myoid jaws (in fact, they are never known

ancestors that had four-crested lower pre- to have been replaced by the permanent

molars. Lavocat (1973: 169-170) has ar- premolar after the early Oligocene of the

gued that, in general, the cheek teeth of lower fossil wood zone), they occur at

Platypittamys were not primitive. We about the nomial mammalian frequency

disagree with him on this. Slightly later among specimens of Deseadan cavio-

{op. cit.: 170), Lavocat stated that the morphs. Neither upper nor lower decidu-

structure of P4 of Deseadomys and Platy- ous teeth are known for Asteromys, Chu-

pittamys is fundamentally identical to butomys, Deseadomys, Litodontomys,

that oi Phiomys andrewsi and, hence (as Lurihayomys, Migraveramus, Palmira-

we interpret his meaning), primitive. But mys, Phitypittamys, Protosteiromys, Scot-

we do not see how it is possible for the amys and Xylechimys, or 11 of the 15

pattern of P4 of Platypittamys to be con- named Deseadan genera. Only the lower

sidered to be primitive without also ac- is known in Cephalomys and Sallamys;

cepting the upper and lower molars and both uppers and lowers are preserved in

the upper premolar of Platypittamys as Branisamys and Incamys. We have

primitive. If this be done, the patterns of therefore used materials from the Col-

P4 might suggest close caviomorph- huehuapian and Santacruzian to supple-

thryonomyoid relationships, but the pat- ment our discussion of the Deseadan

terns of the other cheek teeth would specimens.

demonstrate the absence of such affini- Hoffstetter (1975: 520) cited, as a com-

ties. We know of nothing that would in- mon character shared by the Caviomor-

terfere with the derivation of P4 of A*- pha and Phiomorpha [ = Old World hystri-

teromys, Cephalomys, Deseadomys, Mi- cognaths'P], the frequent retention of "uno

graveramus, Platypittamys or Sallamys o dos premolares y/o los dientes de leche

from those of cylindrodonts (Wood, 1973: correspondientes." We are not sure

Fig. 5A) or, ultimately, of Franimys which if any of these fomis retain both

(Wood, 1962a: Fig. 49A) or Reithropara- dm^ and P^, since these teeth are very

mys {op. cit.: Fig. 42B). difficult to distinguish in rodents (Wood,

Thus, while we agree with Lavocat 1969; 1970b: 245-246), but this is a minor

(1973: 168) that there is the "meme type point. As Hoffstetter added, this reten-

de [lower] premolaire primitive dans cer- tion of antemolar teeth excludes from

tains Phiomyidae du Fayoum et certaines their ancestry those fonns that had lost

formes du Deseado," we feel that using these teeth. But, so far as we are aware,

the similarity of premolars (based on a no one has ever proposed that either the

total Oligocene to Recent thryonomyoid caviomorphs or any of the Old World hys-

sample of three teeth) and disregarding tricognaths were descended from, e.g..

Oligocene Rodents of Bolivia • Patterson and Wood 499

cricetids, and we believe that Hoffstetter crests, the third of which is interrupted,

here was merely setting up a straw man with no indication of a fifth crest,

in order to knock it down. We believe these crests to be the an-

Upper Deciduous Teeth. Among the teroloph, including the protocone; the

Oligocene Thryonomyoidea, dm'* is paracone-protoconule-hypocone; the

known from 11 specimens oi Metaphio- metacone-metaconule-hypocone; and the

7mjs schaubi and one o{ Metaphiomijs or posteroloph. There is no suggestion of a

Phiomys sp. indet. (Wood, 1968: Table mesoloph, which we feel is because the

2). In all of these, the tooth is five-crest- Caviomorpha were derived from North

ed, with an anteroloph, protoloph, me- (or Middle) American franimorphs in

soloph and a posterior marginal postero- which there never was a mesoloph. The

loph, to the middle of which is connected dm"* of Branisamys, figured by Hoffstet-

the end of the metaloph via the lingual ter and Lavocat (1970: Fig.) is both bro-

end of the metaconule (Fig. 30D,F; ken and badly worn, but suggests that

Wood, 1968: Figs. 11, 13B). This pattern there had been a four-crested pattern,

is essentially identical to that of the ad- However, one specimen (Fig. 24B) of

jacent molars. dm^ of this genus clearly shows increas-

Hoffstetter stated (1975: 520) that ing complexity. There is an incipient an-

"Phiomys presenta un reemplazo normal terior crest arising from the buccal end of

de dP por P." So far as we know, this the anteroloph, and the beginning of a

statement is entirely without factual ba- neoloph, arising by forward growth of

sis, and he has apparently completely ridges from the middle and the lingual

misunderstood the situation in Phiomys tip of the posteroloph. The tooth of In-

(Wood, 1968: diagnosis of Phiomys; 39- camys seems to us to be clearly the more

40, 83-84). No upper premolars or decid- primitive.

uous molars have with certainty been re- This tooth is known from a somewhat
ported for Phiomys; the only possible ex- larger number of Santacruzian genera. In
ample of such a tooth is one specimen of Neoreomys (Scott, 1905: PI. 65, Fig. 3),
dm^, identified as ''Phiomys or Metaphio- the anterior part of the tooth seems to be
7nys sp. indet." (Wood, 1968: 67, Fig. similar to that of /ncam?/^. There is a pos-
13B). We suspect that dm^ was replaced, terior prism of enamel, with a lake in its
late in life, by P, simply because the cor- center, which we would interpret as
responding replacement occurred in the being homologous to the two crests in
lower teeth. But this would be a long, Branisamys that we have called the be-
long way from "un reemplazo nonnal." ginnings of a neoloph. In Eocardia

In caviomorphs, dm^ is usually very (Scott, 1905: PI. 68, Fig. 27), the tooth has

similar to the molars in crown features, acquired the typical eocardiid outline,

but may be somewhat smaller and longer but would seem to be composed of the

and have a larger anteroloph and para- same features as in Incamys. Scott also

flexus. Exceptionally, as in living dasy- figured a worn dm"* of Neoreomys aus-

proctids, the anterior half of the tooth tralis (1905: PI. 65, Fig. 4), Stichomys

may elongate and neomorph crests de- regularis (1905: PI. 65, Fig. 19) and

velop within the enlarged parafossette. Schistomys erro (1905: PI. 68, Fig. 26);

Among Deseadan caviomorphs, the these, however, show essentially no de-
tooth is best known in Incamys (Fig. tails of the pattern. They could be (and
17A,H,I), one specimen (Fig. 17H) being probably were) based on the same basic
essentially unworn, and another (Fig. pattern as in Incamys and Branisamys.
171) only slightly worn. Both of these It is clear that the fifth crest that seems
teeth clearly consist of four transverse to be developing in the caviomorph dm"*

500 Bulletin A/uAt'u/;i of Comparative Zoology, Vol. 149, No. 7

(our neoloph) is a forward outgrowth The pattern of dm4 of Sallamy.s pas-

troni the posteroh)ph — absent in Inca- ciiali posed a number of problems that

niijs, developing in Branisaiiujs and fully we have touched on above (p. 390). This

developed in the Santacruzian genera, tooth (Figs. 5D, 31F) is obviously very

We see no possibility of homologizing complex, and closely resembles that of

this crest with the metaloph of the Cephalomys (Wood and Patterson, 1959:

thryonomyoid dm"*, since the neoloph of Fig. 20A,B), the only previously de-

the caviomorphs was clearly in the pro- scribed unbroken Deseadan dm4. In

cess of development, in South America, each, the talonid is like that of the molars,

during the Oligocene. the hypoconid being continuous back-

A tooth that is probably dm^ rather than ward into the posterolophid and forward
P^ is represented by an alveolus in the into the hypolophid. It is in the very com-
type of Branisamys luribayensis (Hoff- plex trigonid that the differences lie.
stetter and Lavocat, 1970: Fig.) and in PU There is a continuous anteroposteriorly
no. 21955 (Fig. 24B); an intra-alveolar directed ectolophid; a cusp that we be-
root in Incamys boliviamis, PU no. 20980 lieve to be the protoconid extends back-
(Fig. 171); and a worn nubbin on both ward toward the entoconid. In Cephalo-
sides of the skull of I. /?oZiuiani/s, PU no. mys (Fig. 31G), the metaconid lies
21726 (Fig. 17A). These tell us nothing behind the protoconid, whereas in Sal-
other than that a minute tooth was pres- lainys (Figs. 5D, 31F) it is farther for-
ent. One was also present in Metaphio- ward. In both, the anterior end of the
mys schaubi (Fig. 29D; Wood, 1968: Fig. tooth is fomied of a multicuspidate com-
IIB, Table 2); no trace of such a tooth is plex that we consider to have been de-
preserved in Gaudeamus. The presence rived from the anteroconid. A complex
of this tooth is merely a primitive rodent anteroconid area is a rather frequent evo-
character. At present, we do not know, for lutionary trend in dm4 of a wide variety
any of the genera involved, whether this of unrelated rodents,
tooth was dm^ or P^ (Wood, 1970b: 245- An attempt to unravel the homologies
246), or whether it had a pattern or was of the parts of the teeth in these two
merely a conical cusp. fonns has led us into an investigation of

Lower Deciduous Teeth. The lower the lower deciduous teeth of later ca-

milk molar (dm4) is better represented viomorphs.

among the Deseadan Caviomorpha than In the simplest type of caviomorph dm4

is the upper, but still is known only from that we have seen, the trigonid consists

four genera: Branisamys, Cephalomys, of buccal (protoconid) and lingual (meta-

Incamys and Sallamys. All three avail- conid) cusps, connected posteriorly

able teeth of Incamys are highly worn; through a metalophid and anteriorly

there is no way of being certain whether through an anterolophid (in which the

the pattern was like that in the other gen- anteroconid may or may not possess some

era, or whether it was different. The individuality), surrounding a trigonid ba-

Princeton collections include a specimen sin or anterofossettid {Erethizon dorsa-

of Branisamys with a broken dm4 (Fig. tum. Fig. 31A). Due to the greater length

24G), and we believe that GB no. 014, of the tooth, this fossettid is larger than

the type of Hartenberger's Villarroelo- its serial homolog in the molars. This

mys boliviamis (Figs. 26A, 31H) is also complex is united with the talonid by the

dm4 of Branisamys. According to our in- ectolophid.

terpretations, the pattern in both of these Deciduous lower molars of this sort are

Branisamys teeth was essentially like known in such octodontids as Sciamys,

that of Sallamys (Figs. 5D, 31F) or of in early erethizontids, and in various

Cephalomys (Figs. 26B, 31G). echimyids, such as Prospaniomys (Figs.

Oligocene Rodents of Bolivia • Patterson and Wood 501

Figure 31. Sketches of lower deciduous molars (dm4) of selected caviomorphs, not to scale; all drawn as if from the
right side. A. Erethizon dorsatum, MCZ no. 51367. Simple pattern. B. Protacaremys prior, AMNH no. 29707. C. P. prior,
FMNH no. P 13295. D. P. prior. AMNH no. 29692, reversed. E. Prospaniomys priscus. AMNH no. 29697, reversed.
F. Sallamys pascuali. PU no. 20909. G. Cephalomys arcidens. ACM no. 3013. H. Branisamys luribayensis (type of
Villarroelomys bolivianus), GB 014, reversed and restored. I. Erethizon dorsatum, MCZ no. B 7752. Complex pattern.

Abbreviations: A — protoconid; B — metaconid; C — hypoconid; D — entoconid.

5H, 31E), as variants in Acarechimys (for
validation of this name, see Appendix 2),
in Eumysops (Kraglievich, 1965: Fig. 2)
and in the Antillean heteropsomvines
(Miller, 1929: PL 2, Figs. 3a, 4a; PL 3,
Fig. la; PL 4, Fig. 2a).

There have been various modifications
of this pattern, ranging from relatively
simple to highly complex. As examples of
simple changes, we may cite: Paradel-
phomijs, in which the ver>' marked obliq-
uity of the lophids has converted the cir-
cular anterofossettid into a slit and has
greatly shortened the ectolophid without
completely obscuring the primitive pat-
tern; and Spaniomys, Stichomys, Cerco-
mys, etc., in which an anteroflexid has
developed between metaconid and an-
terolophid, thus converting the tooth into
as good a replica of the molars as the up-
per milk tooth usually is.

Considerably greater complexity is
present in Protacaremys prior, although
AMNH nos. 29692 and 29707 and FMNH
no. P 13295 are all different in details.
Here there are three crests from the pro-
toconid to or toward the metaconid (Figs.
5G, 31B-D). The anterior of these is

clearly an anterolophid; the nature of the
others is more uncertain. However, on
the basis of variants seen in Erethizon
dorsatum, as discussed below, we be-
lieve the posterior crest to be a metalo-
phid resembling the mesolophid of
thryonomyoids and the middle crest to be
a neomorph crossing the anterofossettid.
We believe that the tooth pattern of this
tooth oi Prospaniomys priscus (Fig. 31E)
is basically the same.

This stage is continued, morphologi-
cally, in Sallamys pascuali (Figs. 5D,
3 IF), where the protoconid has become
separated from the metalophid, and a
deep buccal valley has developed as a
neomorph, as has also the anteriorly fac-
ing opening, leading deep into the an-
teroconid. The teeth of Cephalomys
(Figs. 26B, 31G; Wood and Patterson,
1959: Fig. 20A) and Branisamys (Figs.
26A, 31H) demonstrate continuations of
this process, with the anteroconid now
separated from the metaconid.

This description is, of course, an at-
tempt to present a morphological series
and not a phylogenetic proposal. It is ob-
vious that there is a great deal still to be

502 Bulletin Museum of Coniparatiic Zoology, Vol. 149, No. 7

learned of deciduous tooth evolution in the rapid spread of the genus over a
the early cavioinorphs. We have pre- broad expanse of North America in post-
sented our interpretation (Fig. 31) of the glacial time. The more complex examples
patterns of these teeth, suggesting one of dm4 discussed above surely are in the
manner by which the pattern of Salla- process of developing a variety of neo-
imj.'i, Cephalomys and Branisamys could moqih crests, which, although sometimes
have been derived. We feel no great con- resembling mesoconids and mesolo-
fidence in the complete accuracy of our phids, can only have developed as par-
interpretation, but, with our present lack allelisms to such structures in any other
of knowledge, it is as logical as any other, rodents.

Any interpretation of the actual evolu- The pattern of dm4 of the Oligocene
tion of dm4 in the Caviomorpha is com- petromurids (see below, p. 523, for our
plicated by the extreme variability that usage of this term) is quite different. The
we have observed in this tooth among rear of the tooth is the same as in the ca-
Recent specimens of Erethizon dorsa- viomorphs and other rodents, formed of
turn. Some individuals of this species the hypoconid connected anteriorly with
(e.g., MCZ no. 51367, Fig. 31A) have the the entoconid and posteriorly with the
simplest caviomorph dm4 that we have posterolophid. There is also a continuous
seen. There may be an anterior spur into anteroposterior ectolophid. But the an-
the anterofossettid (MCZ no. 39006), a terior portion of the tooth is strikingly
small swelling on the ectolophid (MCZ different. The anteroconid is a very mi-
no. 39010), or a short spur in this location nor feature of dm4 (Wood, 1968; Figs.
(MCZ no. B9975). The structures on 1F,G; 2B; 3A,B; 4A; 5C; 8A; 9A-D;
these last two specimens look for all the 10A,E; 12F); either the trigonid basin
world like an incipient mesoconid and an opens forward, or there is a deep valley
incipient mesolophid, respectively. The between the protoconid and metaconid,
spur into the anterofossettid may form a which are near the front of the tooth; and
continuous lophid that divides the an- there is usually a strong mesolophid
terofossettid into two (MCZ no. B7752, (Wood, 1968: Figs. 1F,G; 2B; 8C; 9B-D;
Fig. 31 1), and may be joined by a second- lOE; 14D,G; 15E,G), similar to the struc-
ary spur to produce three divisions of the ture that apparently developed very re-
anterofossettid (MCZ no. 8846). The ap- cently in this tooth of Erethizon. Occa-
parent mesolophid may extend to the lin- sionally, the mesolophid is absent in the
gual margin of the tooth, where it may be Fayum genera, although there is still
connected to the posterior tip of the usually a mesoconid (Fig. 30E; Wood,
metaconid (Fig. 31 1). There may be a sec- 1968: Figs. 3A,B; 4A; 5C,F).
ond spur behind the apparent mesolo- Among the Miocene thryonomyoids,
phid (MCZ no. 7800). All intennediate Neosciiiromys africaniis (Fig. 30G), £/-
stages in these developments are present merimys and Myophiomys (Lavocat,
in the MCZ collections, often within 1973: PI. 28, Figs. 11-13) have very sim-
what were surely interbreeding popula- pie dm4, with a protoconid and metaco-
tions. We feel complete confidence that nid connected through the anterior cin-
dm4 of MCZ no. 51367 (Fig. 31A) is a gulum, but not through a metalophid.
primitive tooth, and that dm4 of MCZ no. This is very similar to dm4 of the Oligo-
B7752 (Fig. 311) is a highly specialized cene Neosciuromys sinionsi (Wood, 1968:
one. Fig. 5F), and is somewhat simpler than

Although we do not know, we strongly dm4 of Paraphiomys stromeri or P. pig-

suspect that this condition of extreme otti (Lavocat, 1973: PI. 26, Figs. 7-10).

variability in dm4 of Erethizon is a very The anterior portion of dm4 is unknown

recent development, perhaps related to in Sitnonimys (Lavocat, 1973: PI. 27, Fig.

Oligocene Rodents of Boli\ l\ • Patterson and Wood 503

1) and does not show in the figure of Ken- various ctenodactylids (Wood, 1977a:

ijamijs mariae (Lavocat, 1973: PI. 27, Fig. Fig. 5). Needless to say, we do not con-

6). sider either the entoptychines or the

Only in Diamantomijs luederitzi (La- theridomyids closely related to the ca-

vocat, 1973: PI. 25, Fig. 6) and Pomono- viomorphs, nor the theridomyids and

mys dubius (Fig. 301) is din4 complicated, ctenodactylids to the thryonomyoids. We

Here there are two or three cusps in the have merely selected examples of similar

metaconid-anteroconid area, although deciduous teeth in unrelated fonns.

they are very different from anything We do not see that the patterns of the

found in the Caviomorpha. deciduous teeth give any support to

However, in the thr>'onomyid Gaudea- thryonomyoid-caviomorph relationships.

mus, dm4 is formed of a series of parallel The great variability in dm4 within the

transverse ridges, more like what is seen single living species Erethizon dorsa-

in dm4 of T/iendom f/.s (Wood and Patter- turn, as well as that in Phiomys and

son, 1959: Fig. 35B) than like that in any Metaphiomys (Wood, 1968: Figs. 1F,G;

caviomorph, where the closest approach 2B; 3A,B; 4A; 5C; 8C; 9A-D; 10A,E; 12F)

we have seen to it is the complex Recent shows what extremes of variability may

Erethizon dorsatum (Fig. 311) or Brani- occur when, for some reason, there is

samys (Fig. 31H). As in the Caviomor- either strong selection toward increasing

pha, the talonid is of the type usually (but the complexity of a tooth, or a sudden re-

not in Gaudeamus) found in the molars, duction in selective pressures of any sort

but there has been extensive modifica- acting on it. We believe that this situation

tion of the anterior portion. demonstrates our ignorance of many of

Although there is similarity between the causes of evolutionary change in dm4.

dm4 of Gaudeamus or of Thryonomys It is obvious from the above that not

(Wood, 1962b: Fig. 2A) and those of ca- enough is known about lower deciduous

viomorphs, similarities as great or greater teeth of caviomorphs to be of great sig-

can be found elsewhere among the ro- nificance in unravelling their interrela-

dents. There is little or no difference be- tionships. As an aside, the evolutionary

tween dm4 of these thryonomyids and history of the deciduous teeth has been

that of Anomalurus (Wood, 1962b: Fig. sadly neglected, not only among rodents,

IB); and dm4 of the entoptychine geo- but in all of the Mammalia,

myids Gregor?/7?i(y,s curti/5' (Wood, 1936a: (18) Multiserial Incisor Enamel (by

Fig. 13) and Entoptychus cf. minor A.E.W.). Among rodents, there are three

(Rensberger, 1971: PI. 6, Fig. i) or of types of incisor enamel: the pauciserial,

Theridomys (Vianey-Liaud, 1972: Fig. 1) the primitive type, found in most Eocene

seem to us to be much closer structurally rodents; and two more evolved types,

to those of the Deseadan Caviomorpha uniserial and multiserial, found in later

than are those of the thryonomyids. The rodents (Korvenkontio, 1934). The tem-

pattern of dm4 of thryonomyids is also poral distribution of the three types

reminiscent of that of Theridomys, but makes any other relationship among the

again we do not believe that it indicates three very difficult to visualize,

thryonomyoid-theridomyoid affinities, as "Pauciserial enamel is a good structur-

Gaudeamus is quite advanced for an al predecessor for uniserial and multise-

Oligocene thryonomyoid, and the simi- rial enamel. It would appear to be the

larities of its dm4 to that of Theridomys ancestral condition .... Presumably ac-

must be pure parallelisms, Gaudeamus quisition of two layers in which the or-

having evolved from something like ganization of prisms is different served in

Phiomys. The patterns of dm4 of the pet- some way to strengthen the enamel ....

romurids are very suggestive of those of There is no available evidence to suggest

504 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

that one of the two surviving kinds is su- and bathyergids as well as in Pedetes and
perior to the other or that one of them has Oligocene to Recent ctenodactylids. In
ever evolved from the other" (Wahlert, the anomalurids, the enamel is uniserial.
1968: 13). As indicated, there are no The enamel of the upper incisor of Pro-
known cases of transition between uni- toptychus "appears to be pauciserial.
serial and multiserial enamel, but, in a Pauciserial and multiserial enamels are
few cases, transitional stages between similar, and a transverse section is not
pauciserial and the advanced types have ideal for distinguishing them" (Wahlert,
been found. 1973b: 4). In Prolapsus, both upper and

In pauciserial enamel, the inner layer lower incisors were pauciserial, but the

consists of prisms that may or may not be bands of the lower incisors were inclined

organized into lamellae. The bands are of at an angle of 19° and those of the uppers

variable width, the widths lying in gen- at an angle of 15° (Wood, 1973: 27). That

eral between those of the uniserial and is, the enamel of Prolapsus was in the

multiserial types (Wahlert, 1968: 4). "The process of making the transition from

structure appears similar to multiserial pauciserial to multiserial.

enamel but lacks the unifonnity of orga- Recently, a group of isolated teeth of

nization. Bands are usually not inclined ctenodactyloids has been described

.... Some pauciserial enamels have few- from the middle Eocene of Pakistan, as

er prisms per band, and thus the bands members of a primitive ctenodactyloid

are narrower, suggesting uniserial enam- family, the Chapattimyidae (Hussain et

el in appearance." al., 1978). Isolated incisors, referred to

The bands of the inner layer of multi- this family, were sectioned (op. cit.\ PI.

serial enamel are many prisms wide. 7, Figs. 1-3) and studied with the scan-

They are inclined at considerably greater ning electron microscope. The figures

angles toward the wear surface of the clearly show a diagrammatical pauciser-

tooth; the values given by Korvenkontio ial pattern (cf. Hussain et al., 1978: Figs.

(1934: 116-123) vary from 10° to 45° in 1-3 and Wahlert, 1968: Fig. IC), in spite

the upper incisor and from 22° to 50° in of their having been identified as multi-

the lower. serial (Hussain et al., 1978: 104). Sahni

In uniserial enamel, the bands of the (1980: 23) also concluded that these in-
inner layer are one enamel prism wide, cisors possessed pauciserial enamel.
These are inclined toward the tip of the The presence of pauciserial enamel in
tooth at angles that vary more than do what are undoubtedly ctenodactyloids
those of multiserial enamel, ranging from (very possibly ctenodactylids) from the
0° to 50° in both upper and lower incisors Eocene of Pakistan is of considerable im-
(Korvenkontio, 1934: 116-123). The uni- portance in demonstrating that, even if
serial type is more highly organized than there should turn out to be a distant re-
either of the others, and rather strikingly lationship between ctenodactyloids and
different in appearance from the pauciser- hystricognaths (which is an unproven hy-
ial, whereas "the structure [in pauciser- pothesis), multiserial enamel must have
ial enamel] appears similar to multiserial originated in the ctenodactylids indepen-
enamel but lacks the uniformity of orga- dently of its origin in the various hystri-
nization" (Wahlert, 1968: 4). For illustra- cognath groups. The suggestion by Hus-
tions of the three types of enamel, see sain ef a/, that their Chapattimyidae were
Korvenkontio, 1934: Pis. 3-27; Wahlert, ancestral to the Hystricognathi, including
1968: Fig. 1. the cylindrodonts, as well as the Cteno-

The multiserial type is present, so far dactylidae, is discussed below (p. 515)

as has been determined, in all cavio- and rejected,

morphs, all thryonomyoids, hystricids As with so many of the important mor-

Oligocene Rodents of Bolivta • Patterson and Wood 505

phologic features that charaeterize the
hystricognaths, there is a shortage of fos-
sils representing the intemiediate stages
between the primitive (pauciserial) and
advanced (multiserial) condition. In fact,
the closest things to such intemiediates
that we have encountered are the incisors
of Prolapsus and Cyclomijlus.

We believe the presence of multiserial
incisor enamel in all living hystrico-
gnaths cannot be taken as evidence of the
descent of the Caviomorpha from the
Thryonomyoidea, especially since the
only known possible transitions from typ-
ical pauciserial to multiserial enamel are
those found in the franimorph Prolapsus
and in the primitive tsaganomyid Cyclo-
tiiylus.

(19) Ectoparasites (by A.E.W.). Traub
has recently (1980) reviewed the zooge-
ography of mammalian Heas (Order Si-
phonaptera), chewing lice (Order Mal-
lophaga) and sucking lice (Order
Anoplura), and its implications for the re-
lationships and past geographic distri-
butions of their hosts. Although his con-
clusions are stated as being supportive of
an African origin for the Caviomorpha, a
careful study of his text indicates that his
data do not pennit one to choose between
transatlantic rafting of caviomorph ances-
tors and descent of both the Caviomorpha
and the Old World hystricognaths from
Eocene franimorphs with a Holarctic dis-
tribution.

The Anoplura infest sciurids, hetero-
myids, cricetids including gerbils and
microtines, murids, dipodids, caviids,
chinchillids, echimyids, bathyergids, Pe-
detes and Thryonomys, as well as lorises,
lemurs and tupaiids. Traub considered
(1980: 127-133) that these lice originated
in North America, infested rodents as
soon as the latter appeared, and spread
through most groups of the order. He be-
lieved that the anopluran families prob-
ably date from the initial differentiation
of the rodents. In fact, even the genus
Polyplax may well be that old, since it
"infests rodents (and occasionally other

hosts as well) in most parts of the
world, infesting such theraphions as phi-
omorphs like bathyergids and thryono-
myids; caviomorphs like caviids,
chinchillids and echimyids; dipodids;
pedetids; the only true cricetids in the
Ethiopian region; other cricetid rodents
elsewhere; xerines and other sciurids;
heteromyids; loricids; lemuroids; tu-
paioids; and leporids. The majority of
these associations are surely primary"
(Traub, 1980: 131). This suggests the pos-
sibility that Polyplax as a genus may date
from the Eocene or Paleocene.

The distribution of the rodent Ano-
plura fits well with a North American (or
Asiatic) origin of the order, and a world-
wide distribution on and with the ro-
dents. There are some peculiarities in
their known distribution, probably due to
a careful search not having been made of
all possible hosts, but the data do not
seem to me to give any support to an Af-
rican origin of the Caviomorpha in pref-
erence to a North (or Middle) American
origin.

One suborder of Mallophaga, the Am-
blycera, are primarily parasites of mar-
supials, with one family, the Boopidae,
restricted to Australia and another, the
Trimenopomidae, largely South Ameri-
can but extending into Central America
(Traub, 1980: 122-127). A third family,
the Gyropidae, are also South American,
and perhaps were derived from the Tri-
menopomidae. The trimenopomids in-
fest members of the Caviidae, Chin-
chillidae and Echimyidae, as well as
marsupials, and the gyropids are found
on these caviomorph families and the
Dasyproctidae (Traub, 1980: 122), but
both families are apparentl>' absent on
the Erethizontidae, Octodontidae, Abro-
comidae, Hvdrochoeridae, Dinomyidae
and Cuniculidae. Clay (1970: 94-96) sug-
gested that when the ancestral cavio-
morphs reached South America they had
no amblyceran parasites; that they ac-
quired trimenopomids from South Amer-
ican marsupials; that these then evolved

506

Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

into a distinct family, Gyropidae, on the
cavionioriDhs; and that this was followed
by another cavioniorph infestation by
other trimenopomids from South Ameri-
can marsupials. Certainly the Amblycera
have no bearing on the origin of the Ca-
viomorpha other than to suggest that
these rodents came to South America
from somewhere else.

"While members of the family Tricho-
dectidae (suborder Ischnocera [Order
Mallophaga]) are legion on fissiped Car-
nivora, Perissodactyla, Artiodactyla and
Hyracoidea, they are found on only two
groups of rodents, viz. an isolated branch
occurring on the caviomorph Erethizon-
tidae (New World porcupines) and a sec-
ond distinctive taxon on the geomyids
. . ." (Traub, 1980: 122). Traub gives no
indication as to whether these lice were
found on North American or South Amer-
ican erethizontids. However, the family
clearly is primarily non-rodent and
would seem to have a fundamentally
Holarctic distribution. The erethizontids
presumably acquired the lice from car-
nivores, perissodactyls or artiodactyls
either in South or in North America, and
the geomyids would have done the same
in North or Middle America. The tricho-
dectids therefore would seem to have no
bearing on the origin of the Caviomor-
pha, except that they ultimately had a
northern origin.

The fleas are a more complicated mat-
ter, both as to their own classification and
as to their import for the relationships of
their hosts. Traub believed (1980: 157)
that fleas were in existence on Pangea
and were represented from their separa-
tion on both Gondwanaland and Laura-
sia. He believed that the major flea fam-
ilies originated on the various continents
presumably early in the Cretaceous, and
that the subfamilies of modem fleas prob-
ably originated in the Cretaceous (1980:
113). The presence of the modern genus
Palaeopsylla in the Baltic Amber of Eu-
rope shows that flea genera are very long-
lasting, and thus infestation by members

of a single genus of fleas does not indi-
cate that two populations of mammals
may not have been separated also for a
very long time. Furthermore, another
modem genus, C tenophthalmus , accord-
ing to Traub (1980: 143) is apparently the
immediate ancestor of Palaeopsylla, so
that it must be even older. It does not
seem surprising, in that case, that Cte-
nophthalmus has a geographic distribu-
tion that is not at once readily under-
standable.

The large flea family Heteropsyllidae
was apparently of northern (Laurasian)
origin (Traub, 1980: 157). Some of its de-
rivative subfamilies and tribes are South
American and one is also Australian. One
subfamily, the Ctenophthalminae, in-
cludes four tribes, two Neotropical, one
Nearctic and one Holarctic and Ethiopi-
an. Traub has suggested that these are in-
dications of the African origin of the Ca-
viomorpha, but, if one accepts the
antiquity of flea taxa mentioned above,
these data seem as readily explicable by
having the infestations of the Caviomor-
pha and the African hystricognaths date
from their common Eocene franimorph
ancestors, who reached South America
and Africa by independent routes.

Therefore, although Traub's conclu-
sions read as though it is necessary to as-
sume a transatlantic migration of ca-
viomorph ancestors to explain the present
distribution of the ectoparasites, I do not
believe that this necessarily follows from
his data. Traub infomis me (letter dated
Feb. 27, 1981) that "We are not as far
apart on African-South American connec-
tions as our earlier correspondence may
have indicated."

(20) Other features (by A.E.W.). There
have been certain other supposed evi-
dences of direct caviomorph-thryonomy-
oid relationships that were not cited by
Lavocat (1973), but which were specified
by Lavocat (1971a) or Hoffstetter (1975:
521, 522). These include the presence of
a sacculus urethralis, serology, and fetal
membranes, subjects concerning which

Oligocene Rodents of Bolivia • Patterson and Wood 507

we feel that we cannot speak with any
authority. We do not discuss these here,
since we feel that we have previously
(Wood and Patterson, 1959: 391-392
413—414 and 414, n. 34; Wood, 1974b
47-49; 1975a: 76-77 and Fig. 1; 1977a
135) said everything that can be said on
these topics until further evidence is pro-
duced by specialists in the various fields.

Recently, Sarich and Cronin (1981)
have discussed rodent molecular system-
atics, concentrating on cricetids and hys-
tricognaths. They mention, as one prob-
lem in this connection (op. cit.: 407) the
"poor fossil record" of rodents, which
might have been true a half century ago,
but certainly is not today. They feel
(1981: 419) that cricetids, "given their to-
tal absence in the appropriate North
American fossil record, are, then, almost
certainly of African origin . . . we see no
great problem in having several cross the
Atlantic in the late Eocene and early
Oligocene from Africa to . . . South Amer-
ica." This total absence of cricetids from
the Oligocene and Miocene of North
America will come as a surprise to most
mammalian paleontologists, as will the
suggestion that they originated in Africa,
where they are unknown in the Oligo-
cene and rare in the early Miocene. Sar-
ich and Cronin (1981: Fig. 7) illustrate
their ideas of how the cricetids moved
from Africa to South America to North
America.

In discussing the hystricognaths, Sar-
ich and Cronin mention (1981: 410) the
"total lack of hystricognaths ... in North
America (except, of course, for the recent
migrant Erethizon," apparently disre-
garding the description of hystricognathy
in Prolapsus (Wood, 1972; 1973) and
Guanajuatomys (Black and Stephens,
1973). They develop a molecular phylog-
eny of the Hystricognatha, with two
groups, one including Hydrochoerus,
Cavia, Erethizon, Dasyprocta and Chin-
chilla, and the other Hystrix, Bathyergus
and Petromus (these two closely related),
Hoplomys, Proechimys, Myocastor, Oc-

todon and (probably) Capromys, Geocap-
romys and Plagiodontia. These two
groups, they believed, crossed from Af-
rica to South America separately, proba-
bly (although this point is not specifically
made by them) after the invasion of South
America by the cricetids.

A final surprise to a rodent paleontol-
ogist is Sarich and Cronin's mention
(1981: 419) of "ancient murids in Austra-
lia .. . ."

In view of all the misconceptions that
these authors were laboring under as to
the fossil record of rodents, and the im-
possibility that their ideas of cricetid re-
lationships are correct, there seems to me
to be no reason why their data on hystri-
cognath relationships could not fit equal-
ly well with our thesis of the origin of the
suborder in some part of Holarctica and
their subsequent invasion of South Amer-
ica from Middle America and of Africa
from southwestern Asia, as with their
proposal of multiple transatlantic hystri-
cognath migrations from Africa to South
America.

(21) Summary (by A.E.W.). As we have
tried to make clear in this section, there
are obvious similarities and obvious dif-
ferences between the Caviomorpha and
the Old World Hystricognathi. There is,
however, not a single feature that char-
acterizes all members of these two
groups and that is found nowhere else
among the rodents.

In our opinion, the most basic features
(hystricognathy and the anterior deep-
ening of the pterygoid fossa) are resul-
tants of the lengthening of M. pterygoi-
deus internus. The former occurs,
incipiently to fully developed, among
franimorphs. The condition of the ptery-
goid fossa is unknown in these fomis; we
suspect that it had deepened variable
amounts. The dorsal opening of the fossa,
in geomyoids, Spalax and Aplodontia, is
presumably the result of parallelism. We
suspect that these two characters (or rath-
er, the lengthening of M. pterygoideus
internus which caused them) are the only

508

Bulletin Museum oj Comparative Zoolo^i/, Vol. 149, No. 7

valid diagnostic features of the Hystri-
cognathi.

There are a number of respects in
which the Bathyergoidea are quite dif-
ferent from most or all other hystrico-
gnaths. These include their protrogomor-
phy (except Bathijergoides, which was
hystricomorphous in a manner sui gene-
ris); the rather distinctive ear ossicles
and their much greater fusion than in
any other rodents; the absence of gyren-
cephaly; and their very different nema-
todes.

Three features separate the Erethizon-
tidae from all other modern hystrico-
gnaths — the structure of the middle ear,
the presence of an internal carotid artery,
and the serology. These last two are so
distinctive that the students of these sub-
jects felt that the Erethizontidae deserve
a separate sub- or infraorder of their own.

A number of features are shared with
the completely sciurognathous ctenodac-
tylids — hystricomorphy, fusion of mal-
leus and incus, multiserial incisor enam-
el, and perhaps a sacculus urethralis. The
ctenodact>'lids, however (Wood, 1977a),
clearly have an ancestry entirely inde-
pendent of that of the hystricognaths.

In several features, some or all of the
Caviomorpha are clearly more primitive
than any known Old World hystrico-
gnath, supporting the derivation of the
Old World fomis from the New, but mak-
ing the reverse derivation difficult to ac-
cept. These include the rather wide-
spread occurrence of non-fusion ot the
ear ossicles among caviomorphs, the
presence of an internal carotid artery in
the erethizontids, the four-crested molars
of the primitive caviomorphs, -^nd the
universal replacement of dnV in the De-
seadan caviomorphs. Platypittamys has
less well-developed hystricomorphy than
any known thryonomyoid or hystricid.

The ascending process of the alisphe-
noid has been reduced in living hystrico-
gnaths, though to a lesser extent in the
bathyergids than elsewhere; its condition
is unknown among franimorphs. A simi-
lar reduction characterizes muroids.

The amount of gyrencephaly is greater
among hystricognaths than elsewhere in
the order. However, as pointed out above
(p. 490), gyrencephaly is merely a sec-
ondary result of increase of brain size,
and a greater proportion of hystrico-
gnaths than of sciurognaths have attained
large size.

A number of the features cited by La-
vocat and Hoffstetter do not seem to us
to have any significance whatsoever.
These include the jugal fossa, generally
absent and highly variable when present;
the heaviness of the malar and the sin-
uosity of its anterior suture, both ex-
tremely variable and not confined to
hystricognaths; the lateral strip of the
supraoccipital, randomly present
throughout the Rodentia; the cleft be-
tween the bulla and the squamosal, vary-
ing in its developnient from large to ab-
sent; the development of the maxillary on
the floor of the orbitotemporal fossa,
which does not differ in any respect from
the primitive rodent condition; and the
exceedingly variable retention of an in-
dependent interparietal.

It seems to us to have been demonstrat-
ed that the known thryonomyoids were
morphologically incapable of having giv-
en rise to the Caviomorpha, and that the
evolutionary trends within the thryon-
omyoids were such that it would be dif-
ficult to envisage any near ancestor that
could have had such capabilities.

Some authors (e.g., Sarich and Cronin,
1981) have adopted the point of view that
any evidence indicating a closer rela-
tionship of Old World and New World
hystricognaths to each other than to any
sciurognaths is evidence in support of
transatlantic rafting of caviomorph ances-
tors, ignoring the fact that such a closer
relationship would be just as expectable
under our hypothesis of the derivation
of the Caviomorpha from Middle Ameri-
can Eocene franimoiphs.

We feel that these results indicate that
the Suborder Hystricognathi is a reason-
able concept; that the Caviomorpha and
Bathyergoidea originated from the Fran-

Oligocene Rodents of Bolivia • Patterson and Wood 509

imorpha independently of the Hystrici- availability of numerous fomis, some of

dae and Thr>'onomyoidea; and that there them critical, to one or another of these

has been extensive parallelism among authors, much new information pertinent

the later hystricognath groups after their to caviomorph higher taxonomy has been

separation from their Holarctic frani- made available by their studies,

morph Eocene ancestors. The distinctness of the erethizontids is

fully confimied, all the new evidence

The Classification of the Suborder separating them clearly from the other

Hystricognathi caviomorphs. Bugge, as did Moody and

THE SUPERFAMILIES OF THE P^^^^ff/l^'^^^ ^'^'^^^ ^u^""'^- V"'' ^'''

Caviomorpha ^V^^Xf d. ^hat they merit a sub- or mfraorder

oi their own, and has proposed the Sub-
It is rather generally agreed that there order Erethizontomorpha for them (1974a:
are four major groups within the South 56, 70, 71, 73). He did so primarily on the
American Caviomorpha. This was first basis of their retention of an internal ca-
clearly recognized by Simpson (1945: rotid artery, a primitive structure that has
94-99, 209-213), who proposed four been lost in all other living hystrico-
superfamilies: Erethizontoidea (Erethi- gnaths. It is difficult for us to assess the
zontidae), Octodontoidea (Octodonti- significance of this. Of the Deseadan ro-
dae, Ctenomyidae, Abrocomidae, Echi- dents, the basicranial region is known
myidae, Capromyidae), Chinchilloidea only in Incamys, and here there is what
(Chinchillidae) and Cavioidea (Eocardi- might be a carotid foramen (p. 411). Ap-
idae, Caviidae, Hydrochoeridae, Dasy- parently this area is preserved in a skull
proctidae, Dinomyidae and Heptaxodon- of Branisamys, SAL 102, but is still bur-
tidae). Apart from various shufflings (the ied in matrix (Lavocat, 1976: PI. 2, Fig.
Acaremys group to the Octodontidae, 1). The skull of the Fayum thryonomy-
the Eumegamys group to the Dinomyi- oids is unknown except for maxillary
dae) and changes in rank (elevation of fragments. It is possible, therefore, that
the Cuniculinae to family, lowering of Oligocene hystricognaths possessed this
the Ctenomyidae and Capromyidae to artery (Reithroparamys has what is, to-
subfamily), about which there is inevita- pographically, a carotid canal — Wood,
bly less agreement, his arrangement has 1962a: 122), and that only the erethizon-
stood in its essentials. In our study of the tids have retained it to the present. To
Patagonian Deseadan rodents (Wood and further complicate matters, as indicated
Patterson, 1959) we stressed that the er- above (p. 485), possession of a "carotid"
ethizontids occupied an isolated position foramen by a rodent does not prove pos-
in relation to the rest, and supported the session of an internal carotid artery (Wah-
removal of the dasyproctids, dinomyids, lert, 1974a: 373).

cuniculids and heptaxodontids from the In view of all of these uncertainties —
Cavioidea, transferring them to the Chin- and we believe that we are in very close
chilloidea. agreement with Lavocat (1976: 82-83) re-
in recent years several publications garding them — we doubt if possession of
dealing wholly or in part with Recent an internal carotid by erethizontids can
hystricognaths have appeared: Gorgas safely be adduced to prove their separa-
(1967) on the stomach and intestinal ca- tion at a sub- or infraordinal level. For
nal; Bugge (1971, 1974a, 1974b) on ce- the present at least, we retain them as a
phalic arteries; Woods (1972, 1975) on superfamily of the Caviomorpha, stress-
Jaw, hyoid and fore limb musculature; ing, as we have before, that they were in
Woods and Rowland (1977) on the skin all probability a very early offshoot from
musculature; and George and Weir the ancestral stock.
(1974) on chromosomes. Despite the un- Lavocat (1976: 82) wrote: "En raison

510 Bulletin Museum of Comparative Zoology, Vol. 149, So. 7

des traveaux de J. Bugge . . . Wood (1975) were of considerable importance in the
admet actiiellement que les Erethizonti- Antillean fauna before the arrival of man,
dae pourraient etre assez distants des and strengthens the possibility that they
autres Caviomorphes." This was "admit- had been there for a considerable time,
ted" much earlier (Wood and Patterson, Rodriguez et al. (1979) have discussed
1959: 377, 391-392). What Wood did the Cuban capromyine (capromyid in
state (1975a: 77) was that, if Bugge was their terminology) genus or subgenus
correct in his concept of Erethizontomor- Mesocapromys as a capromyine whose
pha, two invasions of South America numerous primitive, echimyid-like fea-
would be required, the implication being tures demonstrate the derivation of the
that visualizing multiple successful trans- capromyines from the echimyids. We ac-
atlantic raftings became exceedingly dif- cept their identification of the signifi-
ficult, and that North America was hence cance of this fomi as a primitive capro-
the more likely source area. myine, and believe that this reinforces
The coherence of the Octodontoidea is our opinion that the capromyines deserve
not called in question by the new studies only subfamilial status,
of Recent fonns. Unfortmiately, too few Following Patterson and Pascual (1972:
of the numerous echimvid genera were Fig. 12) we place the Ctenomyidae (as a
available to the various authors to pro- subfamily) in the Octodontidae.
vide evidence with which to test the ar- The "classical" cavioid families, Cavi-
rangement of the echimyid subfamilies idae and Hydrochoeridae, differ sharply
suggested above (Fig. 6). Woods (1972: from other hystricognaths in being su-
190) found Proechimys (Heteropsomyi- perficially non-h\stricognathous, the an-
nae) and Echimys (Echimyinae) to be gle leaving the ramus close to the plane
"myologically more different from each of the incisor. The masseteric crest is
other than would be expected if these high on the jaw and heavily built, with a
genera were in the same subfamily." To deep groove extending back to the vicin-
judge from Gorgas' descriptions and fig- ity of the condyle between the crest and
ures, Cercomys and Proechimys (Heter- the side of the ramus; the crest does not
opsomyinae), Kannahateomys (Dact>'lo- reach the angle, the insertions of parts of
myinae) and Echimys (Echimyinae) do the masseter in that region being marked
show certain differences each from the onl\ by faint ridges. The lower incisor is
others in stomach structure, but a larger short, ending under M2. On the basis of
representation is obviously needed. As these distinctive features some authors
indicated above (p. 394), we have trans- (e.g.. Miller and Gidley, 1918; Ellerman,
ferred Chaetomys from the Erethizonti- 1940) have separated the hystricognaths
dae to be a subfamily of the Echimyidae into two grand divisions, one for these
because of its retention throughout life of two families and another for all the rest.
dnV, a characteristic of later echimyids This, of course, was going too far; cavies
and otherwise (now that we have made and capybaras are specialized in these
this transfer of Chaetomys) unknown respects. The extinct and broadly ances-
among the Caviomorpha. tral famil\- Eocardiidae (Scott, 1905: 465)
Our knowledge of fossil capromyines possessed a more nomial hystricognath
has been limited to Recent or sub-Recent angle and a slighter masseteric crest that
material. The description of a number of is lower on the side of the jaw and runs
new subgenera and species of capro- back and down to the angle. The short-
myines from Cuba (Varona and Arredon- ened incisor, however, had been ac-
do, 1979) brings the record of this group quired b\' Deseadan time (Wood and Pat-
in the Pleistocene or Recent of Cuba to terson, 1959: 371). Hartenberger (1975)
8 genera or subgenera and 19 species, suggested that the Hydrochoeridae de-
which suggests that the capromyines serve to be ranked as a superfamily,

Oligocene Rodents of Bolivi^\ • Patterson and Wood 511

largely on the basis of "ViUarroelomijs,"
which turned out not to be a hydrochoe-
rid (p. 444). Capybaras clearly are good
cavioids.

The principal uncertainties in cavio-
morph classification arise over the affin-
ities of the Dasyproctidae, Dinomyidae
and Cuniculidae. Are they cavioids, as
Simpson and others maintained, or
should they be placed elsewhere? The
evidence advanced to support removal of
dasyproctids, dinomyids and cuniculids
from the Cavioidea was dental (Wood
and Patterson, 1959: 324^29). All mem-
bers of these families possess lower in-
cisors of the usual length, and all of them
have cheek teeth that differ in structure
from those of cavioids s.s. Noting resem-
blances between the Deseadan chinchil-
lid Scotamys and the contemporary dasy-
proctid Cephalomys in the structure of
unworn or little worn molar crowns (the
resemblances are real, no matter what in-
terpretation one may wish to put upon
them) and in general agreement with
Schaub (in Stehlin and Schaub, 1951:
369-370), we transferred the families in
question to the Chinchilloidea.

This proposal has of late met with less
than universal approbation. Nearly all
the new evidence is flatly against it. Gor-
gas (1967: 366-367) has found a close re-
semblance between the Cavioidea s.s.
and the Dasyproctidae, etc., and decided
differences between both and the Chin-
chillidae in stomach structure. Bugge
(1971: 534) came to the same conclusion
on the basis of the cephalic arterial sys-
tem, as did Woods (1972: 190) for the
muscular system and Vucetich (1975:
490) for the middle ear.

The evidence from chromosomes is in-
teresting— and a little more complicated.
George and Weir (1974: 90-93) distin-
guished four major groups of cavio-
morphs. One of them includes the Erethi-
zontidae only, another the Octodontoidea.
So far so good; it is in the composition
oi the other two that the surprise comes,
especially in view of the virtually unan-
imous agreement reached — quite in-

dependently— by the other authors just
mentioned. One of George and Weir's
groups includes the Caviidae and Dasy-
proctidae, and the other the Cuniculi-
dae, Hydrochoeridae and Chinchillidae
{Dinomys was not available to them).
They did, however, indicate {op. cit.:
100-101) a possible means of attaining
(their Fig. 8) the Hydrochoerus and
Cuniculus karyotypes "if it is assumed"
that these forms "are cavioids with affin-
ities with the Caviidae and Dasyprocti-
dae respectively." We lay no claim to
competence in karyology, but this possi-
bility seems reasonable to us considering
the weight of the other evidence.

Subjected to so heavy a barrage, we
feel compelled to retreat and to replace
the Dasyproctidae, Dinomyidae and
Cuniculidae in the Cavioidea, although
we admit to doing this without enthusi-
asm. Members of the group thus reunited
agree in the total lack of a lateral process
of the supraoccipital, in contrast to other
caviomorphs, and most of them have a
descending flange of the lacrimal, the di-
nomyids and Cuniculus lacking it. We
wish to stress, however, that, within the
Cavioidea as thus re-expanded, there is
a split between the Dasyproctidae, Di-
nomyidae and Cuniculidae on the one
hand and the Cavioidea s.s. on the other,
and that this split occurred well prior to
the Deseadan. We believe that this
should be recognized in classification,
and propose two subdivisions of the su-
perfamily, that we call Series A and Se-
ries B (see below, p. 523). We offer no
names for these subdivisions, since we
consider the recent proposal by Mc-
Kenna (1975a: esp. 40^2), according to
which every identified stage in phyletic
branching deserves special taxonomic
status and a special name, to be a fright-
ening example of the inflatio ad ahsur-
dum of higher categories that must inev-
itably result if cladistic classification is
unthinkingly pushed to its logical limits.
For a comment on a very similar proposal

512 Bulletin ^Juset^m of Comparative Zoolog,\i, Vol. 149, No. 7

made nearly half a century ago, see Simp- is isolated in time. It is separated from

son, 1945: 25-26, footnote 4. the one following, the Colhuehuapian,

The position of the Antillean Heptax- by perhaps 10 m.y., and it itself followed

odontidae, usually grouped with the Das- upon a hiatus in the record that surely

yproctidae, Dinomyidae and Cuniculi- lasted as long or longer. The ancestral

dae, is ver\' uncertain. Elasmodontomys caviomorphs arrived in South America

(of which Heptaxodon is a synonym) is during this hiatus — just when within it is

not an echimyid, as are most West Indian not known, although we suspect near its

genera, since it replaces dnV (Wood and beginning — but it is very clear that the

Patterson, 1959: 326), as does Ambhjrhi- rodents encountered in the Deseadan are

za. It possesses a very large lateral pro- at a not inconsiderable remove from the

cess of the supraoccipital, a structure ab- beginning of the caviomorph radiation,

sent even in rudimentary form in the The initial burst, however, had not yet

families under discussion and also in all spent itself by the Deseadan. As a result,

cavioids s.s. The cheek teeth are hypso- we encounter fomis that are difficult to

dont, but not hypselodont, whereas in the classify with assurance. We have done

Deseadan chinchillid Scotamys the root- our best to indicate in the pages above

less condition had already been achieved, and in our previous contributions such

Having thus been maneuvered, so to doubts and reservations as we have on

speak, into a corner, we end by offering these scores. We are reasonably sure that

the very unsatisfactory suggestion that in the systematic positions of Platypitta-

our present state of ignorance the family inys, Migraveramus, Deseadomys, Inca-

may best be listed as Caviomorpha iti- mys, Scotamys, Asteromys,Chubutomys,

certae sedis as to superfamily. and Protosteiromys are as we have stat-

Previously (1959: 321-322) we were ed. We have presented the evidence for
skeptical as to the distinctness of the placing Sallamys, Xylechimys, Cephalo-
Family Perimyidae of Landry (1957: 55, mys, Litodotitomys, and Branisamys, but
59), which he placed in the Chinchilloi- we realize that there is room for discus-
dea between the Cephalomyidae and sion concerning them, as by Lavocat
Chinchillidae. Vucetich (1975), however, (1976: 72-73). We agree with him that
has brought forward evidence from mid- much is indeed "provisoire." Hoffstetter
die ear structure that has convinced us of (1976: 8) has stated that "la classification
its distinctness, as a family related to the des Caviomorphes, et surtout celle des
Chinchillidae. Other reasons for the ac- fomies ante-Miocenes, demande une re-
ceptance of this group as a valid chin- vision d'ensemble." Presumably he was
chilloid family were presented by Bon- referring to classification at higher rather
desio et al. (1975: 441-442). As pointed than lower taxonomic levels. Better ma-
out by these authors, the family name, terial of the Deseadan fomis will help
however, should be Neoepiblemidae, here to some extent, but we suspect that
dating from Kraglievich, 1926. These even if we knew all of them from com-
now join the Chinchillidae, relieving the plete skeletons, problems might still re-
loneliness of their otherwise solitary main. What is needed is the filling of the
splendor within their superfamily. The hiatuses so that lineages can be traced
resemblances between Cephalomys and rather than inferred.
Scotamys in the cheek teeth suggest the

possibility of a cM3mmon ancestry for the ^^^^^ Taxonomic Problems of the
Cavioidea and Chuichiiloidea earlv ni tt

, i.rr 4.- 4.- ' HySTRICOGNATHI

caviomorph dirierentiation.

To conclude this section, we mention (1) General (by A.E.W.). One of the

a well known fact. The Deseadan fauna difficulties in trying to reach an accept-

Oligocene Rodents of Bolivia • Patterson and Wood 513

able classification of the hystricognath- should be credited to Waterhouse, 1839,

ous rodents is that, in our opinion, a num- rather than to Brandt, 1855 (who merely

ber of authors shift between talking about gave names to the groups that Water-

hystricomorphs and hystricognaths as house had identified), we believe that

they go along, as though these two terms this statement by Lavocat is in complete

were synonymous. Recently, this ap- contradiction of a vast amount of evi-

proach seems to have been adopted by dence. No return to the classification as

Lavocat (1978: 85). The hystricognathous proposed by Brandt is possible. One may

rodents are those in which the angular assume that all the hystricomorphous

process of the lower jaw arises from the fomis are related, or that all the hystri-

lateral surface of the alveolus of the lower cognathous ones are, or that only those

incisor. The hystricomorphous ones are fonns that are both hystricomorphous and

those in which the origin of M. masseter hystricognathous belong together. But a

rnedialis has spread forward through the combination of all three is impossible. It

infraorbital foramen to arise from the is a fundamental requisite that, in talking

snout, resulting in great enlargement of about these complicated matters, all au-

the infraorbital foramen. These two fea- thors should be careful to use precisely

tures are not universally associated, defined tenns.

There are sciurognathous hystrico- (2) Infraorder Franimorpha (by
morphs (Anomaluridae, Ctenodactylidae, A.E.W.). Members of the Infraorder
Dipodoidea, Pedetidae and Therido- Franimorpha, varying from incipiently to
myoidea) as well as protrogomorphous fully hystricognathous, are known from
hystricognaths (Bathyergidae, Tsagano- the late Paleocene to early Oligocene
myidae, Cylindrodontidae, many reithro- (and possibly late Oligocene — Wood,
paramyids). There is no possible way 1980b: Table 1) of western United States
that the two of these features can be con- and the late Eocene and earliest Oligo-
sidered as both being always associated cene of Mexico (Ferrusquia, 1978: 206).
in a single rodent taxon and neither ever We believe them to have reached Asia at
being present anywhere else. As stated an unknown time or times during the
elsewhere (Wood, 1980c: 270), hystrico- Eocene (probably both early and late
gnathy in rodents began to appear in the Eocene; see above, p. 449). Recent work
late Paleocene, and became gradually (Wood, 1980b: 6; 1981: 86) has indicated
more fully developed. No trace of hystri- that the North American Eocene-Oligo-
comorphy is known before the middle cene Family Cylindrodontidae (repre-
Eocene (Theridomyoidea: Wood, 1974b: sented in Central Asia, according to the
Fig. 4; Ctenodactylidae: Shevyreva, classification we present here, by the
1972b: 134-135, Fig. 1; 1976: Figs. 4-5, Oligocene genera Pseudocylindrodon
PI. 1, Fig. 3b; Prolapsus: unpublished and Morosomys and by the Eocene to
data). Oligocene Ardynomys) were at least sub-
As an illustration of this confusion of hystricognathous and hence were frani-
hystricognathy and hystricomorphy, we morphs.

may quote Lavocat (1969: 1497) who stat- The basic members of the infraorder

ed that "Du point de vue systematique, are the rodents united by Wood (1962a:

nos conclusions confirment et complet- 117-118) as the subfamily Reithropara-

ent I'intuition fondamentale de Brandt myinae of the Family Paramyidae. As he

en montrant que le sous-ordre des Hys- pointed out (1962a: 117), the reithropara-

tricomorphes est une veritable unite na- myines were subhystricognathous; at

turelle, fondee sur des relations phyle- least one genus (Rapanjjys; Wood, 1962a:

tiques precises et etroites." Aside from 148) seems to have had an enlarged in-

the fact that the "intuition fondamentale" fraorbital foramen. Dawson (1977: 197,

514

Bulletin Museum of Comparative Zoologij, Vol. 149, No. 7

n.) has (lueried the justification of calling
the Reithroparaniyinae subhystrico-
gnathous, saying "The 'very incipient'
hystricognathous condition of Reithro-
paramys (Wood, 1973, p. 30-31) seems to
me to be still a sciurognathous jaw, 'in-
cipient' only a posteriori.^' This may be
true, although the hystricognathy of Rei-
throparamijs huerfanensis is quite well
developed (Wood, 1962a: Fig. 46B), but
as pointed out elsewhere (Wood, 1980c:
270) it is inconceivable that hystricogna-
thy appeared instantaneously; it must
have developed gradually, and the forms
in which it was developing deserve to be
called subhystricognathous or incipiently
hystricognathous. Current taxonomic
usage would seem to require them to be
placed in the Suborder Hystricognathi.

The Franimorpha may be defined
(modified from Wood, 1975b: 79) as: sub-
hystricognathous to hystricognathous;
protrogomorphous to hystricomorphous;
with or without a distinct interparietal;
alisphenoid does not extend much if any
above level of glenoid cavity; at least
sometimes a post-condyloid process on
the mandible indicating the differentia-
tion of the M. masseter lateralis profun-
dus, pars posterior, deep division; mo-
lars cuspidate or crested, never a
mesocone or mesoloph, sometimes a me-
soconid but rarely a mesolophid; normal
replacement of deciduous molars; and in-
cisor enamel pauciserial or transitional
from pauciserial to multiserial.

When Wood (1975b: 78-79) erected
the Franimorpha, he placed the Reithro-
paraniyinae in it, but left the other
subfamilies of the Paramyidae in the
sciurognath Infraorder Protrogomorpha,
and left the Reithroparamyinae in the
Paramyidae, which thus straddled the
boundary between the two suborders.
Wood still likes this idea — it indicates
that the earliest known sciurognaths and
hystricognaths were very closely related,
but that they had already started to di-
verge in directions that would ultimately
lead to such differences that their de-
scendants were clearly very distinct from

each other. Certainly no taxonomist liv-
ing in Eocene time would have separated
them at a level higher than that of
subfamilies. Such a solution, of course,
does not make for a neat set of cubby-
holes. Patterson has now persuaded
Wood that, although this is perhaps rea-
sonable, it would not be generally ac-
ceptable.

We therefore recognize the Reithropar-
amyidae as a distinct family, that we de-
fine as follows: franimorjohs with incip-
ient hystricognathy; protrogomorphous
in earlier and best known fomis, infraor-
bital foramen enlarging in at least some
later genera; bulla not co-ossified with
the skull in earliest fonns, but large and
fused to the skull by middle Eocene; cor-
onoid processes large and high; post-con-
dylar process present at least sometimes;
cheek teeth relatively simple, primitively
tritubercular but with a distinct hypo-
cone in upper molars and usually in P*;
metalophid incomplete, trigonid basin
opening posteriorly; entoconid separated
from posterolophid; transverse crest from
entoconid toward hypoconid developing
progressively, ultimately becoming a
complete hypolophid; posterolophid pro-
gressively angulate at point of attachment
to hypoconid; incisors narrow and com-
pressed, enamel thin; skeleton, in the
two cases where it is known, with elon-
gate hind legs; size small to medium.

Guanajuatomys from the later Eocene
of Mexico, and Prolapsus from the mid-
Eocene of the Big Bend area of Texas,
are both fully hystricognathous. Prolap-
sus, at least, had an enlarged infraorbital
foramen. Both have been described, so
far, on fragmentary materials. For both,
better but undescribed materials are now
available. The cheek teeth do not indi-
cate close relationships between the two,
or between either and any other well
known group. We suspect that they be-
long to one or more Middle American
franimorph families, but we know of no
justification for the erection of such a
family or families at the present time.

(3) Cylindrodontidae and Tsagano-

Oligocene Rodents of Bolivia • Patterson and Wood 515

mijidae (by A.E.W.). The Cylindrodonti-
dae have long been recognized as a pro-
trogomorph family, but the angular
processes have been broken off from all
described material. Recently a specimen
was found that clearly had a hystrico-
gnathous angle and a small post-condy-
loid process (Wood, 1981: 86; in press a),
which had led to the transfer of the
family from the Protrogomorpha to the
Franimorpha. The range of the family
is middle Eocene (Bridger and Big Bend
of Texas) through early Oligocene (Chad-
ronian) and possibly late Oligocene of
western United States, early Oligocene
of Canada, earliest Oligocene of Mexico
(Wood, 1980b: 6), as well as late Eo-
cene and early (Ardyn Obo) and middle
(Tatal Gol) Oligocene of central Asia.
The incisor enamel was pauciserial in
middle Eocene specimens of Mysops
(Wahlert, 1968: 15), but uniserial in the
early Oligocene Cylindrodon {op. cit.:
15).

Hussain et al. (1978: 101) suggested
that the Chapattimyidae, from the middle
Eocene of Pakistan, were ancestral,
through Petrokozlovia of the middle to
upper Eocene of central Asia (Shevyreva,
1976: 31) to the middle Eocene and later
cylindrodonts of North America. This
seems, on the face of it, most improbable.
The tooth pattern of the Chapattimyidae
does not suggest to me any affinities with
later rodents other than the Ctenodactyl-
idae. Structurally, the teeth of Chapatti-
tnys and Saykanomys from Pakistan
(Hussain et al., 1978: Pis. 1-6) are dis-
tinctly more advanced in their patterns
than those of Mysops (Wood, 1973: Figs.
3, 4), as seems also to be true of Petro-
kozlovia (Shevyreva, 1976: Fig. 6). As in-
dicated above (p. 504), the incisor enamel
of the Chapattimyidae is pauciserial,
which would allow them to occupy any
position one wished.

For many years. Wood has argued
(1937: 199; 1955: 171; 1970a: 17; 1974a:
50-53), in accord with the suggestion by
Burke (1935: 3^), that the cylindrodonts
were ancestral, through the Asiatic Ar-

dynomys, to the Mongolian genera Tsa-
ganomys, Cyclomylus and their relatives.
This view was accepted by Simpson
(1945: 213), Wilson (1949: 93) and
Schaub (1958: 1537). These genera were
originally referred to the Bathyergidae
(xMatthew and Granger, 1923: 4; Teilhard
de Chardin, 1926: 29-30), an assignment
with which Landry (1957: 72-73) con-
curred. We here refer them to a distinct
family, the Tsaganomyidae, related to
both the Cylindrodontidae and the
Bathyergidae.

There are many features in common
between the cylindrodonts and the tsa-
ganomyids, including both tooth and
skull characteristics (Wood, 1974a: 51-
52). The teeth of both are clearly four-
crested in both upper and lower jaws,
with no suggestions of any possible fifth
crest or of a mesocone or mesoconid.
The skull of the American species of Ar-
dynomys (it is unknown in the Asiatic
species) shows, among other features, a
backward movement of the growing base
of the upper incisor, so that it protrudes
into the orbit in exactly the same area
(although not not quite so far) and in the
same manner as in the Tsaganomyidae.
The arrangement in Ardynomys could
probably also have given rise to the con-
dition in the bathyergids, by an increase
in the radius of curvature of the upper
incisor, that would have pushed the
growing base of the incisor upward and
backward, so that it would have been en-
abled to grow backward. In the Tsagan-
omyidae, however, the upper incisor has
extended its growing base back in the
direction that had already been estab-
lished in Ardynomys, as a result of which
the large ridge, marking the course of the
incisor across the anterior part of the or-
bit, fomis a pronounced knob, lateral to
M*~^. There is much individual variation
in the extent of this knob; sometimes
there is an appreciable space below it
above the alveolar border; in other spec-
imens, the posteroventral end of the knob
lies ventral to the occlusal surface of the
upper cheek teeth. In these latter cases,

516 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

the ridge impinged on the course of the
infraorbital artery, vein and nerve to the
infraorbital foramen; we are not sure, but
presume that they passed lateral to the
incisive ridge.

We are infonned by Dr. J. H. Wahlert
(in lit., 15 March 1980) "that the enamel
of both Cyclomylus and Tsaganomys is
intermediate between pauciserial and
multiserial. Cyclomylus is closer to pau-
ciserial and Tsaganomys to multiserial."
Later, he added (in lit., 28 March 1980)
"The incisor enamel of My sops ... is
pauciserial with a slight inclination and
is very much like that of Cyclomylus.
Relationship is allowed but certainly not
demonstrated by this similarity."

We define the Tsaganomyidae as fol-
lows: protrogomorphous and hystrico-
gnathous rodents of large size; skull mas-
sive; occiput forwardly sloping, very
wide; bulla flattened anteriorly; zygo-
matic arch extending posteriorly well be-
yond posterior root of zygoma; incisors
procumbent, presenting nearly straight
forward as in specialized bathyergids;
upper incisors extending back into orbit,
bases forming bulges on side of maxillae,
sometimes ending lateral and ventral to
occlusal surface of M^"^; lower incisors
extending back to beneath condyles; in-
cisor enamel transitional from uniserial
to multiserial; cheek teeth hypselodont,
four-crested; deciduous teeth replaced in
nomial manner, dm^ small, peg-like, dnV
essentially molariform; numerous bur-
rowing adaptations in skeleton and body
musculature.

Genera. Tsaganomys, Cyclomylus,
Pseudotsaganomys if not a synonym of
Cyclomylus (Kowalski, 1974: 159), Se-
pulkomys and Beatomus if valid.

Distribution. Middle to late Oligo-
cene. Central Asia.

(4) Bathyergidae. In the Bathyergidae,
the elongation of the upper incisors has
carried the growing base nearly straight
backward. The incisors of Bathyergoides
(Lavocat, 1973: PI. 7, Fig. 5) and Prohe-
liophobius (Lavocat, 1973: PI. 5, Fig. 6)

make ridges part way across the orbit,
directed about as in Ardynomys and
Bathycrgus. In the other bathyergid gen-
era {Cryptomys, Georychus and Helio-
phohius) the growing base lies posterior
to M^. We do not believe that it would be
possible to evolve the structure of the ad-
vanced Bathyergidae from that of Tsa-
ganomys, or vice versa, although both
could be derived from that of Ardyno-
mys.

One of us (A.E.W.) believes that the
franimorph family Cylindrodontidae was
directly ancestral to the Tsaganomyidae
and, probably, also to the Bathyergidae,
although the Oligocene cylindrodonts
are clearly too specialized to be bathyer-
goid ancestors, in that they had acquired
uniserial incisor enamel. However, ''Cy-
clomylus has inclined pauciserial enamel
that looks a lot like [that of] Mysops but
is thicker" (Wahlert, in lit., 13 February
1980).

At the time of Patterson's death, we
had not completely agreed on how we
wanted to treat the cylindrodonts and tsa-
ganomyids. At least at one time, Patter-
son felt that a relationship such as that
described above would require the trans-
ferral of the Cylindrodontidae from the
Franimorpha to the Bathyergoidea. Wood
objected, because the Bathyergoidea are
fully hystricognathous and have a ptery-
goid fossa that opens into the orbitotem-
poral fossa, whereas the cylindrodonts
are subhystricognathous with a closed
pterygoid fossa. The classification given
below (p. 572) follows Wood's opinion.

Stromer (1923: 263; 1926: 134) de-
scribed Bathyergoides neotertiarius from
the Miocene of Southwest Africa (Nami-
bia) and indicated that it was a definite
bathyergid. He implied, although he was
not certain (1926: 148), that it was in-
volved in the ancestry of the Bathyergi-
dae. Lavocat (1973: 109), in his descrip-
tion of the East African Miocene rodents,
gave the date of publication of Bathyer-
goides neotertiarius as Stromer, 1924, a
paper that he did not include in his bib-

Oligocene Rodents of Bolivia • Patterson and Wood 517

liography. Both Stromer (1926) and Ro-
mer et al. (1962: 1341) give the date of
pubhcation of the paper in question as
1923, a figure that we accept.

Lavocat (1973: 109-133) also adopted
the view that Bathyergoides was ances-
tral, at least structurallv, to the Bathver-
gidae. He (1973: 150-153) made a case
for considering that Bathyergoides was
an intermediate stage between the
thryonomyoids and the bathyergids, but
we believe that the evidence of probable
relationships of the bathyergids to the
Oligocene tsaganomyids and of the latter
to the Cylindrodontidae, rules out that
possibility. As noted above (p. 473), if
nematode evidence is of any value in
rodent taxonomy, it should be a strong
support for the absence of relationship
between the bathyergids and thr>'onomy-
oids. It is also significant in this
connection that the bathyergids and tsa-
ganomyids are all protrogomorphous,
with no forward extension of M. masseter
medialis, except in Bathyergoides and,
perhaps, in Cryptomys. Tullberg (1899:
PL 2, Fig. 17), cited' by Lavocat (1973:
151) as conclusive support for the exis-
tence of hystricomorphy in bathyergids,
shows the jaw muscles oi Cryptomys (cit-
ed as Georychus) coecutiens with M.
masseter medialis just visible in the lat-
eral view of the infraorbital foramen.
Comparison of Tullberg's lateral view of
the cleaned skull (1899: PI. 2, Fig. 14)
with the figure of the jaw muscles (1899:
PI. 2, Fig. 17) suggests that the muscle
was merely visible through the infraor-
bital foramen, and that there was no ac-
tual penetration — in other words, no ori-
gin of any of the fibers from the snout in
front of the foramen. However, in his
text, Tullberg states (1899: 79) that a dif-
ference of this fomi from Georychus ca-
pensis is ". . . dass Foramen infraorbitale
. . . bedeutend weiter ist, als bei jener
Art, und dass ein kleiner Teil des Mas-
seter medialis . . . dieses Foramen hier in
der That durchsetzt."

Lavocat's figure of the skull of Bathy-

ergoides (1973: PI. 7, Fig. 5) shows the
fossa of origin of M. masseter medialis
lying in front of, but largely above, the
infraorbital foramen. The direction of the
fibers of the muscle must have been near-
ly vertical throughout, rather different
from the condition in the other hystrico-
gnaths. We interpret these conditions to
mean that there has been significant se-
lective pressure acting on the Bathyer-
gidae to lengthen M. masseter medialis,
but that true hystricomorphy has never
been generally attained in this family.

Lavocat has demonstrated (1973: PI.
23, Fig. 2; PL 26, Figs. 1, 6; PL 29, Figs.
1-9; PL 30, Figs. 1-6) that the cheek
teeth of Bathyergoides (and probably of
other bathyergids) are based on four
transverse crests, with no trace of meso-
loph or mesolophid (Fig. 30K), as is also
true in the Tsaganomyidae. We have
been unable to detennine, either from
Lavocat's description or from his pic-
tures, whether there was nonnal replace-
ment of the deciduous teeth or whether
they were retained. It seems clear that
the anterior tooth of the series generally
had the same pattern as the subsequent
teeth (Lavocat, 1973: PL 26, Fig. 1; PL
29, Figs. 1, 7; PL 30, Figs. 1, 2, 5), which
suggests that they were P4^ rather than
dm4^. There clearly was rapid wear of the
cheek teeth, so that one tooth was con-
siderably worn before the one behind it
came into use. This, of course, is normal
among burrowing rodents.

Lavocat has suggested (1973: 109, 122,
123) that there were four upper and five
lower cheek teeth in Bathyergoides. This
latter number was based on the jaw of no.
SO 736, which has three cheek teeth in
use, a fourth that had not as yet erupted,
and what Lavocat identified as an alveo-
lus at the front of the series (1973: PL 30,
Fig. 4). We are somewhat hesitant in ac-
cepting four upper and five lower cheek
teeth as being the correct formula for
Bathyergoides, as it would be unique
among rodents.

518 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

The whole matter of what teeth are
present in hathyergids is still unsettled,
but it is extraordinarily unlikely that P3
is one oi them (Lavocat, 1973: 123), since
this tooth is otherwise unknown among
the Rodentia. It seems probable that the
increase of the cheek-tooth fonnula of
bathyergids (up to e**) involves either the
retention of dm/ along with P/, or the
duplication of one or more teeth as in
Otocijon (Wood and Wood, 1933: 43), or
both. Certainly, if Bathyergoides pos-
sesses P3, it could not have been de-
scended from any known or reasonably
anticipated Eocene or Oligocene rodent.

It seems improbable to us that the
Bathyergidae secondarily became protro-
gomorphous, as Lavocat (1973: 151) pro-
posed. We know of no cases where such
a secondary development has been dem-
onstrated. If, as seems clear, the forward
growth of the origins of the branches of
the masseter was an adaptation that in-
creased the functional strength of the
muscles that moved the incisors, it would
be most amazing to find a secondary re-
duction of the jaw muscles in animals
that had highly modified the incisors to
pemiit their extensive use in burrowing.
Such a reduction would have been much
more remarkable than the retention of
the primitive protrogomorphous condi-
tion by rodents that might have been
helped had they evolved modified jaw
muscles. The retention of protrogomor-
phy would simply mean that (as in the
case oi Aplodontia) the pertinent muta-
tions had not occurred under the proper
conditions. Reversion to protrogomorphy
would require not only the reverse mu-
tations, but also selection for less effi-
cient jaw muscles.

All the features discussed above seem
to us to demonstrate that the Bathyergi-
dae could not have been descended from
the Thryonomyoidea as Lavocat be-
lieved, and as Wood was once willing to
accept (1975b: 942). We agree with the
conclusions of Hill et at. (1957: 511) that
the available data "point . . . emphatical-

ly towards a complete severance of the
bathyergids from all" other living ro-
dents.

Bathyergoides apparently was a prim-
itive member of the Oligocene stock from
which the Bathyergidae were derived,
but was on a side line as shown by the
fact that it had begun to develop hystri-
comorphy. Although we are not certain of
the precise relationships of Bathyer-
goides, we see no valid reason for sepa-
rating it from the Bathyergidae. The def-
inition of the Family Bathyergoididae, as
given by Lavocat (1973: 109) — "Famille
de Bathyergoidea dans laquelle la struc-
ture des dents jugales est bien conser-
vee." — is perhaps a valid generic diag-
nosis, but hardly seems worthy of familial
separation.

We therefore define the Bathyergidae
as follows: protrogomorphous to occa-
sionally slightly hystricomorphous hys-
tricognaths; skull and skeleton highly
modified for burrowing; occiput broad,
inclined forward; incisors hypertrophied
and procumbent, with growing bases ex-
tended unusually far to the rear; cheek-
tooth fonnula variable, with additional
teeth of unknown nature; cheek teeth
hypsodont to hypselodont, with crown
patterns based on four transverse crests
in both upper and lower teeth, with no
trace of mesoloph or mesolophid; incisor
enamel multiserial.

As we have tried to indicate, there are
many similarities between the Tsagano-
myidae and the Bathyergidae, as well as
differences. It seems probable, however,
that the two are related, both being de-
scended from late Eocene or early Oli-
gocene Asian cylindrodonts (but not
from chapattimyids). We therefore unite
the two families in a Superfamily Bathy-
ergoidea. We have no feelings as to
whether one or two lines crossed the
boundary from the cylindrodonts to the
Bathyergoidea.

(5) Hystricidae. Lavocat (1973: 160) in-
cluded the Hystricidae, together with the
Bathyergoidea and the Thryonomyoidea,

Oligocene Rodents of Bolivia • Patterson and Wood 519

in his Iniraorder Phiomorpha, and be-
lieved the common ancestry of all to have
been primitive thryonomyoids. He stated
(1969: 1496) that some East African forms
"correspondaient parfaitement a ce que
Ton pouvait attendre" of hystricid ances-
tors, but neither at this time nor later
(1973: 153-158) was he more specific nor
did he demonstrate how the hystricids
could possibly be descended from any-
thing that had begun to acquire the fea-
tures that characterize the Thryonomy-
oidea. Lavocat noted (1978: 74), for the
first time so far as we are aware, that, in
the Hystricidae, there is "replacement of
dP by P*," without apparently realizing
that this makes all known thryonomyoids
too advanced to be hystricid ancestors.
Incidentally, it should be emphasized
that, in the hystricids (in contrast to all
known thryonomyoids) there is also uni-
versal replacement of dnij (dP4) by P4 at
the ontogenetic stage nomial for mam-
mals.

As pointed out elsewhere (Wood and
Patterson, 1970: 632-633; Wood, 1974b:
38) derivation of hystricids from Oligo-
cene or later thryonomyoids is impossi-
ble, because all hystricids have a nonnal
mammalian replacement of the decidu-
ous teeth. The reverse direction, deri-
vation of thryonomyoids from Eocene
hystricids, would be possible, although
we know of no evidence in support of
such a derivation. Since no hystricids are
known before the late Miocene, one can
use one's imagination rather freely in de-
temiining the characteristics of Eocene
hystricids. Lavocat (1978: 79) stated that
"There does not exist in the lower Mio-
cene of Africa any known fomi which
could be said with certainty to be ances-
tral to the Hystricidae . . . ." In fact, one
can go farther, and state with complete
confidence that all known Oligocene and
early Miocene African rodents can be
completely ruled out as possible ances-
tors of the Hystricidae.

Where the hystricids lived in the mid-
Tertiary and from what they were de-

rived are pure guesses. Our guess would
be that they lived in southern Asia, and
that they were derived from the same hy-
pothetical franimorph population that we
believe to have been ancestral to the
thryonomyoids. It is clear to us that there
are no known rodents that could be con-
sidered to have been immediately ances-
tral to the Hystricidae.

If Lavocat was correct that the Hystrici-
dae and the Thryonomyoidea should be
united at the infraordinal level, the term
"Hystricomorpha" ought to be used for
the combination, since it has both long
priority and very general acceptance.
Lavocat's concept, "Phiomorpha," is es-
sentially identical to the concept "Hys-
tricomorpha" of those authors who are
willing to accept the Caviomorpha as a
distinct entity. The general acceptance of
the temi "Hystricomorpha," even when
it is used in no more modern a sense than
that of Brandt (1855) was shown by its
usage, by all who were not taxonomists,
at the symposium on the Hystricomor-
pha, held in London in 1973 (Rowlands
and Weir, 1974).

We here consider the Infraorder Hys-
tricomorpha to include, for the time
being, only the Family Hystricidae. We
believe that future discoveries will not
demonstrate that the Bathyergoidea be-
long here. We have no opinion as to
whether it will be shown that the hystric-
ids and the thryonomyoids had a common
ancestor that had evolved beyond the
franimorph stage. If they did, they should
be united in the Hystricomorpha; if not,
the temi Phiomorpha may be retained for
the Thryonomyoidea.

(6) Thryonomyoidea. As indicated, we
here restrict the Infraorder Phiomorpha
to the Thryonomyoidea. Throughout this
paper, we have used the latter term to
avoid confusion, since our usage of
"Phiomorpha" is so different from that of
Lavocat. It is perhaps not necessary to
retain three infraorders (Bathyergomor-
pha, Hystricomorpha and Phiomorpha)
for the Old World Hystricognathi, but

520 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

this at least would emphasize our current Paraphiomys pifj^otti and P. stromeri

complete ignorance of the interrelation- (Lavocat, 1973: PI. 26, Figs. 7-10; PI. 27,

ships of these groups. It may even have Figs. 3-4, 9-10; PI. 28, Figs. 6-7) the

been the case that the Eocene Irani- teeth are very brachyodont and there are

morph ancestors of the caviomorphs and always elements of the metalophulid II

thryonomyoids were more closely allied present.

to each other than either was to the In his table of tooth measurements
ancestors of the Hystricidae and Bathyer- (1973: 272, Table 14), Lavocat gives no
goidea. measurements of lower teeth of Para-
It is perhaps presumptuous on our part phiomijs pigotti except one of 15.0 for
to differ with Lavocat on the classifica- what we believe is the length of the tooth
tion of the Thryonomyoidea, but there row. He gives lengths of individual teeth
are a number of points on which we have of three specimens of Paraphiomys
come to differ with his conclusions, and stromeri (including the type); these sug-
we believe that we should present these gest that, in the latter species, the teeth
and explain our reasoning. are about 2 mm long. The lengths of the
We do not believe that Thryonomys lower cheek teeth in Stromer's three
and Petromus should be united in a sin- specimens of Neosciuromys africanus
gle family, since we believe (as appar- range from 3.40 to 4.48 mm, so that they
ently Lavocat did likewise) that they be- would seem to be appreciably larger than
long to lines that were already separate those o^ Paraphiomys pigotti. Finally, in
in the lowest levels of the Jebel el Qat- Neosciuromys, M2 is significantly larger
rani Formation. Wood (1968: 68-73) com- than any of the other cheek teeth (Fig.
pared Gaudeamus with Thryonomys, 30G). Neosciuromys clearly was a mem-
and pointed out many similarities in ber of the Thryonomyidae.
tooth and jaw structure (cf. Wood, 1968: The specimen from the Jebel el Qat-
Figs. 14-15 and Wood, 1962b: 317-320, rani Fomiation of Egypt, described by
Fig. 2A-C); Wood's phylogenetic chart Wood (1968: 48-51, Fig. 5F) as Para-
(1968: Fig. 17) showed Gaudeamus as phiomys simonsi, agrees with Neosciuro-
approximately ancestral to Thryonomys. mys africanus in the simplification of its
Lavocat (1973: 162) considered Gaudea- pattern (although it has not gone quite so
mus as much too specialized to occupy far), in its appreciable height of crown,
such a position, although he kept the two and in the large size of M2 (and M3).
together in the Thryonomyinae. We see There seems no basis, now that illustra-
no evidence of such precocious special- tions of all the Miocene forms are avail-
ization, and believe that Gaudeamus was able (which they were not when Wood's
very close to, if not actually, the ancestor 1968 paper was written) for not including
of Thryonomys. We believe that these this species in Neosciuromys as N. si-
genera, belong to a group that early ac- monsi, a species somewhat smaller and
quired highly crested cheek teeth, with more primitive than the early Miocene
the crests reduced essentially to three in N. africanus.

all teeth except dm4 (Wood, 1962b: Fig. Lavocat (1973: 21) defined the subfam-

2A-B). ily Thryonomyinae ( = our Thryonomyi-

We do not believe that Neosciuromys dae) simply as "Thryonomyidae dont la

africanus Stromer is a synonym of Para- structure des dents jugales n'est pas tres

phiomys pigotti, since Neosciuromys simplifiee. " On the contrary, we believe,

clearly had higher crowned teeth than these fonns clearly have very simplified

Paraphiomys and had reduced the pat- tooth patterns, at least in comparison to

tern of the lower cheek teeth to three the other Oligocene and Miocene thryon-

transverse crests with no trace of subsid- omyoid genera,

iary elements (Fig. 30G), whereas in Petromus (Wood, 1962b: Fig. IC-F)

Oligocene Rodents of Bolivia • Patterson and Wood 521

has a ver>' different tooth pattern from nae, but that it was inadvertently omitted

that of the Thryonomyidae, it being from his table on page 160. We believe

based on four transverse crests rather that Andrewsimys also belongs here, on

than three. This type of pattern is found the basis of Lavocat's definition of the

in a considerable number of the Oligo- family, rather than in the Phiomyinae

cene and Miocene African genera. The where he placed it. For the present, but

oldest family name available for these is with great uncertainty, we accept the

Petromuridae of Tullberg, 1899 (as Petro- union of Lavocat's Subfamilies Phiocri-

myidae). We divide these, in general fol- cetomyinae and Myophiomyinae in the

lowing Lavocat's opinions of the relation- Family Myophiomyidae.

ships of the various fonns, but not his Lavocat (1973: 163) reported the dis-

fonnal taxonomy (1973: 160), into three covery, in the Eocene of Pakistan, of an

subfamilies: the Phiomyinae to include undescribed genus, which he suspected

Phiomijs, Paraphiomys, Kenyamys, Si- "puisse avoir deja presente toutes les

moninujs and Epiphiomys; the Diaman- structures essentielles des Phiomorpha

tomyinae (we can see no justification for . . . ." This citation was quoted by Wood

considering these a separate family) for (1975a: 75-76, 78) as an occurrence of the

Metaphiomys (Fig. 30E), Diamantomys Phiomorpha in the Eocene of Pakistan.

(Fig. 30J) and Pomonomys (Fig. 30H-I); Although we are not sure, we believe that

and the Petromurinae for Petromus. La- this is probably a specimen of the mate-

vocat placed Metaphiomys in a separate rial described by Hussain et ah (1978) as

subfamily from the Miocene genera that members of the Chapattimyidae, a group

he recognized (1973: 52, 160) as its prob- that we do not believe had anything to

able descendants; this seems unwarrant- do with the ancestry either of the Thryon-

ed to us, and we agree with his statement omyoidea or of the Hystricidae. As those

in the descriptive part of his paper (1973: authors noted (Hussain et ah, 1978: 75),

71) that he felt he should adopt a vertical the only rodents they refer to this family

classification and consider Metaphiomys whose lower jaws are known were sciuro-

the direct ancestor of Diflmanfomjys' and gnathous. These seem most unlikely

Pomonomys. ancestors for the thryonomyoids. Lavo-

Lavocat's Family Myophiomyidae cat has infomied Wood {in lit., 25 Jan.

unites all the highly cuspidate genera, 1979) that he no longer believed "que les

but it may not be a natural group. Phio- Rongeurs de I'Eocene du Pakistan soi-

cricetomys differs widely from the others ent apparentes aux Phiomorphes. . . . ces

in the very large size of the cingulum and Rongeurs ont tons un tres fort metaco-

the length of the anterior tooth, in the nule, ce qui ne s'accorde pas du tout avec

roundness of the individual cusps, and in la structure de Phiomys andrewsi."

the reduction of the dentition in the only Recently, Shevyreva and Chkhikvadze

known specimen to three cheek teeth (1978: 418) have suggested, in a brief ab-

(Wood, 1968: Fig. 16). Wood (1968: 74- stract, that a fomi represented by a lower

75) concluded that the teeth were more jaw ("Sp. 3") from the middle Eocene of

probably Mi_3 than dm4-M.2; however Kazakhstan, might have something to do

(p. 496), perhaps they are dm4-M2. The ap- with the ancestry of the "Phiomorpha."

parent reduction of M3 in Elmerimys We doubt whether any fossils yet report-

(Lavocat, 1973: 105; PI. 28, Fig. 11) could ed have any bearing on the immediate

be a clue pointing in the same direction, if ancestry of the thryonomyoids.

Lavocat's two subfamilies (Phiocricetomy- Lavocat (1973: 20) cited the Infraorder

inae and Myophiomyinae) really belong Phiomorpha as dating from Lavocat,

together. It is clear from Lavocat's text 1967, but listed, in his Bibliography, two

(1973: 103, 108) that he intended to in- papers of that date, Lavocat 1967A and

elude Phiomyoides in the Myophiomyi- 1967B. In the fonner paper, which dis-

522 Bulletin Museum of Coinpurative Zoology, Vol. 149, No. 7

cussed the Miocene rodents of East Af-
rica and their relationships, he neither
used nor defined the temi "Phiomor-
pha. " The expression "Superfamille des
Phioniyoides" was used on page 60, and
the tenn "Phiomyoidae" [sic] in the ab-
stract (1967a: 57). The second paper, dis-
cussing the ear region of the therido-
myoids, used the terms "Phiomorphes"
and "Pliiomorphs" in the French and En-
glisli abstracts, respectively (1967b: 491),
and the text makes comparisons of the
theridomyoid ear with that of "Phio-
morphes," not otherwise identified, the
term being used three times (1967b: 498).

The only attempts by Lavocat at a def-
inition of the Infraorder that we have
found were in his East African mono-
graph (1973: 20), and in his chapter
(1978: 71) on the Rodentia and Lagomor-
pha in "Evolution of African Mammals."
In the first of these papers, his definition
seems hardly adequate: ''Definition:
Rongeurs d' introduction ancienne en Af-
rique, de provenance actuellement in-
connue, de distribution actuelle tropicale
a tempere chaud, dans I'Ancien Monde
exclusivement. Structure infraorbitaire
primitivement hystricomorphe, tres mo-
difiee et regresse dans la famille des
Bathyergidae. 4 ou 5 dents jugales, ou
plus chez certains Bathyergoidea." His
later definition was more illuminating:
"Infraorder Phiomorpha Lavocat 1962.
DIAGNOSIS. Lower jaw always hystri-
cognath. Infraorbital structure primarily
hystricomorph, greatly modified and
secondarily reduced in some Bathyergoi-
dea. Molar teeth morphologically tetra- or
pentalophodont; greatly simplified in the
Bathyergidae. Generally four cheek teeth,
rarely three, sometimes five" (1978: 71).
Even this definition does not separate the
"Phiomorpha" from the Caviomorpha.

Lavocat concluded his monumental
work on the early Miocene rodents of
East Africa with the remark that above all
we now need rodent faunas from the
Eocene of South America and Africa
(1973: 258). We heartily agree, and add

Middle America to the list, along with
more extensive rodent faunas from the
Eocene of both north and south Asia.

Classification of the Suborder
Hystricognathi

We now summarize, in tabular form,
our current ideas as to the classification
of the Suborder Hystricognathi.

Suborder Hystricognathi Tullberg 1899
Infraorder Franimorpha Wood 1975a
Family Reithroparamyidae (new
rank; Wood, 1962a as a
subfamily)
Family Protoptychidae Wood

1937
Family Cylindrodontidae Miller
and Gidley 1918 (excluding
the Subfamily Tsaganomyi-
nae; otherwise as given by
Wood, 1974a)
Incertae sedis as to family —
Guanajuatomys and Pro-
lapsus
Infraorder Bathvergomorpha Roberts
1951
Superfamily Bathyergoidea Osborn
1910

Family Bathyergidae Water-
house 1841 (Bathyergidae of
authors, plus Bathyer-
goides, Proheliophobius and
Paracryptomys, if the last is
not a nomen vanum)
Family Tsaganomyidae Mat-
thew and Granger 1923 (as
a subfamily)
Infraorder Hystricomorpha Brandt 1855
(as a suborder; Tullberg, 1899 as
an Infraorder)

Family Hystricidae Burnett 1830
Infraorder Phiomorpha Lavocat (first
fomial use as an infraorder by La-
vocat, 1973: 20; first undefined use
of the term "Phiomorphes" by
Lavocat, 1967b: 491, 498; Lavocat,
1978: 71 dates the infraorder from
Lavocat, 1962, in which paper nei-
ther "Phiomorpha" nor "phio-
morphes" is used)

Oligocene Rodents of Bolivia • Patterson and Wood 523

Superfamily Thryonomyoidea Wood
1955

Family Thryonomyidae Pocock
1922

Genera: Gaudeamus, Neo-
sciuromijs, Thryonomys
Family Petromuridae Tullberg
1899 (as Petromyidae)
Subfamily Petromurinae El-
lemian 1940
Genus: Petronius
Subfamily Phiomyinae Schlos-
ser 1911 (including Fam-
ily Kenyamyidae Lavocat
1973)
Genera: Phiomys, Para-
phiomys, Kenyamys, Si-
monimys, Epiphiomys
Subfamily Diamantomyinae
Schaub 1958 (as a family;
including Subfamily Meta-
phiomyinae Lavocat 1973)
Genera: Metaphiomys, Dia-
mantomys, Pomonomys
Family Myophiomyidae Lavocat
1973
Subfamily Myophiomyinae
Lavocat 1973
Genera: Myophiomys, El-
merimys, Andrewsi-
mys, Phiomyoides
Subfamily Phiocricetomyinae
Lavocat 1973
Genus: Phiocricetomys
Infraorder Caviomorpha Wood and
Patterson 1955 (in Wood, 1955, as
a suborder)
Superfamily Erethizontoidea Simp-
son 1945

Family Erethizontidae Thomas
1897 (Erethizontidae of au-
thors less Chaetomys)
Superfamily Octodontoidea Simpson
1945

Family Octodontidae Water-
house 1839 (including as
a subfamily, Ctenomyidae
Tate 1935)
Family Echimyidae Gray 1825
(as Echimyna)

Subfamily Echimyinae Mur-
ray 1866
Subfamily Adelphomyinae Pat-
terson and Pascual 1968
Subfamily Dactylomvinae Tate

1935
Subfamily Chaetomyinae Thom-
as 1897
Subfamily Heteropsomyinae

Anthony 1917
Subfamily Plagiodontinae El-
lerman 1940 (as Plagio-
dontiinae)
Subfamily Capromyinae Smith

1842 (as a family)
Subfamily Myocastorinae Ame-
ghino 1904 (as a family)
Family Abrocomidae Miller and
Gidley 1918
Superfamily Cavioidea Kraglievich
1930b
Series A, new

Family Eocardiidae Ameghino

1891
Family Caviidae Waterhouse

1839
Family Hydrochoeridae Gill
1872
Series B, new

Family Dasyproctidae Smith

1842
Family Dinomyidae Alston 1876
Family Cuniculidae Miller and
Gidley 1918
Superfamily Chinchilloidea Krag-
lievich 1940
Family Chinchillidae Bennett

1833
Family Neoepiblemidae Krag-
lievich 1926
Incertae sedis as to superfamily

Family Heptaxodontidae Miller
and Gidley 1918

CONCLUSIONS (by A.E.W.)

The skulls, lower jaws and dentitions
of rodents from the Deseadan Oligocene
of Bolivia are described. Migraveramus

524 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

is referred to the Octodontidae, Sallaniys
to the Echimyidae, Incami/s to the Dasy-
proctidae and Branisamys to the Dino-
myidae. Two undeseribed dasyproctid
genera are briefly discussed, as well as
material referred b\ Lavocat (1976: 59-
62) to the Patagonian genus Cephalomys.
It is concluded that the Bolivian genus
Lurihayomys and the Umguayan genus
Palmiramys are nomina vami.

The largest part of the paper is the dis-
cussion section. The two contemporary
theories of the origin of the Caviomor-
pha, namely transatlantic transportation
of African waif migrants or island hop-
ping from Middle America directly or
through the Antilles, are discussed. It is
pointed out that, whichever hypothesis is
accepted, there has been independent
evolution of the Old and New World hys-
tricognath rodents for at least the last 40
m.y.

During the Eocene, the Atlantic Ocean
was narrower than at present. In the ear-
liest Eocene, there was a high latitude
North Atlantic connection between Eu-
rope and North America that enabled
paramyid rodents from the latter to reach
the fonner. Europe at this time was sep-
arated from Asia by a wide waterwa\' (at
least, there is no evidence that any mam-
mals were able to cross this waterway in
either direction). The waterway closed
during the early Oligocene (the "grand
coupure"), pemiitting the invasion of Eu-
rope by a wide variety of Asian mammals.

South Asia (Pakistan, India, southeast-
ern Asia) had, quite certainly, attained its
present relationship with northern Asia
at least by the middle Eocene, from
which time a very considerable fauna of
terrestrial mammals, almost all of north
Asiatic aspect, are known. The complete
absence of any reported autochthonous
terrestrial south Asiatic faunal elements,
other than possible moeritheres, suggests
rather strongly that these southern areas
may have been part of the Asian conti-
nent for a considerable period of time.

The exact Eocene relationships be-

tween Asia and North America are not
known. There almost certainly was a
high-latitude land connection in the late
Paleocene or early Eocene, that permit-
ted the transfer of early Eocene rodents
from one continent (we believe North
America) to the other. There was also a
connection later in the Eocene that per-
mitted the Cylindrodontidae to reach
Asia and the Lagomorpha to reach North
America.

Africa had long been isolated from the
rest of the world. The African Eocene
mammalian fauna was largely different
from that of any other part of the world,
although species questionably referred to
the Palaearctic genera Apternodon and
Brachyodiis are present (Simons, 1968:
15). By the end of the Eocene, as these
two forms indicate, Africa had estab-
lished sufficiently close relationships
with Palaearctica (presumably through
southwestern Asia) to have permitted a
few invasions. These must have included
the thryonomyoid rodents (still unknown
from the Asian Eocene or Oligocene).

South America, like Africa, was an is-
land continent during the Eocene. There
had been a situation earlier that permit-
ted some waif dispersal between North
and South America, whether in one di-
rection or both is uncertain. At some un-
known time after the mid-Eocene iMus-
ters, caviomorph rodents reached South
America. And, at some pre-Deseadan
date probably later than the arrival of the
rodents, primates reached South Ameri-
ca.

Rodents that show the initial stages of
the development of hystricognathy (the
migration of the angular process a short
distance laterad of the incisive alveolus)
appear in the late Paleocene of western
United States. These and their later but
generally similar descendants are united
in the Infraorder Franimorpha. Hystri-
cognathy increases among North Ameri-
can rodents during the Eocene, being
completely developed in the middle
Eocene Prolapsus from Texas. Hystrico-

Oligocene Rodents of Bolivia • Patterson and Wood 525

morphy also appears in a number of fran-
imorph lines. Although hystricomorphy
appears in middle Eocene theridomyoids
of Europe and ctenodactylids of Asia, all
evidence indicates that neither of these
groups ever developed, or gave rise to
fomis that developed, hystricognathy.

The Eocene fauna of Middle America
is very poorly known. But there were a
number of hystricomorphous and hystri-
cognathous lines of rodents evolving in
this area. Among these are the rodents of
the Family Cylindrodontidae, which
were hystricognathous but protrogomor-
phous. We believe that Middle American
members of this family were ideal can-
didates, both in their geographic location
and their structural features, to have been
the ancestors of the Caviomorpha.

The Eocene relationships of North and
South America are poorly known. There
would seem, however, to have been pe-
riods of orogeny, fonning chains of vol-
canic islands more or less along the
course of the present Isthmus of Panama,
in the middle Eocene, and there was un-
questionably a chain of islands connect-
ing Middle America and South America
along the course of the present Antilles.
Either of these would have provided a
much easier invasion route to South
America than the South Atlantic would
have.

The Deseadan rodent fauna was com-
pared with that of present-day northeast-
ern Brazil. The frequency make-up at the
various localities is strikingly similar in
the two cases, with (usually) one species
per locality the predominant one, but
with relative abundances varying widely
from localit\' to localit>', even between
those in close proximity. Rodents are one
of if not the most abundant order in all
Deseadan localities that have produced
extensive collections.

In contrast, the early Oligocene ro-
dents of Africa are much less diverse, tax-
onomically, than are the South American
ones. In nearly every collecting locality,
they are both individually and by taxa

much less abundant than was the case
with the Deseadan rodents. The only ex-
ception (Yale Quarry G) was a small quar-
ry, producing only small mammals; here
there was an overwhelming preponder-
ance of one species of rodent.

The African Oligocene rodents were
contrasted with their contemporaries, the
rodents of the Porvenir local fauna of the
early Oligocene Chambers Tuff Forma-
tion of Trans-Pecos Texas. Here six fam-
ilies of rodents are present, one with 3
and two with 2 subfamilies each, 13 gen-
era and 15 species, all represented by a
total of 44 specimens. Obviously, the di-
versity of North American early Oligo-
cene rodents was vastly greater than that
of their African contemporaries (only two
closeK' related families, 3 genera and 4
named species plus 2 possible additional
ones, are represented among 59 speci-
mens from the early Oligocene). This
clearly disproves the hypothesis that ro-
dents reached Africa in the Paleocene or
early Eocene.

The early Miocene rodents of East Af-
rica present a very different picture. Here
the numbers and diversit>' of the rodents
are very similar to those in the South
American Deseadan. We feel that this
very clearly indicates that, in the early
Oligocene, rodents had been in South
America about as long as, in the early
Miocene, they had been in Africa. The
East African rodents consist of two
groups — the descendants of the Oligo-
cene forms, and a group of ver\' recent
immigrants, presumably from southwest-
ern Asia (Bathyergidae, Anomaluridae,
Pedetidae, Cricetidae, Sciuridae). Al-
though the latter group is represented by
more taxa than are the thrvonomyoids,
the\' make up less than 10% of the indi-
viduals that have been collected, and
were obviously just beginning to estab-
lish a foothold on the continent.

The Oligocene rodents of Africa are
compared with the Recent murids of Aus-
tralia. These last apparently reached that
continent from the East Indies about 4.5

526 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

niA'.b.p., and have since diversified in a nematode with that host. In any case,

continent where they had no competi- Evaginuris is an exceedingly doubtful

tors. The morj^hologic diversification of support for thryonomyoid-caviomorph re-

the Australian Pseudomyinae seems lationships.

somewhat greater than that of the African Similarities occur between the fleas

Oligocene thryonomyoids (separated into and lice of Old World and New World

three families not because they are that hystricognaths (Traub, 1980). These, we

different but because their Miocene de- believe, indicate that the caviomorphs

scendants achieved greater diversity). and the Old World fonns are more closely

That is, rodents would seem to have related to each other than either is to any
reached Africa no more (and probably other living rodent, but there seems to be
less) than 4.5 m.y. before the deposition just as much support from these pests for
of the lower fossil wood zone, but to have a North or Middle American origin of the
reached South America two or three caviomorphs as for an African one.
times as much before the Deseadan. Al- A detailed analysis was made of the
lowing for the various uncertainties, this morphologic features that have been cit-
would indicate that rodents had reached ed by various authors, especially Lavocat
Africa in the latest Eocene, but that they (1973: 168) as indicative of close relation-
arrived in South America about the mid- ships between the Oligocene and later
die of the Eocene. caviomorphs and the Old World hystri-

The evidence for a close relationship cognaths, similarities supposedly so great

between the Caviomorpha and the Old as to require the rafting of thryonomyoids

World hystricognaths, based on parasitic from Africa to South America. Hystrico-

nematodes, has been milked for all it is gnathy is common to both groups, but is

worth. There apparently has been as likewise present in the North American

much parallelism among heligmosomid Eocene-Oligocene franimorphs. This fea-

nematodes as Wood thinks there has ture developed to permit lengthening of

been among rodents; the entire basis for M. ptenjgoideus internus. This length-

heligmosomid support of relationship be- ening also resulted in deepening and

tween thryonomyoids and caviomorphs eventual perforation of the dorsal margin

is, apparently, based on the genus Para- of the pterygoid fossa. Due to damage in

heligmonella, present in certain echi- all available specimens, it is not known

myids. If we interpret her meaning cor- whether or not such deepening and per-

rectly, Durette-Desset (1971: 50) was foration had occurred in the Eocene fran-

very uncertain as to the accuracy of her imorphs. Hystricognathy, certainly, could

assignment of this genus to the otherwise have been inherited by both thryono-

exclusively Old World family Heligmo- myoids and caviomorphs from Holarctic

somidae. franimorph ancestors.

The oxyurine nematode Evaginuris is Some features (structure of the middle

primarily known from caviomorphs. Two ear, presence of an internal carotid artery,

species have been reported from Hijstrix, and serology) separate the Erethizonti-

but both records deserve to be ques- dae from all other post-Eocene hystrico-

tioned. In one case, the host was a zoo gnaths. These features would seem to in-

specimen that could have been infected dicate that they are more primitive than

from contact with an American porcu- any of the Thryonomyoidea, and would

pine. The other species might also have permit them to have given rise to the Old

been in a zoo specimen, and, in addition. World fonns (which we do not believe

the author of the species seemed to ques- happened) but would seem to be a com-

tion either the accuracy of the identifi- plete barrier to their being descendants

cation of the host or the association of the of thryonomyoids.

Oligocene Rodents of Bolivia • Patterson and Wood 527

The Bathyergoidea are similarly iso-
lated from all other post-Eocene hystri-
cognaths, because of their retention of
protrogomorphy, their rather distinctive
ear ossicles with much greater fusion
than occurs in any other hystricognaths,
their smooth brains and their distinctive
nematodes.

A number of the features cited by La-
vocat and Hoffstetter are not, as they im-
plied, hystricognath characters, but are
widely distributed among rodents in gen-
eral.

We conclude that the evidence indi-
cates that the Caviomorpha and Old
World hystricognaths are related; that it
is morphologically impossible for the
Thryonomyoidea to have given rise to the
Caviomorpha or to the Hystricidae, but
that evolution in either of the reverse di-
rections would have been possible; and
that the most probable scenario is for
both the Thryonomyoidea and the Ca-
viomorpha to have been derived from
middle Eocene northern hemisphere
franimorphs, even though these are still
unknown in the Old World.

A classification of the Suborder Hystri-
cognathi is presented. Among the Ca-
viomorpha, several changes are made
from our 1959 classification. These in-
volve: transferral of the Chaetomyinae
and Capromyinae to the Echimyidae, and
of the Ctenomyinae to the Octodonti-
dae; transferral of the Dasyproctidae,
Dinomyidae and Cuniculidae from the
Chinchilloidea to the Cavioidea; divi-
sion of the Cavioidea into two Series,
one for these three families and the other
for the Eocardiidae, Caviidae and Hydro-
choeridae; transferral of the Heptaxodon-
tidae out of all recognized superfamilies
to become incertae sedis; and recognition
of the Neoepiblemidae as a second family
of chinchilloids.

Among the Franimorpha, the Reithro-
paramyidae are raised to family status,
and the Cylindrodontidae are included
because they have recently been discov-
ered to have been hystricognathous. The

northern Asiatic Tsaganomyidae are con-
sidered to have been descended from the
cylindrodonts but to be bathyergoids
rather than franimorphs. It is concluded
that there are no relationships whatso-
ever between the rodents from Pakistan
included in the Chapattimyidae by Hus-
sain et al. (1978) and either the Thryon-
omyoidea or the Cylindrodontidae.

We doubt that the modern Bathyergi-
dae are secondarily protrogomorphous,
as Lavocat (1973: 151) has proposed, and
consider that Bathij ergo ides was special-
ized, not primitive, in its possession of
partial hystricomorphy. The totality of
the evidence seems to us to show that the
Bathyergoidea were of northern Asiatic
origin, evolving from Oligocene cylin-
drodonts, and that they have no relation-
ship (except descent from primitive fran-
imorphs) with other hystricognaths.

There continues to be no information
as to the ancestry of the Hystricidae. We
have no opinion as to whether they were
ancestral to the Thryonomyoidea or
whether the two groups evolved inde-
pendently from the Franimorpha. We
feel certain that the hystricids are not de-
scended from thryonomyoids.

If the Thryonomyoidea and Hystrici-
dae are related at the infraordinal level,
they should be united in an Infraorder
Hystricomorpha rather than Phiomorpha,
since the fomier tenu has long priority
and essentially universal recognition.

We recognize three families of the
Thryonomyoidea — the Thryonomyidae,
Petromuridae and Myophiomyidae. In
the first family we place Thrijonomijs,
Gaudeamus and Neosciuromys, the last
being quite distinct from Paraphiomys
with which Lavocat (1973: 22) synony-
mized it. This genus also includes the
Oligocene species N. simonsi, described
by Wood (1968) as Paraphiomys simonsi.
We divide the Petromuridae into three
subfamilies: the Petromurinae for Petro-
niiis; the Phiomyinae for Phiomys, Para-
phiomys, Kenyamys, Simonimys and
Epiphiomys; and the Diamantomyinae

528 Bulletin Museiiin of Coniixirdtivc Zof>/o^;/, Vol. 74.9, No. 7

tor Met(ij)Jiiomys, Diamantomijs and Po-
nionoiuys. The Myophioiiiyidae include
two subfamilies: the Myophioniyinae for
Myophioniys, Elmerimys, Audrewsimys,
and Pfiioniyoides; and the Phiocriceto-
niyinae for Phiocricetomys. We feel no
conviction that these last two subfamilies
deserve to be united at the familial level.
We assume that Lavocat's citation
(1973: 163) of an undescribed Eocene ge-
nus trom Pakistan that might be a thryon-
omyoid was a reference to one of the
fomis described by Hussain et al. (1978),
in which case it is a ctenodactyloid and
has no bearing on the problem of hystri-
cognath relationships.

APPENDIX 1: LIST OF REFERRED
SPECIMENS OF SALLAMYS PASCUALI,
INCAMYS BOLIVIANUS AND
BRANISAMYS LURIBAYENSIS

Sallamys pascuali. Eleven specimens in the
Princeton collection: PU nos. 20906, a fragment of
the right maxilla with P-M'; 20907, a portion of the
right ramus with part of the incisor and P4-M2;
20908, a portion of the right ramus with part of the
incisor, roots of P4 and much worn Mi_2; 20909, a
fragment of the right ramus with part of the incisor
and dm4-M2; 20910, a fragment of the right ramus
with M, and the roots of P4; 20911, a fragment of
the left ramus with M2 much worn and roots of P4-
M,; 20912, a portion of the left ramus with part of
the incisor and M,_2; 20913, a portion of the left
ramus with a broken incisor and the roots of P4-M3;
20982, an upper left molar, probably LM'; 21727,
a portion of the right ramus with M2_3 and the roots
of P4-iVI,; and 21950, part of the right ramus with
M,_2. PU nos. 20908 and 20911 could be parts of
one individual.

Incainys bolivianus. One hundred forty specimens
in the MNHN, Princeton and MACN collections, as
follows: SAL 119, Ldm^-M'; SAL 120, (R or L?) P-
M^; SAL 121, (R or L?) P"; SAL 122, (R or L?) P^
SAL 123, (R or L?) P; SAL 124, Ldm4-M2; SAL
12.5, Rdm4-M2; SAL 126, RP4-M,; SAL 127, LP4;
SAL 128, (R or L?) dm4-M3; "33 mandibules plus
ou moins completes SAL 129 a SAL 161" (Lavocat,
1976: 50); SAL 162, partial skull, and SAL 162A and
162B, associated rami of the lower jaw, type of /.
pretiosus. The Princeton material includes PU no.
21726, skull and incomplete mandible found in
connection, skull somewhat distorted but nearly
complete, damaged in the orbital region and lacking

the tips of the premaxillaries and incisors, the arch-
es and antorbital bars, and part of the right bulla,
mandible incomplete posteriorly on both sides and
lacking tips of incisors and symphysis, but includ-
ing all deciduous and pennanent molars; PU no.
21945, skull with articulated mandible, missing the
tip of the snout and left zygomatic arch, the ventral
part of the mandible and the angles, but otherwise
nearly complete and uncrushed; PU no. 20944, the
anterior part of a skull, damaged and lacking all the
cheek teeth; and 91 other specimens — PU no.
20953, incomplete premaxilla with broken I'; PU
no. 21728, maxilla with LM'"^ and partiallv erupted
LP^ PU nos. 20929, 20938-9, 20947, 20950-2,
20954-68, 20979-80, 21949, 21956, 21958-9, 21970-
1, 21982 and 21985, fragmentar\- maxillae and iso-
lated upper cheek teeth; PU nos'. 20916-28, 20930-
7, 20940-,3, 20945-6, 20948-9, 20969-78, 20981,
21729-31, 21735, 21941, 21946-7, 219.52-4, 21957,
21981, 21983 and 21987, fragmentary lower jaws
and isolated lower cheek teeth. Of these, PU nos.
20965 and 20979, R and L P, mav represent one
individual; PU nos. 20954 and 20960, R and L P"-
M^ another; and PU nos. 20958 and 20963, R and
L P, a third. As indicated above (p. 429), MACN
no. A 52-113, an isolated upper cheek tooth, iden-
tified by Wood and Patterson (1959: 362-4) as
"Pdasyproctid gen. et sp. indet.," belongs here.

Branisamijs lurihaijensis. Thirty-one specimens in
the MNHN and Princeton collections, as follows:
a partial mandible with RP4-M, and LM,_3 (SAL
103) was referred here by Hoffstetter and Lavocat
(1970). Lavocat (1976: 33) referred SAL 104, (R or
L?) M'"^; SAL 105, fragment of the ramus with
RP4-M3; SAL 106, erupting P4 and M, (R or LP);
SAL 107, LMi and posterior part of Ldm4; SAL 108-

115, eight partial mandibles with teeth; and SAL

116, jaw fragment (R or L?) with roots of cheek
teeth. Our material consists of PU no. 20914, incom-
plete anterior halt of a skull with broken incisors
and damaged LP-M'; PU no. 21960, a partial snout
with both incisors, showing the naso-frontal contact
and damaged ventral and lateral areas back to near
the rear of the diastema; PU nos. 21732 and 21733,
right and left maxillary fragments with little worn
P-M\ almost certainly representing the same in-
dividual; PU no. 21943, maxillar>- fragment with
worn and broken RP-M^; PU no. 21955, maxillary
fragment with Rdm'*-M' and the root of dm^; PU no.
20915, small fragment of maxilla with LP; PU no.
21944, mandible with RM,_3; PU no. 21951, frag-
mentary mandible with RP4-M2; PU no. 21980,
fragment of right mandible with P4; PU no. 21987,
fragmentarv mandible with Rdm4-M,; PU no.
21986, isolated LM2; PU no. 217,34, fragment of
mandible with RM3 and damaged incisor; and PU
nos. 21978, 21979, 22172 and 22173, fragmentary
mandibles with incisors and roots of the cheek
teeth. GB 014, Ldnij, the type of Villurroelonnjs
bolivianus, belongs here.

Oligocene Rodents of Bolivl^ • Patterson and Wood

529

APPENDIX 2 (BY B.P.): VALIDATION
OF ACARECHIMYS

The name Acarechimi/s was quoted ex
ms. by J. L. Kraglievich (1965: 258) and
credited to Patterson. The formalities,
which it becomes necessary to supply,
are as follows.

Acarechimys Patterson
{in J. L. Kraglievich, 1965)

Diagnosis. Cheek teeth somewhat
higher crowned than in Protacaremys.
Upper molars with transverse lophs; an-
terolophs and posterolophs of M^~^
closely approaching buccal extremities of
proto- and metalophs, uniting early in
wear; posteroloph of \P reduced. Dm'*
tending toward reduction and loss of
metaloph. Lower molars with variable
metalophid, interrupted or absent in
Mi_2, absent or interrupted in Mg, meta-
flexid deeper than mesoflexid in Mg. Dm4
with metalophid united to metaconid or,
rarely, falling short if it; anterior fossettid
usually divided by anteroposteriorly or
obliquely aligned neolophid. Masseteric
crest stronger and more projecting than
in Protacaremys.

Distribution. Santacruzian Miocene,
Patagonia.

Type Species. Acaremys minutus
Ameghino 1887 (Synonym: Stichomys
gracilis Ameghino 1891).

Referred Species. Acaremys minutis-
simus Ameghino 1887 (Synonyms: Sti-
chomys diminutus Ameghino 1891, Scia-
mys tenuissimus Ameghino 1894 and,
provisionally, Stichomys constans
Ameghino 1887).

The generic name commemorates the
long-standing confusion of early octodon-
tids (as "Acaremyinae" or "Acaremyi-
dae") and echimyids, which is reflected
in the synonymy above.

LITERATURE CITED

Ahlfeld, F., and L. BraniSa. 1960. Geologia de
Bolivia. La Paz: Institut. Boliv. del Petrol., pp.

1-245.

Alston, E. R. 1876. On the classification of the
order Glires. Proc. Zool. Soc. Lond., 1876: 61-
98.

Ameghln'O, F. 1887. Enumeracion sistematica de
las especies de mamiferos fosiles coleccionados
por Carlos Ameghino en los terrenos eocenos
de la Patagonia austral y depositados en el Mu-
seo de La Plata. Bol. Mus. La Plata, 1: 1-26.

. 1889. Contribucion al conocimiento de los

maniiteros fosiles de la Repiiblica argentina.
Act. Acad. Nac. Cienc. Cordoba, 6: i-xxxii + 1-
1027.

. 1891. Nuevos restos de mamiferos fosiles

descubiertos por Carlos Ameghino en el
Eoceno inferios de la Patagonia austral. Espe-
cies nuevas, adiciones v correcciones. Rev. Arg.
Hist. Nat., 1: 289-328.

. 1894. Enumeration synoptique des especes

de mammiferes fossiles des formations eocenes
de Patagonie. Bol. Acad. Nac. Cienc. Cordoba,
13: 2.59^4,5.

. 1897. Mammiferes cretaces de I'Argentine.

Deuxieme contribution a la connaissance de la
faune mammalogique des couches a Pyrothe-
rium. Bol. Inst. Geogr. Argent., 18: 406^29,
4.31-521.

. 1904. Nuevas especies de mamiferos, cre-

taceos y terciarios de la Repiiblica Argentina.

An. Soc. Cienc. Argent, 56-58: 1-142.
Anthony, H. E. 1917. New fossil rodents from

Porto Rico, with additional notes on Elasmo-

dontomys obliquus Anthony and Heteropso-

mys insulans Anthonv. Bull. Amer. Mus. Nat.

Hist, 37: 183-189.
. 1922. A new fossil rodent from Ecuador.

Amer. Mus. Novitates No. 35, 4 pp.
Archer, M., AND A. Bartholomai. 1978. Tertiary

mammals of Australia; a synoptic review. Al-

cheringa, 2: 1-19.
Arcoverde de Freitas, C. 19.57. Noticia sobre a

peste de Noreste. Rev. Bras. Malariol., Doengas

Trop., 9: 123-133.
Baird, D., M. Woodburne, and a. Lawrence.

1966. Notice oi "Pyrotherium and other mam-
mals from Bolivia." Soc. Vert. Pal. News Bull.,

No. 77, p. 18.
Baskin, J. A. 1978. Bensonomys, Calomys and the

origin of the Phyllotine group of Neotropical

Cricetines. Jour. Mammal., 59: 125-135.
Beddard, F. E. 1892. On the brain and muscular

anatomv of Aulacodus. Proc. Zool. Soc. Lond.,

1892: .520-527.
Bennett, E. T. 1833. On the family of Chinchil-

lidae, and on a new genus referrible [sic] to it.

Proc. Zool. Soc. Lond., 1833: 57-60.
Berggren, W. a., M. C. McKenna, J. Hardenbol,

AND J. D. Obradontch. 1978. Revised Paleo-

gene polaritv time scale. Jour. Geol., 86: 67-

81.
Berggren, W. A., and J. A. Van Couvering.

1974. The late Neogene. Biostratigraphy, Geo-

530 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

chronology and Paleoclimatology of the last 15
million years in marine and continental se-
quences. Developments in Paleontology and
Stratigraph\- 2. Palaeogeography, Palaeoclima-
tologv and Palaeoecology, New York, Elsevier,
16 (l'-2): xi + 216 pp.

Bishop, W. W. 1967. The Later Tertian,' in East
Africa — volcanics, sediments, and faunal inven-
tory, pp. 31-56. In W. W. Bishop and J. D.
Clark (eds.). Background to evolution in Africa.
S>stematic investigation of the African Later
Tertiary and Quaternary. Chicago, London,
Univ. Chicago Press, x + 935 pp.

Bishop, W. W., and F. Whyte. 1962. Tertiary
mammalian faunas and sedients in Karamoja
and Kavirondo, East Africa. Nature, 196: 1283-
1287.

Black, C. C, and J. J. Stephens, III. 1973. Ro-
dents from the Paleogene at Guanajuato, Mex-
ico. Texas Tech Univ., Mus., Occas. Paper No.
14, 10 pp.

BoNDESio, P., R. Pascual, and G. Vucetich.
1975. Los Neoepiblemidae (Rodentia, Cavio-
morpha); su caracterizacion y sus relaciones
filogeneticas con los Chinchillidae. Actas Pri-
mero Congr. Argent. Paleontol. Bioestrat., Tu-
cuman, Arg., 2: 431^447.

BowDlCH, T. E. 1821. An analysis of the natural
classifications of Mammalia for the use of stu-
dents and travellers. Paris, J. Smith, 115 + [31]
pp.

Brandt, J. F. 1855. Beitrage zurnahem Kenntniss
der Saugethiere Russlands. Mem. Acad. Imp.
St.-Petersbourg, (6) 9: 1-^75.

Bugge, J. 1971. The cephalic arterial system in
New and Old World hystricomorphs, and in
bathyergoids, with special reference to the sys-
tematic classification of rodents. Acta anat., 80:
516-536.

. 1974a. The cephalic arterial system in In-

sectivores. Primates, Rodents and Lagomorphs,
with special reference to the systematic clas-
sification. Acta anat., 87, Suppl. 62: 1-160.
1974b. The cephalic arteries of hystrico-

morjDh rodents. Zool. Soc. Lond., Symposium
34, pp. 61-76; discussion, pp. 76-78.

Burke, J. J. 1935. Pseudocylindrodon, a new
rodent genus from the Pipestone Springs Oli-
gocene of Montana. Ann. Carnegie Mus., 25:
1-1.

Burnett, G. T. 1830. Illustrations of the Quad-
rupeda, or quadrupeds, being the arrangement
of the four-footed beasts indicated in outline.
Quart. Jour. Sci. Lit. Art, 28: 336-,353.

Chovv, M. M., and C.-k. Li. 1965. Homogalax and
Heptodon of Shantung. Vertebr. PalAsiat., 9:
15-22 (English summary, pp. 19-22).

Chow, M. M., C.-k. Li, and Y.-p. Chang. 1973. Late
Eocene mammalian faunas of Honan and Shan-
si with notes on some vertebrate fossils col-

lected therefrom. Vertebr. PalAsiat., 11: 165-
181 (English summary, pp. 179-181).
Clay, T. 1970. The Amblycera (Phthiraptera: In-
secta). Bull. Brit. Mus. (Nat. Hist), No. 25, pp.
73-98.

COCKERELL, T. D. A., L. 1. MiLLER, AND M.
Printz. 1914. The auditory ossicles of Amer-
ican rodents. Bull. Amer. Mus. Nat. Hist., 33:
347^80.

CoNROY, G. C. 1976. Primate postcranial remains
from the Oligocene of Egypt. Contrib. Prima-
tol., 8: 1-134.

Croizat, L. 1979. Review of: Biogeographie: Fau-
na und Flora der Erde und ihre geschichtliche
Entwicklung, by P. Banarescu and N. Bo§caiu.
Syst. Zool., 28: 250-252.

Dareste, C. 1855. Note sur le cerveau des ron-
geurs et particulierement sur le cerveau du ca-
biai. Ann. Sci. natur., Zool., (4), 3: 355-365.

Dashzeveg, D., and M. C. McKenna. 1977. Tar-
sioid primate from the early Tertiary of the
Mongolian People's Republic. Acta Palaeontol.
Pol., 22: 119-137.

Dawson, M. R. 1961. The skull of Sciuravus ni-
tidus, a middle Eocene rodent. Peabody Mus.
Nat. Hist, Postilla No. 53, 13 pp.

. 1964. Late Eocene Rodents (Mammalia)

from Inner Mongolia. Amer. Mus. Novitates
No. 2191, 15 pp.

. 1977. Late Eocene rodent radiations: North

America, Europe and Asia. Geobios, Mem.
Spec, 1: 195-209.

DORAN, A. H. G. 1879. Morphology of the Mam-
malian Ossicula auditus. Trans. Linn. Soc.
Lond., Zool., (2), 1: 371-497.

Draseke, J. 1929. Gehirn und Schadel von Pede-
tes coffer (Pall.). Ein Beitrag zur makrosko-
pischen Anatomic der Nagetiere. Morph.
Jahrb., 62: 39-57.

. 1942. Schadel und Gehirn der Capromyi-

dae. Zur vergleichenden Anatomic der Nage-
tiere. Zeit. f Saugetierk., 16: 238-244.

Durette-Desset, M.-C. 1971. Essai de classifi-
cation des nematodes heligmosomes. Correla-
tions avec la paleobiogeographie des botes.
Mem. Mus. Nat. Hist. Nat, (A), 69: 1-136.

ELLERNtAN, J. R. 1940. The families and genera of
living rodents. 1. London, Brit. Mus. (Nat.
Hist), pp. 1-689.

Engesser, B. 1979. Relationships of some insec-
tivores and rodents from the Miocene of North
America and Europe. Bull. Caniegie Mus. Nat.
Hist, No. 14, 68 pp.

EVERNDEN, J. F., S. J. KrIZ, AND C. ChERRONI.
1966. Correlaciones de las formaciones terci-
arias de la cuenca altiplanica a base de edades
absolutas, detenuinadas por el metodo Potasio-
Argon. Serv. Geol. Bolivia, Hoja infomi.. No. 1.

Ferrusquia-Villafranca, I. 1978. Distribution
of Cenozoic vertebrate faunas in Middle Amer-

Oligocene Rodents of Bolivia • Patterson and Wood 531

ica and problems of migration betAveen North
and South America. In I. Femisquia-Villafran-
ca (ed.), Conexiones Terrestres entre Norte y
Sudamerica, Univ. Nal. Auton. Mexico, Inst.
Geol., Bol. 101, Chap. 13, pp. 193-321.

Fields, R. W. 1957. Hystricomorph rodents from
the late Miocene of Colombia, South America.
Univ. Cal. Publ. Geol. Sci., 32: 273^04.

Gazin, C. L. 1955. A review of the Upper Eocene
Artiodactvla of North America. Smiths. Misc.
Coll., 128 (8): i-iii + 1-96.

Gazin, C. L., and J. M. Sullivan. 1942. A new
titanothere from the Eocene of iMississippi,
with notes on the correlation between the ma-
rine Eocene of the Gulf Coastal Plain and con-
tinental Eocene of the Rockv Mountain region.
Smiths. Misc. Coll., 101 (13): 1-13.

George, W., and B. J. Weir. 1974. Hystricomorph
chromosomes. Zool. Soc. Lond., Svmposium
34, pp. 79-108.

Gill, T. 1872. Arrangement of the families of
mammals with analvtical tables. Smiths. Misc.
Coll., 11: vi + 98pp.

GiNGERICH, P. D., AND G. F. GUNNELL. 1979. Sys-
tematics and evolution of the genus Esthonyx
(Mammalia, Tillodontia) in the early Eocene of
North America. Univ. Mich., Contr. Mus. Pa-
leontol., 25: 125-153.

GoLZ, D. J., AND J. A. LiLLEGRAVEN. 1977. Sum-
mary of known occurrences of terrestrial ver-
tebrates from Eocene strata of southern Cali-
fornia. Univ. Wyo., Contr. Geol., 15: 43-65.

GORGAS, M. 1967. Vergleichend-anatomische Un-
tersuchungen am Magen-Darm-Kanal der Sci-
uromorpha, Hystricomorpha und Caviomorpha
(Rodentia). Fine Studie iiber den Einfluss von
Phylogenie, Spezialisation und funktioneller
Adaptation auf dem Saugetierdami. Zeitschr.
wiss. Zool., Abt. A, 175: 237-404.

Grasse, p. -P., AND P. L. Dekeyser. 1955. Ordre
des Rongeurs. In P. -P. Grasse (ed.), Traite de
Zoologie. Anatomic, Systematique, Biologic.
Paris, Masson et Cie., 17(2): 1321-1525.

Gray, J. E. 1825. Outline of an attempt at the dis-
position of the Mammalia into tribes and fam-
ilies with a list of the genera apparently apper-
taining to each tribe. Ann. Philos., n.s., 10:
337-344.

GuiMARAES, L. R. 1972. Contribuicao a epide-
miologia da peste endemica no Noreste do Bra-
sil e Estado da Bahia. Estudo das pulgas en-
contradas nessa regiao. Rev. Bras. Malariol.,
Doengas Trop., 24: 95-163.

Guthrie, D. A. 1963. The carotid circulation in
the Rodentia. Bull. Mus. Comp. Zool., 128:
455^81.

Haffer, J. 1970. Geologic-climatic history and
zoogeographic significance of the Uraba region
in northwestern Colombia. Caldasia, 10: 603-
636.

Hag.mann, G. 1907. Uber das Gebiss von Coelo-
genijs und Dasyprocta in seinen verschieden-
en Stadien der Abkauung. Z. Morphol. u. An-
thropol., 10: 464-480.

Hartenberger, J.-L. 1975. Nouvelles decou-
vertes de Rongeurs dans le Deseadien (Oligo-
cene inferieur) de Salla Luribay (Bolivie). C. R.
Acad. Sci. Paris, (D), 280: 427-430.

. 1980. Donnees et hypotheses sur la radia-
tion initiale des Rongeurs, pp. 285-.301. In
J. Michaux (ed.), Paheovertebrata, Mem. Jubil.
R. Lavocat, 1980, 382 pp.

Hershkovitz, p. 1972. The Recent mammals of
the Neotropical Region: a zoogeographic and
ecological review. In A. Keast, F. C. Erk, and
B. Glass (eds.), Evolution, Mammals and
Southern Continents. Albany, State Univ. N.Y.
Press, pp. 311-431.

Hill, J. E. 1935. The cranial foramina in rodents.
Jour. Mammal., 16: 121-129.

Hill, W. C. O., A. Porter, R. T. Bloom, J. Seago,
AND M. D. Southwick. 1957. Field and la-
boratory studies on the Naked Mole Rat, Het-
erocephalus glaber. Proc. Zool. Soc. Lond.,
128, Part 4: 455-514.

Hoffstetter, R. 1952. Les mammiferes pleisto-
cenes de la Republique de I'Equateur. Mem.
Soc. Geol. France, n.s., 31: 1-391.

. 1968. Un gisement de mammiferes desea-

diens (Oligocene inferieur) en Bolivie. C. R.
Acad. Sci. Paris, (D), 267: 1095-1097.

. 1971. Le peuplement mammalien de

I'Amerique du Sud. Role des continents aus-
traux comme centres d'origine, de diversifica-
tion et de dispersion pour certains groupes
mammuliens. An. Acad. Bras. Cienc, 43 (Supl.):
125—144.

. 1972. Origine et dispersion des rongeurs

hvstricognathes. C. R. Acad. Sci. Paris, (D),
274: 2867-2870.

. 1973. Origine, comprehension et signifi-
cation des taxons de rang superieur: quelques
enseignments tires de I'histoire des mammi-
feres. Ann. Paleontol. (Vertebres), 59: 137-169.

— . 1975. El origen de los Caviomorpha y el

problema de los Hystricognathi (Rodentia). Ac-
tas Primer. Congr. Argent. Paleontol. Bioestrat.,
Tucumiin, Arg., 2: 505-528.

. 1976. Rongeurs caviomorphes de

lOligocene de Bolivie. 1. Introduction au
Deseadien de Bolivie. Palaeovertebrata, 7: 1-
14.
-. 1977. Primates, filogenia e historia biogeo-

graphica. Studia Geologica (Salamanca), 13:
211-253.
Hoffstetter, R., and R. Lavocat. 1970. Decou-
verte dans le Deseadien de Bolivie de genres
pentalophodontes appuyant les atfinites afri-
caines des Rongeurs Caviomorphes. C. R.
Acad. Sci. Paris, (D), 271: 172-175.

532 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

HOFFSTETTER, R., C. MARTINEZ, M. MaTTAUER,
AND P. Thomasi. 1971. Lacayani, un nouveau
gisement bolivien de maiiiiiiiferes deseadiens
(Oligocene inferieur). C. R. Acad. Sci. Paris,
(D), 273: 2215-2218.

HussAiN, S. T., H. De Bruijn, and J. M. Leiders.
1978. Middle Eocene Rodents from the Kala
Chitta Range (Punjab, Pakistan). Proc. Kon.
Ned. Akad. Wetenschappen, Ser. B, 81: 74-
112.

Hyrtl, J. 1845. Vergleichend-anatoniische Unter-
suchungen iiber das innere Gehororgan des
Menschen und der Saugethiere. Prague, F.
Ehrlich, viii + 139 pp.

Jacobs, L. L. 1977. Rodents of the Heniphillian
Redington Local Fauna, San Pedro Valley, Ar-
izona. Jour. Paleontol., 51: 505-519.

KORVENKONTIO, V. A. 1934. Mikroskopische Un-
tersuchungen an Nagerincisiven unter Hinweis
auf die Schmelzstruktur der Backenzahne. His-
tologisch-phyletische Studie. Ann. Zool. Soc.
Zool.-Bot. Fennicae Vanamo, 2: 1-274.

KOWALSKI, K. 1974. Middle Oligocene Rodents
from Mongolia. Results of the Polish-Mongo-
lian Palaeontological Expeditions — Part V. Pa-
laeontol. Pol., 30: 147-178.

Kraglievich, J. L. 1965. Speciation phyletique
dans les rongeurs fossiles du genre Eumysops
Amegh. {Echimyidae, Heteropsomyinae).
Mammalia, 29: 258-267.

Kraglievich, L. 1926. Los grandes roedores ter-
ciarios de la Argentina y sus relaciones con
ciertos generos pleistocenos de las Antillas. An.
Mus. Nac. Hist. Nat., Buenos Aires, 34: 121-
135.

. 1930a. Descripcion de un interesante roe-
dor eumegamido descubierto en el Uruguay:
Gyriabrus Teisseirei n. sp. Rev. Soc. Amigos
Arqueol., Montevideo, 4: 219-224.

. 1930b. Diagnosis osteologico-dentaria de

los generos vivientes de la subfamilia "Cavi-
inae." An. Mus. Nac. Hist. Nat., Buenos Aires,
36: 59-96.

. 1930c. La Fomiacion Friaseana del Rio

Frias, Rio Fenix, Lagima Blanca, etc. y su fauna
de Mamiferos. (Extracto de un extenso trabajo
del autor.) Physis (Rev. Soc. Arg. Cienc. Nat.),
10: 127-161.

. 1931. Cuatro notas paleontologicas (Sobre

"Octomylodon aversus" Amegh., "Argyrolagiis
palmerf Amegh., "Tetrastylus montanus"
Amegh., y "Miinizia paranensis" N. Gen., N.
Sp.). Physis, 10: 242-266.

. 1932a. Diagnosis de nuevos generos y es-

pecies de roedores cavidos y eumegamidos fo-
siles de la Argentina. Rectificacion generica de
algunas especies conocidas y adiciones al cono-
cimiento de otras. An. Soc. Cient. Arg., 114:
155-181,211-237.

. 1932b. Nuevos apuntes para la geologia y

paleontologia del Uruguay. An. Mus. Hist. Nat.
Montevideo, (2), 3: 355^38.

— . 1934. La antigiiedad Pliocena de las faunas
de Monte Hemioso y Ghapadmalal deducidas
de su comparacion con las cjue le precedieron
y succedieron. Montevideo: El Siglo lllustrado,
938: 1-136.

1940. Los roedores extinguidos del grupo

Neoepiblemidae. Obras Geol. Paleontol., La
Plata, 3: 739-764.

Ladd, J. W. 1976. Relative motion of South Amer-
ica with respect to North America and Carib-
bean tectonics. Bull. Geol. Soc. Amer., 87:
969-976.

Landry, S. C., Jr. 1957. The interrelationships of
the New and Old World hystricomorph rodents.
Univ. Calif. Publ. Zool., 56: 1-118.

. 1977. Review of: The Biology of Hystri-
comorph Rodents, I. W. Rowlands and B. J.
Weir (eds.). Jour. Mammal., 58: 459-461.

Lavocat, R. 1951. Le parallelisme chez les ron-
geurs et la classification des pores-epics. Mam-
malia, 15: 32-38.

. 1954. Sur I'importance des phenomenes de

parallelisme dans revolution. C. R. 19th Inter-
nat. Geol. Congr., 1952, Pt. 19, pp. 125-129.

. 1955. Sur les relations systematiques entre

les Theridoniyides et divers Rongeurs fossiles
d'Afrique. C. R. Acad. Sci. Paris, (D), 240: 634-
635.

. 1962. Reflexions sur I'origine et la structure

du groupe des Rongeurs. Colloques Intemat.
C. N. R. S., No. 163, pp. 491-501.

. 1967a. Les microfaunes du Neogene

d'Afrique orientale et leurs rapports avec celles
de la region palearctique, pp. 57-66. In W. W.
Bishop and J. D. Clark (eds.), Background to
Evolution in Africa. Chicago, Univ. Chicago
Press, X + 935 pp.

. 1967b. Observations sur la region auditive

des rongeurs Theridomorphes. Problemes ac-
tuels de Paleontologie (Evolution des Ver-
tebres), Colloques Intemat. C. N. R. S., No.
611, pp. 491-501.

. 1969. La systematique des Rongeurs hys-

tricomorphes et la derive des continents. C. R.
Acad. Sci. Paris, (D), 269: 1496-1497.

. 1971a. Affinites systematique des cavio-

morphes et des phiomorphes et origine Afri-
caine des caviomorphes. An. Acad. Bras.
Cienc, 43 (Supl.): 515-522.

. 1971b. Essai sur les relations du maxillaire

et du palatin dans la region orbito-temporale
des Rongeurs. Abh. hess. L.-Amt Bodenforsch.,
60: 117-120.

— . 1973. Les rongeurs du Miocene d'Afrique
Orientale. 1. Miocene inferieur. Ecole Pratique
des Hautes Etudes, Inst, de Montpellier, Mem.
1: iv -I- 284 pp.

— . 1974a. The interrelationships between the
African and South American rodents and their
bearing on the problem of the origin of South
American monkeys. Jour. Human Evol., 3:
323-326.

Oligocene Rodents of Bolivia • Patterson and Wood

533

— . 1974b. What is an hystricognath? Zool. Soc.

Lond., Symposium 34, pp. 7-19; discussion pp.

55-60.
. 1976. Rongeurs caviomorphes de

roligocene de Bolivie. II. Rongeurs du Bassin

Deseadien de Salla-Luribay. Palaeovertebrata,

7: 15-90.
— . 1978. Rodentia and Lagomorpha, pp. 69-

89. In V. S. Maglio and H. B. S. Cooke (eds.),

Evolution of African Mammals. Cambridge,

Harvard Univ. Press, ix + 641 pp.

1981. (Book published in March 1981, but

dated 1980; reprints dated 1981.) The impli-
cations of rodent paleontology and biogeogra-
phy to the geographical sources and origin of
platyrrhine primates, pp. 93-102. In R. L. Cio-
chon and A. B. Chiarelli (eds.), Evolutionary
Biologv' of the New World Monkeys and Con-
tinental Drift. N.Y., Plenum Publ. Corp., xvi +
528 pp.

Lavocat, R., and J. -P. Parent. 1971. Valeur sys-
tematique de la region de I'oreille moyenne
des Rongeurs. C. R. Acad. Sci. Paris, (D), 273:
1478-1480.

Li, C.-k. 1975. Yuomys, a new ischyromyoid rodent
genus from the Upper Eocene of North China.
Vertebr. PalAsiat., 13: 58-70 (English sum-
mary, pp. 68-70).

. 1977. Paleocene eurymylids (Anagalida,

Mammalia) of Qianshan, Anhui. Vertebr.
PalAsiat., 15: 103-118 (English summary, pp.
116-118).

Li, C.-k., C.-s. Chiu, D.-f. Yan, and S.-h. Hsieh.
1979. Notes on some early Eocene mammalian
fossils of Hengtung, Hunan. Vertebr. PalAsiat.,
17: 71-80 (English summary, pp. 79-80).

VON LiNSTOW, O. 1899. Nematoden aus der Ber-
liner Zoologischen Sammlung. Mitt. zool.
Samml. Mus. Naturkierk. Berlin, 1(2): 1-28.

LOOMIS, F. B. 1914. The Deseado Fomiation of
Patagonia. Rumford Press, Concord, N.H., 232
pp.

Malfait, B. T., and M. G. Dinkelman. 1972.
Circum-Caribbean tectonic and igneous activ-
ity and the evolution of the Caribbean Plate.
Bull. Geol. Soc. Amer., 83: 251-271.

Marshall, L. G., R. Pascual, G. H. Curtiss, and
R. E. Drake. 1977. South American geochro-
nology. Radiometric time scale for middle to
late Tertiary mammal-bearing horizons of Pata-
gonia. Science, 195: 132.5-1328.

Matthew, W. D., and W. Granger. 1923. New
Bathyergidae from the Oligocene of Mongolia.
Amer. Mus. Novitates, No. 101, 5 pp.

McKenna, M. C. 1975a. Toward a phylogenetic
classification of the Mammalia., pp. 21-46. In
W. P. Luckett and F. S. Szalay (eds.), Phylog-
eny of the Primates. Plenum Publ. Corp., New
York, xiv -I- 483 pp.

. 197.5b. Fossil mammals and early Eocene

North Atlantic land continuity. Ann. Missouri
Bot. Garden, 62: 335^.53.

. 1981. (Book published in March 1981, but

dated 1980; reprints dated 1981.) Early history
and biogeography of South America's extinct
land mammals, pp. 43-77. In R. L. Ciochon and
A. B. Chiarelli (eds.). Evolutionary' Biology of
the New World Monkeys and Continental
Drift. N.Y., Plenum Publ. Corp., xvi + .528 pp.

Merriam, R. 1972. Evidence for 200 miles of right
lateral displacement on faults of the San Andre-
as zone in southwest Sonora, Mexico. Geol.
Soc. Amer., Abstracts with Programs, 4: 199.

MiCHAUX, J. 1964. Diagnoses de quelques Para-
myides de 1 Eocene inferieur de France. C. R.
Somm., Soc. Geol. France, 1964 (4): 153.

. 1968. Les Paramyidae (Rodentia) de

I'Eocene inferieur du Bassin de Paris. Palaeo-
vertebrata, 1: 1.35-193.

Miller, G. S., Jr. 1929. A second collection of
mammals from caves near St. Michel, Haiti.
Smiths. Misc. Coll., 81 (9): 1-30.

Miller, G. S., Jr., and J. W. Gidley. 1918. Syn-
opsis of the supergeneric groups of rodents.
Jour. Wash. Acad. Sci., 8: 431-448.

MONES, A., AND L. R. Castigliom. 1979. Addi-
tions to the knowledge on fossil rodents of
Uruguav (Mammalia: Rodentia). Palaontol. Z.,
53: 77^7.

MoNES, A., AND M. Ubilla. 1978. La edad De-
seadense (Oligoceno Inferior) de la Formacion
Fray Bentos y su contenido paleontologico, con
especial referencia a la presencia de Probor-
hyaena cf. gigantea Ameghino (Marsupialia:
Borhyaenidae) en el Uruguay. Nota Preliminar.
Comun. Paleontol. Mus. Hist, natur. Montevi-
deo 1: 151-158.

Moody, P. A., and D. A. Doninger. 1956. Sero-
logical light on porcupine relationships. Evo-
lution, 10: 47-55.

Murray, A. 1866. The geographic distribution of
mammals. London, Day & Son Ltd., xvi + 420
pp.

OSBORN, H. F. 1910. The age of Mammals in Eu-
rope, Asia and North America. New York, Mac-
millan Co., xx -I- 63.5 pp.

Parent, J. -P. 1976. Disposition fondamentale et
variabilite de la region auditive des rongeurs
hvstricognathes. C. R. Acad. Sci. Paris, (D),
283: 243-245.

Pascual, R. 1967. Los roedores Octodontoidea
(Caviomorpha) de la fomiacion Arroyo Chasico
(Plioceno inferior) de la Provincia de Buenos
Aires. Rev. Museo La Plata (n.s.), Secc. Paleon-
tol., 5: 259-282.

Patterson, B. 1954. The geologic history of non-
human primates in the Old World. Human
Biol., 26: 191-209.

Patterson, B., and R. Pascual. 1968. New echi-
myid rodents from the Oligocene of Patagonia,
and a synopsis of the family. Mus. Comp. Zool.,
Breviora No. 301, 14 pp.

. 1972. The Fossil Mammal Fauna of South

America, pp. 247-309. /;i A. Keast, F. C. Erk,

534 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

and B. Glass (eds.). Evolution, Mammals and
Southern Continents. Albany, State Univ. N.Y.
Press, 543 pp.

Perfit, M. R., and B. C. Heezen. 1978. The ge-
()log\' and evolution ot the Cayman Trench.
Bull Geol. Soc. Amer., 89: 1155-1174.

PiLLERi, G. 1959a. Beitrage zur vergleichenden
Morphologie des Nagetiergehirnes. Makrosko-
pische und vergleichend-anatomische Betrach-
tungen iiber das Zentralnervensystem der Na-
getiere. 1. Beitrag: Sciuromorpha. Acta anat.,
39, Suppl. 38: 1^2.

. 1959b. Beitrage zur vergleichenden Mor-
phologie des Nagetiergehirnes. Makrosko-
pische und vergleichend-anatomische Betrach-
tungen iiber das Zentralnervensystem der
Nagetiere. 2. Beitrag: Hystricomorpha. Acta
anat., 39, Suppl. 38: 43-95.

. 1959c. Beitrage zur vergleichenden Mor-
phologie des Nagetiergehirnes. Makrosko-
pische und vergleichend-anatomische Betrach-
tungen iiber das Zentralnervensystem der
Nagetiere. 3. Beitrag: Das Gehini der Wasser-
nageni [Castor canadensis. Ondatra zibethica,
Myocastor coijpus). Acta anat., 39, Suppl. 38:
96-123.

. 1959d. Das Gehirn der Chinchillas und

vergleichend-anatomische Betrachtungen mit
verwandten Nagerarten (Rodentia, Hystrico-
morpha). Acta Zool., 40: 23-il.

. 1959e. Das Gehirn des kanadischen Bi-

bers — erste zusammenfassende Ergebnisse nach
makroskopischen Untersuchungen. Z. Anat.
Entwicklungsgesch., 121: 179-187.

. 1959f. Uber das Gehirn einiger Dasyproc-

tinae (Rodentia — Hystricomorpha, Dasyprocti-
dae). Rev. Suisse Zool., 66: 633-639.

. 1959g. Das Gehirn von Paradolichotis sal-

inicola und Dolichotis patagona (Rodentia,
Hystricomorpha). Z. Anat. Entwicklungsgesch.,
121: 142-148.

. 1959h. Zur Morphologie und postembrion-

alen Entwicklung des Gehirns von Dasyprocta
aguti Lin. (Rodentia, Hystricomorpha). Rev.
Suisse Zool., 66: 545-553.

1959i. Ontogenese und Cerebralisation

beim Biber (Castor canadensis, Kuhl). Rev.
Suisse Zool., 66: 165-174.

— . 1959j. Das Zentralnervensystem des Cas-
tor canadensis, Kuhl. Acta Zool., 40: 105-140.

— . 1960a. Beitrage zur vergleichenden Mor-
phologie des Nagetiergehirnes. 4. Beitrag:
Zentralnervensystem, Korperorgane und stam-
mesgeschichtliche Verwandschait der Aplo-
dontia rufa Rafinesque (Rodentia, Aplodon-
toidea). Acta anat, 42, Suppl. 40: 5-35.

— . 1960b. Beitrage zur vergleichenden Mor-
phologie des Nagetiergehirnes. 5. Beitrag: Ver-
gleichendmorphologische Untersuchungen
iiber das Zentralnervensystem nearktischer

Sciuromorpha und Bemerkungen zum Problem
Himform und Ta.xonomie. Acta anat., 42,
Suppl. 40: .36-68.

. 1960c. Beitrage zur vergleichenden Mor-
phologie des Nagetiergehirnes. 6. Beitrag: Ma-
terialien zur vergleichendes Anatomic des Ge-
hirnes der Myomorpha. Acta anat., 42, Suppl.
40: 69-88.

— . 1960d. Comparative anatomical investiga-
tions on the central nervous system of rodents,
and relationships between brain fonn and tax-
onomy. Rev. Suisse Zool., 67: 373-386.

— . 1960e. Zur Morphologie des Zentralner-
vensystems von Erethizon dorsatum, Linnaeus
(Rodentia, HystricomoriDha). Z. Anat. Entwick-
lungsgesch., 121: 369-375.

— . 1960f. Uber das Zentralnervensystem von
Heterocephaliis glaber (Rodentia, Bathyergi-
dae). Acta zool., 41: 101-111.

— . 1961a. Uber das Gehirn von Hypogeomys
antimena Grandidier 1869 (Rodentia, Neso-
myinae). Rev. Suisse Zool., 68: 425^31.

— . 1961b. Das Gehirn der Macrotarsomys
(Rodentia, Nesomyinae). Rev. Suisse Zool., 68:
431-442.

1962. Vergleichend-anatomische Unter-

suchungen am Gehirn von Lophiomys, Tatera,
Brachyuromys und Petromus (Mammalia, Ro-
dentia). Rev. Suisse Zool., 69: 481-496.

Plane, M. D. 1967. Stratigraphy and Vertebrate
Fauna of the Otibanda Formation, New
Guinea. Comm. Australia Dept. Natl. Develop.,
Bull. Bur. Min. Resources, Geol. and Geophy.,
No. 86, pp. v-vi, 1-64.

POCOCK, R. I. 1922. On the external characters of
some hvstricomorph rodents. Proc. Zool. Soc.
Lond., i922: 365-127.

QUENTIN, J.-C. 1973a. Affinites entre les Oxyures
parasites des rongeurs Hystricides, Erethizon-
tides et Dinomyides. Interet paleobiogeogra-
phique. C. R. Acad. Sci. Paris, (D), 276: 2015-
2017.

. 1973b. Morphologie et position systema-

tique d'Oxyuris stossichi Setti, 1897. Interet
paleobiogeographique de cette espece. Bull.
Mus. Nat. Hist. Nat., (3), 183: 1403-1408.

1973c. Les Oxyurinae de Rongeurs. Bull.

Mus. Nat. Hist. Nat., (3), 167: 1046-1096.

Rensberger, J. M. 1971. Entoptychine pocket go-
phers (Mammalia, Geomyoidea) of the early
Miocene John Dav Fomiation, Oregon. Univ.
Calif Publ. Geol. Sci., 90: 1-209.

Roberts, A. 1951. The Mammals of South Africa.
Central News Agency, South Africa, 700 pp.

Rodriguez, L., \'. BaruS, and J. Kratochvil.
1979. The genus Mesocapromys, a link be-
tween the Families Echimvidae and Capro-
myidae. Folia Zool., 28: 97-102.

Romer, a. S., N. E. Wright, T. Edinger, and R.
Van Frank. 1962. Bibliography of Fossil Ver-

Oligocene Rodents of Bolivia • Patterson and Wood 535

tebrates Exclusive of North America, 1509-
1927. Geol. Soc. Amer., Mem. 87, 2 Vols.
Ixxxix + 1544 pp.

Rowlands, I. W., and B. J. Weir (eds.) 1974. The
biology of the Hystricomorph Rodents. Zool.
Soc. London, Symposium 34, xix + 482 pp.

Sahni, A. 1980. SEM studies of Eocene and Si-
walik rodent enamels. Geoscience Jour., 1, No.
2: 21^0.

Sahni, A., AND V. P. MiSHRA. 1975. Lower Tertiar>
vertebrates from western India. Palaeontol.
Soc. India Mon. 3: 48 pp.

Sahni, a., AND V. C. Srivastava. 1976. Eocene
rodents and associated reptiles from the Suba-
thu Formation of northwestern India. Jour. Pa-
leontol., 50: 922-928.

Sarich, V. M., AND J. E. Cronin. 1981. (Book pub-
lished in March 1981, but dated 1980; reprints
dated 1981.) South American mammal molec-
ular systematics, evolutionary' clocks and con-
tinental drift, pp. 399^21. In R. L. Ciochon
and A. B. Chiarelli (eds.). Evolutionary- Biology
of the New World Monkeys and Continental
Drift. N.Y., Plenum Publ. Corp., xvi + 528 pp.

Savage, R. J. G. 1969. Early Tertiary mammal lo-
cality in southern Libya. Proc. Geol. Soc.
Lond., 1969 (1657): 167-171.

. 1971. Review of the fossil mammals of Lib-
ya. Symposium on the Geology of Libya, Fac.
Sci. Univ. Libya, pp. 217-225.

SCHAUB, S. 1953a. Remarks on the distribution and
classification of the "Hystricomorpha." Verb,
naturforsch. Ges. Basel, 64: 389-400.

. 1953b. La trigonodontie des rongeurs sim-

plicidentes. Ann. Palaeontol., 39: 29-57.

1958. Simplicidentata ( = Rodentia). In J.

Piveteau (ed.), Traite de Paleontologie. Paris,
Masson et Cie., 6 (2): 659-818.

Schlosser, M. 1884a. Nachtrage und Berichti-
gungen zu: die Nager des europaischen Ter-
tiars. Palaeontographica, 31: 323-328.

. 1884b. Die Nager des europaischen Ter-

tiars nebst Betrachtungen iiber die Organisa-
tion und die geschichtliche Entwicklung der
Nager iiberhaupt. Palaeontographica, 31: 1-
140.
-. 1911. Beitrage zur Kenntnis der oligozanen

Landsaugetiere aus deni Fayum (Agypten)
Beitr. Palaontol. Geol. Osterr.-Ung. u. d. Ori-
ents, 24: 51-167.

Scott, W. B. 1905. Glires. Part 111 of Reports of
The Princeton University Expeditions to Pata-
gonia, 1896-1899, Vol. 5, Palaeontology, pp.
384^90.

Shevyreva, N. S. 1968. Novie rod Grizhunov Se-
meistva Ctenodactylidae izh Srednego Oligo-
zena Kazakhstana i Mongolii. Bull. Moskov.
Obshtch. Ispytat. Prirodi, Otd. Geol., 43 (1):
155-156.

. 1971a. Novie Sredneoligozenovie Grizhu-

ni Kazhakstana i Mongolii. Akad. Nauk SSSR,
Paleontol. Inst, Trudi, 130: 70-86.

— . 1971b. Pervayah Naxodka Eozenovix Gri-
zhunov V SSSR. Soobshtch. Akad. Nauk Gru-
zhinskoi SSR, 61: 745-747.

— . 1972a. Ob Evolutsii Grizhunov Semeistva
Sciuravidae. Dokladi Akad. Nauk SSSR, 206:
1453-1454.

— . 1972b. Novie Grizhuni izh Paleogena
Mongolii i Kazhakstana. Akad. Nauk SSSR, Pa-
leontol. Zh., 1972: 134-145.

— . 1972c. Paleogenovie Grizhuni Azhii (Se-
meistva Paramyidae, Sciuravidae, Ischyromyi-
dae, Cylindrodontidae). Akad. Nauk SSSR, Pa-
leontol. Inst. Avtorepherat Dissertatsii Biol.
Nauk, no. 04.00.09, Paleontol. i. Stratigr., pp.
1-25.

1976. Paleogenovie grizhuni Azhii (Se-

meistva Paramyidae, Sciuravidae, Ischyromyi-
dae, Cylindrodontidae). Akad. Nauk SSSR,
Trudi Paleontol. Inst. 158: 1-115.

Shewreva, N. S., and V. M. Chkhikvadze. 1978.
Myomorph rodents (Mammalia) from the Eo-
cene of Zaissan Depression. 2nd Internat. The-
riolog. Congr., Brno, June 20-27, 1978. Ab-
stracts of Papers, p. 418.

Simons, E. L. 1968. African Oligocene Mammals:
Introduction, history of study, and faunal
succession. Part I of: Early Cenozoic Mamma-
lian Faunas, Favum Province, Egypt. Bull. Pea-
body Mus. Nat.' Hist, 28: 1-21. '

. 1976. The fossil record of primate phylog-

eny, pp. 34-62. In M. Goodman and J. Tashin
(eds.). Molecular Anthropology. r''=>w York, Ple-
num Publ. Co.

Simpson, G. G. 1940. Review of the mammal-bear-
ing Tertiary of South America. Proc. Amer.
Phil. Soc, 83: 649-709.

. 1945. The principles of classification and

a classification of mammals. Bull. Amer. Mus.
Nat Hist., 85: xvi + 350 pp.

. 1950. History of the fauna of Latin America.

Amer. Scientist, 38: 361-389.

. 1978. Early mammals in South America:

fact, controversv and mvsterv. Proc. Amer. Phil.
Soc, 122: 318^28.

1980. Splendid isolation. The curious his-

tory of South American mammals. Yale Univ.
Press, New Haven, 266 pp.

Smith, C. H. 1842. Mammalia. Introduction to
Mammals. In W. Jardine (ed.). The Naturalist's
Library. London, Chatto and Windus, 15: 75-
313.

Stehlin, H. G. 1909. Remarques sur les faunules
de mammiferes des couches eocenes et oligo-
cenes du Bassin de Paris. Bull. Soc. Geol.
France, (4), 9: 488-520.

Stehlin, H. G., and S. Schaub. 1951. Die Tri-
gonodontie der simplicidentaten Nager.
Schweiz. palaontol. Abhandl., 67: 1-385.

536 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Stirton, R. a. 1953. Vertebrate paleontology and
continental stratigraphy in Colombia. Bull.
Geol. Soc. Amer., 64: 603-622.

Stromer, E. 1923. Ergebnisse der Bearbeitung
niitteltertiarer Wirbeltier-Reste aus Deutsch-
Siidwest-Afrika. Sitzber. Bayer. Akad. Wiss.
1923: 253-270.

. 1926. Re.ste Land- und Siisswasser-bewoh-

nender Wirbeltiere aus den Diamantenfeldern
Deutsch-Siidvvestatrikas. In E. Kaiser, Die
Diamantenwiiste Siidwestafrikas. Dietrich Rei-
mer, Berlin, 2: 107-153.

SzALAY, F. S. 1976. Systematics of the Omoniyidae
(Tarsiifoniies, Primates), taxonomy, phylogeny
and adaptations. Bull. Amer. Mus. Nat. Hist.,
156: 157^49.

Tate, G. H. H. 1935. The taxonomy of the genera
of neotropical hystricoid rodents. Bull. Amer.
Mus. Nat. Hist., 68: 295-147.

. 1951. The rodents of Australia and New

Guinea. Bull. Amer. Mus. Nat. Hist, 97: 183-
430.

Teilhard de Chardin, p. 1926. Description de
Mammiferes Tertiaires de Chine et de Mon-
golie. Ann. Paleontol. 15: 52 pp.

Thenius, E. 1959. Wirbeltierfaunen. Pt. 2 of Ter-
tiar, by A. Papp and E. Thenius. In F. Lotze
(ed.), Handbuch der stratigraphischen Geolo-
gic, 3 (Pt. 2): xi + 328 pp.

Thomas, O. 1897. On the genera of rodents: an
attempt to bring up to date the current arrange-
ment of the order. Proc. Zool. Soc. Lond.,
1896: 1012-1028.

Traub, R. 1980. The Zoogeography and Evolution
of some Fleas, Lice and Mammals, pp. 93-172.
In R. Traub and H. Starcke (eds.). Fleas. Proc.
Intemat. Conference Fleas, Ashton Wold, Pe-
terborough, U.K., 21-25 June 1977. A. A. Bal-
kema, Rotterdam, 480 pp.

TULLBERG, T. 1899. Ueber das System der Nage-
thiere: eine phylogenetische Studie. Nova Acta
Reg. Soc. Scient. Upsala, (3), 18: i-v, 1-514.

TURNBULL, W. D., AND E. L. LUNDELIUS. 1970.
The Hamilton fauna: a late Pliocene mamma-
lian fauna from the Grange Bum, Victoria, Aus-
tralia. Fieldiana, Geol., 19: 163 pp.

Van Kampen, P. N. 1905. Die Tympanalgegend
des Saiigetierschadels. Morph. Jahrb., 34: 321-
722.

Varona, L. S., and O. Arredondo. 1979. Nuevos
taxones fosiles de Capromyidae (Rodentia: Ca-
viomorpha). Poeyana, No. 195, 51 pp.

VlANEY-LlAUD, M. 1972. L'evolution du genre
Theridomijs a I'Oligocene moyen. Interet bio-
stratigraphique. Bull. Mus. Nat. Hist. Nat., (3),
98: 295-371.

Viera, C. da C. 1955. Lista remissiva dos mami-
feros do Brasil. Arq. Zool. Est. Sao Paulo, 8:
341^74.

Vinogradov, B. C, and P. P. Gambarian. 1952.
Oligotsenovie Tsilindrodonti Mongolii i Ka-

zhakstana (Cylindrodontidae, Glires, Mam-
malia). Akad. Nauk SSSR, Trudi Paleontol. Inst.,
41: 13-42.

VUCETICH, M. G. 1975. La anatomia del oido me-
dio como indicadora de relaciones sistematicas
y filogeneticas en algunos grupos de roedores
Caviomorpha. Actas Primero Congr. Argent.
Paleontol. Bioestrat., Tucuman, Arg., 2: 477-
494.

Wahlert, J. H. 1968. Variability in rodent incisor
enamel as viewed in thin section, and the mi-
crostructiire of the enamel in fossil and recent
rodent groups. Mus. Comp. Zool., Breviora No.
309, 18 pp.

. 1973. Protoptychus, a hystricomorphous

rodent Irom the late Eocene of North America.
Mus. Comp. Zool., Breviora No. 419, 14 pp.

. 1974a. The cranial foramina of Protrogo-

morphous rodents, an anatomical and phylo-
genetic study. Bull. Mus. Comp. Zool., 146:
363-410.

1974b. [Work on microstructure of rodent

enamel reported.] Soc. Vert. Paleontol. News
Bull. No. 100, p. 18.

Waterhouse, G. E. 1839. Observations on the
Rodentia, with a view to point out the groups,
as indicated by the structure of the crania, in
this order of Mammals. Mag. Nat. Hist., n.s., 3:
90-96.

. 1841. Observations on the Rodentia. Ann.

Mag. Nat. Hist., 8: 81-84.

Webb, S. D. 1978. A history of savannah verte-
brates in the New World. Part II: South Amer-
ica and the Great Interchange. Ann. Rev. Ecol.
Syst., 9: 393^26.

West, R. M. 1980. Middle Eocene large mammal
assemblage with Tethyan affinities, Ganda Kas
region, Pakistan. Jour. Paleontol., 54: 508-533.

Wilson, R. W. 1937. Two new Eocene rodents
from the Green River Basin, Wyoming. Amer.
Jour. Sci., (5), 34: 447^56.

. 1949. Early Tertiary rodents of North

America. Carnegie Inst. Wash., Publ. 584: 67-
164.

. 1972. Evolution and extinction in early

Tertiary Rodents. 24th Intemat. Geol. Congr.,
1972, Sec. 7, pp. 217-224.

1980. The stratigraphic sequence of North

American rodent faunas, pp. 278-283. In J.
Michaux (ed.), Paheovertebrata, Mem. Jubil.
R. Lavocat, 1980, 382 pp.

Winge, H. 1887. Jordfundne og nulevende Gnav-
ere (Rodentia) fra Lagoa Santa, Minas Geraes,
Brasilien. Med Udsigt over Gnavernes ind-
byrdes Slaegtskab. Museo Lundii (Copenha-
gen), 1, No. 3: 1-178.

Wood, A. E. 1935. Evolution and relationships of
the Heteromyid Rodents, with new forms from
the Tertiary of westem North America. Ann.
Carnegie Mus., 24: 73-262.

Oligocene Rodents of Bolivia • Patterson and Wood 537

. 1936a. Geomyid rodents from the middle

Tertiary. Amer. Mus. Novitates No. 866, 31 pp.

. 1936b. Parallel radiation among the Geo-

myoid Rodents. Jour. Mammal., 18: 171-176.

. 1937.^ Rodentia. Part II of: The Mammalian

Fauna of the White River Oligocene, by W. B.
Scott, G. L. Jepsen, and A. E. Wood. Trans.
Amer. Phil. Soc. (n.s.), 28: 155-269.

— . 1940. Lagomorpha. Part 111 of: The Mam-
malian Fauna of the White River Oligocene, by
W. B. Scott, G. L. Jepsen, and A. E. Wood.
Trans. Amer. Phil. Soc. (n.s.), 28: 271-362.

— . 1942. Notes on the Paleocene lagomorph,
Eurymylus. Amer. Mus. Novitates No. 1162, 7
pp.

— . 1947. Rodents — a study in evolution. Evo-
lution, 1: 154-162.

— . 1949. A new Oligocene rodent genus from
Patagonia. Amer. Mus. Novitates No. 1435, 54
pp.

— . 1950. Porcupines, paleogeography and par-
allelism. Evolution, 4: 87-98.

— . 1955. A revised classification of the ro-
dents. Jour. Mammal., 36: 165-187.

— . 1957. What, if anything, is a rabbit? Evo-
lution, 11: 417-425.

— . 1962a. The early Tertiary rodents of the
Family Paramyidae. Trans. Amer. Phil. Soc.
(n.s.), 52: 1-261.

— . 1962b. The juvenile tooth patterns of cer-
tain African rodents. Jour. Mammal., 43: 310-
322.

— . 1965. Grades and Glades among Rodents.
Evolution, 19: 115-130.

. 1968. The African Oligocene Rodentia.

Part II of: Early Cenozoic Mammalian Faunas,
Fayum Province, Egypt. Bull. Peabody Mus.
Nat. Hist, 28: 23-105.

. 1969. The third upper premolars of the

woodchuck (Marmota). Jour. Mammal., 50:
608-609.

. 1970a. The early Oligocene rodent Ardyn-

omys (Family Cylindrodontidae) from Mongo-
lia and Montana. Amer. Mus. Novitates, No.
2418, 18 pp.

. 1970b. The European Eocene paramyid

rodent, Plesiarctomys. Verhandl. naturf Ges.
Basel, 80: 237-278.

. 1972. An Eocene hystricognathous rodent

from Texas: its significance in interpretations
of continental drift. Science, 175: 1250-1251.

. 1973. Eocene rodents, Pruett Formation,

southwest Texas; their pertinence to the origin
of the South American Gaviomorpha. Texas
Mem. Mus., Pearce-Sellards Series No. 20, 40
pp.

. 1974a. Early Tertiary vertebrate faunas,

Vieja Group, Trans-Pecos Texas: Rodentia.
Texas Mem. Mus., Bull. 21, iv + 112 pp.

— . 1974b. The evolution of the old world and
new world hystricomorphs. Zool. Soc. London,

Symposium 34, pp. 21-54; discussion, pp. 55-
60.

— . 1975a. The problem of the hystricognath-
ous rodents. Iti G. R. Smith and N. E. Friedland
(eds.), Studies on Genozoic Paleontology and
Stratigraphy in Honor of Claude W. Hibbard,
Glaude W. Hibbard Memorial Vol. 3, Univ.
Mich., Papers on Paleontol., No. 12, pp. 75-80.

— . 1975b. Review of: Les Rongeurs du Mio-
cene d'Atrique Orientale. 1. Miocene Inferieur,
by R. Lavocat. Jour. Paleontol., 49: 942-943.

— . 1977a. The evolution of the rodent Family
Gtenodactylidae. Jour. Palaeontol. Soc. India,
20: 120-137.

— . 1977b. The Rodentia as clues to Genozoic
migrations between the Americas and Europe
and Africa. In R. M. West (ed.). Paleontology
and Plate Tectonics, Milwaukee Public Mus.
Spec. Publ. Biol, and Geol., 2: 9.5-109.

— . 1980a. The earliest South American ro-
dents: a comment. Syst. Zool., 29: 96.

— . 1980b. The Oligocene Rodents of North
America. Trans. Amer. Phil. Soc, 70: 3-69.

— . 1980c. Problems of classification as applied
to the Rodentia, pp. 263-272. In J. Michaux
(ed.), Palyeovertebrata, Mem. Jubil. R. Lavocat
1980, 382 pp.

— . 1981. (Book published in March 1981, but

dated 1980; reprints dated 1981.) The Origin of
the Gaviomorph Rodents from a source in Mid-
dle America: a clue to the area of origin of the
Platyrrhine Primates, pp. 79-91. In R. L. Gio-
chon and A. B. Ghiarelli (eds.), Evolutionary
Biology of the New World Monkeys and Gon-
tinental Drift. N.Y., Plenum Publ. Gorp., xvi +
528 pp.

— . In press, a. Hystricognathy in the North
American Oligocene rodent Cylindrodon and
the origin of the Gaviomorpha. In R. Mer-
gel (ed.), Garnegie Mus. Nat. Hist., Special
Publ.

In press, b. The radiation of the Order Ro-

dentia in the Southern Continents; the dates,
numbers and sources of the invasions. In
Schriftenreihe Geologische Wissenschaften,
Berlin.

Wood, A. E., and B. Patterson. 1959. The ro-
dents of the Deseadan Oligocene of Patagonia
and the beginnings of South American rodent
evolution. Bull. Mus. Gomp. Zool., 120: 279-
428.

. 1970. Relationships among hystricognath-
ous and hystricomorphous rodents. Mammalia,
34: 628-639.

Wood, A. E., and R. W. Wilson. 1936. A sug-
gested nomenclature for the cusps of the cheek
teeth of rodents. Jour. Paleontol., 10: 388-391.

Wood, A. E., and H. E. Wood, 2nd. 1933. The
genetic and phylogenetic significance of the
presence of a third upper molar in a modern
dog. Amer. Midi. Nat., 14: 36^8.

538 Bulletin Museum oj Coinpanitiic Zoology, Vol. 149, No. 7

Woods, C. A. 1972. Comparative inyolog) of jaw,
hyoid and pectoral appendicular regions of
New and Old World hystriconiorpli rodents.
Bull. Amer. Mus. Nat. Hist., 147: 117-198.

. 1975. The hvoid, laryngeal and pharyngeal

regions of bathyergid and other selected ro-
dents. Jour. Morph., 147: 229-250.
Woods, C. A., and E. B. Howland. 1977. The
skin musculature of Hystricognath and other
selected rodents. Zbl. Vet. Med. C. Anat. Hist.
Embryol., 6: 240-264.

ADDENDUM (by A.E.W.)

Since this manuscript was sent to the
printer, several papers pertinent to our
discussion have been received.

Evidence of the distinct nature of the
Tertiary Middle American rodent fauna
(p. 457) is given by Slaughter (1981), who
described a rather distinctive new genus
of geomyoid from the Miocene of Texas
and Panama. He suggested (1981: 114)
"that the Coastal Plain, perhaps all the
way to Florida, was a northern portion
of the Central American tropical biotic
province . . . ."

Presence of the Hesperomyinae in Chi-
huahua in the early Pliocene (Jacobs and
Lindsay, 1981), well before their earliest
known occurrence in South America,
should put an end to suggestions that they
had an African origin. This also adds to
the documentation of the Middle Ameri-
can fauna.

Vucetich (1980: 99-105) redescribed
Simplimus indivisus and agreed with
Fields that it is a dinomyid (see above,
p. 427). She differs somewhat from our
caviomorph classification. She transferred
(1980: 10) the Dasyproctidae to the Cavi-
oidea, as do we, but left the Dinomyidae
in the Chinchilloidea, and did not men-
tion the Cuniculidae. We continue to
believe (see above, p. 511) that the
Dasyproctidae and Dinomyidae belong
together, in whatever superfamily they
are placed.

Parent (1980: 166-169) described the
middle ear of erethizontids. He conclud-

ed (1980: 253) that they were quite dis-
tinct from the other caviomorphs and that
they perhaps diverged from the Cavio-
morpha before their arrival in South
America. He also noted (op. cit.: 255)
that the erethizontid ear region is the
most primitive among all the hystrico-
gnaths he studied (which did not include
the Eocene franimorphs Reithroparamijs
and Frotoptijchus). I believe that this
monograph demonstrates the correctness
of our interpretation (see above, p. 482)
that he has shown that it would be im-
possible to derive the Erediizontidae from
any Old World hystricognath. His conclu-
sions support those we present above (p.
509). As previously, we accept the thesis
that the Hystricognathi are a natural group.

ADDENDUM, LITERATURE CITED

Jacobs, L. L., and E. H. Lindsay. 1981. Prosig-
modon oroscoi, a new sigmodont rodent from
the late Tertiary of Mexico. Jour. Paleontol.,
55: 425-430.

Parent, J. P. 1980. Recherches sur I'oreille moy-
enne des rongeurs actuels et fossiles. Anatomic.
Valeur systematique. Ecole Pratique des
Hautes Etudes, Inst, de Montpellier, Mem.
11, V + 286 pp.

Slaughter, Bob H. 1981. A new genus of geo-
myoid rodent from the Miocene of Texas and
Panama. Jour. Vert. Paleontol., 1: 111-115.

Vucetich, M.G. 1980. Los Roedores Caviomorpha
de la Edad Friasense (Colloncurense + Fria-
sense). Su evolucion y aportes al conocimiento
de los cambios ambientales del Terciario sup-
erior de Patagonia. Unpublished doctoral thesis,
Facultad de Ciencias Naturales, Univ. Nac. La
Plata, iv + 145 pp.

Oligocene Rodents of Bolivia • Patterson and Wood 539

INDEX

All hixa mentioned in this paper are included in the index. Bold-faced citations are to the descriptions
of the respective taxa. Starred pages refer to illustrations or tables.

Abrocoma 403, 484

Abrocomidae 472, 505, 509, 523

Acarechimys 392*, 501, 529

A. minutus 480

Acaremyidae 529

Acaremyinae 448, 529

Acaremy.s 403, 509

A. niinutisniinus 529

A. minutus 529

Aconaemys 483

Adelphomyinae 391, 392, 393*, 395, 523

Adelphomys 393*

Advenimus 4.54

Afrocricetodon songhori korui 468*

A. s. songhori 468*

Afrocricetodon sp. 1 468*

Afrocricetodon sp. 2 468*

Agouti 482

Ailuravinae 453

Akodon a. arviculoides 462*, 464*

Altonius 452

Amblycera 505, 506

Amblyrhiza 512

Andrewsimys 493, 521, 523, 528

A. parvus 468*

Anomaluridae 460, 469*, 481, 491, 513, 525

Anomalurus 482, 503

A. pusiUus 491

Anoplura 505

Anthracobune 453

Aplodontia 479, 486, 487, 490, 491, 507, 518

Apternodon 524

Ardynomys 455, 513, 515, 516

A. occidentcdis 486

Artiodactyla 466*, 470*, 473, 506

Asteromys 448, 467, 493, 496, 497, 498, 512

A. punctus 461*
Asteromys sp. 461*
Astrapotheria 466*
Atherurus 481, 490

Bathyergidae 450, 480, 481, 483, 485, 513, 515, 516,

517, 518, 518,522,525,527
Bathyergoidea 449, 450, 452*, 455, 469*, 473, 477,

489, 508, 516, 518, 519, 520, 522, 527
Bathyergoides 508, 516, 517, 518, 522, 527

B. neotertiarius 468*, 469, 478, 494*, 516
Bathyergoididae 518
Bathyergomorpha 378, 519, 522
Bathyergus 507, 516

Beatomus 516

Boopidae 505

Brachyodus 524

Branisamys 374, 377, 384, 390, 398, 403, 421, 422,
424, 425, 426, 428, 431-445, 446*, 447, 448,
467, 493, 495, 496, 498, 499, 500, 501, 502, 503,
509, 512, 524

B. luribatjensis 431-445, 432*, 433*, 434*, 436*,

438*, 440*, 442*, 461*, 500, 501*, 528
Branisella 471
Brevistriatinae 472, 473
Cactaceae 462
Calomys callosus expulsus 462*, 464*

C. laucha tener 462*, 464*
Calomys sp. 462*, 464*
Capromyidae 393, 473, 509

Capromyinae 392*, 393, 395, 472, 474, 523, 527

Capromys 472, 473, 491, 507

Carnivora 506

Castor 482, 486, 487, 489, 490

C. fiber 485

Castoridae 492

Castorimorpha 491

Cavia 483, 484, 485, 487, 490, 507

Caviella 473

Caviidae 472, 473, 487, 505, 509, 510, 511, 523, 527

Cavioidea 395, 448, 480, 482, 509, 511, 512, 523,
527, 538

Caviomorpha 380, 382, 388, 393, 428, 449, 450, 451,
452*, 458, 459, 471, 475, 476, 477, 482, 483, 484,
486, 487, 489, 491, 492, 493, 495, 496, 497, 498,
499, 500, 502, 503, 505, 506, 507, 508, 509, 510,
512, 519, 522, 523, 524, 525, 526, 527, 538

Caviomorpha inc. sed. 448, 512

Caviomorpha indet. 461*

Cephalomyidae 512

Cephalomys 374, 375, 392, 398, 400, 403, 408, 409,
416, 421, 422, 425, 426, 427, 428, 430, 431, 432,
437, 444, 445, 448, 463, 467, 493, 495, 496, 497,
498, 500, 501, 502, 511, 512, 524

C. arcidens 430, 444*. 444, 461*, 501*

C. bolivianus 374, 430-431, 445, 461*

C. plexus 430,431,461*

Cephalomys sp. 461*

Cercomys 472, 473, 501, 510

C. cuniculariiis 462*, 463, 464*

C. c. inermis 464*

C. c. laurentius 464*

Chaetomyinae 393*, 394-395, 523, 527

Chaetomys 379, 394, 481, 510, 523

C. subspiriosus 394

Chapattimyidae 4,54, 497, 504, 515, 521, 527

Chapattimys 4.54, 515

Chasichimys 393

Chinchilla 403, 411, 483, 490, .507

Chinchillidae 448, 472, 482, 485, .505, .509, 511, 512,
523

Chinchilloidea 377, 447, 448, 480, .509, 511, 512,
523, 527, 538

Chiroptera 466*

Choeromys 490

C. harrisoni 490, 491

Chubutomys 448, 467, 493, 496, 498, 512

540 Bulletin Museum of Comparative Zoolofiy, Vol. 149, No. 7

C. simpsoni 461*

Citrllus 4^J()

Coetuhu 474, 478, 485, 486, 491

C pri'heiisilis 485

C. p. prehcnsilis 462*, 464*

colomhianus 427

Condylarthra 466*

Conilurus 470

Cordicauda 472

Creodontii 466*

Cricetidae 469*, 491, 525

Cryptomys 516, 517

C. coecutiens 517

Ctenodactylidae 454, 481, 483, 504, 513, 515

Ctenodactylus 473

Ctenomyidae 509, 510, 523

Ctenomyinae 527

Ctenomys All

Ctenophthalminae 506

Ctenophtludmus 506

Cuniculidae 395, 448, 472, 487, 505, 511, 512, 523,

527, 538
Cuniculinae 509
Cuniculus 482,491, 511
Cyclomyhis 505, 515, 516
Cylindrodon 458, 515
Cylindrodontidae 513, 514-515, 515, 516, 517,

522, 524, 525, 527
Cynomys 467

Dactylomyinae 392, 393*, 394, 395, 510, 523
Dasyprocta 398, 403, 406, 408, 409, 411, 412, 427,

428, 429*, 429, 446*, 473, 477, 483, 484, 485,

490, 491, 507

D. croconata 429*
D. noblei 429*

D. prymnolopha 462*, 464*

D. punctata isthmica 429*

Dasyproctidae 377, 395, 406, 426, 429, 445, 447,

448, 472, 487, 505, 509, 511, 512, 523, 524, 527,

538
Dasyproctidae aff. Neoreomys 430, 448, 461*
Dasyproctidae inc. sed. 431
PDasyproctidae gen. et sp. indet. 429
Dasyptodon atavus 373
Dermoptera 473
Deseadomys 385, 392, 393*, 421, 448, 467, 493, 495,

496, 497, 498, 512
D. arambourgi 375, 385, 386, 388, 391, 461*, 496
D. loomisi 375, 385, 391, 461*, 496
Diamantomyinae 521, 523, 527
Diamantomys 408, 480, 481, 486, 488, 493, 521, 523,

528
D. luederitzi 406, 468*, 468, 469, 478, 486, 489,

494*, 503
Dinomyidae 377, 395, 429, 431, 445, 447, 448, 472,

487, 505, 509, 511, 512, 523, 524, 527, 538
Dinomyinae 448
Dinomys 431, 432, 433, 434, 435, 437, 439, 440, 441,

445, 446*, 474, 477, 481, 483, 511
Dinomys sp. 434*
Dipodoidea 481, 491, 513
Dipus 483

Dolichotis 485, 487, 490
Drytomomys 427, 446*, 447, 448, 483

D. aeqiiatoriulis 441, 483, 484*

Echimyidae 377, 382, 391, 392*, 392, 393, 394, 472,

480, 505, 509, 510, 523, 524, 527
Echimyinae 392, 393*, 394, 395, 510, 523
Echimyna 382, 523
Echimys 406, 477, 481, 483, 510

E. armatus 464*

E. dusythrix lamurum 462*

Echimys sp. 462*

Echinoprocta 478

Edentatii 466*

Elasmodontomys 487, 512

Elmerimys 502, 521, 523, 528

E. woodi 468*

Embrithopoda 466*

Entoptychus cf. minor 503

Eocardia 499

Eocardiidae 474, 509, 510, 523, 527

Eocardiidae indet. 461*

Eomyidae 491

Eosteiromys 491

Epiphiomys 493, 521, 523, 527

E. coryndoni 468*

Erethizon 439, 474, 478, 485, 490, 502, 507

£. dorsatum 474, 500, 501*, 501, 502, 503

E. d. dorsatum 491

E. epixanthum 474

Erethizontidae 382, 394, 448, 472, 480, 486, 505,

506, 508, 509, 510, 511, 523, 526, 538
Erethizontoidea 449, 509, 523
Erethizontomorpha 485, 509, 510
Euniegamyinae 448
Eumegamys 509
Eumysops 501
Eurymylidae 4.54
Evaginuris 473, 474, 475, 526

E. compar 474
£. evoluta 474
£. stossichi 474
Floresomys 457

Franimorpha 452*, 459, 473, 477, 482, 484, 487, 489,

508, 513, 514, 514, 515, 516, 522, 524, 527
Franimys 428, 476, 478, 482, 498

F. amherstensis 382
Franimys? 454
Galea 485, 487

G. spixii wellsi 462*, 464*

Gaudeamus 420, 426, 467, 495, 496, 497, 500, 503,

520, 523, 527
G. aegyptius 465*, 492, 495
Geocapromys 507
Geomys 490
Georychus 516
G. capensis 517
G. coecutiens 517
Graphiurus 481
Gregorymys curtus 503
Guanajuatomys 458, 507, 514, 522
Guerlinguetus sp. 462*
Gyropidae 505, 506

Oligocene Rodents of Bolivia • Patterson and Wood 541

Hegetotheria 466*

Heligmonella 472

Heligmonellinae 472, 473

Heligmosomidae 471, 472, 526

Heligmostrongylus 472

Heliophobiits 516

Helminthoxijs 473, 474

Heomys 454

H. orientalis 454

Heptaxodon 512

Heptaxodontidae 448, 509, 512, 523, 527

Heptodon 452

Hesperomyinae 538

Heterocephalus 473, 490, 491

Heteromyidae 474

Heteromyoxyuris 473, 474, 475

Heteropsomyinae 382, 391, 392*, 393, 395, 510, 523

Heteropsomys 394

Heteropsyllidae 506

Hexalobodon 393

Hilgertia 473

Holochilus b. brasiliensis 462*, 464*

Homogalax 452

Hoplomys 507

Hydrochoeridae 377, 472, 487, 505, 509, 510, 511,
523, 527

Hydrochoerus 408, 477, 478, 484, 489, 507, 511

Hydromyinae 470

Hypogeoniys 490

Hyracoidea 466*, 506

Hystricidae 449, 450, 452*, 453, 477, 480, 482, 485,
489, 492, 509, 518, 519, 520, 521, 522, 527

Hystricognathi 378, 379, 380, 449, 450, 453, 455,
460, 475, 476, 481, 483, 484, 485, 487, 492, 504,
507, 508, 509, 512, 514, 519, 522, 527, 538

Hystricoidea 449

Hystricomorpha 378, 450, 451, 490, 492, 513, 519,
522, 527

Hystrix 481, 484, 490, 507, 526

H. brachyura 474

H. brachyura (?) 474

H. cristata 474

Impalaiinae 472, 473

Iticamys 374, 377, 383, 384, 388, 390, 395-426,
427, 428, 429, 430, 431, 432, 433, 437, 439, 440,
445, 446*, 448, 449, 464, 467, 477, 480, 481, 482,
486, 487, 488, 493, 495, 496, 497, 498, 499, 500,
509, 512, 524

7. boliviamis 375, 376*, 377*, 377, 395-426, 397*,
399*, 401*, 402*, 405*, 407*, 412*, 413*, 415*,
417*, 418*, 419*, 420*, 421*, 423*, 429, 430,
449, 461*, 478, 500, 528

7. pretiosus 396, 396, 404, 411, 414, 418, 528

Insectivora 466*

Ischnocera 506

Ischyromyidae 490

Ischyrotomus 406, 488

7. horribilis 479, 481

7. oweni 412, 479

Jaculus 404

Kannabateomys All , 510

Kenyamyidae 523

Kenyamys 493, 521, 523, 527

K. mariae 468*, 503

Kerodon 487

K. rupestris 462*, 464*

Lagidium 403, 473

Lagomoriiha 522, 524

Lagostomus 490

Lemuroidea 473

Leporidae 473

Lepus 485

Litodontomys 394, 427, 448, 467, 493, 496, 498, 512

L. chtibutensis 461*

Litopterna 466*

Luribayomys 374, 448, 493, 498, 524

L. masticator 403, 448, 461*

Mallophaga 505, 506

Manitsha 406

Marmot a 490

M. flaviventris 490

M. marmota 485

M. monax 485

Marsupialia 452, 466*, 473

Mastacomys 470

Megapedetes pentadactylus 468*

Megapedetes sp. 468*

Meldimys 453

Meliakrouniomys 382

Mesembriomys 470

Mesocapromys 510

Metaphiomyinae 523

Metaphiomys 383, 403, 465*, 491, 495, 496, 497,

503, 521, 523, 528
M. beadnelli 465*

M. schaubi 420, 465*, 466, 492, 494*, 499, 500
Metaphiomys sp. indet. 499
Cf. Metaphiomys sp. 465*
Metkamys 4.54
Microcavia 411, 473, 487
Microparamys 452, 4.54
M. Ungchaensis 453, 4.55
Migraveramus 377, 380-382, 448, 449, 459, 467,

493, 496, 497, 498, 512, .523
M. beatus 377*, 380-382, 381*, 449, 461*
Moeritherium 466
Morosomys 513
Muridae 455, 472
Murinae 470
Muroidea 488, 489
Mus 464*
Muscardinus 486

Myocastor 391, 395, 406, 481, 483, 487, 490, 507
Myocastorinae 391, 392*, 395, 523
Myomorpha 490

Myophiomyidae .521, .523, 527, 528
Myophiomyinae 521, 523, 528
Myophiomys 493, .502, 523, 528
M. arambourgi 468*
Myoprocta 398, 403, 406, 408, 409, 410, 414, 426,

427, 428, 429*, 429, 446*, 490
M. acouchy 426, 429*
Myosciurus 404
Mysops 4,58, 515, 516

542 Bulletin Museum of Comparative Zoology, Vol. 149, No. 7

Nectomtfs squamipes 462*, 464*

Neoepihk'inidat' 512, 523, 527

Neoreomys 398, 400, 403, 404, 406, 408, 410, 411,

412, 414, 416, 424, 425, 426, 427, 428, 430, 432,

433, 437, 445, 446*, 447, 478, 499
N. australis 499

Neoscitiromy.^ 391, 467, 496, 520, 523, 527
N. ufricunus 494*, 502, 520
N. simon.si 465*, 467, 502, 520, 527
Nesodon 373
Nippostrongylinae 472
Notiotitanops 458
Notocricetodcyn petteri 468*
Notomys 470
Notoungulata 466*
Octodon All, 483, 491, 507
Octodontidae 377, 380, 391, 392*, 472, 480, 505,

509, 510, 523, 524, 527
Octodontoidea 377, 380, 480, 485, 509, 510, 511, 523
Octodontoidea nov. gen. 1 374, 375, 461*
Octodontoidea nov. gen. 2 374, 461*
Odocoileus 414
Olenopsis 394, 427, 447
O. typicus 447
O. uncinus 447
Ondatra 467
Ortleppstrongylus 472
Oryctolagus 485

Oryzomys capita laticeps 462*, 464*
O. lamia 462*, 464*
O. nigripes 462*, 464*
O. subflavus 462*, 462, 464*
Oryzomys sp. 462*
Otocyon 518

Oxymycterus angularis 462*, 464*
Oxyuridae 471
Oxyurinae 471, 473
Oxyuris 474

Pachynodon validus 373
Palaeopsylla 506

Palmiramys 448, 448-449, 449, 493, 498, 524
P. waltheri 448-449, 449, 461*
Paracryptomys 522
Paradelphomys 392, 393*, 395, 501
Paradolichotis 490
Paraheligmonella 472, 473, 526
Paraheligmonina 472
Paramyidae 382, 398, 435, 437, 452*, 453, 489, 513,

514
Paramys 406, 482
P. copei 479, 488
P. delicatus 479, 482, 488
Paranamys 447
Paranomalurus hishopi 468*
P. soniae 468*
P. ivalkeri 468*
Paraphiomys 391, 408, 409, 477, 480, 481, 482, 483,

488, 489, 491, 493, 520, 521, 523, 527
P. pigotti 391, 406, 440, 468*, 468, 469, 478, 481,

483, 489, 494*, 502, 520
P. simonsi 391, 467, 496, 520, 527 /

P. stromeri 391, 406, 478, 481, 489, 502, 520 ^

P. s. hopwoodi 468*

P. s. .stromeri 468*

Pectinator 490

P. specki 490

Pedete.s 482, 487, 489, 490, 504, 505

P. caffer 490

Pedetidae 469*, 481, 513, 525

Pentalophodonta 492

Perimyidae 512

Perimy.s 482, 483

Perimys sp. 483

Perissodactyla 506

Petrokozlovia 515

Petromuridae 391, 521, 523, 527

Petromurinae 521, 523, 527

Petromus 403, 406, 472, 473, 477, 480, 481, 483, 484,

490, 497, 507, 520, 521, 523, 527
P. typicus 490
Petromyidae 521, 523
Phiocricetomyinae 521, 523, 528
Phiocricetomys 496, 521, 523, 528
P. minutus 465*

Phiomorpha 378, 475, 489, 498, 519, 521, 522, 527
Phiomyidae 465, 491, 495, 498
Phiomyinae 521, 523, 527
Phiomyoidae 522
Phiomvoi'des 522
Phiomy aides 521, 523, 528
Phiomys 442, 465*, 491, 492, 493, 495, 496, 497,

498, 499, 503, 521, 523, 527
P. andrewsi 420, 426, 465*, 494*, 496, 497, 498, 521
P. cf. andrewsi 468*
P. lavacati 465*, 496
P. paraphiamyoides 465*, 466, 496
P. aff. paraphiamyoides 465*
Phiomys sp. indet. 465*, 499
Plagiodontia 507
Plagiodontiinae 523
Plagiodontinae 392*, 393, 395, 523
Platypittamys 374, 375, 381, 382, 385, 386, 388, 398,

403, 410, 411, 414, 421, 428, 429, 432, 448, 459,

467, 477, 480, 481, 487, 488, 489, 493, 495, 496,

497, 498, 508, 512
P. brachyadon 461*
Platyrrhini 473
Plesiarctamys 439
Polyplax 505

Pomonomys 521, 523, 528
P. duhius 494*, 503
Potamarchinae 448
Primates 466*, 470*
Proborhyaena 448
Proboscidea 466*, 470*
Proechimys 395, 406, 459, 483, 507, 510
P. a. albispinus 464*
P. guyannensis arescens 462*, 464*
P. iheringi denigratus 464*
Proheliaphabius 516, 522
P. leakeyi 468*
Prolapsus 458, 476, 481, 504, 505, 507, 513, 514,

522, 524
Propachyruchos 448

Oligocene Rodents of Bolivia • Patterson and Wood 543

Prospaniumys 391, 392*, 392, 393, 500

P. priscus 387*, 501*, 501

Protacaremys 391, 392*, 392, 393, 529

P. prior 387*, 501*, 501

Protadelphomys 391, 392*, 392

Protar.somys macinnesi 468*

Protoptvchidae 458, 522

Protoptychus 403, 428, 429, 458, 478, 481, 487, 488,

495, 504, 538
Protosteiromys 421, 467, 491, 493, 496, 498, 512
P. asmodeophilus 448, 461*
P. medianus 376*, 448, 461*
Protrogomorpha 514, 515
Pseudocylindrodon 455, 513
P. texanus 491
P.seudoheligmosomum 472
Pseudomyinae 470, 526
Pseudosciuridae 492
Pseudotomus 406
Pseudotsaganomys 516
Pteromys 487
Pudicci 472
Pudicinae 472, 473
Raparnys 481, 513
Riittus 464*
Reithroparamvidae 382, 398, 435, 437, 450, 453,

479, 514, 522,527
Reithroparamyinae 513, 514
Reithroparamys 409, 453, 477, 478, 487, 488, 498,

509, 514, 538
R. delicatissimus 477, 479, 481
R. huerfanensis 411, 514
Rhipidimys masticalis cearanns 462*, 464*
Rodentia 416, 453, 466*, 470*, 472, 473, 476, 508,

518, 522
Sallamys 374, 377, 382-390, 391, 392*, 392, 393,

403, 421, 439, 445, 448, 467, 481, 493, 495, 496,

497, 498, 500, 502, 512, 524
S. pascuali 379, 383, 383-390, 384*, 387*, 389*,

396, 461*, 500, 501*, 501, 528
Saykanomys 515
Schist oniys 411
S. erro 499
schurmanni 427
Sciamys 403,416, 500
S. principalis 391
S. tenuissimus 529
Sciuravidae 454
Sciuravus 403
S. nitidus 479
Sciuridae 469*, 525
Sciuridae gen. et sp. indet. 468*
Sciurognathi 485
Sciurus 409
S. niger 490

Sciurus (Guerlitiguetus) sp. 462*
Scleromys 403, 426, 427, 428, 429
Sclcromys? 427
Scotcimys 374, 421, 422, 424, 425, 448, 467, 493,

495,496,498,511, 512
S. antiquus 461*
Scotamys sp. 461*

Sepulkomys 516

Sicista 482

Simunimys 493, 502, 521, 523, 527

S. genovefae 468*

Simplimus 427, 445

S. colombianus 427

S. indivisus 421, 538

?SimpliTnus 445, 448

?S. angustus 446*

?S. schurmanni 446*

Siphonaptera 505

Sirenia 466*

Spalacopus 481, 483

Spa/ax 479, 507

Spaniomys 391, 392*, 392, 393, 394, 501

Stichomys 393*, 501

S. constans 529

S. diminutus 529

S. gracilis 529

S. regularis 499

Stilestrongylus 472

Tamias 487

Thecurus 481

Theridomyidae 449, 492

Theridomyoidea 460, 491, 492, 513

Theridomys 503

Theridomys-Trechomysplan 491, 492

Thisbemys 406

T. corrugatus 480

Thryonomyidae 391, 520, 521, 523, 527

Thr>'onomyinae 520

Thryonomyoidea 378, 449, 450, 452*, 453, 467,

468*, 468, 469, 473, 475, 477, 481, 487, 489, 492,

493, 496, 497, 499, 505, 509, 518, 519, 520, 521,

523, 526, 527
Thryonomyoidea indet. 468*
Thryonomys 406, 472, 473, 477, 480, 481, 483, 484,

485, 486, 487, 488, 489, 490, 491, 497, 503, 505,

520, 523, 527
T. harrisoni 490, 491
Trichodectidae 506
Trichostrongylidae 472, 473
Trichys 411, 481
Trimenopomidae 505
Tsaganomyidae 449, 481, 513, 514, 515, 516, 517,

518, 522, 527
Tsaganomyinae 522
Tsaganomys 455, 476, 478, 515, 516
Tvpotheria 466*
Vianella 472
Vianniinae 472

Villarroelomys 374, 377, 431, 444
V. holivianus 431, 444*, 444, 500, 501*, 528
"Villarroelomys" 511
Vulcatiisciurus africanus 468*
Wiedomys pyrrhorhinus 462*, 464*
Xylechimys 385, 391, 393*, 448, 467, 493, 496, 498,

512
X. ohliquus 461*
Yuomys 455

Zenkerella wintoni 468*
Zygodontomys lasiurus pixuna 462*, 462, 464*

IP^

ij

!^ 0

DATE DUE

c^

-J

f^ ^ 7-/.^

7^

Jjnv r ..

i^ IflflT

'^^=7ml

Printed
In USA

HIGHSMITH #45230

'L

1 ^r-r,^