The auditory region of Reithroparamys delicatissimus (Mammalia, Rodentia) and its systematic implications

Novitate MUSEUM

vitates

PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY

CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y.

Number 2972, 35 pp., 7 figs., 3 tables

10024
June 26, 1990

The Auditory Region of
Reithroparamys delicatissimus
(Mammalia, Rodentia) and its

Systematic Implications

JIN MENG'

ABSTRACT

The significance and phylogenetic position of
the rodent subfamily Reithroparamyinae is con-
troversial. This taxon has been included variously
in Ischyromyidae (Paramyidae, infraorder Protro-
gomorpha) or in the infraorder Franimorpha. Its
placement in the Franimorpha was based on an
interpretation of the mandible as incipiently hys-
tricognathous; reithroparamyines were therefore
regarded by some workers as at least part of the
ancestral stock of the Hystricognathi. Others con-
sidered the Reithroparamyinae to be the most
primitive ischyromyids, giving rise to later North
American forms. Reithroparamys delicatissimus
discussed herein presents several derived features
of the auditory region. These include: (1) inflated

bullae with internal septa, (2) apparent loss of the
promontory artery, (3) the internal carotid artery
crossing over the fenestra rotunda, (4) the presence
of a meato-cochlear bridge, (5) a somewhat swol-
len promontorium, (6) bony tube for the stapedial
artery and facial nerve partially formed, (7) fe-
nestra ovalis large and tilted, and (8) epitympanic
recess dorsally expanded. Analysis of these de-
rived characters allows the ancestor-descendant
relationship between reithroparamyines and hys-
tricognathous rodents to be rejected. For the same
reason, Reithroparamyines do not represent the
most primitive ischyromyids; instead, a close re-
lationship of reithroparamyines with sciurids,
aplodontids, and glirids is proposed.

INTRODUCTION

The genus Reithroparamys was first pro-
posed in 1920 by W. D. Matthew. Although
Matthew designated AMNH 12561 the ge-
notype of Reithroparamys, the type was in

fact Paramys delicatissimus (Leidy 1871)
(Wood, 1962; Korth, 1984). Wood (1962)
based a new subfamily, the Reithroparamyi-
nae, on the genus and placed it in the family

' Graduate Student, Department of Vertebrate Paleontology, American Museum of Natural History; Department

of Geology, Columbia University, New York, N.Y. 10027.

Copyright © American Museum of Natural History 1990

ISSN 0003-0082 / Price $3.75

Paramyidae, superfamily Ischyromyoidea.
Later, Black (1971) grouped the Reithropara-
myinae, Ischyromyinae, Paramyinae and
Prosciurinae in the family Ischyromyidae AI-
ston (1876). Wood (1975) proposed a new
infraorder, the Franimorpha, under the
suborder Hystricognathi, and placed the Rei-
throparamyinae within it. Subsequent au-
thors have either retained the Reithropara-
myinae as a subfamily of the Paramyidae or
Ischyromyidae (e.g., Korth, 1984) under the
suborder Protrogomorpha or Scituromorpha,
or placed it as a subfamily or family in the
infraorder Franimorpha under Hystricogna-
thi (e.g., Chaline and Mein, 1979; Patterson
and Wood, 1982).

It has long been argued that the North
American protrogomorphs are the most
primitive rodents. Their relationships with
later groups remain problematic (e.g., Har-
tenberger, 1980: fig. 2; Wood, 1985: fig. 1),
because few if any derive characters are known
for these rodents. Among the North Ameri-
can protrogomorphous rodents, reithropara-
myines play a crucial role in the reconstruc-
tion of rodent phylogeny. They are central to
the debate on fundamental problems in the
intraordinal relationships among rodents,
such as the definition, composition, and or-
igin of the Hystricognathi and the early di-
vergence of rodents in North America and
Asia. The controversial systematic position
of the reithroparamyines has been more or
less a direct result of divergent opinions about
their mandible structure.

A major problem in the origin of hystri-
cognaths is finding a temporal and geographic
link between the Early Oligocene hystrico-
gnaths of Africa and South America, and the
earliest Hystricidae from southern Asian
Middle Miocene (Flynn et al., 1986). Wood
(1974) presumed that Eocene reithropara-
myines present in North America, and pre-
sumably present in Asia, provide the linkage,
while Lavocat (1980) endorsed direct trans-
oceanic dispersal of hystricognaths from Af-
rica to South America. Recently, Korth (1984)
proposed a hypothesis that reithropara-
myines are the most primitive protrogomor-
phous stock in North America, and that they
give rise to all other ischyromyids and later
groups. Auditory features of Reithroparamys

AMERICAN MUSEUM NOVITATES

NO. 2972

delicatissimus described in this report present
evidence that sheds new light on these prob-
lems.

Reithroparamys delicatissimus (AMNH
12561), the best and most completely pre-
served specimen of the subfamily, has been
studied by many authors (most extensively
by Wood, 1962). No agreement has been
reached about the nature of the angular pro-
cess of the mandible, and therefore its sys-
tematic position remains uncertain. Further
preparation of this specimen, especially its
ear region, has revealed morphological attri-
butes previously unknown in Reithropara-
mys and known in few other paramyids.

ACKNOWLEDGMENTS

I am grateful to Dr. M. C. McKenna
(American Museum of Natural History) for
suggesting that I study this subject; he en-
couraged me throughout this study and in-
cisively criticized an earlier version of this
paper. I am also grateful to Dr. M. J. Novacek
(AMNH) for patiently introducing mam-
malian auditory structures to me and for per-
mitting me to study the specimens under his
care. For further critical and detailed reading
I thank Drs. P. Luckett (University Puerto
Rico), L. Flynn (Peabody Museum), A. Wyss
(AMNH) and M. Norell (AMNH); this paper
is greatly improved and more readable with
their valuable comments. I profited enor-
mously from discussion with Dr. J. H. Wah-
lert (AMNH) on rodent anatomy and system-
atics. I have also benefited greatly from
discussions with J. R. Wible (University Chi-
cago) on the internal carotid complex in var-
ious groups of mammals. Thanks are given
to Dr. C.-k. Li (Institute of Vertebrate Pa-
leontology and Paleoanthropology, Beijing)
for permitting me access to specimens under
his care. J. Shumsky (AMNH) is gratefully
acknowledged for ably preparing the speci-
men described here. I wish to express my
deep appreciation to many people at the De-
partment of Vertebrate Paleontology,
AMNBH, for their help in various ways. Fi-
nally, I thank my wife, Yu Liu, for her assis-
tance in preparation of manuscripts. This
work was supported by a faculty fellowship
from Columbia University and by the Frick
Laboratory Endowment Fund, AMNH.

1990

11cm

Fig. 1.

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 3

‘S
Ge ES
ek Pa ET a ah ity 4,
he ~—
Sgt
NG

X

JF

Ventral view of the basicranial region of Reithroparamys delicatissimus (AMNH 12561). B,

bulla; BO, basioccipital; BS, basisphenoid; EAM, external auditory meatus; EPER, ectotympanic part
of the epitympanic recess; FM, foramen magnum; FR, fenestra rotunda; FSM, fossa for stapedius muscle;
HF, hypoglossal foramen; I, incus; JF, jugular foramen; MP, mastoid process; OC, occipital condyle;
PPER, petrosal part of the epitympanic recess; SMF, stylomastoid foramen.

DESCRIPTION OF
THE EAR REGION OF
REITHROPARAMYS
DELICATISSIMUS

Reithroparamys delicatissimus (AMNH,
12561) is represented by an almost complete
skull, articulated mandibles, and some post-
cranial skeleton (see Wood, 1962). The bullae
on both sides of the skull are preserved al-
though both are broken at their posterior end
(fig. 1). The bullae are inflated and slightly
flat lateroventrally, completely covering the
tympanic cavity. The ectotympanic does not
expand laterally to form an external auditory
canal. The suture between the bulla and the
surrounding cranial elements is distinct, in-
dicating that the bulla is formed entirely by
the ectotympanic. The bulla is tightly joined
but not yet fused to the petrosal. The bullar
wall is thin but gradually thickens toward the
external auditory meatus. Internally a verti-
cal bony septum is present at the anterome-
dial corner.

On the lateral side of the bulla, the contact
between the ectotympanic and the squamosal
is limited to the posterodorsal margin of the
bulla. The squamosal is completely excluded

from the tympanic cavity by the ectotym-
panic. A marked imprint on the ventral mar-
gin of the squamosal indicates the overlap-
ping of the dorsally expanded ectotympanic
above the external auditory meatus (fig. 2).
The anterolateral part of the bulla is sepa-
rated from the squamosal by a narrow band
of the petrosal (fig. 2). The medial side of the
bulla is lacking foramina, fissures, or grooves
that might be interpreted as an entrance or
pathway for a medially placed internal ca-
rotid artery or even venous drainage. At the
posteromedial end of the bulla, immediately
ventral to the jugular foramen, is a foramen
which was identified as the stapedial foramen
by Wahlert (1974) because of its posterior
position, but is herein referred to as the in-
ternal carotid foramen (see discussion be-
low). The jugular and carotid foramina are
separated by a narrowly exposed ridge of the
petrosal. The carotid foramen is emarginated
ventrally by the ectotympanic and dorsally
by the petrosal. Breakage at the posterior end
of the bulla has permitted removal of most
of the matrix filling the tympanic cavity.
The internal carotid artery enters the tym-
panic cavity through the carotid foramen, and

SAT PPER

TTF

AMERICAN MUSEUM NOVITATES

NO. 2972

5mm

MCB FR HE

Fig. 2. Lateral (slightly posteroventral) view of the basicranial region, emphasizing the auditory
region, of R. delicatissimus. B, bulla; BO, basioccipital; PPER, petrosal part of the epitympanic recess;
FO, fenestra ovalis; FR, fenestra rotunda; FSM, fossa for the stapedius muscle; HF, hypoglossal foramen;
IB, imprint left by dorsal extension of the bulla; JF, jugular foramen; M, mastoid; MCB, meato-cochlear
bridge (only base preserved); MP, mastoid process; OC, occipital condyle; P, petrosal; PF, postglenoid
foramen; PM, promontorium; SAG, groove left by the stapedial artery; SAT, tube for stapedial artery;
SMF, stylomastoid foramen; SQ, squamosal; TTF, fossa for tensor tympani muscle.

runs in an open groove until it crosses the
fenestra ovalis. The large size of the carotid
foramen and associated groove suggest that
the internal carotid and stapedial artery also
may have been fairly large. The promonto-
rium is somewhat inflated. There is no groove
or any other trace on the surface of the pro-
montorium to indicate the existence ofa pro-
montory artery. The fenestra rotundum is
small and nearly circular. It is located at the
posterior end of the promontorium and faces
posteroventrally. The internal carotid artery
runs laterally along the posterior end of the
promontorium, shielding most ofthe fenestra
rotundum. After passing the fenestra rotun-
dum, the internal carotid artery continues as

the stapedial artery, while the promontory
artery 1s assumed to have been lost (see char-
acter analysis section for more details). One
of the most distinctive features of the tym-
panic cavity 1s the pathway of the stapedial
artery. Lateral to the fenestra rotundum, a
marked trough for the stapedial artery leads
to the fenestra ovalis. In mammals where it
is present, the stapedial artery normally trav-
els along the ventral rim of the fenestra ro-
tundum and approaches the fenestra ovalis
from its ventral side. In this case, the long
axis of the fenestra ovalis is usually oriented
horizontally. The stapedial artery in R. de-
licatissimus, however, approaches the fenes-
tra ovalis from its posterior rim. Therefore,

1990

the fenestra ovalis is somewhat tilted so that
the stapedial artery can easily run through
the stapes so oriented. The fenestra ovalis is
relatively large and its rim forms a slightly
raised lip.

Another distinctive feature of the tympan-
ic cavity is a prominent osseous process de-
veloped on the promontorium at the ventral
rim of the fenestra ovalis. This process oc-
cupies the normal pathway of the stapedial
artery. In the left tympanic cavity, this pro-
cess is broken at its base, but on the right side
it is a distinctly cone-shaped structure pro-
jecting posterolaterally. Due to breakage, it
does not touch any other structure but in life
the distal end of this process probably con-
tacted the ectotympanic at the posterior end
of the external auditory meatus. This struc-
ture is here referred to as the meato-cochlear
bridge. The stapedial artery runs posterodor-
sal to this bridge. In ventral view, the fenestra
ovalis is concealed by the osseous process. A
small process posterior to the fenestra ovalis,
but dorsal to the groove for the stapedial ar-
tery, is also developed. The stapedial artery
is thus confined by the meato-cochlear bridge
ventrally and this process dorsally before it
passes over the fenestra ovalis.

After crossing the fenestra ovalis, the sta-
pedial artery enters a bony canal. This canal
is formed by the petrosal and is excavated
ventrally in the tympanic roof. Its ventral
side is cracked, indicating that either the ca-

‘nal was not completely formed in life or was
broken in preservation. The canal extends
anterolaterally, and the stapedial artery ex-
ited the tympanic cavity at its anterolateral
corner. There is no indication of the bifur-
cation of the superior and inferior ramus of
the stapedial artery within the tympanic cav-
ity. Posterior to the fenestra ovalis, the lateral
wall of this canal continues posteriorly to the
stylomastoid foramen; the medial wall of the
canal is incomplete so that it opens medially
into the fossa assumed for the stapedius mus-
cle. Thus, the facial nerve is exposed medially
into the tympanic cavity after crossing the
stapedial artery dorsally. In the right tym-
panic cavity, the bony tube for stapedial ar-
tery and facial nerve is not preserved; the
tympanic aperture of the facial nerve is vis-
ible at the anterior end of the fenestra ovalis;
and a groove originating from it runs poste-

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 5

riorly to the stylomastoid foramen. At the
exit of the facial nerve on the right, the ec-
totympanic curves inward and covers the
ventromedial side of the facial nerve. Because
most of the posterior region of the bulla is
broken, it is difficult to estimate the extent
to which the facial nerve was covered by the
ectotympanic. The posterior portion of the
path for the facial nerve may have consisted
of a bony tube formed by the petrosal later-
ally and the ectotympanic medially.

Posterior to the fenestra ovalis, a deep fos-
sa, posteriorly bounded by a tubercle of the
petrosal, is the fossa assumed for the stape-
dius muscle. In life this fossa would have
been covered ventrally by the bulla, and the
stapedius muscle would have been complete-
ly enclosed in the tympanic cavity. Posterior
to the stapedius fossa, the mastoid process of
the petrosal is small and not horizontally or
ventrally expanded.

The epitympanic recess in R. delicatissi-
mus, if examined through the external au-
ditory meatus, lies between the tegmen tym-
pani of the petrosal and the ectotympanic.
The petrosal portion of the epitympanic re-
cess is located on the lateral surface of the
ventrally projecting tegmen tympani of the
petrosal and is well separated from the fe-
nestra ovalis. Immediately lateral to the pe-
trosal portion of the epitympanic recess is the
expanded ectotympanic portion of the epi-
tympanic recess, which forms a distinct el-
liptical fossa in the ectotympanic. This fossa
is situated dorsal to the external auditory
meatus but is not distinctly separated from -
it.

Lateral to the stapedial canal and antero-
dorsal to the epitympanic recess, there is a
deep depression roofed by the petrosal and
medially bounded by the stapedial artery tube.
This depression probably implies a dorsal ex-
pansion of the epitympanic recess. Because
of this depression, the bony tube for the sta-
pedial artery appears even more prominent.
A tiny bone, possibly the displaced incus, lies
within it. A broad depression assumed for
the insertion of the tensor tympani muscle
lies at the anterolateral side of the promon-
torium and medial to the stapedial tube. An-
terior to this depression is a breakage which
cannot be ascertained to be the piriform fe-
nestra.

CHARACTER ANALYSIS

In this section, attention is given to the
determination of character polarity. Focus
centers on the derived auditory characters
found in early rodents in reconstructing their
phylogeny. Whether a character is primitive
or derived depends on the systematic level
being considered. Because R. delicatissimus
has been regarded either as a member of the
most primitive ischyromyids (Korth, 1984),
or in the ancestral stock of the Hystricognathi
(Wood, 1985), its derived auditory features
are of special significance at the intraordinal
level of rodents. In the following discussion,
the Reithroparamys auditory region is com-
pared with those of Paramys, Sciuravus, Co-
comys, and the Theridomorpha. The ear re-
gions of these groups are well known and
generally accepted as representing the prim-
itive morphotypes in rodents (Lavocat and
Parent, 1985; Li et al., 1989). Comparison is
also made between Reithroparamys and mu-
roids, especially when characters are not
available in early fossil forms due to frag-
mentary materials. This is because muroids
are thought to have the most primitive au-
ditory region among living rodents (Lavocat
and Parent, 1985). The primitiveness of the
auditory characters in the groups mentioned
above is obtained through outgroup com-
parison (for instance, Cocomys, Lietal., 1989).
The character polarity of Reithroparamys is
basically determined by comparison with
these rodent groups. However, when the po-
larity cannot be ascertained on comparison
with these groups, or when the character states
are unclear in these groups, other mamma-
lian groups, especially those that may be
closely related to rodents such as lagomorphs
and macroscelidids (Novacek, 1985; Nova-
cek and Wyss, 1986; Novacek et al. 1988) or
primitive eutherians such as Leptictis, are
employed to determine polarity.

ECTOTYMPANIC BULLA INFLATED WITH
INTERNAL SEPTUM. Novacek (1977, 1980)
pointed out that the primitive condition of
the auditory bulla in eutherians was probably
one similar to that of monotremes; there the
bony bulla is not present and the ectotym-
panic rests at a low angle to the horizontal
plane of the skull. An auditory bulla formed
completely by the ectotympanic is present in

AMERICAN MUSEUM NOVITATES

NO. 2972

Rodentia and Lagomorpha. It is widely agreed
that an ectotympanic bulla represents a de-
rived condition in eutherians. In rodents a
small ectotympanic bulla loosely attached to
the petrosal, such as occurs in Sciuravus and
Cocomys, is taken as the primitive condition.

In early rodents, an osseous bulla is not
commonly found associated with the skull.
An osseous bulla has not yet been found in
Eocene ischyromyids except Reithroparamys
(Korth, 1984). Sciuravus may have a poorly
developed bulla, preserved in one specimen
(USNM 22477) and illustrated in a recon-
struction by Dawson (1961) and Wahlert
(1974). The bulla of Sciuravus must have been
loosely attached to the petrosal because no
markings on the petrosal are detected. The
bulla was not inflated and was completely
confined to the ventral side of the skull. It
even failed to cover the middle lacerate fo-
ramen. Cocomys, from the early Eocene of
South China, displays a similar bullar con-
dition (Li et al., 1989). A bulla co-ossified
with the skull, like that of the R. delicatissi-
mus, has been regarded as a derived condi-
tion in rodents. In R. delicatissimus, how-
ever, the bulla is not only co-ossified with the
skull but also inflated to a considerable de-
gree. The squamosal of R. delicatissimus is
blocked, in ventral view, by the inflated bulla
(fig. 1); the bulla was also tightly attached,
though not completely fused, to the petrosal.

‘Additionally, a vertical septum is present in

the bulla of R. delicatissimus; this is definitely
a derived condition. This condition is very
much like that of sciurids, where a few septa
are present in the bulla and one of them is
always located at its anteromedial corner
within the bulla.

A bony bulla is assumed to have been pres-
ent in at least some paramyines (Wood, 1962).
If such a bulla was present, it was probably
similar to that of Sciuravus and Cocomys,
and more primitive than that of R. delicatis-
simus. This may be inferred on two grounds.
When viewing the lateral side of the R. de-
licatissimus skull, one can see a distinct im-
print on the ventral margin of the squamosal
(fig. 2), that accommodates the dorsal expan-
sion of the ectotympanic. Such an imprint
was not observed on any available skulls of
Paramys. This implies that the lateral wall
of the bulla in Paramys, if indeed the bulla

1990

was present, must have been entirely con-
fined to the ventral side the skull.

In addition, Paramys (AMNH 12508, for
example) has a distinct tympanohyal, or a
mastoid tubercle (which may be represented
in part by an outgrowth of the mastoid and
may fuse with the tympanohyal) (Novacek,
1986). The tympanohyal is the most cranial
element of the hyoid bar, which often fuses
to the petrosal in late developmental stages
(van der Klaauw, 1931; De Beer, 1937;
McDowell, 1958; MacPhee, 1981). Similar
to the condition in Lepftictis, there is a round-
ed fossa on the ventral surface of the tym-
panohyal in Paramys. A prominent tympa-
nohyal is generally associated with a poorly
developed bulla, as in marsupials, Leptictis,
creodonts, and primitive carnivorans. In ro-
dents where the bulla is considerably devel-
oped, the tympanohyal, or even the mastoid
process where the tympanohyal is attached,
fuses with the bullar wall so that a distinct
tympanohyal is not usually present. If Para-
mys had a bulla, it was most likely very small
and loosely attached to the petrosal. Lavocat
and Parent (1985) claimed that the most
primitive rodent auditory region is that of
the Theridomorpha. Theridomorphs have a
somewhat inflated bulla with the external au-
ditory meatus slightly projected laterally
(Lavocat, 1967, pl. 2). This bulla is clearly
more derived than that of Sciuravus and Co-
comys, and possibly that of paramyines.

Loss OF THE PROMONTORY ARTERY (PA).
The classic consideration of the primitive
morphotype of the internal carotid artery
(ICA) in eutherians is that the ICA system
has two main trunks: the medial ICA courses
medially to the tympanic cavity while the
lateral ICA runs along the ventral surface of
the promontorium and later gives off the sta-
pedial artery (Matthew, 1909; Van Valen,
1965, 1966; McKenna, 1966; Szalay, 1975;
Archibald, 1977; Parent, 1980, 1983). An in-
ternal carotid artery and stapedial artery can
be traced back in various other tetrapods
(Goodrich, 1930; Romer 1956; Romer et al.,
1977). An alternative hypothesis about the
ICA system has been proposed recently by
several workers (Presley, 1979; Cartmill and
MacPhee, 1980; MacPhee, 1981; Wible,
1983, 1986, 1987). Based on the observation
of the auditory ontogeny and comparative

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 7

anatomy of living mammals, it is argued that
there is only a single ICA trunk in all fossil
and living mammals, and that the single ICA
simply takes different positions in the tym-
panic region in different groups of mammals
(fig. 3). The artery may be either in a medial
or lateral position; there is no example in
extant mammals showing the simultaneous
presence of the ICA in both positions. This
interpretion is followed in this paper, even
though the more traditional position was tak-
en by Lavocat and Parent (1985).

The terminology of the internal carotid ar-
tery system needs to be clarified because dif-
ferent terms have been used in the literature.
Besides the usage in the two-trunk model of
Matthew and others (see above), the internal
carotid artery is regarded as the portion that
terminates at the point where it bifurcates
into stapedial and promontory arteries (No-
vacek, 1986). This portion is taken as equiv-
alent to the lateral internal carotid artery by
MacPhee (1981). In the usage of others (e.g.,
Bugge, 1985; Wible, 1986), the internal ca-
rotid artery includes the portion before the
stapedial-promontory bifurcation and the
promontory artery. In this paper (fig. 3), in-
ternal carotid artery system (ICAS) is used to
indicate the proximal internal carotid artery,
after it stems off the external carotid artery,
and all its terminal branches including the
stapedial artery. Proximal internal carotid ar-
tery (PICA) is used to indicate the portion
before the stapedial-promontory bifurcation,
while distal internal carotid artery (DICA) is
used for the portion beyond the bifurcation
point, no matter which position it is in. When
DICA is in the promontory position, it is
referred to as promontory artery (PA); when
it is in the medial position, it is referred to
as medial distal internal carotid artery (MDI-
CA).

It is widely agreed that the reduction of the
ICA system is a general evolutionary tenden-
cy in eutherians, apparently occurring inde-
pendently, in various lineages. Similarly, it
is a common evolutionary pattern in euthe-
rian mammals that the stapedial artery was
reduced and then lost while the MDICA or
PICA may still exist. The final stage of this
hypothesized transformation series is the
complete loss of the ICA system, and its func-
tional replacement by the external carotid

8 AMERICAN MUSEUM NOVITATES

NO. 2972

EC PA
P sQ
PP SA
EOFN EAM
BP FO
FR ER
PICA FN
FSM
A

MDICA

SAT

FNC

FOT

MCB

Fig. 3. Block diagrams showing character states in the ventral view of the tympanic cavity. A,
Primitive eutherian condition; B, muroid condition; C, reithroparamyine condition. See text for details.
Abbreviations: BP, branching point of SA and DICA; PA, promontory artery; EAM, external auditory
meatus; EC, ectotympanic; EOFN, external opening of the facial nerve; ER, epitympanic recess; FN,
facial nerve; FNC, facial nerve canal; FO, fenestra ovalis; FOT, fenestra ovalis tilted; FR, fenestra
rotunda; FSM, fossa for stapedius muscle; MCB, meato-cochlear bridge; MDICA, medial distal internal
carotid artery; P, petrosal; PICA, proximal internal carotid artery; PP, promontorium proper; SA,
stapedial artery; SAT, tube for stapedial artery; SQ, squamosal.

system (e.g., Carnivora, Artiodactyla, and
some Primates). Persistence of a functional
stapedial artery while the promontory artery
is completely absent is a rare occurrence in
eutherians, recorded in some microchirop-
terans (Buchanan and Arata, 1969) and some
rodents. A similar situation may (Szalay,
1975; MacPhee, 1981) or may not occur
(Conroy and Wible, 1978) in lemurs.

In living rodents, muroids are the group
that presents the most primitive ICA pattern,
i.e., a MDICA and stapedial artery are both
present (Bugge, 1974b, 1985) although the
bifurcation of the DICA and stapedial artery
is medial to the bulla (Goodrich, 1930; Wi-

ble, 1987), a rare condition in eutherians (fig.
3B). Among early rodents, sciuravids surely
had a promontory and stapedial artery (Wah-
lert, 1974; personal obs. on AMNH 11614).
Theridomorpha and some paramyines may
have both arteries (Lavocat and Parent, 1985),
but the promontory artery has not been clear-
ly observed in Paramys. No matter which
position the DICA is in, the condition in
which both the distal internal carotid artery
and the stapedial artery are present is prim-
itive relative to the condition in which either
is lost.

It may be asked from which condition, sci-
uravid-like or muroid-like, the R. delicatis-

1990

simus condition was derived? If derived from
the muroid condition, then the DICA would
have been lost as a MDICA and the stapedial
artery within the tympanic cavity would be
homogeneous. If derived from the sciuravid
condition, the vessel within the tympanic
cavity is not completely the stapedial artery
because its proximal end is the terminal end
ofthe proximal internal carotid artery. It may
further be asked which position of the inter-
nal carotid artery, PA or MDICA, is primi-
tive for rodents? In answering this question,
the single ICA hypothesis encounters some
difficulty. Recently, Wible (1986) has at-
tempted to clarify this problem. Several lines
of evidence suggest that a promontory artery
(i.e., an artery in the transpromontoral po-
sition) is more primitive than a medially
placed distal internal carotid artery for pla-
centals. Other workers argued a contrary view
(e.g., Presley, 1979; Novacek, 1985, 1986).
Wible’s hypothesis is adopted here. The
promontory artery is probably present in ear-
ly rodents such as Sciuravus, Cocomys (Li et
al., 1989), and possibly Paramys. The prom-
ontory artery is also present in macroscel-
ids and possibly in early lagomorphs. Based
on the present evidence, it seems likely that
the promontory artery is primitive in rodents
and the reithroparamyine condition is de-
rived from an ICA system which had the dis-
tal internal carotid artery in the promontory
position. Therefore, the proximal portion of
the vessel in the tympanic cavity is part of
the proximal internal carotid artery (fig. 3).
In other words, the vessel in the tympanic
cavity is not completely homologous to a sta-
pedial artery and the foramen through which
the artery enters the tympanic cavity is more
precisely called the carotid foramen instead
of the stapedial foramen.

Interestingly, it has been observed with in
sciuroids (Tandler, 1899; Bugge, 1971b,
1974b; Wible, 1984) and gliroids (Bugge,
1971b, 1974b) the internal carotid artery does
enter the tympanic cavity, as evidenced by
its accompanying internal carotid nerve (Wi-
ble, 1984). In front of the cochlear fossula,
the artery and nerve diverge: the artery (now
the proximal stapedial artery) runs laterally
toward the fenestra ovalis, and the nerve runs
forward. Wible (1984) pointed out that Guth-
rie (1963) mistakenly identified the sciuroid

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 9

and gliroid internal carotid as the stapedial
artery and the posterior carotid foramen as
a stapedial foramen.

An extratympanic bifurcation of the DICA
and stapedial artery has been poorly dem-
onstrated (Ting and Li, 1984; Li and Ting,
1985) in the eurymylid Rhombomylus. The
authors (Ting and Li, 1984: 98) stated: “‘the
stapedial artery, after stemming from the ex-
ternal carotid artery, enters the tympanic
cavity through the stapedial foramen at the
posteromedial corner of the bulla.” and “‘The
ICA, after stemming from the external ca-
rotid artery, enters the cranial cavity through
a fissure-like foramen at the anteromedial
corner of the bulla (originally in Chinese).”
The systematic implications of this feature
are profound. In placental mammals, an ex-
tratympanic bifurcation of the DICA and sta-
pedial artery is found only in some muroids
(Wible, 1987) and probably in Heteromyidae
(Bugge, 1974b, 1985; Lavocat and Parent,
1985). This means that these rodents may
have evolved from an ancestor that had an
ICA pattern similar to that of Rhombomylus.
If this interpretation is further confirmed, it
may serve as additional evidence for eury-
mylid-rodent relationship. Nevertheless, the
condition as described in Rhombomylus by
Ting and Li (1984, 1985) is questioned be-
cause the “internal carotid canal’’ seems too
anteriorly located and too small to be a ca-
rotid foramen.

In the skull of Paramys, there is a very large
canal between the petrosal and the basioc-
cipital, that was interpreted by Lavocat and
Parent (1985) as a possible course for the
medial internal carotid artery. Moreover,
none of the specimens of Paramys in the col-
lection of the American Museum of Natural
History displays a groove on the promon-
torium obviously branching from the groove
for the stapedial artery to indicate the pres-
ence of the promontory artery. A medial ca-
nal is not present in R. delicatissimus. An
internal carotid artery located between the
petrosal and the basioccipital has never been
recorded (Wible, 1983, 1984, personal com-
mun.). The medial canal in Paramys likely
housed the inferior petrosal sinus. It is pos-
sible that the promontory artery in Paramys
was reduced or lost.

THE INTERNAL CAROTID ARTERY SHIELD-

10 AMERICAN MUSEUM NOVITATES

ING THE FENESTRA ROTUNDA. An internal ca-
rotid artery with the stapedial artery has been
reported in many groups of living and fossil
mammals (Wible, 1987). Its shielding the fe-
nestra rotunda is uncommon. The common
and primitive condition is that the proximal
internal carotid artery enters the tympanic
cavity through the carotid foramen and then
bifurcates into the promontory and stapedial
arteries on the ventral surface of the prom-
ontorium. The promontory artery travels
anteriorly while the stapedial artery runs lat-
erally along the ventral rim of the fenestra
rotundum toward the fenestra ovalis (fig. 3A).
This pattern has been reported to be present
in a few fossil mammals (MacIntyre, 1972;
Szalay, 1975; Cifelli, 1982; Coombs and
Coombs, 1982; Novacek, 1986), and in some
early rodents (Wahlert, 1974; Parent, 1980;
Bugge, 1974a; Lavocat and Parent, 1985; Li
et al., 1989). Lavocat and Parent (1985) re-
ported the stapedial artery of aplodontids as
being lost, but in their character distribution
chart (ibid., ch. 13, fig. 4), they coded the
fenestra rotunda crossed by the stapedial ar-
tery. Nevertheless, the stapedial artery is
present in some primitive aplodontids, such
as prosciurines (Wahlert, 1974; Luckett and
Hartenberger, 1985). In Allomys nitens
(AMNH 6997) evidence of this vessel also
exists: a bony tube partially crosses the fe-
nestra rotunda, similar to the condition in
sciurids. The proximal end of the stapedial
artery is probably the distal end of the prox-
imal internal carotid artery as postulated in
this paper.

It has been reported that the lateral internal
carotid artery (=PICA in this paper) shields
the ventral part of the fenestra rotundum in
a few placental groups. This condition is re-
garded as primitive in placentals (Szalay,
1972, 1975; Archibald, 1977). Novacek
(1980, 1986), however, provided an alter-
native explanation for the shielding in certain
groups. He argued that in many groups this
shielding may result from specialized modi-
fications including the marked enlargement
of the stapedial artery or its enclosure in bony
tubes or both. In R. delicatissimus, the shield-
ing clearly results from the development of
the meato-cochlear bridge (see below), and
likely represents a derived condition. If the
condition of the internal carotid artery in R.

NO. 2972

delicatissimus is derived from a sciuravid
pattern, the shielding is at least partly formed
by the proximal internal carotid artery (fig.
3C).

THE OSSEOUS MEATO-COCHLEAR BRIDGE.
This structure is very narrowly distributed in
rodents; it is reported only in Sciuridae and
Aplodontidae (Lavocat and Parent, 1985). No
such condition has been observed in other
mammalian groups. A similar structure may
be formed by the fusion of the tympanohyal
with the mastoid tubercle in some mammals.
In the adult crania of some mammals, the
proximal end of the tympanohyal fuses to the
mastoid of the petrosal and its distal end may
extend to the promontorium, ventrally arch-
ing over the exit groove of the facial nerve.
In some cases, this process may touch or fuse
to the promontorium, to form a bridge which
looks superficially like the meato-cochlear
bridge. These two conditions are clearly not
homologous. The tympanohyal fuses with the
mastoid and touches the promontorium at
the ventral side of the fenestra rotundum as
in the insectivore Apternodus, or at the pos-
teroventral side of the fenestra ovalis as in
some creodonts and primitive carnivorans.
In all these cases, this osseous connection does
not block the stapedial artery at the ventral
side of the fenestra ovalis. The stapedial ar-
tery runs anteroventrally to the connection.
If the meato-cochlear bridge is present, how-
ever, it joins the ectotympanic at the external
auditory meatus and the stapedial artery runs
posterodorsal to it (figs. 2, 3). Cifelli (1982)
has proposed, alternatively, that in primitive
carnivorans, the tympanohyal fuses to a ven-
tromedial elongation of the squamosal. I have
reexamined several specimens of primitive
carnivorans and creodonts and found that
specimens with clear sutures show a mastoid-
tympanohyal connection instead of a squa-
mosal-tympanohyal connection.

BONY TUBE FOR FACIAL NERVE AND STA-
PEDIAL ARTERY. A bony tube for the stapedial
and promontory artery has been reported in
various groups of placentals and has been
regarded as a derived condition (Szalay, 1975;
Archibald, 1979; Novacek, 1980). The prim-
itive condition is a sulcus or groove on the
promontorium, i.e., the stapedial artery is ex-
posed ventrally into the tympanic cavity.

As described by Lavocat and Parent (1985),

1990

the stapedial artery, after passing the fenestra
ovalis, enters the facial canal and exits the
tympanic cavity by several possible routes.
In fact, the stapedial artery does not enter the
facial canal but just crosses it. In rodents where
the stapedial tube is present, the facial nerve
and the stapedial artery are enclosed in two
separate tubes (fig. 3). The stapedial artery
crosses the facial nerve ventrally in all eu-
therians, whether or not it is enclosed in a
bony tube. The facial nerve runs posteriorly
and usually parallel to the lateral side of the
promontorium while the stapedial artery runs
anterolaterally away from it. Their intersec-
tion occurs at a very short distance inside the
tympanic cavity. If the bony tubes for the
stapedial artery and facial nerve are present,
they may be confluent with each other at the
crossing point, but this does not mean the
stapedial artery enters the facial canal.

After it exits the petrosal and enters the
middle ear cavity posteriorly, the facial nerve
is covered by a thin bony lamina for a short
distance. The tensor tympani muscle usually
inserts on the ventral surface of this bony
lamina. Primitively, the course of the facial
nerve in the tympanic cavity is open. In a
more derived condition, however, the facial
nerve is completely enclosed in a bony tube.
In rodents most of this tube is formed by the
petrosal, and only its posteromedial portion
is formed by the ectotympanic. When the ec-
totympanic is well developed, it sinks dor-
sally into the tympanic cavity, and finally
covers the stapedius muscle fossa and the
ventromedial side of the facial nerve. When
the ectotympanic fuses to the petrosal, the
facial nerve is thus completely enclosed in a
bony tube.

Similarly, the primitive condition of the
stapedial artery in placentals is that the sta-
pedial artery is conveyed in a sulcus and is
thus exposed ventrally in the tympanic cav-
ity. After it crosses the fenestra ovalis and
then the facial nerve, the stapedial artery bi-
furcates into the ramus superior and ramus
inferior within the tympanic cavity. In a few
groups (rodents, lagomorphs, bats, and ele-
phant shrews) the ramus superior/inferior bi-
furcation is within the cranial cavity, which
is accepted as derived condition for Eutheria
(Wible, 1986).

In primitive rodents, such as Paramys, Sci-

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 11

uravus, Cocomys, and Theridomorpha, the
facial nerve, after coursing through the in-
ternal auditory meatus, appears in the tym-

panic cavity at an opening anterior or lateral

to the fenestra ovalis. After passing the sta-
pedial artery, the lateral side of the facial nerve
is bounded by a bony wall formed by the
petrosal until it exits the tympanic cavity
through the stylomastoid foramen. Com-
pared to the condition in R. delicatissimus,
this wall in those forms is very short. In ad-
dition, the wall in the primitive forms is much
lower so that the fenestra ovalis and the epi-
tympanic recess is not significantly separated
by it. In all these forms, then, the stapedial
artery is completely exposed through the
tympanic cavity. After passing by the fenestra
ovalis, the proximal stapedial artery imme-
diately pierces the petrosal and exits the tym-
panic cavity.

R. delicatissimus displays a more derived
condition. Initially, the stapedial artery is ex-
posed in the tympanic cavity as it courses
over the fenestra rotundum, but after passing
through the stapes it immediately runs into
a bony tube. As in sciurids, the bony tube for
the stapedial artery projects ventrally to a
remarkable degree from the ventral roof of
the tegmen tympani. The facial nerve enters
the tympanic cavity after it passes the fenes-
tra ovalis, but only its medial side is exposed.
Before it exits the tympanic cavity, the facial
nerve may well be enclosed medially by the
ectotympanic. The lateral wall for the facial
nerve projects so much that it separates the
fenestra ovalis from the epitympanic recess
as in sciurids.

ORIENTATION OF THE FENESTRA OVALIS.
The orientation of the fenestra ovalis in ro-
dents is generally classified as either vertical
or horizontal (Parent, 1980, 1983; Lavocat
and Parent, 1985). A vertical fenestra ovalis
may refer to its vertical orientation relative
to the frontal (horizontal) plane of the skull,
while a horizontal fenestra ovalis is parallel
to the frontal plane. The vertical fenestra
Ovalis is regarded as the primitive condition
(ibid.). The orientation of the fenestra ovalis
can also be described by the orientation of
its long axis. The fenestra ovalis in mammals
is more or less oval. A small and somewhat
rounded fenestra ovalis occurs in some groups
as a primitive condition (Archibald, 1979;

12 AMERICAN MUSEUM NOVITATES

Prothero, 1983). Orientation of the long axis
of the fenestra ovalis parallel to both frontal
and sagittal planes is likely to be primitive.
This condition is present in primitive euthe-
rians and most living mammals, whether or
not the fenestra ovalis is vertical or horizon-
tal. Paramys, Sciuravus, and Cocomys have
this condition. In R. delicatissimus, however,
the fenestra ovalis is somewhat tilted to the
frontal plane possibly because of the presence
of the meato-cochlear bridge and the course
of the stapedial artery approaching from its
posterior side (figs. 2, 3A). In addition, the
fenestra ovalis is larger than that in para-
myines and sciuravids. A large fenestra ovalis
is also believed to be a derived condition
(Lavocat and Parent, 1985).

EXPANDED EPITYMPANIC REcEsS. Unlike
many mammals where the epitympanic re-
cess is formed medially by the petrosal and
laterally by the squamosal, the epitympanic
recess in rodents is laterally bounded by the
ectotympanic, which completely excludes the
squamosal from the tympanic region (figs. 2,
3). Among living rodents, muroids retain a
primitive ear region that compares closely
with that of the Theridomorpha (Lavocat and
Parent, 1985). The epitympanic recess in mu-
roids is located primarily in the tegmen tym-
pani of the petrosal and the ectotympanic
contributes only a narrow process to the lat-
eral wall of the recess. The recess is not ex-
panded into the ectotympanic nor is it sep-
arated from the external auditory meatus. In
other living rodents the ectotympanic, which
makes up the lateral part of the epitympanic
recess, is greatly excavated dorsally to form
a deep fossa or notch. A bony lamina is also
developed from the ectotympanic, separating
the epitympanic recess from the external au-
ditory meatus. In R. delicatissimus there is a
very marked fossa in the ectotympanic (figs.
1, 3A). This fossa may function as the lateral
part of the epitympanic recess. Primitively,
however, it is confluent with the external au-
ditory meatus. Anterodorsal to the epitym-
panic recess and lateral to the stapedial artery
canal, a deep depression is formed in the pe-
trosal. This depression provides the space for
a further expansion of the epitympanic re-
cess. In other primitive rodents such as Para-
mys and Sciuravus, a complete bulla is not
known and the condition of the epitympanic

NO. 2972

recess is difficult to ascertain. As discussed
in the section on the bulla above, the ecto-
tympanic bulla in these forms was possibly
only poorly developed and not yet dorsally
expanded, so that a deep epitympanic recess
probably did not exist. In Cocomys, the epi-
tympanic recess in the petrosal is distinct and
there is no expanded fossa in the ectotym-
panic (Lietal., 1989). In other living rodents,
the epitympanic recess in the tegmen tym-
pani becomes less important while its ecto-
tympanic part dominates. In addition, with
the projecting facial canal, the epitympanic
recess in the tegmen tympani is well sepa-
rated from the fenestra ovalis.

INFLATION OF THE PROMONTORIUM. In
primitive mammals, such as Leptictis (No-
vacek, 1986), ?Protungulatum (MacIntyre,
1972), and marsupials (Clemens, 1966), the
promontorium proper is generally low or flat.
In most mammals, the elongated cochlea is
coiled into a spiral; presumably this reflects
a need for retaining an elongated labyrinth
within the confined space of the ear region.
Although monotremes have less than one
cochlear turn (Griffiths, 1978), most mam-
mals have at least one or two (MaclIntyre,
1972). The shape of the promontorium prob-
ably reflects to some extent the number of
cochlear coils and the orientation of the co-
chlea. However, there appears to be no direct
relationship between the height of the cochlea
and the number of half-turns in some mam-
mals (Pye, 1979), nor such a relationship in
rodents, although in general, the greater the
cochlear height, the greater the number of
half-turns in selected groups of rodents (Pye,
1977).

The cochlea coils around a central axis and
this axis may point ventrally or anteriorly,
indicating a rotation of the cochlear orien-
tation. Generally, when the axis points an-
teriorly, the promontorium appears more in-
flated, while the promontorium looks low and
flat when the axis points lateroventrally. The
cochlear capsules undergo certain rotation and
shift during ontogeny. Different groups of
mammals present various patterns of rota-
tion and shift (Zeller, 1987). The detailed
knowledge of the cochlear rotation and shift
during rodent ontogeny is unfortunately not
yet available. Whether the cochlea coils tight-
ly or loosely may also affect the shape of the

1990

promontorium, such as in some rodents where
the last part of the cochlea remains uncoiled
(Lavocat and Parent, 1985). In fossil rodents,
Paramys, Sciuravus, and Cocomys have alow
and flat promontorium, while R. delicatissi-
mus has a more inflated one. Nevertheless,
we do not know whether or not the latter has
more cochlear turns or has a more anteriorly
pointed cochlear axis than the former, unless
broken cochlear specimens are available.

THE STAPEDIUS MUSCLE. Lavocat and Par-
ent (1985) believed that Paramyinae are
somewhat derived in that the stapedius mus-
cle lies completely within the bulla, and the
ear region is located anteriorly in the skull.
They considered Sciuravus and Theridomor-
pha to be more primitive because their ear
region is located at the very posterior part of
the skull and the stapedius muscle extends
outside the tympanic cavity. The ear region
is located less posteriorly in Paramys than in
other rodents because its mastoid process is
horizontally and posteriorly elongated. There
is a distinct trough on the ventral side of the
mastoid process. This condition is very sim-
ilar to that of Adelomys (Lavocat, 1967: fig.
2; pl. 2), although in Paramys this trough is
probably longer. A corresponding structure,
for the insertion of the posterior belly of the
digastric muscle, is present in Leptictis and
such a trough was regarded as a primitive
condition in various groups (Novacek, 1986).
Early erinaceomorphs, such as Pholidocercus
hassiacus and Diacodon alticuspis, have a
similar trough on the ventral side of the mas-
toid process (MacPhee et al., 1988). This
trough can also be found in early carnivorans
and creodonts. In living rodents, the mastoid
process is horizontally shortened or in some
cases may be vertically protruded. When the
mastoid process is reduced, the auditory re-
gion appears to be more posteriorly located.
It is an open question whether the condition
in Paramys is primitive or derived.

The stapedius muscle in Paramys likely ex-
tended outside the tympanic cavity. This is
because the fossa for the stapedius muscle in
Paramys is less restricted posteriorly than in
R. delicatissimus, and the bulla is too small
to completely cover the tympanic cavity. R.
delicatissimus has a reduced mastoid process
and the stapedius muscle is assumed to be
within the tympanic cavity because the fossa

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 13

is deep and its posterior side is sharply
bounded by a tubercle formed by the petro-
sal. In addition, the bulla is expanded pos-
teriorly, covering the tympanic cavity and the
stapedius muscle fossa.

OTHER CHARACTERS. There are some other
characters that are basically primitive in
Reithroparamys and will not be discussed in
detail. Such discussions may be found in Par-
ent (1980, 1983) and Lavocat and Parent
(1985). These characters include small and
circular fenestra rotundum, presence of the
stapedial artery, the external auditory meatus
not separated from the epitympanic recess,
and tensor tympani muscle not covered.

METHODS AND RESULTS

Table 1 lists the auditory features consid-
ered in this analysis. Except for those per-
taining to reithroparamyines, most features
in other selected rodent groups have been
taken from Parent (1980, 1983), Lavocat and
Parent (1985), Wahlert (1974), Bugge (1985),
George (1985), and Luckett (1985). The dis-
tribution of these characters in selected groups
of rodents is listed in table 2. This data set

' was analyzed using both PAUP (3.0) provid-

ed by D. Swofford to M. Novacek and HEN-
NIG 86 (version 1.5) by J. Farris. The
McClade (2.87c) Test was used for data input
and character tracing. Data were run on the
heuristic and branch-and-bound searches in
the PAUP and on ie calculations in HENNIG
86. The PAUP program is convenient in many
aspects (see also Novacek, 1989), although
its branch-and-bound search is slow in han-
dling this particular data set. Calculation un-
der the command ie in HENNIG 86 gener-
ates trees by implicit enumeration algorithms
and the results are certain to be of minimal
length (Farris, 1988; see Platnick, 1989, for
more details).

Originally, most of the multistate charac-
ters have been coded as open, diagonally lined
or solid squares for primitive, intermediately
derived and derived conditions, respectively
(Lavocat and Parent, 1985). The coding is
replaced in this paper by 0, 1, and 2 for the
purpose of calculation. Apparently, when an
intermediately derived condition is specified,
transformation of character states is as-
sumed, i.e., from primitive to intermediately

14

AMERICAN MUSEUM NOVITATES

TABLE 1
Selected Auditory Characters and Character
Polarity as Discussed in Text

. Bullae are small (0), moderately inflated (1), or in-

flated (2).

. Bulla-petrosal contact is very loose (0), tight (1), or

fused (2).

. Promontorium proper is low (0), slightly expanded

(1), or swollen (2).

. Epitympanic recess is a small and shallow fossa (0),

or expanded dorsally into the ectotympanic (1).

. Epitympanic recess overlies the roof of external au-

ditory meatus (0), or is separated from the external
auditory meatus (1).

. Coils of cochlea are uniform (0), or its last part un-

coiled (1).

. Cochlea is in a normal shape (0), slightly bent (1),

or strongly bent (2).

. Internal septa of the bullae are absent (0), weakly

developed (1), or well developed (2).

. Distal internal carotid artery is present (0), or absent

(1).

. Stapedial artery is present (0), reduced (1), or absent

(2).

. Fenestra ovalis is vertical (0), inclined (1), or hori-
zontal (or tilted) (2).

. Fenestra ovalis is small (0), moderately large (1), or
large (2).

. Fenestra rotunda is small (0), moderately large (1),
or large (2).

. Fenestra rotunda is regular (rounded) (0), or twisted
(1).

. Proximal internal carotid artery runs ventral to the
fenestra rotunda (0), partially shielding (1), or greatly
shielding the fenestra rotunda (2).

. Stapedius muscle is large (0), reduced (1), or absent
(2).

. Stapedius muscle is exposed in the tympanic cavity
(0), partially covered (1), or in a closed fossa (2).

. Posterior part of the stapedius muscle is outside the
bulla (0), or completely inside the bulla.

. Tensor tympani muscle is uncovered (0), partially
covered (1), or in a closed fossa (2).

. Facial nerve is exposed in the tympanic cavity (0),
partially enclosed by bony element (1), or complete-
ly enclosed in a bony tube (2).

. Hypotympanic recess is absent (0), present but small
(1), or large (2).

. Petrosal orientation is horizontal (0), tilted (1), or
more vertical (2).

. Meato-cochlear bridge is absent (0), or present (1).

. Maleus and incus are separated (0), tightly jointed
(1), or fused (2).

. Stapedial tube is absent (0), partially developed (1),
or complete (2).

NO. 2972

derived and then to derived state (or 0-1-2),
although an intermediately derived state is
more or less arbitrarily determined, such as
the fenestra rotunda being small (0), mod-
erately large (1), or large (2). Calculations are
done based on two character-type sets: all
characters unordered (nonadditive) and all
characters ordered (additive). As an ordered
character type, the character states are or-
dered as 0-1-2. When a character changes in
a cladogram from state 0 to 2, or vice versa,
it will be counted as two steps. In other words,
the character must proceed progressively
through state 1. However, there is no re-
quirement that state 0 be the ancestral state
(Swofford, 1989). As an unordered character
type, any state is capable of transforming di-
rectly to any other state, and character state
1 in a multistate character is no longer treated
as intermediately derived but simply a de-
rived state. When a character transforms from
state 0 to 2, or vice versa, it will be counted
as only one step.

Figure 4 is the strict consensus (nelson) tree
derived from six equally most parsimonious
trees that result from a heuristic search of
PAUP and ie algorithms of HENNIG 86,
based on unordered (nonadditive) characters.
Branch-and-bound search of PAUP is too
slow to be completed in this particular data
set. All the six trees have 77 steps of tree
length and 0.56 overall consistency index. The
consensus tree has 77 character changes, equal
to the tree length.

Figure 5 is the most parsimonious tree from
branch-and-bound search of PAUP and ie
algorithms of HENNIG 86, based on ordered
(additive) characters. It has 93 steps and 0.46
overall consistency index, but 81 character
changes. Character changes are not equal to
tree length because when a character changes
from state 0 to 2, or vice versa, it is counted
as two steps.

Figure 6 is the strict consensus of four
equally most parsimonious trees from branch-
and-bound search, based on irreversible
characters of PAUP. The irreversible char-
acters are equivalent to ordered characters
with the additional constraint of irreversi-
bility being imposed, i.e., transformations
from a more derived state to a less derived
state are prohibited. In this analysis, the “‘ir-

1990 MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 15

n
3 BloocococoNONANDCOCOCOeH
3
> alorerooconoocoCOnNANS
s fae)
= al}oooooooconmnn CooooCoo-
~~
% AQ
2 Nicoccocooo-~coccoNnNnoo
+=
= ue,
5 nl/ocoooonmoCONSCOAOeR
g Slocooeooonnnnnoonanaae
Z On
is SBIODOOCH ANN DOR FRA NNO
‘E
a Plon DOO OH HHH ORR AHS
Dr
=a) “locooooocONcDOoOACOCCOCO
2.8
che SPlooeoocoonoNnNnonOoNnean
SES C1900 6-06 Nanooooon
eoUs
eaTU +
Be aollg|tlocoocooo one Toe
q J
<3 ll € oO
2a g[|=|ocooooo- SA oA CS
ota
Zee \lé
2 A, OIN|[oceo oO cC OH ONAN NA Ae BAe
W > I
oe “lon ooooCoOHn HANAN ORR AAN
Pe OD .n
nea
wor Oo
as § & SlcoceccocoNnonHONNNANS
aEBaz
taect y
ae ES) AlODDCCCDCOCOH HH DOOR Bee
ee
PEE r~mloooonoononnnoonnaHn
evs
—™o ° m™LOCODSOADANNANNANAN HANS
sav
qsap
oS9
hs ee) wmwlooddod © ew wt wt wt mt wt we et et = OO
Sm 3
ga Ti oocOoOO CO ws) wt mm mm mt met et ee
Mo)
g° mlocooocoORHNNNANANNANANA]
os
As NlLODCDOOR NN AN eR ee NANA
ll
Q a=l1O0ONOnaRNNANNANANAAANAN
be)
a o
_ ES
on 3 oOo wo om
os =
2 rs br cm) a SUaé
oS Ee 3 a3
s s » 5 Y so) 5s 3S FREES
7) a 8 6 eso 5 oy ee
é SECCEEL ELIT EPEEE
F ° wae Of8 i] gvg Ss
£ BEES ESESESR RE SSSE
_— —_ —
& BS eS SR ORS AS AVE SRS
I OZGEASNTOSCKOKAORSEe
Il
2. CmQOAwWHUDHnAMHNAZOLO

16 AMERICAN MUSEUM NOVITATES NO. 2972
wu
C4 Ai
z Lu
= uw a: y ¢
= r a < + <q o.
ue re) ao 69S - © 5 w & w
aS < w E > £u & & € «< @g gs Ww
Oo a = e544 @¢2 o 2 2 z~Z2 g <
<@eqe24¢co6o2282 928 2 8 &€ 2 B
f£zeSogt zt 28 FS Ee =z 8 9
So a@ 5a go 2 eke ER ee SS
” cc ” < © < Sg a. © O = © b= a. = NY
2-2
5-0 1-1 | 10-1) [17-1 11-0 7-1 | 411-2 7.1
13-0 15-1 |18-0 13-0 14-1 | 12-2 14-1
20-1 25-1 19-1 16-2 17-2
25-1 14.1 |21-2 19-1
23-0
ies 16-2
rane 16-0
13-2
3-2,7-2
, 2461
6-1,12-2 21-1
® 1-2
2-1,9-1,11-2 22-2, 2-1 8-1 2-1
15-2,19-0,23-1 14-1,19-1 9-] 71
3) 20-0 6) 24-2 22-
3-1, 8-2 oe 1
10-0,11-0,25-2 73
5) 16-1
12-1,13-1 2
1-2,2-2,3-2,4-1
5-1,10-2,19-2 Il
Bhi 19-1
11-1,18-2

Q)

Fig. 4. Cladogram depicting the strict (nelson) consensus tree for selected rodent groups, derived
from six equally most parsimonious trees resulting from heuristic search of PAUP and implicit enu-
meration algorithms of HENNIG 86. All characters unordered. Tree length = 77; overall consistency
index = 0.56; character changes = 77. Character changes result from accelerated transformation (ACCT-
RAN) optimization of PAUP. Number preceding dash = character; number after dash = character state.
* only recent members of the family; **, Erethizontidae not included. See text for more details.

rev.up” command is used to specify that states
higher in the symbols order are derived rel-
ative to states lower in the symbols order,
such as state 2 being derived relative to state
1. The four trees have 102 steps and 0.42
overall consistency index. The consensus tree
has 83 character changes.

The character changes in figures 4, 5, and
6 are also results of the accelerated transfor-
mation (ACCTRAN) optimization of PAUP.

If a derived state is not consistently present
in all members of a given group, ACCTRAN
prefers reversals over parallelisms whereas
DELTRAN (delayed transformation) can be
thought of as preferring parallelisms over re-
versals. Both optimizations produce clado-
grams with the same topography and equal
tree length, but with different explanations of
character changes. For example, as a result
of ACCTRAN, the derived condition (2) of

1990 MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 17
r
> WwW
= uw = a % r Ww
— *
a, < <t < 3; =< © <
< 6 w § ; eb ee 28 cE
Oo 2 ¢g6<a«25u4 & es 09 @ &# § § Y@ =z
<a@285998 842 8 2 5 §S zg 2k
ec = O oc ao = re) s
ea Ww z= Y > F- <
a a ee ee
on £- = N = 8 eae © ee & 6 -E 6 Se OF Fg
2-2
10-1 | 17-
11-1 | 45-1 | 18-0 ms 1-2
25-4 | 19-1
141 | 12-2
@ 21-2 es 11-0 | 43-4
14-1
23 16-1 | 24-4
16-2 5&0 [19-1
oo 419) 13-0
11-0 init me MLAS)
11-1 ee 16-2 ae 25-1 3-2 12-0
21-1 21-1 17-2
cau Ore 6-1
Q 16-0,19-0,23-1 as 7-2
3-1,7-0,8-2 7-2 22-2
10-0,25-2 vee
8-1,9-1,13-1
14-0, 16-1
2-2,19-2
20-2,22-0
1-2,3-2,4-1,5-1
10-2,19-1,12-1
2-1,7-1 14-1,18-1
(3) 22-1
19-1

Fig. 5. Cladogram depicting the most parsimonious result from branch-and-bound searches of PAUP
and implicit enumeration algorithms of HENNIG 86, based on ordered characters. Overall consistency
index = 0.46; tree length = 93; character changes = 81. Character changes are not equal to the tree
length because when a character changes from state 0 to 2, or vice versa, it is counted as two steps.
Character changes are results of ACCTRAN optimization of PAUP.

character 10 occurs at node 4 in figure 4,
reverses to 10-0 at node 7, and then appears
in Aplodontidae as 10-1. Asa result of DEL-
TRAN, however, character state 10-2 will
occur at node 6 and in Castoridae, and state
10-1 in Aplodontidae. These character states
can be interpreted as acquired independently

in these taxa and groups. Nonetheless, char-
acter 10 displays three steps of changes in
both ACCTRAN and DELTRAN optimi-
zations. Because the results (the cladograms)
are the same, only those of ACCTRAN op-
timization are illustrated.

Table 3 lists the character consistency in-

18 AMERICAN MUSEUM NOVITATES NO. 2972
<
Z w
= wi
Tow Ww = o a % w
oO. <x q < onl a < <
wu f£ 4 ra) cw er Ff} = a ud
< oOo § = <q <« 5 s; 2 ff w wu <<
Pt i22uy58 e963 ¢ & gz
fem a) = Q —
qa?aqge<s¢éE = a = 2 S
ao zr & € 8 G@etegsgsg¢geoges&§&ss8 @
5 Pb st Go 2k Ee 2G 5 & SZ
7 ra crc on» <€ © £ OF 0 F OO € a2 S&S N
2-2
7-4 11-2
11-1 | 10-1
14-1 12-2
15-1 147-1179 [60 Jig. [19-1 ay
2-2 25-1 178.0 | 179 | 74 | 49, [21-0
20-1 11-0 2-1
ie oe 14-1,16-2)19-1 | 94, | 120 16-1 7-1
18-1 11-0 23-4 21-2 aE 17-2 8-1,9-1 wo
15-0 13-2,14-0 (3)
16-2
3-2,6-1 16-1,21-1
ie 742,122 Gy — 242 20-0
(10) 2-2,19-2 22-2
8-2,9-1,11-2
13-1,15-2,25-2 vai e
6) 19-1 )
2-1,3-1,4-1,5-1
11-1,12-1,18-1,20-2 19-1

1-1

Fig. 6. Cladogram depicting the strict consensus tree derived from four equally most parsimonious
trees. The four trees are obtained through branch-and-bound search based on irreversible characters of
PAUP. Tree length = 102; overall consistency index = 0.42; character changes = 83.

dices for all characters in figures 4, 5, and 6.
Some of these characters have very low con-
sistency indices (CI), especially characters 7,
11, 14, 16, and 19. The CI of these characters
are consistently lower than 0.5 in all the three
algorithms. Figure 7 provides the consensus
results of the most parsimonious trees re-
sulting from the branch-and-bound searches
of PAUP on unordered, ordered, and irre-
versible characters, respectively, with the
above five characters excluded from the al-
gorithms. Figure 7A is the consensus of 81
equally most parsimonious trees of unor-

dered characters. The branch-and-bound
search of PAUP is possible in this case, al-
though it took 20 hours to complete the cal-
culation. The 81 trees have 51 steps tree length
and 0.68 overall CI. Figure 7B is the strict
consensus result from eight equally most par-
simonious trees of ordered characters, each
of them having 61 steps and 0.56 overall CI.
Figure 7C is the strict consensus of four
equally most parsimonious trees of irrevers-
ible characters. The four trees have the same
67 steps tree length and 5.07 overall CI.

As one can see, Figures 4—7 present similar

1990

results. A few groups are stable in all these
cladograms. These are Anomaluridae-Pedet-
idae, Ctenocactylidae-Thryonomyidae-Ca-
viomorpha, and the group consisting of
Sciuridae, Aplodontidae, Gliridae, Reithro-
paramyinae, and Heteromyidae. The occur-
rences of Anomaluridae-Pedetidae and Cas-
toridae vary considerably in these cladograms,
indicating uncertain systematic positions of
these taxa (see discussion).

The auditory features do not contribute
anything to the monophyly of Rodentia, i.e.,
no derived character appears at node 1 in
figures 4, 5, and 6. Auditory features are only
used to reconstruct the relationships within
the group. Rodent monophyly is, however,
recognized elsewhere by some other charac-
ters. In this analysis, Rodentia (ingroup) is
assumed to be monophyletic and a hypo-
thetical ancestor (outgroup) is employed for
the purpose of analysis. This outgroup is not
indicated in the cladograms but is implied by
the rooted trees.

- DISCUSSION

A cladistic analysis of rodent phylogeny
based on the auditory features is attempted
in this paper, although such an analysis is
generally regarded as difficult, as cautioned
by Wilson (1986), because parallelism is
thought to be an important factor in rodent
evolution. Character distribution in rodents
is so inconsistent that it is almost impossible
to obtain a single shortest cladogram on any
data set by a manual procedure. This is prob-
ably why some workers (e.g., Parent, 1980;
Lavocat and Parent, 1985) have provided a
table of character distributions but failed to
present a cladogram out of their distribution
data.

I agree with Luckett and Hartenberger
(1985) that those phylogenetic hypotheses
corroborated by data from several different
(and preferably unrelated) organ systems are
more likely to reflect the true phylogeny ofa
group, than are those hypotheses corrobo-
rated by single character complexes. As has
already been mentioned, previously pro-
posed phylogenetic relationships of reithro-
paramyines to other groups of rodents have
been based primarily on dental and a few
cranial features. Data from the auditory re-

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS

19

TABLE 3
Consistency Indices for Characters in figure 4 (CI-
A), figure 5 (CI-B), figure 6 (CI-C)
(Asterisk indicates character that has CI consis-
tently lower than 0.5 in all three algorithms)

Character CI-A CI-B CI-C
1 0.667 0.667 1.000
2 0.400 0.500 0.400
3 0.667 0.500 0.667
4 1.000 1.000 1.000
5 0.500 0.500 0.500
6 0.500 0.333 0.333
7* 0.400 0.286 0.400
8 1.000 1.000 1.000
9 0.500 0.500 0.333

10 0.667 0.400 1.000
11* 0.333 0.286 0.296
12 0.667 0.500 0.500
13 0.500 0.500 0.400
14* 0.250 0.250 0.250
15 1.000 0.667 0.667
16* 0.400 0.286 0.400
17 1.000 0.667 1.000
18 0.500 0.333 0.333
19* 0.333 0.333 0.333
20 0.667 0.667 0.667
21 0.667 0.500 0.500
22 1.000 0.667 1.000
23 0.500 0.500 0.500
24 1.000 0.667 1.000
25 0.667 0.550 0.667

gion in major groups of rodents can be used
to test previous hypotheses. Although most
of these auditory features are known by some
authors, a more explicit resolution of rela-
tionships based on these auditory features is
provided in this paper. This data set in turn
produces hypothetical relationships, which
are open to further testing. The following dis-
cussion will focus on the phylogenetic posi-
tion of reithroparamyines and related groups.

REITHROPARAMYINAE
CONTROVERSY

The controversy about the relationship of
reithroparamyines to other rodent groups be-
gan when they were implicated in the origin
of the South American Caviomorpha. The
Caviomorpha are first known from the Oli-
gocene (Deseadan) of South America, and
were already fully hystricognathous and hys-

20 AMERICAN MUSEUM NOVITATES NO. 2972
<
wi

ley a

w Ew < 5 23 +9

oq ius, © Seu5a34

SEC ELEE FEEEEER

< 26°88 Sca =z

e235 SheSoESE SSS

Soe epee shee Hs ee

®RQENIHDZCOLEROGEO

wi uw
< <
Ww

a. a< a. < adtac
oe: ag ieua us we 18 8 EwEEeo yw
SEP EE Tt EEE SPE EEEEEEEE
zQae e a7ae a20a caofzaa>
zQalfoa<oSaszaza *aQ5OOR Sc S825 SFe
e=-oSf&eaQ 9090sEs CrQOouSorerzez&>2uc
SGRTOU SCOR ZS =E sw 2 HSOce > cc
5 wu Sarat aee23ae ee a ee eee
BESNTHOOtrOOrOdoG OQEFENTOCOcCOOOr doa

Cc

Fig. 7. Consensus results after characters 7, 11, 14, 16, and 19 are excluded (see table 3). All trees
are obtained through branch-and-bound searches of PAUP. A, Strict consensus of 81 equally most
parsimonious trees of unordered characters; tree length = 51; CI = 0.67. B, Strict consensus of 8 equally
most parsimonious trees of ordered characters; tree length = 61; CI = 0.56. C, Strict consensus of 4
equally most parsimonious trees of irreversible characters; tree length = 67; CI = 5.07. A, B, and C are
similar in configuration to figs. 4, 5, and 6, respectively, but with a higher overall consistency index.

tricomorphous (Wood, 1985). Two widely
differing hypotheses concerning the origin of
Caviomorpha are held by Wood and Lavo-
cat. Views of these authors have been strong-
ly and consistently expressed over the last
quarter century. According to Lavocat (1973;
1974a, 1974b; 1976; 1980), caviomorphs
originated from the African Phiomorpha,
a hystricomorphous and hystricognathous
group. Phiomorpha, in the usage of Lavocat,
includes Old World hystricognaths but is re-
stricted to the Thryonomyoidea by others
(Patterson and Wood, 1982; Wood 1985).
Lavocat’s hypothesis is based largely on the

morphological similarities between Cavio-
morpha and Phiomorpha, particularly of the
middle ear region (Lavocat, 1973, 1976; Par-
ent, 1980, 1983; Lavocat and Parent, 1985).
According to Wood (1974, 1975, 1977, 1980,
1981, 1983, 1984, 1985; Patterson and
Wood, 1982), the South American cavio-
morphs are descended from the reithropara-
myines or franimorphs, a North or Middle
American group with incipient to full hys-
tricognathy.

Beyond simply representing the ancestral
stock of the Caviomorpha, Reithroparamyi-
nae, included in Franimorpha, have also been

1990

postulated as progenitors for all Hystrico-
gnathi. As early as 1975, Wood suggested that
all the special similarities shared by the living
hystricognaths that were absent in the Frani-
morpha must have evolved independently,
by parallelism (except the incipient to full
hystricognathy). He also believed that from
Late Paleocene or early Eocene common
ancestors similar to Franimys, the frani-
morphs evolved independently in both North
America and Asia, and that the nearest ap-
proach to Eocene hystricognaths were the
Reithroparamyinae. The New and Old World
hystricognaths were regarded as indepen-
dently derived from New and Old World sub-
hystricognathous rodents, presumably all
members of the Reithroparamyinae, that had
reached North America and Asia by the Early
Eocene. Furthermore, Patterson and Wood
(1982: 453) stated: ““The Order Rodentia was
certainly of northern origin. Members of it
first appear in the latest Paleocene of western
North America. These forms, which had al-
ready acquired all the basic ordinal charac-
ters, are referable to two very closely related
families, the scilurognathous Paramyidae and
the incipiently hystricognathous Reithro-
paramyidae, the latter, in our opinion, the
ancestral stock of all later members of the
Hystricognathi.”

Hystricognathi monophyly has been sup-
ported by evidence including blood vascular,
reproductive, chromosomal, and skeletal fea-
tures (George, 1985), fetal membranes and
placenta (Luckett, 1980, 1985), internal ca-
rotid pattern (Bugge, 1985), and features at
the molecular level (Sarich, 1985; Shoshani
etal., 1985; Beintema, 1985; De Jong, 1985).
As to the content of the suborder Hystrico-
gnathi, however, there are two different view-
points. As defined by Wood (1975, 1985),
Hystricognathi includes the Hystricidae, the
Thryonomyoidea, the Bathyergoidea, the
Caviomorpha, and the Eocene-Oligocene
Franimorpha. Recently, Wood (1985) argued
that the Hystricognathi are a natural group,
but only if one includes the basic hystricogna-
thous stock, the Franimorpha, in which many
of the features secondarily associated with
hystricognathy had not yet developed. In
contrast to this, the inclusion of Frani-
morpha has been rejected by others (Korth,
1984; Luckett and Hartenberger, 1985).

Clearly, the grouping of franimorphs and

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 21

the transition of hystricognaths from frani-
morphs relies heavily on the dubious feature
of incipient histricognathy which is believed
to be present in reithroparamyines by Wood.
That the condition seen in reithroparamyines
is indeed incipient histricognathy has been
questioned by others (e.g., Dawson, 1977)
who regarded it instead as sciurognathy. Korth
(1984) has even argued that none of the species
included in the Franimorpha appears to have
attained histricognathy as defined by Tull-
berg (1899), that the “incipient’’ histricogna-
thous condition appears quite common
among early rodents, and that it may in fact
represent the primitive condition for rodents.
Apparently, if this feature is disregarded, there
remains little else to suggest that reithropara-
myines are particularly closely related to oth-
er groups of rodents.

REITHROPARAMYIDAE AND
HYSTRICOGNATHI

As introduced above, two views are held
about the inclusion of reithroparamyines in
the Hystricognathi, based on differing inter-
pretations of mandible structure. Two ques-
tions addressing this problem may be posed:
(1) are reithroparamyines the ancestral stock,
or at least within the ancestral stock, of the
histricognaths? (2) are franimorphs a mono-
phyletic group? The answers in both cases
appear to be no.

While the mandible of Reithroparamyinae
can be viewed as either “incipiently” histri-
cognathous or sciurognathous, derived fea-
tures of the ear region of Reithroparamys
clearly favor a closer relationship of reithro-
paramyines to sciurids, aplodontids, and
glirids, shown in figures 4—7. In other words,
reithroparamyines are more closely related to
sciurognathous rodents than to the Hystri-
cognathi. In figures 4-7, reithroparamyines
are far distantly separated from histricogna-
thous groups. There is no shared derived
character in the auditory region demonstrat-
ing any close relationship between reithro-
paramyines and histricognaths. Based on the
features of the auditory region, reithropara-
myines as the ancestral stock of histricogna-
thous rodents is rejected.

This raises the question of “‘Franimorpha”’
monophyly, particularly considering the fact
that some of the members included in the

22 AMERICAN MUSEUM NOVITATES

*Franimorpha”’ do develop typical hystri-
comorphy and histricognathy, but others have
features typical of sciuromorphous and sci-
urognathous rodents. Franimorpha was pro-
posed by Wood (1975) as an infraorder that
first included Reithroparamyinae, Prolapsus,
Protoptychus, and Guanajuatomys. After-
ward, Wood (1980, 1985; Patterson and
Wood, 1982) added Cylindrodontidae to this
infraorder and the Reithroparamyinae was
elevated to the family Reithroparamyidae.
Among the “‘Franimorpha,”’ the late Eocene
Protoptychus (Scott, 1895; Wilson, 1937) is
recognized as fully histricomorphous (Wah-
lert, 1973), although differing views have been
published concerning its mandibular struc-
ture (Wood, 1975; Dawson, 1977; Korth,
1984). This may indicate that Reithropara-
mys and Protoptychus actually represent two
different evolutionary lineages, i.e., Protopty-
chus evolved to a hystricomorph while the
protrogomorphous Reithroparamys to a sci-
uromorph because the latter bears some de-
rived characters only found in sciuromorphs.
Wood (1977, 1981) described the skull of
Prolapsus, another member of “Franimor-
pha,” as hystricomorphous. But Korth (1984)
believed that the skull of Prolapsus was sim-
ilar to that of the Bridgerian Sciuravus and
that both are clearly protrogomorphous. Un-
certainty also exists about the structure of
the mandibles of Prolapsus and Mysops,
although Flynn et al. (1986) thought that
histricognathy occurs in the latter Eocene
Prolapsus. Unfortunately, basicranial mor-
phology in these forms is poorly known.
The ear region in cylindrodontids seems
quite different from that of Reithroparamys.
As described by Wahlert (1974), there is no
stapedial artery in cylindrodontids but the
internal carotid artery is probably present.
The posterior opening of the carotid artery is
very small and is separate from the jugular
foramen. In reithroparamyines, as in sci-
urids, the carotid foramen occurs roughly in
the same depression with the jugular fora-
men. The carotid canal, at least in Ardynomys
and Cylindrodon, enters and runs anteriorly
through the periotic. The circulation system
indicates that Reithroparamys and cylindro-
dontids are widely divergent: the stapedial
artery is retained but the DICA is lost in the
former while the reverse is true of the latter.
All these features suggest that Franimorpha

NO. 2972

is not a monophyletic group, and there is not
any shared and derived feature to support
such a grouping. Korth (1984) has actually
returned the Reithroparamyinae back to the
Ischyromyidae and relegated other members
of Franimorpha to various rodent groups.

Recently, a new genus, Marfilomys, from
Central Mexico has been described by Fer-
rusquia-Villafranca (1989). According to the
author, this new genus shows greatest resem-
blance to 7 groups among 27 rodent families
compared: reithroparamyine ischyromyids,
cylindrodontids, protoptychids, octodontids,
cocomyids, yuomyids, and chapattimyine
ctenodactylids. The author concluded that
“the resemblances were interpreted as phy-
logenetically related in the case of the reithro-
paramyines, cylindrodontids, protoptychids,
and the early octodontids, because these four
taxa are both hystricognathous and histri-
comorphous (at least incipiently) and the first
three, which are united as the Franimorpha,
have an Eocene record and inhabit North and
Middle America” (ibid., p. 114). Therefore,
the discovery of Marfilomys in Middle Amer-
ica lends strong support to Wood’s hypoth-
esis that the franimorphs were essentially a
Middle American group, and that from this
group stemmed the ancestral caviomorphs.
Apparently, the conclusion reached by the
author is based on nothing but the same as-
sumption provided by Wood that the Frani-
morpha are of at least incipient histricog-
nathy and of Middle or North America
distribution. Additionally, some of the char-
acters believed to be shared by Marfilomys
and franimorphs by the author (Ferrusquia-
Villafranca, 1989: 112) may not be unique to
these groups, but have a wide distribution in
rodents (for instances: upper cheek teeth 1-0-
2-3; deciduous premolars normally replaced;
premaxillae forming a substantial part of the
rostral dorsum, and meeting the frontals; and
the relatively posteriorly set incisive foram-
ina transected by the premaxillo-maxillary
suture or being limited posteriorly by this
suture).

REITHROPARAMYINAE AND
OTHER ISCHYROMYIDS

Differentiating taxonomic groups is much
easier than specifying their similarities, es-
pecially their synapomorphies. Recognizing

1990

the Franimorpha as a nonmonophyletic group
leaves open the question of reithroparamyine
relationships to other groups. Although the
family Reithroparamyidae has been included
in the Franimorpha, many authors usually
regard it as a subfamily within the Paramyi-
dae (Wood, 1962) or Ischyromyidae (Black,
1971; Korth, 1984). In these cases, the as-
sociations have been based for the most part
on primitive characters. The diagnosis of the
Ischyromyidae proposed by Black (1971) and
adopted by Korth (1984), illustrates this point:
““Cheek teeth basically low-crowned and tri-
tubercular with hypocone, when present, sec-
ondary in importance to protocone; lophate
condition rare, found only in a few advanced
forms; talonid basins generally large and un-
divided; infraorbital foramen generally small,
rounded, not compressed; zygoma heavy;
masseter arises from ventral surface of zygo-
ma; skull quite narrow in postorbital region;
nasals usually long; temporal fossa large;
braincase small, not inflated; bulla co-ossified
only in a few species; tibia and fibula separate;
humerus with entepicondylar foramen.”
Recently, it has become widely accepted
that Heomys from the Middle or Late Paleo-
cene of China is the mammal closest to the
ancestry of the Rodentia and in particular to
the early Eocene ctenodactyloids (Li, 1977;
Li and Ting, 1985; Li et al. 1987; Li et al.
1989; Korth, 1984; Flynn et al. 1986). Korth
postulated that, if the ctenodactyloids rep-
resent the primitive condition, then reithro-
paramyines are the most primitive ischyro-
myids because they maintain a hypocone on
P4, a distinct hypoconulid on lower molars,
and relatively larger hypocones on the upper
molars than in paramyines. The simpler mo-
lar pattern in paramyines would be consid-
ered more derived. Korth (1984) has also
considered four features of skull and man-
dible to be primitive for rodents: (1) nasal
bones extended posteriorly to a level even
with the posterior margin of the premaxil-
laries; (2) double mental foramen on the
mandible; (3) posterior margin of the anterior
root of the zygoma even with the posterior
margin of P4; and (4) auditory bulla not os-
sified to the skull. Korth believed that the
nasals extending farther posteriorly than the
premaxillaries in the Paramyinace is a derived
condition, because in several eurymylids (un-
known in Heomys) the premaxillaries extend

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 23

farther posteriorly than the nasals. A similar
situation, i.e., the premaxillaries extending to
or beyond the posterior edge of the nasals, is
present in reithroparamyines, pseudopara-
myines, sciuravids, and early ctenodacty-
loids. Therefore, Korth concluded that all
paramyines could be derived from a primi-
tive reithroparamyine stock by lengthening
of the nasal bones and simplification of the
occlusal pattern of the cheek teeth. In con-
trast, the ear region of Reithroparamys deli-
catissimus tells a very different story—that
reithroparamyines are probably a more de-
rived group than paramyines, as shown in
figures 4-7.

It is to be noted that some members in-
cluded in the genus Acritoparamys Korth,
1984, such as A. atavus, A. atwateri, and A.

francesi, had been placed in Paramys by oth-

ers (Jepsen, 1937; Wood, 1959, 1962;
McKenna, 1961; Rose, 1981). Acritoparamys
is included in Reithroparamyinae by Korth
(1984), and therefore, reithroparamyine ro-
dents become the earliest known in North
America. As pointed out by Korth (ibid.), the
most primitive paramyine is Paramys taurus,
which possesses some of the primitive rei-
throparamyine characters such as a hypocon-
ulid on the lower molars and hypocone on
P4. This makes the separation of the two
groups difficult.

Double mental foramina on the mandible
are present in the Heomys and nearly all early
ischyromyids and sciuravids and so are likely
primitive. R. delicatissimus has only one
mental foramen.

According to Korth (1984), the posterior
margin of the anterior root of the zygoma in
primitive eutherians is level with the poste-
rior molars, and may even be farther poste-
rior than M3. In Heomys it is level with the
posterior margin of M1. In early ctenodac-
tyloids and most primitive paramyines, it is
level with the posterior margin of P4, while
in reithroparamyines it progressively moves
forward until it is in line with the anterior
margin of P4 as in Microparamys and Apa-
tosciuravus. Reithroparamyines therefore
have a more derived zygomatic condition. If
paramyines were derived from reithropara-
myines, one would have to assume a reversal
in this character.

Among Eocene ischyromyids, a bulla co-
ossified with the skull is only found in reithro-

24 AMERICAN MUSEUM NOVITATES

paramyines. Korth suggested that this might
bar the Reithroparamyinae from ancestry of
paramyines, but that the earliest reithropara-
myines may not have possessed this char-
acter. At least, it may not have developed in
the Reithroparamyinae until the Paramyinae
had already split off. This statement is a spec-
ulation for which there is no evidence, al-
though the systematic position of Reithro-
paramys and Acritoparamys may be separate
issues.

Among the four features of the skull and
mandible listed by Korth, only the first seems
to be primitive in reithroparamyines. How-
ever, whether the nasal bones extending pos-
teriorly to a level with the posterior margin
of the premaxillaries is primitive or derived
remains an open question. First, “eurymy-
loids” may have premaxillaries extending to
or more posterior than the posterior margin
of the nasals, but this does not necessarily
imply a primitive condition for rodents be-
cause some eurymyloids (Rhombomylus, for
instance) are already too specialized (Li and
Ting, 1984, 1985; Li et al., 1989) to present
the primitive morphotype for rodents. Sec-
ondly, in Cocomys lingchaensis, the nasal
bones extend more posteriorly than the pos-
terior margin of the premaxillaries (Li et al.,
1989). Thirdly, and most important, the con-
ditions of a small premaxillary and the pos-
terodorsal process of premaxillary not ex-
tending to the frontals are widely distributed
in mammals and are generally regarded as
primitive (Novacek, 1985, 1986; Li and Ting,
1985). In most mammalian groups, the pre-
maxillary is a small element much less ex-
tended posteriorly than the nasal. It is more
acceptable that a more posteriorly extended
premaxillary in some rodents represents a de-
rived condition.

Simplification of the occlusal pattern of the
cheek teeth does not seem well defended. P3
in Heomys and Cocomys is relatively larger,
with two cusps and two roots (at least in Heo-
mys). The evolutionary tendency is more
likely the simplification of this tooth, because
in all ischyromyids the P3 is single-cusped
and single-rooted. Comparatively, however,
Reithroparamys has a more reduced P3 than
does Paramys. As has been pointed out by
various workers (Li, 1977; Li et al., 1989;
Dawson et al., 1984), the P4 in both Heomys
and Cocomys is nonmolariform, and the non-

NO. 2972

molarized premolar, essential in pointing to
the relationship between Cocomys and Heo-
mys, 1S another important plesiomorphous
feature. Such a P4 has only a single buccal
cusp (paracone), and lacks a metacone and
hypocone. Flynn (personal commun.) prefers
to call the P4 of Heomys “submolariform.”
P4 is submolariform in Paramyinae and mo-
lariform in Reithroparamyinae (Korth, 1984).
If Korth is correct, i.e., paramyines are de-
rived from reithroparamyines, the transfor-
mation of P4 must be from non- or submo-
lariform to molariform and then back to
submolariform.

Several other features may also be more
derived in reithroparamyines than in para-
myines. In Reithroparamys, the masseteric
fossa of the mandible is bounded by much
heavier ridges and terminates more anterior-
ly. The postglenoid foramen in Reithropara-
mys is more reduced than in Paramys. In
Cocomys this foramen is also very large (Li
et al., 1989). The upper and lower incisors in
Reithroparamys are more laterally com-
pressed, and the anterior surface of the upper
incisors is flat. The snout is shorter and more
tapered than that of Paramys.

These features, plus the derived auditory
region represented by R. delicatissimus, sug-
gest that the reithroparamyines are more de-
rived than the paramyines. The earliest is-
chyromyoids, such as A. atavus (or P. atavus),
may truly represent the ancestral stock of the
ischyromyoids, but there is little evidence to
suggest their placement among reithropara-
myines and therefore to suggest that reithro-
paramyines are ancestral to paramyines. In-
stead, a possible relationship that requires
mention is that reithroparamyines may be
derived from a morphotype similar to Para-
mys. Besides the characters discussed above,
the possible absence of the promontory artery
in Paramys may suggest such a possibility.

REITHROPARAMYINAE, SCIURIDAE,
APLODONTIDAE, AND GLIRIDAE

As shown in the cladograms (figs. 4 and 5),
one interesting grouping consists of Reith-
roparamyinae, Gliridae, Sciuridae, and
Aplodontidae. A unique character at this node
is character 15, i.e., the proximal internal ca-
rotid artery shields the fenestra rotunda.
[Wahlert (personal commun.) considers that

1990

this feature derives independently in glirids
because it does not occur in all of them.]
Characters supporting this grouping but also
occurring in other groups are 2-1, 11-2, and
23-1.

Although aplodontids, sciurids, geo-
myoids, and castoroids were previously in-
cluded in Sciuromorpha by Simpson (1945),
not all these groups appear to be closely re-
lated (Wood, 1937, 1955, 1959; Stehlin and
Schaub, 1951; Schaub, 1953; Lavocat 1956).
An exception seems to be the Sciuridae and
Aplodontidae, where a close sister-group re-
lationship has been proposed by many re-
searchers (Vianey-Liaud 1985; Wahlert,
1985b; Lavocat and Parent, 1985), based on
the dental and cranial evidence of early fos-
sils. This sister-group relationship is also sup-
ported by molecular evidence (Sarich, 1985),
although both forms share the most primitive
fetal membrane complex in rodents (Fisher
and Mossman, 1969; Luckett, 1971, 1985;
Luckett and Mossman, 1981). Moreover,
Lavocat and Parent (1985) pointed out that
these two groups share a few derived char-
acters in the auditory region, including (1)
cochlea bent; (2) meato-cochlear bridge; (3)
fenestra rotunda twisted; (4) absence of the
internal carotid artery, although some of these
characters are also shared with other groups.
These authors also believed that the auditory
region of living Aplodontidae is more ad-
vanced than that of sciurids in two regards:
loss of the stapedial artery, and loss of the
stapedial bony tube crossing the fenestra ro-
tunda (Lavocat and Parent, 1985). However,
it has been noted by Luckett and Hartenber-
ger (1985) that the stapedial artery is present
in primitive prosciurine aplodontids as de-
scribed by Wahlert (1974). As mentioned
above, in a specimen of Allomys nitens
(AMNH 6997), the ear region displays a very
similar condition to that in Palaeosciurus and
other sciurids. In this specimen, the stapedial
is apparently present and enclosed in a bony
tube along its entire course through the tym-
panic cavity. The pathway of this artery is
similar to that of sciurids, i.e., the stapedial
artery (probably the portion of the proximal
internal carotid artery) partially crosses over
the fenestra rotunda in a bony tube. This in-
dicates that the other aplodontids may have
their stapedial artery secondarily reduced
from a condition present in Allomys. It is

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 25

likely that sciurids and aplodontids form a
sister-group based on the known evidence.

Sciurids appeared in Europe immediately
after the “Grande Coupure’’ (Lopez and
Thaler, 1974; Vianey-Liaud, 1979, 1985;
Hartenberger, 1983), and cannot have orig-
inated in Europe. Wilson (1949) suggested
that the Aplodontidae, Sciuridae, and Cas-
toroidea were derived directly from Eocene
paramyines. Lavocat and Parent (1985) held
that the Sciuroidea appeared simultaneously
in the Oligocene of North America, Europe,
and Asia, suggesting a previous Asiatic his-
tory, but information showing their connec-
tion to earlier forms is lacking. In general,
sciurids and aplodontids have been regarded
as derived from the Eocene protrogomor-
phous ischyromyids or sciuravids of North
America (Korth, 1984; Vianey-Liaud, 1985),
but no specific lineage has been proposed to
support this.

Emry and Thorington (1982) described a
specimen (USNM 243981), which they
thought to be closely related to, if not con-
specific with, Protosciurus jeffersoni. They re-
garded this specimen from the Oligocene
White River Formation of Wyoming as not
only the oldest fossil squirrel, but as almost
certainly a tree squirrel. Although this spec-
imen is basically protrogomorphous, the au-
ditory region contains some derived features
of modern sciurids: periotic and tympanic
bulla fused into a single unit, bulla enlarged,
transbullar septae present, and stapedial ar-
tery enclosed in a bony conduit through the
middle ear. Vianey-Liaud (1985) pointed out
these authors’ failure to recognize the Euro-
pean sciurid Palaeosciurus goti (Vianey-
Liaud, 1974a, 1974b, 1975), from Mas de
Got, Quercy. This species is also known in
the Early Oligocene locality of Aubrelong 1
(Quercy). Summarizing some of the dental,
cranial, and postcranial features which define
the family Sciuridae, Vianey-Liaud acknowl-
edged that the earliest Oligocene European
‘‘squirrel’’ was clearly a sciuromorphous sci-
urid and rejected Protosciurus jeffersoni as a
sciurid.

Wood (1980b) held that the transition from
protrogomorphy to sciuromorphy should
mark the boundary between the Paramyidae
and the Sciuridae; and this assumption was
followed by Vianey-Liaud (1985). However,
sciuromorphy as presently conceived is not

26 AMERICAN MUSEUM NOVITATES

unique to Sciuridae, being independently at-
tained in geomyoids and castorids (Emry and
Thorington, 1982), although geomyoids may
differ in some details, such as position of the
infraorbital foramen. The specializations of
the auditory region of sciurids could have
taken place before the appearance of sciuro-
morphy, and may represent a good diagnosis
for the group. Lavocat and Parent (1985),
based on characters of the ear region, pro-
posed that the Sciuroidea (family Sciuridae)
are monophyletic and recognizable by two
rare auditory features that are always asso-
ciated: (1) the stapedial artery (PICA used
herein) crossing the middle part of the fe-
nestra rotunda within a bony canal; and (2)
a strong osseous bridge connecting the prom-
ontorium to the auditory meatus (osseous
meato-cochlear bridge) and hiding the ossi-
cles in ventral view. These two character were
also regarded as derived characters in Sciuri-
dae by Vianey-Liaud. Whether or not the
meato-cochlear bridge exists in Protosciurus
jeffersoni was not mentioned by Emry and
Thorington (1982).

It has been argued (Vianey-Liaud, 1985)
that if the sciuromorphy of sciurids were
transformed from the protrogomorphy of
North American rodents, there is no record
of such a morphological transition. She ad-
mitted that if the sciuromorphous sciurids
and the protrogomorphous aplodontids
probably originated from primitive protro-
gomorphous rodents, it might be possible to
find a protrogomorphous “‘squirrel.” It seems
that the sciuravids and paramyids have ear
regions so primitive that they could be related
to almost anything. It is impossible to trace
relationships with later rodents by looking at
these primitive characters. Reithroparamys is
not sciuromorphous, but its auditory region
displays conditions of typical sciurids, sug-
gesting a divergence of at least some North
American protrogomorphs toward the sci-
urid-aplodontid rodents. If Lavocat and Par-
ent (1985) are correct, one can simply call
Reithroparamys a “pro-sciurid”’ or a protro-
gomorphous “‘squirrel’’ as used by Vianey-
Liaud (1985) because it has both a meato-
cochlear bridge and the internal carotid artery
crossing the fenestra rotunda, although
the bony tube for the stapedial artery has not
yet completely developed. This may lend

NO. 2972

support to the assumption of Black and Sut-
ton (1984) that sciurids certainly appear to
be North American in origin.

The systematic position of Gliridae is also
controversial. Glirids have been considered
to be related to the muroid-dipodoid clade
by some authors (e.g., Wahlert, 1978; 1985a,
1985b), based on zygomatic structure. Other
authors (Wood, 1980a; Dawson and Krish-
talka, 1984; Flynn et al., 1985; Vianey-Liaud,
1985) believed that muroids and glirids ac-
quired these features independently and that
glirids are not myomorphous. According to
Vianey-Liaud, the ““myomorphy”’ of glirids
is only “‘pseudo-myomorphy,” derived from
a primitive protrogomorphy, in contrast to
an ancestral state of histricomorphy for mu-
roids. Auditory features provided by Lavocat
and Parent (1985) and Bugge (1985) support
the grouping of glirids and sciurids, although
such a relationship is more dubious. This
group is supported in the present analysis (figs.
4-7).

Interestingly, Reithroparamyinae has been
grouped with Gliroidea based on the “‘gli-
roid” tooth pattern (Hartenberger, 1985). It
was also shown that the glirids could have
originated from the middle Eocene European
Microparamys (Hartenberger, 1971); the lat-
ter was included in the subfamily Reithro-
paramyinae (Korth, 1984). In addition, rei-
throparamyines and glirids share the
condition of the proximal internal carotid ar-
tery crossing the fenestra rotunda. This may
hint at a special relationship of reithropara-
myines and glirids. Moreover, because glirids
likely acquired their ““myomorphy” from a
protrogomorphous condition, it is possible
that Reithroparamys represents the ancen-
stral morphotype of glirids, retaining protro-
gomorphy but sharing some derived char-
acters with glirids.

SOME OTHER SELECTED
GROUPS OF RODENTS

CASTORIDAE. The castorids are placed
among the Sciuromorpha in Simpson’s (1945)
classification. Hartenberger (1985) main-
tained a similar view, considering the cas-
torids and sciurids a monophyletic assem-

1990

blage on the basis of sciuromorphy. A close
relationship between castorids and sciurids
has been rejected by others (e.g., Schaub,
1953; Wood, 1955). The affinities of casto-
rids to other rodents still remain problematic
(Wood, 1959; Wahlert, 1977; Bugge, 1985;
Lavocat and Parent, 1985). Bugge empha-
sized the difference of the cephalic arterial
system between castorids and sciurids and
believed that a close relationship of these two
groups is improbable. As demonstrated by
Bugge, castorids retain the medial distal in-
ternal carotid artery but lose the stapedial
artery while sciurids lose the promontory ar-
tery but retain the stapedial artery. Luckett
and Hartenberger (1985) argued, citing Wah-
lert (1977), that the stapedial artery does oc-
cur in some fossil castoroids. However, the
existence of the stapedial artery is a primitive
character in rodents. In addition, the pathway
of the stapedial artery in the castoroid Eu-
typomys (Wahlert, 1977) is primitive, cours-
ing along the ventral rim of the fenestra ro-
tunda, instead of across the fenestra rotunda
as in sciurids. Therefore, if castorids share
any affinity to sciurids and aplodontids, as
suggested by Wilson (1949), they must have
diverged well before the origin of reithropara-
myines. In other words, castorids and sci-
urids, if related in some way, must have a
very distant relationship. The presence of the
medial distal internal carotid artery in cas-
torids may turn out to be a clue for castorid
phylogeny. It is more likely that castorids are
derived from an ancestral stock giving rise to
muroids than to sciurids. Nonetheless, the
uncertain position of Castoridae is well re-
flected in figures 4-7.

CTENODACTYLOIDEA. The living Ctenodac-
tylidae have remained until recently a group
of uncertain position relative to other rodents
(Simpson, 1945; Dawson et al., 1984; Li et
al., 1989). On the other hand, Luckett (1980,
1985) and George (1985) suggested a sister-
group relationship of recent ctenodactylids
with histricognathi. This relationship is
strongly supported by auditory features as
shown in figures 4—7, although this grouping
may go too far by combining Ctenodactylidae
with Caviomorpha as a sister group of
Thryonomyidae (figs. 4, 5, and 7). However,
I would like to see this as a sign for a more
general scheme that Ctenodactylidae are

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 27

closely related to Hystricognathi as shown in
figure 6. As an alternative view, Flynn et al.
(1986) suggested that Ctenodactyloidea plus
Pedetidae are the sister group of Hystrico-
gnathi.

Problems arise when early ctenodactyloids
are included. Fossil ctenodactyloids have been
extensively described from the Tertiary of
Asia (Shevyreva, 1972; Dawson, 1977; Sahni
and Srivastava, 1977; Hartenberger, 1977,
1980, 1982; Hussain et al., 1978; Li et al.,
1979; Dasheveg, 1982; de Bruijn et al., 1982;
Dawson et al., 1984). It has been noted that
the early ctenodactyloids are the rodents most
similar to the eurymylid Heomys, the mam-
mal which is closest to the direct ancestor to
the rodents (Li, 1977; Gingerich and Gun-
nell, 1979; Chaline and Mein, 1979; Harten-
berger, 1980; Dawson et al., 1984; Dawson
and Krishtalka, 1984; Korth, 1984; Harten-
berger, 1985; Li et al., 1987; Li et al., 1989).

Cocomys, the best preserved ctenodacty-
loid, has almost all those primitive character
conditions in its ear region, except for the
stapedius muscle which was probably en-
closed within the tympanic cavity and the
facial nerve that is partially ossified (Li et
al., 1989). Although theridomorphs also have
primitive ear regions, they are not so prim-
itive as Cocomys. For example, therido-
morphs (Adelomys, for instance) have an in-
flated bulla (Lavocat, 1967: pl. 2), whereas
Cocomys has only a small elliptical bulla
which is completely confined to the ventral
side of the skull. Because the bulla is very
poorly developed, it is not impossible that
the stapedius muscle in Cocomys was partly
exposed at the posterior end of the tympanic
cavity. Moreover, Cocomys possesses some
primitive characters that are either present
in the genus alone or retained from ancestral
eutherian condition but never occur in other
rodents (Li et al. 1989). Among these are two
auditory features: large epitympanic wing of
petrosal and large pyriform fenestra. Coco-
mys has perhaps the most primitive auditory
region among rodents. When Cocomys (fam-
ily Cocomyidae) and Ctenodactylidae are
placed in the same superfamily Ctenodacty-
loidea, it becomes difficult to discuss the re-
lationship of the superfamily to other ro-
dents, because it contains the most primitive
as well as some of more derived rodents. Re-

28 AMERICAN MUSEUM NOVITATES

cently, Flynn et al. (1986) excluded Cocomys
from Ctenodactyloidea because it is protro-
gomorphous and lacks hypolophids. Li et al.
(1989) considered that Cocomys shares only
a few cranial features (deep pterygoid fossa
and palatal process of palatine extending more
anteriorly) with later ctenodactylids, such as
Tataromys and Ctenodactylus. They there-
fore conclude (ibid.) that this may be taken
as evidence to support Wilson’s assumption
that “recent Ctenodactylidae are still rather
much of an incertae sedis.”

In considering the primitiveness of the Co-
comys and the close relationship of the
Ctenodactylidae with the Hystricognathi, an
early dichotomy in rodent phylogeny be-
tween Asiatic cocomyids and North Ameri-
can-European paramyids (Hartenberger,
1980; Luckett and Hartenberger, 1985) ap-
pears to gain support. The cocomyids may
have given rise to the later ctenodactyloids
in the Asiatic area, from which the Cteno-
dactylidae and the Hystricognathi were likely
descended. The cocomyid stock also may
have given rise to the ischyromyids, the basal
stock for some later North American and Eu-
ropean groups. Such a dichotomy emerges in
figures 4-7, which can be compared to that
of Luckett and Hartenberger (1985: fig. 2).
Cocomys itself is probably too derived to be
the direct ancestor of ischyromyids in at least
one respect, the relatively larger infraorbital
foramen. Vianey-Liaud (1985) termed this
““pre-hystricomorphy” and believed that my-
omorphy is derived from this condition.

ANOMALURIDAE AND PEDETIDAE. These two
families are grouped together in this analysis.
This is consistent with the result derived from
analysis of the arterial pattern (Bugge, 1974b,
1985; George, 1981). Affinities of pedetids
and anomalurids are neither supported nor
contradicted by fetal membrane data (Luck-
ett, 1985). Hartenberger (1980) suggested a
possible relationship of Pedetidae and
Anomaluridae with Hystricognathi, but re-
cently (1985) he provided a theridomyid-
anomalurid grouping and placed Pedetidae
as incertae sedis. Flynn et al. (1986) placed
Pedetidae with Ctenodactyloidea. A close re-
lationship of anomalurids and pedetids with
Phiomorpha was suggested by Lavocat and
Parent (1985), but Wood (1985) held that the

NO. 2972

cheek-tooth patterns of anomalurids and
pedetids have nothing in common and that
there is no valid basis for assuming any re-
lationships between the anomalurids and
pedetids on the one hand and the Hystricog-
nathi on the other. George (1985) distin-
guished a sciuromorph-myomorph clade and
a ctenodactylid-hystricognath clade, and sug-
gested that anomalurids and pedetids do not
associate readily with either of them. Alter-
natively, Jaeger (1988) tentatively considered
Anomaluridae as a sister group of the Ther-
idomyidae. Auditory features support an
anomalurid-pedetid sister group, but its re-
lationships with other groups vary consid-
erably (figs. 4—7).

MuRrRoIpEa. A close relationship of Muroi-
dea and Dipodoidea has been suggested on
various grounds (Klingener, 1964; Bugge,
1971a; Emry, 1981; Vianey-Liaud, 1985;
Flynn et al., 1985; Hartenberger, 1985; Luck-
ett, 1985; Sarich, 1985). Due to their prim-
itiveness, auditory features contribute little
to the discussion of this proposed relation-
ship. It has been argued that muroids may
be derived from North American sciuravids
(Wood, 1959; Korth, 1984). According to
Luckett and Hartenberger (1985), however,
the unique pattern of apparent “‘absence”’ of
the medial internal carotid artery, and the
occurrence of a promontory artery in sci-
uravids (Wahlert, 1974), would seem to con-
tradict an ancestral-descendant relationship
between sciuravids and muroids. The prem-
ise of this statement is that the classical con-
sideration of the internal carotid system is
correct (see above). If the internal carotid ar-
tery is but a single vessel and if its lateral
placement is primitive in rodents, as as-
sumed in this paper, then the ancestral-de-
scendant relationship between sciuravids and
muroids cannot be ruled out by this partic-
ular character.

An alternative relationship is one between
ctenodactyloids and muroids (Vianey-Liaud,
1985; Flynn et al., 1985). This is based on
the observation that early muroids display a
change of the infraorbital region from the
hystricomorphous pattern to the typical my-
omorphous condition (Vianey-Liaud, 1979,
1985). Hystricomorphy (or pre-hystricomor-
phy) characterizes early ctenodactyloids.

1990

CONCLUSIONS

Some of the points discussed above are
long-standing problems in rodent systemat-
ics. Features from the auditory region of R.
delicatissimus may offer data to be incorpo-
rated in future discussion of rodent phylo-
genetics. The salient systematic conclusions
may be summarized as follows.

1. ‘““Franimorpha” is not a monophyletic
group because no derived character supports
such a group. The incipient hystricognathy
of “‘Franimorpha” is ambiguous and not
widely accepted. Even if this condition exists,
it is probably not unique to “‘Franimorpha.”
On the other hand, reithroparamyines pos-
sess some shared derived characters with sci-
urids and aplodontids, such as the meato-
cochlear bridge and the internal carotid
artery over the fenestra rotundum. In addi-
tion, Protoptycus acquired hystricomorphy
independently and Cylindrodontidae have an
ear region that is derived in a different di-
rection. These diverse morphologies indicate
that these “franimorph” groups are more
likely on different phylogenetic lines. Con-
sequently, reithroparamyines can hardly be
ancestral to hystricognaths nor can they be
included in Hystricognathi.

2. Reithroparamyines are protrogomor-
phous rodents that have several derived fea-
tures compared to paramyids and other is-
chyromyids and share only primitive
characters with the latter. Little evidence sup-
ports the placement of Reithroparamyinae,
as the ancestral stock, in Paramyidae or Is-
chyromyidae. Instead, reithroparamyines
share derived characters with Sciuridae,
Aplodontidae, and Gliridae. Although a re-
vised classification is not attempted because
the auditory features are only part of the total
evidence, these features do reveal certain spe-
cialization in reithroparamyines. It is not im-
possible that reithroparamyines are included
in a clade consisting of more derived groups
such as Sciuridae and Aplodontidae, instead
of in a grade of primitive protrogomorphs.

3. Finally, it is herein suggested that rei-
throparamyines be separated from ischyro-
myids and recognized as a family Reithro-
paramyidae (Patterson and Wood, 1982),
including only the genera and most of the

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 29

species recognized by Korth (1984). This
family may be defined as protrogomorphous
rodents with some derived features: P3 re-
duced and P4 molariform; incisors more lat-
erally compressed, with the anterior surface
flat; masseteric fossa bounded by ridges and
terminating more anteriorly on the side of
the mandible; snout short and tapered; pre-
maxillaries extended to or beyond the pos-
terior margin of nasals; zygoma progressively
moved forward until it is in line with anterior
margin of P4; postglenoid foramen reduced;
bulla co-ossified with skull, with internal sep-
ta formed; bony tube for stapedial artery and
facial nerve partially formed; stapedial artery
crossing fenestra rotunda; meato-cochlear
bridge developed; epitympanic recess ex-
panded in the ectotympanic.

REFERENCES

Archibald, J. D.

1977. Ectotympanic bone and internal carotid
circulation of eutherians in reference to
anthropoid origins. J. Hum. Evol. 6:
609-670.

Oldest known eutherian stapes and a
marsupial petrosal bone from the Late
Cretaceous of North America. Nature
281: 669-670.
Beintema, J. J.
1985. Amino acid sequence data and evolu-
tionary relationships among histrico-
gnaths and other rodents. Jn W. P.
Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
549-565. New York and London: Ple-
num Press.
Black, C. C.
1971. Paleontology and geology of the Bad-
water Creek area, central Wyoming. Part
7. Rodents of the Family Ischyromidae.
Ann, Carnegie Mus. 43: 179-217.
Black, C. C., and J. F. Sutton
1984. Paleocene and Eocene rodents of North
America. Spec. Publ. Carnegie. Mus.
Nat. Hist. 9: 67-84.
Bruijn, H. de, S. T. Hussain, and J. J. M. Leinders
1982. On some Early Eocene rodent remains
from Barbara Banda, Kohat, Pakistan,
and the early history of the order Ro-
dentia. Proc. Kon. Ned. Akad. We-
tensch., ser. B, 85: 249-258.
Buchanan, G. D., and A. A. Arata
1969. Cranial vasculature of a neotropical

1979.

30 AMERICAN MUSEUM NOVITATES

fruit-eating bat, Artibeus lituratus. Anat.
Anz. 124: 314-325.
Bugge, J.

1971a. The cephalic arterial system in mole-
rats (Spalacidae) bamboo rats (Rhizo-
myidae), jumping mice and jerboas (Di-
podoidea) and dormice (Gliroidea) with
special reference to the systematic clas-
sification of rodents. Acta Anat. 79: 165—
180.

1971b. The cephalic arterial system in sciuro-
morphs with special reference to the
systematic classification of rodents. Acta
Anat. 80: 336-361.

1974a. The cephalic arteries of hystricomorph
rodents. Symp. Zool. Soc. London 34:
61-78.

1974b. The cephalic arterial system in insecti-

vores, primates, rodents and lago-
morphs, with special reference to the
systematic classification. Acta Anat. 87
(Suppl. 62): 1-160.

Systematic value of the carotid arterial
pattern in rodents. Jn W. P. Luckett and
J.-L. Hartenberger (eds.), Evolutionary
relationships among rodents: a multi-
disciplinary analysis, pp. 355-379. New
York and London: Plenum Press.
Cartmill, M., and R. D. E. MacPhee

1980. Tupaiid affinities: the evidence of the
carotid arteries and cranial skeleton. Jn
W. P. Luckett (ed.), Comparative biol-
ogy and evolutionary relationships of
tree shrews, pp. 95-132. New York: Ple-
num Press.

Chaline, J., and P. Mein

1979. Les Rongeurs et l’Evolution. Doin Ed-

iteurs, Paris.
Cifelli, R. L.

1982. The petrosal structure of Hyopsodus with
respect to that of some other ungulates,
and its phylogenetic implications. J. Pa-
leontol. 56: 795-805.

Clemens, W.

1966. Fossil mammals of the type Lance For-
mation, Wyoming. Part 2. Marsupialia.
Univ. California Publ. Geol. Sci., 62: 1-
122%

Conroy, G. C., and J. R. Wible

1978. Middle ear morphology of Lemus var-
iegatus, some implications for primate
paleontology. Folia Primatol. 29(2): 81-
85.

Coombs, M. C., and W. P. Coombs

1982. Anatomy ofthe ear region of four Eocene
artiodactyls: Gobiohyus, ?Helohyus,
Diacodexis and Homacodon. J. Vert.
Paleontol. 2: 219-236.

1985.

NO. 2972

Dashzeveg, D.

1982. La faune de Mammiferes du Paleogene
inferieur de Naran-Bulak (Asie Cen-
trale) et ses correlations avec l’Europe
et l Amerique du Nord. Bull. Soc. Geol.
France, ser. 7, 24(2): 275-281.

Dawson, M. R.

1961. The skull of Sciuravus nitidus, a middle
Eocene rodent. Postilla, no. 53.
1977. Late Eocene rodent radiations: North

America, Europe and Asia. Geobios
Mem. Spec. 1: 195-209.
Dawson, M. R., and L. Krishtalka
1984. Fossil history of the families of recent
mammals. Jn S. Anderson and J. K.
Jones, Jr. (eds.). Orders and families of
recent mammals of the world, pp. 11-
57. New York: Wiley.
Dawson, M. R., C.-K. Li, and T. Qi
1984. Eocene ctenodactyloid rodents (Mam-
malia) of Eastern and Central Asia. Car-
negie Mus. Nat. Hist. Spec. Publ. 9: 138-
150.
DeBeer, G. R.
1937. The development of the vertebrate skull.
Oxford: Clarendon Press, xxiv + 552

pp.
De Jong, W. W.

1985. Superordinal] affinities of Rodentia
studied by sequence analysis of eye lens
protein. Jn W. P. Luckett and J.-L. Har-
tenberger (eds.), Evolutionary relation-
ships among rodents: a multidisciplin-
ary analysis, pp. 211-266. New York
and London: Plenum Press.

Emry, R. J.

1981. New material of the Oligocene muroid
rodent Nonomys, and its bearing on mu-
roid origins. Am. Mus. Novitates 2712:
14 pp.

Emry, R., and R. W. Thorington, Jr.

1982. Descriptive and comparative osteology
of the oldest fossil squirrel, Protosciurus
(Rodentia, Sciuridae). Smithsonian
Contrib. Paleobiol. 7: 1-35.

Farris, J. S.
1988. HENNIG 86 (version 1.5) Reference.

Ferrusquia- Villafranca, I.

1989. A new rodent genus from central Mex-
ico and its bearing on the origin of the
Caviomorpha. Jn C. C. Black and M.
R. Dawson (eds.), Papers on fossil ro-
dents—in honor of Albert Elmer Wood,
pp. 92-117. Science Series no. 33, Nat.
Hist. Mus. Los Angeles County.

Fischer, T. V., and H. W. Mossman

1969. The fetal membranes of Pedetes capen-
sis, and their taxonomic significance.
Am. J. Anat. 116: 29-68.

1990

Flynn, L. J., L. L. Jacobs, and I. U. Cheema

1986. Baluchimyinae, a new Ctenodactyloid

rodent subfamily from the Miocene of
Baluchistan. Am. Mus. Novitates 2841:
58 pp.

Flynn, L. J., L. L. Jacobs, and E. H. Lindsay

1985. Problems in muroid phylogeny: rela-

tionship to other rodents and origin of

major groups. Jn W. P. Luckett and J.-

L. Hartenberger (eds.), Evolutionary re-

lationships among rodents: a multidis-

ciplinary analysis, pp. 589-616. New

York and London: Plenum Press.
George, W.

1981. Blood vascular patterns in rodents:
Contributions to an analysis of rodent
family relationships. Zool. J. Linn. Soc.
73: 287-306.

Reproductive and chromosomal char-
acters of ctenodactylids as a key to their
evolutionary relationships. Jn W. P.
Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
453-474. New York and London: Ple-
num Press.
Gingerich, P. D., and G. F. Gunnell
1979. Systematics and evolution of the genus
Esthonyx (Mammalia, Tillodontia) in
the early Eocene of North America.
Contrib. Mus. Paleontol., Univ. Mich-
igan 25: 125-153.
Goodrich, E.
1930. Structure and development of the ver-
tebrates. London: MacMillan.
Griffiths, M.

1978. The biology of the monotremes. New

York: Academic Press.
Guthrie, D. A.

1963. The carotid circulation in the Rodentia.
Bull. Mus. Comp. Zool. Harvard 128:
445-481.

Hartenberger, J.-L.

1971. Contribution a letude des genres Gii-
ravus et Microparamys (Rodentia) de
Eocene d’Europe. Palaeovertebrata 4:
97-135.

A propos de l’origine des rongeurs. Geo-
bios, Mem. Spec. 1: 183-193.
Données et hypothéses sur la radiation
initiale des rongeurs. Palaeovertebrata
Mem. Jubil. R. Lavocat: 285-301.

A review of the Eocene rodents of Pa-
kistan. Contrib. Mus. Paleontol., Univ.
Michigan 26(2): 19-35.

La Grande Coupure. Pour La Science
67: 26-38.

The Order Rodentia: Major questions
on their evolutionary origin, relation-

1985.

1977.

1980.

1982.

1983.

1985.

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 31

ships and suprafamilial systematics. In
W. P. Luckett and J.-L. Hartenberger
(eds.), Evolutionary relationships among
rodents: a multidisciplinary analysis, pp.
1-33. New York and London: Plenum
Press.

Hussain, S. T., H. de Bruijn, and J. M. Leinders

1978. Middle Eocene rodents from the Kala
Chitta Range (Punjab, Pakistan). Proc.
Kon. Ned. Akad. Wetensch., Amster-
dam, ser. B, 81: 74-112.
Jaeger, J.-J.
1988. Rodent phylogeny: new data and old

problems. Jn M. J. Benton (ed.), The
phylogeny and classification of the tet-
rapods, vol. 2: Mammals. Syst. Assoc.
Spec. Vol. 35B: 177-199.
Jaeger, J.-J., C. Denys, and B. Coiffait
1985. New Phiomorpha and Anomaluridae
from the Late Eocene of North-West Af-
rica: phylogenetic implications. Jn W.
P, Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
567-588. New York and London: Ple-
num Press.
Jepsen, G. L.
1937. A Paleocene rodent, Paramys alavus.
Proc. Am. Philos. Soc. 78: 291-301.
Klaauw, C. J. van der

1931. The auditory bulla in some fossil mam-
mals. Bull. Am. Mus. Nat. Hist. 62: 1-
352.
Klingener, D.
1964. The comparative myology of four di-

podoid rodents (genera Zapus, Napeo-
zapus, Sicista, and Jaculus). Mus. Zool.
Univ. Michigan Misc. Publ. 124: 1-100.
Korth, W. W.
1984. Earliest Tertiary evolution and radia-
tion of rodents in North America. Bull.
Carnegie Mus. 24: 1-71.
Lavocat, R.
1956. Reflexions sur la classification des ron-
geurs. Mammalia 20: 49-56.

1967. Observations sur la région auditive des
Rongeurs Théridomorphes. Problemes
actuels de Paleontologie, Colloque In-
ternat. C.N.R.S. 611: 491-501.

1973. Les Rongeurs du Miocéne d’Afrique

Orientale. I. Miocéne inférieur.
E.P.H.E., Inst. Montpellier, Mem. 1: 1-
284.

1974a. The interrelationships between the Af-
rican and South American rodents and
their bearing on the problem of the or-
igin of South American monkeys. J.
Hum. Evol. 3: 323-326.

32 AMERICAN MUSEUM NOVITATES

1974b. What is an histricognath? Symp. Zool.
Soc. London 34: 7-19.

1976. Rongeurs caviomorphes de l’Oligocéne
de Bolivie. IJ. Rongeurs du Bassin De-
seadien de Salla-Luribay. Palaeoverte-
brata 7: 15-90.

1980. The implications of rodent paleontol-

ogy and biogeography to the geograph-
ical sources and origin of latyrrhine pri-
mates. In R. L. Ciochon and A. B.
Chiarelli (eds.), Evolutionary biology of
the new world monkeys and continental
drift, pp. 93-102. New York: Plenum
Press.
Lavocat, R., and J.-P. Parent
1985. Phylogenetic analysis of middle ear fea-
tures in fossil and living rodents. Jn W.
P. Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
333-354. New York and London: Ple-
num Press.
Li, C.-k.
1977. Paleocene eurymyloids (Anagalida,
Mammalia) of Qianshan, Anhui. Ver-
tebr. PalAsiat. 15: 103-118.
Li, C.-k., C.-s. Chiu, D.-f. Yan, and S.-m. Hsieh
1979. Notes on some Early Eocene mamma-
lian fossils of Hengtong, Hunan. Ver-
tebr. PalAsiat. 17: 71-82.
Li, C.-k., and S.-y. Ting

1984. The Paleogene mammals of China. Bull.
Carnegie Mus. Nat. Hist. 21: 1-98.
1985. Possible phylogenetic relationship of

Asiatic eurymylids and rodents, with
comments on mimotonids. In W. P.
Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
35-58. New York and London: Plenum
Press.
Li, C.-k., R. W. Wilson, M. R. Dawson, and L.
Krishtalka
1987. The origin of rodents and lagomorphs.
In H. H. Genoways (ed.), Current mam-
malogy 1, pp. 97-108. New York: Ple-
num Press.
Li, C-k., J.-j. Zheng, and S.-y. Ting
1989. The skull of Cocomys lingchaensis, an
early Eocene ctenodactyloid rodent of
Asia. In C. C. Black and M. R. Dawson
(eds.), Papers on fossil rodents—in hon-
or of Albert Elmer Wood, pp. 180-192.
Science Series no. 33, Nat. Hist. Mus.
Los Angeles County.
Lopez, N., and L. Thaler
1974. Sur le plus ancien Lagomorphe euro-
peen et la ““Grande Coupure”’ Oligocene
de Stehlin. Palaeovertebrata 6: 243-251.

NO. 2972

Luckett, W. P.

1971. The development of the chorio-allan-
toic placenta of the African scaly-tailed
squirrels (Family Anomaluridae). Am.
J. Anat. 130: 159-178.

Monophyletic or diphyletic origin of
Anthropoidea and Hystricognathi: evi-
dence of the fetal membranes. Jn R. L.
Ciochon and A. B. Chiarelli (eds.), Evo-
lutionary biology of the new world mon-
keys and continental drift, pp. 347-368.
New York: Plenum Press.
Superordinal and intraordinal affinities
of rodents: developmental evidence
from the dentition and placentation. In
W. P. Luckett and J.-L. Hartenberger
(eds.), Evolutionary relationships among
rodents: a multidisciplinary analysis, pp.
227-276. New York and London: Ple-
num Press.
Luckett, W. P., and J.-L. Hartenberger
1985. Evolutionary relationships among ro-
dents: comments and conclusions. Jn W.
P. Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
685-712. New York and London: Ple-
num Press.
Luckett, W. P., and H. W. Mossman
1981. Development and phylogenetic signifi-
cance of the fetal membranes and pla-
centa of the African histricognathous
rodents Bathyergus and Hystrix. Am. J.
Anat. 162: 265-285.
Matthew, W. D.
1909. The Carnivora and Insectivora of The
Bridger Basin, Middle Eocene. Mem.
Am. Mus. Nat. Hist. 9: 289-567.
MaclIntyre, G. T.
1972. The trisulcate petrosal pattern of mam-
mals. In T. Dobzhansky, M. K. Hecht,
and W. Steere (eds.), Evol. Biol. 6: 275-
303.
MacPhee, R. D. E.
1981. Auditory regions of primates and eu-
therian insectivores. Morphology, on-
togeny, and character analysis. Jn F. S.
Szalay (ed.), Contributions to primatol-
ogy, vol. 18. Basel: S. Karger, xii + 282

1980.

1985.

pp.
MacPhee, R. D. E., M. J. Novacek, and G. Storch
1988. Basicranial morphology of early Ter-
tiary erinaceomorphs and the origin of
Primates. Am. Mus. Novitates 2921: 42
pp.
McDowell, S. B.
1958. The Great Antillean insectivores. Bull.
Am. Mus. Nat. Hist. 115: 113-214.

1990 MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS 33

McKenna, M. C.

1961. A note on the origin of rodents. Am.
Mus. Novitates 2037: 5 pp.

1966. Paleontology and the origin of the pri-
mates. Folia Primatol. 4: 1-25.

Novacek, M. J.

1977. Aspects of the problem of variation, or-
igin, and evolution of the eutherian au-
ditory bulla. Mammal Rev., 7: 131-149.

1980. Cranioskeletal features in tupaiids and

selected Eutheria as phylogenetic evi-
dence. In W. P. Luckett, (ed.), Com-
parative biology and evolutionary re-
lationships of tree shrews, pp. 35-93.
New York: Plenum Press.

Cranial evidence for rodent affinities. Jn
W. P. Luckett and J.-L. Hartenberger
(eds.), Evolutionary relationships among
rodents: a multidisciplinary analysis, pp.
59-81. New York and London: Plenum
Press.

1986. The skull of leptictid insectivorans and
the higher-level classification of euthe-
rian mammals. Bull. Am. Mus. Nat.
Hist. 183(1): 1-111.

1989. Higher mammal phylogeny: the mor-
phological-molecular synthesis. Jn B.
Fernholm, K. Bremer, and H. Jornvall
(eds.), The hierarchy of life, pp. 421-
435. Elsevier Science Publishers B. V.
(Biomedical Division).

Novacek, M. J., and A. R. Wyss

1986. Higher-level relationships of the recent
eutherian orders: morphological evi-
dence. Cladistics 2(3): 257-286.

Novacek, M. J., A. R. Wyss, and M. C. McKenna

1988. The major groups of eutherian mam-
mals. In M. J. Benton (ed.), The phy-
logeny and classification of the tetra-
pods, vol. 2: Mammals. Syst. Assoc.
Spec. Vol. no. 35B, pp. 31-71. Oxford:
Clarendon Press.

Parent, J.-P.

1980. Recherches sur loreille moyenne des
Rongeurs actuels et fossiles. Anatomie.
Valeur systématique. Mém. Trav.
E.P.H.E., Montpellier 11: 1-286.

1983. Anatomie et valeur systématique de
loreille moyenne des Rongeurs actuels
et fossiles. Mammalia 47: 93-122.

Patterson, B., and A. E. Wood

1982. Rodents from the Deseadan Oligocene
of Bolivia and the relationships of the
Caviomorpha. Bull. Mus. Comp. Zool.
149: 371-543.

Platnick, N. I.

1989. Anempirical comparison of microcom-
puter parsimony programs, II. Cladis-
tics 5: 145-161.

1985.

Presley, R.
1979. The primitive course of the internal ca-
rotid artery in mammals. Acta Anat.
103: 138-144.
Prothero, D. R.

1983. The oldest mammalian petrosals from
North America. J. Paleontol. 57: 1040—-
1046.
Pye, A.

1977. The structure of the cochlea in some
myomorph and caviomorph rodents. J.
Zool. (London) 182: 309-321.
1979. The structure of the cochlea in some
mammals. J. Zool. (London) 187: 39-
53.
Romer, A. S.
1956. Osteology of the reptiles. Chicago: Univ.
Chicago Press.
Romer, A. S., and S. P. Thomas

1977. The vertebrate body, 5th ed. Philadel-
phia: W. B. Saunders.
Rose, K. D.
1981. The Clarkforkian land-mammal age and

mammalian faunal composition across

the Paleocene-Eocene boundary. Pap.

Paleontol. Univ. Michigan 26: 1-197.
Sahni, A., and M. C. Srivastava

1977. Eocene rodents of India: their palaeo-
biogeographic significance. Geobios,
Mem. Spec. 1: 87-95.

Sarich, V. M.

1985. Rodent macromolecular systematics. In
W. P. Luckett and J.-L. Hartenberger
(eds.), Evolutionary relationships among
rodents: a multidisciplinary analysis, pp.
423-452. New York and London: Ple-
num Press.

Schaub, S.

1953. Remarks on the distribution and clas-
sification of the ““Hystricomorpha.”’
Verh. Naturforsch. Ges. Basel 64: 389-—
400.

Shevyreva, N.

1972. New rodents in the Paleogene of Mon-
golia and Kazakhstan. Paleontol. J.
(Moscow) 3: 134-145.

Shoshani, J.,. M. Goodman, J. Czelusniak, and G.

Braunitzer

1985. A phylogeny of Rodentia and other eu-
therian orders: parsimony analysis uti-
lizing amino acid sequences of alpha and
beta hemoglobin chains. Jn W. P. Luck-
ett and J.-L. Hartenberger (eds.), Evo-
lutionary relationships among rodents:
a multidisciplinary analysis, pp. 191-
210. New York and London: Plenum
Press.

Simpson, G. G.
1945. The principles of classification and a

34

AMERICAN MUSEUM NOVITATES

classification of mammals. Bull. Am.
Mus. Nat. Hist. 85: 1-350.

Stehlin, H. G., and S. Schaub

1951. Die trigonodontie der simplicidentaten
Nager. Schweiz. Palaeontol. Abh. 67: 1-
385.

Swofford, D.

1989. Phylogenetic Analysis Using Parsimony

(PAUP) (version 3.0) for Macintosh PC.
Szalay F. S.

1972. Cranial morphology of the early Ter-
tiary Phenacolemur and its bearing on
primate phylogeny. Am. J. Phys. An-
thropol. 36: 59-76.

1975. Phylogeny of primate higher taxa: the
basicranial evidence. Jn W. P. Luckett
and F. S. Szalay (eds.), Phylogeny of the
primates: a multidisciplinary approach,
pp. 91-125. New York: Plenum Press.

Tandler, J.
1899. Zur vergleichenden anatomie der kop-

farterien bei den Mammalia. Denkschr.
Acad. Wiss. (Wien) 67: 677-784.

Ting, S.-Y., and C.-k, Li

1984.

1985.

Tullberg,
1899.

The structure of the ear region of Rhom-
bomylus (Anagalia, Mammalia). Ver-
tebr. PalAsiat. 22: 92-102.

Possible phylogenetic relationship of
Asiatic eurymylids and rodents, with
comments on mimotonids. Jn W. P.
Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
35-58. New York and London: Plenum
Press.

T.

Uber das System der Nagethiere: eine
phylogenetische Studie. Nova Acta Reg.
Soc. Scient. Uppsala, ser. 3, 18: 1-514.

Van Valen, L.

1965.

1966.

Treeshrews, primates and fossils. Evo-
lution 19: 137-151.

Deltatheridia, a new order of mammals.
Bull. Am. Mus. Nat. Hist. 132: 1-126.

Vianey-Liaud, M.

1974a.

1974b.

1975.

Palaeosciurus goti nov. sp., écureuil ter-
restre de l’Oligocéne moyen du Quercy.
Données nouvelles sur l’apparition des
Sciuridés en Europe. Ann. Paleontol.
Vertebr. (Paris) 60: 103-122.

Les Rongeurs de l’Oligocéne infeérieur
d’Escamps. Palaeovertebrata 6: 197-
241.

Caractéristiques évolutives des Ron-
geurs de l’Oligocéne d’Europe Occiden-
tale. Coll. Internat. C.N.R.S., Paris 218:
765-776.

1979.

1985.

NO. 2972

L’évolution des Rongeurs a |’Oligocéne
en Europe occidentale. Paleontographi-
ca 166: 136-236.

Possible evolutionary relationships
among Eocene and Lower Oligocene ro-
dents of Asia, Europe and North Amer-
ica. Jn W. P. Luckett and J.-L. Harten-
berger (eds.), Evolutionary relationships
among rodents: a multidisciplinary
analysis, pp. 277-309. New York and
London: Plenum Press.

Wahlert, J. H.

1973.

1974.

1977.

1978.

198Sa.

1985b.

Wible, J. R

1983.

1984.

1986.

1987.

Protoptychus, a hystricomorphous ro-
dent from the late Eocene of North
America. Breviora 419: 1-14.

The cranial foramina of protrogomor-
phous rodents, an anatomical and phy-
logenetic study. Bull. Mus. Comp. Zool.
146: 363-410.

Cranial foramina and relationships of
Eutypomys (Rodentia, Eutypomyidae).
Am. Mus. Novitates 2626: 8 pp.
Cranial foramina and relationship of the
Eomyoidea (Rodentia, Geomorpha).
Skull and upper teeth of Kansasimys.
Am. Mus. Novitates 2645: 16 pp.
Skull morphology and relationships of
geomyoid rodents. Am. Mus. Novitates
2812: 20 pp.

Cranial foramina of rodents. Jn W. P.
Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
311-332. New York and London: Ple-
num Press.

The internal carotid artery in early eu-
therians. Acta Palaeontol. Polonica 28
(1, 2): 281-293.

The ontogeny and phylogeny of the
mammalian cranial arterial pattern.
Ph.D. diss., Duke University, Durham,
705 pp.

Transformation in the extracranial
course of the internal carotid artery in
mammalian phylogeny. J. Vertebr. Pa-
leontol. 6(4): 313-325.

The eutherian stapedial artery: charac-
ter analysis and implications for super-
ordinal relationships. Zool. J. Linn. Soc.
91: 107-135.

Wilson, R. W.

1949.

1986.

Early Tertiary rodents of North Amer-
ica. Carnegie Inst. Washington Publ.
584: 67-164.

The Paleogene record of rodents: facts
and interpretation. Jn K. M. Flanagan

1990

Wood, A.
1937.

1955.
1959.

1962.

1973.

1974.

1975.

1977.

1980.

MENG: AUDITORY REGION OF REITHROPARAMYS DELICATISSIMUS

and J. A. Lillegraven (eds.), Vertebrates,
phylogeny and philosophy. Contrib.
Geol., pp. 163-176. Univ. Wyoming.
E.

The mammalian fauna of the White
River Oligocene. Part II. Rodentia.
Trans. Am. Philos. Soc., n. ser., 28: 155-
269.

A revised classification of the rodents.
J. Mammal. 36: 165-187.

Eocene radiation and phylogeny of the
rodents. Evolution 13: 354-361.

The early Tertiary rodents of the family
Paramyidae. Trans. Am. Philos. Soc. 52:
1-261.

Eocene rodents, Pruett Formation,
southwest Texas; their pertinence to the
origin of the South American Cavio-
morpha. Texas Mem. Mus., Pearce-Sel-
lards ser. 20: 1-40.

The evolution of the Old World and
New World histricomorphs. Symp. Zool.
Soc. London 34: 21-60.

The problem of the histricognathous ro-
dents. In G. R. Smith and N. E. Fried-
land (eds.), Studies on Cenozoic pa-
leontology and stratigraphy in honor of
Claude W. Hibbard Memorial Vol., 3.
Univ. Michigan Papers Paleontol. 12:
75-80.

The evolution of the rodent family
Ctenodactylidae. J. Palaeontol. Soc. In-
dia 20: 120-137.

The Oligocene rodents of North Amer-
ica. Trans. Am. Philos. Soc. 70: 1-68.

1981.

1983.

1984.

1985.

Zeller, U.
1987.

35

The origin of the caviormorph rodents
from a source in Middle America: a clue
to the area of origin of the platyrrhine
primates. Jn R. L. Ciochon and A. B.
Chiarelli (eds.), Evolutionary biology of
the new world monkeys and continental
drift, pp. 79-91. New York: Plenum
Press.

The radiation of the Order Rodentia in
the southern continents; the dates, num-
bers and sources of the invasions. Jn W.-
D. Heinrich (ed.), Wirbeltier-Evolution
und Faunenwandel in Kanozoicum, pp.
381-393. Schriftener. Geol. Wiss. 19/
20, Berlin.

Hystricognathy in the North America
Oligocene rodent Cylindrodon and the
origin of the Caviomorpha. In: R. M.
Mengel (ed.), Papers in vertebrate pa-
leontology honoring Robert Warren
Wilson, pp. 151-160, Carnegie Mus.
Nat. Hist., Spec. Publ. 9, Pittsburgh.
The relationships, origin and dispersal
of the histricognathous rodents. In W.
P. Luckett and J.-L. Hartenberger (eds.),
Evolutionary relationships among ro-
dents: a multidisciplinary analysis, pp.
475-513. New York and London: Ple-
num Press.

Morphogenesis of the mammalian skull
with special reference to Tupaia. In H.-
J. Kuhn and U. Zeller (eds.), Morpho-
genesis of the mammalian skull. Mam-
malia Depicta. Berlin: Verlag Paul Parey.

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the
Novitates, Bulletin, and Anthropological Papers published during the last five years are available

free of charge. Address orders to: American Museum of Natural History Library, Department D,
Central Park West at 79th St., New York, N.Y. 10024.

THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER.