NovttatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2890, pp. 1-17, figs. 1-7, tables 1-3 September 4, 1987

Notes on Bolivian 3: A Revised Diagnosis of Andalgalomys (Rodentia, Muridae) and the Description of a New Subspecies

NANCY OLDS,' SYDNEY ANDERSON,2 AND TERRY L. YATES3

ABSTRACT Reported characters ofthe phyllotine genus An- range northward approximately 300 km. Karyo- dalgalomys are reviewed and a revised diagnosis types for both subspecies ofA. pearsoni are: A. p. is presented. Andalgalomys is phenetically more pearsoni has 78 chromosomes and A. p. n. ssp. 76. similar to Graomys than to Calomys or Eligmo- These are the highest reported diploid numbers dontia. A new subspecies ofAndalgalomys pear- for phyllotine . The first specimens from soni is described from Bolivia. This, the first Bo- Bolivia of Thrichomys apereoides and Ctenomys livian record for the genus, extends the known minutus are documented. RESUMEN Los caracteres descritos del genero phyllotino damente 300 km al norte. Cariotipos para ambas Andalgalomys son revisados y un diagnosis cor- subespecies de A. pearsoni estan: A. p. pearsoni regido es presentado. Fenotipicamente Andalga- tiene 78 cromosomas y A. p. n. ssp. 76. Estos son lomys es mas similar a Graomys que a Calomys los informes mas altos del numero de cromosomas o . Una nueva subespecie de An- en los roedores phyllotinos. Los primeros espe- dalgalomys pearsoni de Bolivia es descrita. Este cimenes de Bolivia de Thrichomys apereoides y es el primer informe Boliviano para este genero, Ctenomys minutus estan documentados. extendiendo su distribuci6n conocido aproxima- ' Predoctoral Fellow in the Doctoral Training Program, American Museum of Natural History. 2 Curator, Department of Mammalogy, American Museum of Natural History. 3 Research Associate, Department of Mammalogy, American Museum of Natural History; Curator of Mammals, Museum of Southwestern Biology.

Copyright © American Museum of Natural History 1987 ISSN 0003-0082 / Price $2.40 2 AMERICAN MUSEUM NOVITATES NO. 2890

65 60

NZ1BOLIVIA ~~~0300 km

P ~A. p. dorbignyi

20 20 7'LJA~~. p., pear oni

I ~ P A R AGUA Y

25 ARGENTINA -.

A. N.. olrogi~ %

65 60 Fig. 1. Map of Paraguay and parts of Argentina and Bolivia. Localities shown are inhabited by Andalgalomys and are listed in the Appendix.

INTRODUCTION In August of 1984, two specimens of An- We document these geographic and taxo- dalgalomys were trapped in the eastern part nomic discoveries, add new comparative data, ofSanta Cruz Department, Bolivia, in scrub- and reevaluate relationships at both the spe- by vegetation characteristic of the borders of cific and generic levels. We review the pub- the arid Chaco located to the south. These, lished characters ofthe genus Andalgalomys, the first Bolivian specimens ofAndalgalomys compare specimens, modify Williams and (Anderson, 1985), extend the known range Mares' (1978) original description, and revise of the genus northward from Paraguay (fig. the diagnosis of the genus. Since Williams 1) and provide evidence ofa morphologically and Mares described Andalgalomys, addi- distinct northern population of Andalgalo- tional specimens of A. pearsoni have been mys pearsoni. collected from Paraguay. The number of 1 987 OLDS, ANDERSON, AND YATES: ANDALGALOMYS 3 available specimens has approximately tri- IED lateral incisor exposure pled and thus character variation within the BUD bullar depth genus can be assessed more adequately. SKD depth of skull zPw width of zygomatic plate METHODS Cell preparations of the new subspecies We studied specimens (listed in the Ap- were made in the field from bone marrow of pendix) of Andalgalomys and other phyllo- the femur as described by Baker et al. (1982). tine genera from the collections in the Mu- A slide of the cell suspension was prepared seum of Southwestern Biology at the in the field and later stained with giemsa. A University of New Mexico (MSB), Albu- cell suspension for future banding analyses querque; the Carnegie Museum (CM), Pitts- was preserved in liquid nitrogen and depos- burgh; the Museum of Zoology at the Uni- ited in the frozen tissues division of the Mu- versity of Michigan (UMMZ), Ann Arbor; seum of Southwestern Biology at the Uni- the American Museum of Natural History versity of New Mexico. Dr. Philip Myers of (AMNH), New York; the Museum of Ver- the University of Michigan generously pro- tebrate Zoology, University of California at vided a chromosomal slide from the speci- Berkeley (MVZ); the National Museum of men from Paraguay. Determinations of dip- Natural History (USNM), Washington, D.C.; loid and fundamental numbers were based and the Field Museum of Natural History on examination of at least 20 metaphase (FMNH), Chicago. spreads from each specimen. External measurements (in mm), recorded Statistical analyses (Principal Components from specimen labels, are: Analysis, Canonical Discriminant Analysis, univariate analyses) were done using the pro- TOT total length grams ofthe Statistical Analysis System (SAS HBL head and body length Institute Inc., 1982) at the City University of TAI tail length New York Center. HFL hind foot length Computer EAR ear length We compared specimens ofcomparable age between species, and examined sex and age The weight in grams was also recorded. differences within taxa. We found no obvious The following dimensions were measured sexual dimorphism in species of Andalgalo- with a stage craniometer to the nearest 0.01 mys, but dimorphism might become evident mm (see Anderson, 1969, for definitions of if larger series were studied. the dimensions used): Dental terminology follows Hershkovitz (1962) and Williams and Mares (1978). BUL length of bulla MXL alveolar length of maxillary tooth row PRL palatilar length ACKNOWLEDGMENTS ZAL length of zygomatic aperture IFL length of incisive foramina The two specimens of the new Bolivian CIL condyloincisive length subspecies were obtained on an expedition MSB breadth of mesopterygoid fossa supported in part by the National Science IDB interdental palatal breadth Foundation (Grants BSR 83-16740 and 84- PDB breadth of postdental constriction 08923, to Anderson and Yates, respectively). BRB breadth of rostrum We acknowledge the field collectors J. A. Cook PLB palatal breadth and D. L. Moore, the Carnegie Museum for POB breadth across paroccipital processes lending specimens, curators at MCZ, FMNH, AZB anterior zygomatic breadth USNM, and UMMZ for use of their collec- PZB posterior zygomatic breadth tions, P. Myers for sending the slide with MTB breadth of Ml A. OIL occipito-interparietal length unstudied chromosomes of p. pearsoni ONL occipitonasal length from Paraguay, N. A. Neff for checking lo- NLL nasal length cality records, C. J. Cole and M. Wasserman ROL length of rostrum for use of their laboratories and assistance in IOB interorbital breadth karyotypic analyses, and C. W. Myers for BCB breadth of braincase photographic assistance. We thank A. L. Z *% 4ow,

q, .t., e .T-

~-*z

c>c1

1j.

cis

cn Cq UkO-o tr-

L !.,..4 .L'rA CW1,~r *hi'e'> zk

c _~ 00;,: ci w

A4 i4

i 00

v O U_

4 }R~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1m., )4I-

Cd or

W4~~~~~~~~~~~~~~~~~~~~~~~~~~4

1* ZI

t ItE--

C t

( A;.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Ent e

-. * W i,IW

is- *t'a tgst

C t ^ v Dt's?ss}e~~~wfn 9 *C~~~~~~~~~ AtV..t~~~~~~~~~~~~4Cu

5 6 AMERICAN MUSEUM NOVITATES NO. 2890

Gardner, L. F. Marcus, G. G. Musser, J. L. penicillate (unlike most phyllotines, but Patton, and D. F. Williams for critically re- like , Graomys, Irenomys, and viewing an earlier draft of the manuscript. Eligmodontia). 3. Tail longer than head and body (unlike Genus Andalgalomys Galenomys, Auliscomys, Calomys, Pun- Williams and Mares, 1978 omys, Chinchillula, Andinomys, Eune- omys, Neotomys, Reithrodon). Andalgalomys Williams and Mares, 1978: 197; 4. Pinnae moderately large and sparsely type species Andalgalomys olrogi, op. cit.: 203, covered with fine hairs. True of many by original designation; also included A. pear- phyllotines. soni (Myers, 1977). 5. Soles offeet naked (unlike Eligmodontia CONTEXT: Family Muridae, subfamily Sig- and Galenomys). modontinae, tribe Phyllotini (see Olds and Anderson, in press, for diagnosis of tribe). CRANIAL CHARACrERS (see figs. 2, 3): CoNTENT: Andalgalomys includes A. olrogi (Williams and Mares, 1978) and A. pearsoni 6. Sides of interorbital region diverging (Myers, 1977), the latter now represented by posteriorly (like Graomys and some Cal- two subspecies. omys and unlike Auliscomys, Phyllotis, DiAGNosIs: A genus of Phyllotini distin- Galenomys). guished by the following combination of 7. Supraorbital region ledged (as in Gra- characters: pelage yellowish-brown; small omys and some species of Calomys). whitish subauricular spot; tail equal to or 8. Nasals slender and straight. Also true of longer than head and body length; tail pen- Eligmodontia, however, nasals relative- icillate; soles of feet naked; supraorbital re- ly longer in Andalgalomys. gion with ledged and posteriorly divergent 9. Molar rows parallel (or slightly conver- sides; nasals slender, straight, and long; bul- gent anteriorly). True of some phyllo- lae greatly inflated relative to those of other tines, but rows tending to converge an- phyllotines; palatine with slits; molars teriorly in Neotomys, Reithrodon, brachydont, tuberculate, slightly crested, and Euneomys, Andinomys, Irenomys, and with a tendency toward lamination in the up- Punomys. per molars; postcingulum ofM I with a slight 10. Palatines usually having large slitlike fo- notch or fold; anteroloph of M2 small and ramina. directed nearly anteriorly; major fold of M3 11. Parapterygoid fossa border ledged ante- deep, separating metacone-hypocone from the riorly and medially (as in phyllotines protocone-paracone by a double enamel wall. generally). Williams and Mares (1978) presented a 12. Bullae relatively greatly inflated (infla- general description ofAndalgalomys that in- tion greater in A. olrogi than in A. pear- cludes some characters common to all phyl- soni). Bullae also somewhat enlarged in lotines and some shared with other Sigmo- Graomys and in some species of Elig- dontinae. Comparisons were in a separate modontia. section. We extracted the following charac- ters for further consideration and comment DENTAL CHARACTERS (see fig. 4): as needed. The species ofAndalgalomys are contrasted with representatives of Calomys, 13. Molars brachydont (as in Calomys and Eligmodontia, and Graomys in figures 2 Eligmodontia). and 3. 14. Molar crowns tuberculate and slightly crested (as in Calomys and Eligmodon- tia), some tendency toward lamination EXTERNAL CHARACTERS: in the upper molars (not present in Cal- 1. Size moderate (head and body lengths of omys and Eligmodontia). adults now known to range from 85 to 15. Ml having well-developed anterior me- 125 mm). dian fold and an anteromedian style. Not 2. Tail slightly to moderately haired and always true; fold often weakly developed 1987 OLDS, ANDERSON, AND YATES: ANDALGALOMYS 7

POII

Fig. 4. Views of upper (above) and lower (below) right molar rows (SEM photos), of A. A. pearsoni dorbignyi (MSB 55245, holotype), B. A. p. pearsoni (AMNH 262345), C. A. olrogi (CM 44020), D. Graomys domorum (AMNH 260754), E. Graomys griseoflavus (AMNH 246848), F. Calomys sorellus (AMNH 231697), and G. Eligmodontia typus (AMNH 25702). Scales shown represent 1 mm.

in Andalgalomys, and style present or FURTHER COMMENTS ON DENTITION: In the not. Calomys showing a similar pattern. general description given by Williams and 16. Postcingulum of Ml having slight notch Mares (1978), we found some discrepancies or fold. between their comments and the specimens 17. Anteroloph of M2 small and directed we examined. These are summarized here. nearly anteriorly. The first primary fold of M2 is generally 18. Major fold of M3 deep, separating the shallow, but not always. The first minor fold metacone-hypocone from the proto- was said to be small and directed in an cone-paracone by a double enamel wall anteroposterior plane. This is confusing be- that usually persists in worn teeth (Gra- cause the fold is horizontal along the anterior omys tending to show this; but not Cal- edge of the tooth and is oriented anteropos- omys and Eligmodontia). teriorly along the anteroloph. 19. Procingulum of m2 not apparent. Gen- In the M2, they said that the metacone is erally true of phyllotines. greater than or equal to the hypocone in size. 20. Anteroconulid of m2 not apparent. Be- Myers (1977, in describingpearsoni), said the cause anteroconulid a part of procingu- opposite; to us they seem approximately equal lum, redundant with 19. in size. Williams and Mares also said that the 8 AMERICAN MUSEUM NOVITATES NO. 2890 paracone of the M2 is greater than or equal COMPARISONS: Within the Phyllotini, An- to the protocone; they seem approximately dalgalomys is most similar to Calomys, Elig- equal to us. modontia, and Graomys. Williams and Mares Anteromedian fold of the ml was said to (1978) noted these similarities and suggested form a small notch or deep groove on the (p. 199) that Andalgalomys is more closely anterior surface of procingulum, but also to related to the "Calomys-Eligmodontia com- penetrate deeply as an enamel fold. We note plex than to Graomys." Our interpretation considerable variation in this character; it may of the morphology suggests that Andalgalo- be present or absent; when present it is not mys may be more closely related to Graomys. always an enamel fold. Williams and Mares used two species of Cal- The anterolingual and anterolabial conu- omys in their comparisons, C. callosus (C. lids were said to be usually not well differ- muriculus is a synonym of C. callosus) and entiated from each other and usually ofequal C. musculinus, of which they had only one size. We note that the former may be larger. specimen. Calomys callosus is more like An- The protoconid of the ml was said to be dalgalomys in many respects than Andalga- nearly opposite the metaconid. This varies; lomys is like any other species of Calomys. the protoconid may be posterior to the meta- Graomys is more similar to Andalgalomys conid. The hypoconid of the ml was said to externally and cranially than is Eligmodon- be opposite or behind the entoconid. We note tia. The diagnosis given above will distin- that it is behind. guish Andalgalomys from these other genera. In the m2, the first minor fold was said not We summarize specific generic compari- to be evident, however, it is evident in A. sons that seem meaningful (from Williams pearsoni. and Mares, 1978, and from our own obser- In the m3, the first primary fold was said vations) as follows: to be about midway on the tooth. It is more Andalgalomys differs from Eligmodontia anterior in A. pearsoni. in having a longer tail, naked-soled feet, a BACULAR CHARACrERS FOR ANDALGALOMYS larger, more wedge-shaped skull, longer ros- (characters from Williams and Mares, 1978; trum (figs. 2, 3), and more ovate molar cusps terminology and comparative data from (fig. 4). In Andalgalomys the M3 is larger and Hooper and Musser, 1964): more triangular, the metacone is larger than the hypocone, the procingulum of the m2 is absent or obsolete, and the first primary fold 21. Tridigitate (also true of other phyllotine of the m3 is in the middle of the tooth. and nonphyllotine genera). In comparison to Calomys callosus, An- 22. Medial digit shorter than lateral digits, dalgalomys is larger and has a tail that is curved slightly dorsad (unlike Calomys, longer than the head and body, a more slen- Eligmodontia, Graomys, and Aulisco- der rostrum, larger bullae, a notch on the mys). postcingulum of m 1, and no evident procin- 23. Lateral digits robust (true of others, in- gulum of m2. cluding Calomys and Eligmodontia, but Andalgalomys differs from Graomys in untrue of Auliscomys). being smaller, in lacking an alisphenoid strut, 24. Proximal end of baculum expanded lat- and in having more brachydont molars, more erally and having straight edges. In this, ovate cusps, an anterior fold and anterome- Andalgalomys resembling Eligmodontia dian style in the Ml, no anteroloph of M3, more than the three others in Hooper and a rounded rather than ridged anterocon- and Musser's figure 4. ulid in the m2. 25. Distal end slightly expanded and ball- Andalgalomys differs from these three gen- shaped. era (and other phyllotine genera) in having a 26. Bacular mounds extending beyond glans whitish subauricular patch; a tendency to- hood. ward lamination in the upper molars; a slight 27. Glans penis as long as, or longer than, notch in the postcingulum ofMl; the antero- baculum (generally true of sigmodontine loph of M2 directed nearly anterior; and the rodents). deep fold of M3. 1987 OLDS, ANDERSON, AND YATES: ANDALGALOMYS 9

The ratio ofcrown lengths ofM3/M2 (table TABLE 1 1) is greater in Andalgalomys than in Calo- Means ofCrown Lengths of Upper Molars and the mys, Eligmodontia, or Graomys. The crown Ratio M3/M2 in Andalgalomys, Calomys, Elig- length of the M3 is most reduced in Elig- modontia, and Graomys modontia. In increasing order, the Ml is pro- LM3/ portionally longer in Graomys, Andalgalo- Species n LM1 LM2 LM3 LM2 mys, Eligmodontia, and Calomys. In the lower dentition, the ml is relatively longer C. callosus 28 1.89 1.18 0.87 73.73 C. lepidus 22 1.73 1.04 0.70 67.30 in Andalgalomys than in other genera and C. sorellus 9 1.79 1.09 0.76 69.72 the length of m3 is approximately the same C. tener 32 1.67 1.04 0.70 67.30 proportion of m2 in all. A. p. pearsoni 11 2.11 1.14 0.98 85.96 A. p. subsp. n. 2 2.43 1.29 1.08 84.05 Andalgalomys A. olrogi 6 2.13 1.18 0.95 80.51 pearsoni dorbignyi, E. sp. A 5 1.69 1.09 0.72 66.06 new subspecies E. sp. B 2 1.65 1.05 0.73 69.86 HOLOTYPE: Museum of Southwestern Bi- G. griseoflavus 7 2.43 1.50 1.15 76.60 ology catalog number 55245; adult male; skin G. domorum 6 2.37 1.47 1.05 71.43 and skeleton (plus frozen tissues and cell sus- pension with NK number 12402); prepared by Joseph A. Cook on October 7, 1984. TYPE LOCALITY: Bolivia: Department of gomatic plate; relatively smaller bullae; rel- Santa Cruz; 29.5 km west of Robore, 475 m atively shorter and less hairy tail; shorter and (180l9'S, 60002'W). coarser hair; darker pelage, ears, and tail; and DISTRIBUTION: Known only from the type more chromosomes (76 in A. p. dorbignyi, 78 locality. in A. p. pearsoni, and 60 in A. olrogi). SPECIMENS: Holotype and AMNH 260762 HABITAT: Arid, scrubby, low, partly open (adult female with teeth slightly more worn vegetation. than those of holotype). ASSOCIATED SPECIEs: At the same locality DIAGNOSIS: A subspecies ofAndalgalomys or one of two nearby localities, the first Bo- pearsoni distinguished from both A. pearsoni livian specimens of two species were cap- pearsoni and A. olrogi by larger size, more tured: a young Thrichomys apereoides and pronounced anterior zygomatic spines, three Ctenomys minutus (the former reported broader mesopterygoid fossa, and 76 chro- by name only as Cercomys cunicularius, and mosomes. the latter as Ctenomys minimus by Ander- DESCRIPTION: Any pale reddish brown son, 1985). Other mammals obtained were mouse from southeastern Bolivia that has Bolomys lasiurus, Chaetophractus vellerosus, white spots below its ears, sharply demarked and Tolypeutes matacus. All are listed in the white belly, tail about the same length as head Appendix. and body, total length (of adults) about 250 ETYMOLOGY: Because the previously rec- mm, hind foot about 26 mm, weight about ognized taxa of Andalgalomys were named 45 g, and a diploid number of 76 is probably for naturalists who worked in the area, we Andalgalomys pearsoni dorbignyi. External, follow the same tradition and name this new cranial, and dental characters for the genus subspecies for Alcide d'Orbigny, the first nat- are outlined above. Characters distinguishing uralist to collect in eastern Santa Cruz de- species and subspecies within Andalgalomys partment (in 1830). are in the diagnosis immediately above and KARYOTYPIC DATA: The karyotype of A. the comparisons immediately below. Mea- pearsoni differs considerably from that re- surements of the two specimens are listed in ported for A. olrogi (Williams and Mares, table 2. 1978). Diploid numbers (fig. 5) are 76 in A. COMPARISONS: A. pearsoni dorbignyi and p. dorbignyi and 78 in A. p. pearsoni com- A. pearsonipearsoni can be distinguished from pared to 60 in A. olrogi. Furthermore, in con- A. olrogi by larger size; broader posterior trast to the all-biarmed karyotype ofA. olrogi, frontal region; broader rostrum; wider zy- A. p. dorbignyi and A. p. pearsoni have 10 00 ~ r- -, 1- N ^ 00 o.00 C' 0, -. 0o% N _-_ t -* t o oo o - C) C. o en 00 +I "+I1q +1 +1 +1 tl +I , +1 +1 | 00 00 N - en T00 o o o0 - " - ao r. oo r-o ° n o 00 ur t - C ._

1- 00 N 00 N 00 C, I'D tn oy 0 - ~O Th N e4n 00 -, .00 - 00 0)00 0 09C)C000 - 0 +1 +1- +l +l ? +I "t +I +1 | +1 +1 |' +1 +1 +1 oo t +l _ ^ 12 e o N en N to N +I" _ "0 en ICi Nt.t 0 e+9 .2 _ I - -4 N N PC 0 OF (U 0 00 0 'O N , _ ^ N 'I"o c oo .oo ~. om.o.0o- 0o 00 o= - oENoc,, +l S +1 0) 0 1 +I e +I 1 +I +1 j' +I +06 +0 + en +1+l- el;I +I +I 0 o 00R enIl en N biN N _

- r}

+1 C,4 0 , 00 UN N _ fl en ( oo t ci 0 o 0@ - _ 0 00 C C'4 00 w U,O4 0N +1 +1 °,° +1 +1 tU +1 +1 +1 oo+1 ON +I09N .- -~ ( N O N t_ C 00 Ni

aN 1N tn +I

00 C0 0 q O 1-1 r- , N . Co C) Nl- + - (N 0D6 CD00 C) a 0 N 09- t =, No - +1 +1 +1 +l +1 +l N +1 0+ c0 4 00 00 +11.NNO+1:~~~~0G O N In en.1 _q OA _ r 6~ - S4 NO o +1 +1

2 .E -4 -- -* I" I N .- o* _ I00 00 e 7 O - O~ C5 - 6 - 6 t- No 0 0 -4 0 o N oo 000 , 4 a, N l-^ O > + " +1+ +1I' +I +lI"-+I N +1 t +1' N W) _ r 00~ C +l- +WI V) Ce N. 0o 09 £ F V oo "* r-. UWn ao, 0 tl I 16 I '0 en Co ; 't O; ° C N N - - '0CZ = CNe^ 00o C~ o0) t) 0 00 O N o C~&. ;~, z -N 00 4) _~ ND Oen wi te tl ( r-No 0o00 _ N eN

C. z w> t,

NN N . . Nb £ o £ 000- -- N 'T UN; N- 4 rJa L1: p.4 N 0-

10 -0 0 o Co"o N In 0 o t} cl X ooq 0e"i 0+11N o) io o en 09 60 te i o en -o N 1- o 6 +1 +l - +l -- +I r1 +I cl +I "I +I 00 +1 +1 +l +1 C~)cl E~Z~ -' - en0 A r-00rN °° CA _, rl48t en C0% In _ °1 °1 00C~ C' en B en cr 0° cr-11c 0 ** o j In _ N - ^ O

+l- N t N c. tn (oNr oo °° o N C0 0%enO N oen o= 't _4 06 0C 0(10N 60 0 eq00+1 +1 +1 o +1 - +1 1 +1 j +1 '7+1 9 0N*~ +1 C +1 C + - +1 +l I.+l 0% N I ooo

el t o09'T N+1-+l. --_ Cl4 +1 +1 +1 +1 +1 %- 0i-C4 N 00n 0%Cv 0) 0 Cysk, a- coo en e _ -, -, _ O en In o .00 0 ^ oN 00C -- ~o en +l0 +l +1 - oo - o 1- +l +1 +1 0%9 CD~~ Cl 0% 'i- 00 C *1 Cl tr 09 00*t 0 C~) 0% r)00 _C -- 0% t - +1 C +1 +1 _* N 'IT oo O eb q ^-- oo InRC 6 o1 + te +I m o0b o0 R~00 +1 06 0,_ +1 +1 - 0o -i +1 Cl PC oc N o0o.N + +0 +l+-t r- g o ooo WI o. _ ON_ w 0 £ 00 e enC 00C4 04 0) _,0 CD en_ 0 9G0l0 0 0% _ e - CeN0 m +1 C) o +1 +1 +I +I WI +I (t*n o NN _N. -, Cl 00C Cl4 -0060 Cli 00 6 CN N d +1 +l 0? +1 - +1 ' +1 1 +1 C +1l& +1lr 0-~ - C+14 o < 1.- m 14 6* 06 t- 0~ O. t N 0c% - N Nt N °~ C:) "O A 0 I") It Cc0 0 N +1 -, ontno ) _elON N 0C +1 '7 +1 '7 eno o 16 o~~~~~~~~~~~~~~ l 00 o~~~~~~~~~~e 1 09 '- oRo - ent l 0 0 +1 +1 9 +1- +1 +I~ +I '- +1 C OC 06 I4 N- 0 0 N °°N o~~~~~~~~~~C) - 0- - 00w. 06 efi Colr-_ Clt 1- t t o en C) 0 0 0 Cl4 0 0) o en 1- 00 'IO 1- Cl4 C N 'O If') -_ - tt1 6c; Cl I-f

0 00 0 0 Cl CAl 0 00 'I 00 C~~ n oR e#i 00t Cli 0%1 If') en-c -f

4 4 C *4 0 2 0o z 04 12 AMERICAN MUSEUM NOVITATES NO. 2890

and 12 pairs ofuniarmed and 28 and 27 pairs of biarmed chromosomes, respectively. The sex chromosomes in A. pearsoni are both ac- rocentric rather than biarmed as in A. olrogi. 0 The X in the former is the largest element in 6 the karyotype and the Y is a medium-size o II +1 +1 cq Cl4O acrocentric. . The diploid numbers observed in A. pear- - soni (both subspecies) are the highest re- ported for any phyllotine . Pearson and

0 Patton (1976) hypothesized that a basic trend st C o0o O in chromosomal evolution among these ro- c-, +1 00 +1 dents has been a decrease in diploid number o Zi o6q Cl o- through Robertsonian fusion. If this is cor- N1 o °) rect, A. pearsoni may possess close to the primitive condition for the phyllotines and oo the hypothetical primitive diploid number o en would be about 78. The totally uniarmed au- *N +1 I7 +l -- tosomal complement proposed by these au- ci, t11R 4 C-r +4-. _n *-- _ thors, however, is not consistent with the large _I number of biarmed chromosomal pairs (fig. 5) found in A. pearsoni. The karyotypes ofA.

00 pearsoni may contain a number of hetero- 0.oo0 67o0^O* chromatic arms although this cannot be de- 'ci- PCl Eq E tCl.c o 'f - N termined without banding analysis. Changes L14 +l - +l~ - within chromosomes, as well as changes in e~ i' o number ofchromosomes, must also have oc- curred if the A. olrogi karyotype was derived from the condition found in A. pearsoni. In o diploid number, Andalgalomys is closer to +I 0. Cl +16l " 1"-0 the 70 found in Neotomys ebriosus and Phyl- 011 _1 - +1 o St lotis osilae and the 64 as in Calomys sorellus (as suggested by Williams and Mares, 1978, than to any species of Graomys, Eligmodon- ~C4 tia, or other Calomys studied by them). NUMERICAL ANALYSES: We selected two +Cl 09l multivariate analyses, principal components co and canonical variates: the first was used to explore the general pattern of variation and ,0 to see whether groupings appeared in the data o t when our concept oftaxonomy was not spec- 0,ZN 0 Z~%~ 0%zI e. ified, and the second analysis was used to N _l observe the clustering of individuals within N- and between our specified groups. The latter Cl - relates to the recognizability ofthe genus An- dalgalomys and its affinities with the three other genera represented. There are other analyses that would express the phenetic af- finities more completely or quantitatively. We calculated Mahalanobis distances and decid- ed that these data did not alter our tentative conclusions about relationships and did not warrant inclusion. We performed principal components anal- ysis with the PRINCOMP procedure of SAS 1987 OLDS, ANDERSON, AND YATES: ANDALGALOMYS 13

MIXewKa a. asu 8ua.vz z,e, u." 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

U I A nohiasasa ."D is4 1a 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 XY

AAAS4 AnDSt,Aaas^SaUss as IDaait is a 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

4 UP IA8 2 9 fit-St It II XA U l%% 6 i I*d 1bea 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 XX

Fig. 5. KaryotypesofA. pearsonidorbignyi, AMNH260762(above)andA. pearsonipearsoni, UMMZ 134386 (below).

(SAS Institute Inc., 1982) using the correla- TABLE 3 tion matrix on specimens of Andalgalomys Eigenvectors for the First Three Principal Com- olrogi (n = 6) and A. pearsoni dorbignyi (n = ponents 2), A. p. pearsoni (n = 1 1), and on the fol- (The cumulative percentage ofvariance explained lowing species: Calomys callosus (n = 7), by each component is given at the bottom of the Eligmodontia sp. A (n = 9), E. sp. B (n = 2), table.) Graomys domorum (n = 11), and G. griseo- Dim.a PCI PC2 PC3 = we used 28 flavus (n 12). Initially, cranial BUL 0.21 0.32 -0.16 and dental variables. After examination of MXL 0.22 0.03 -0.02 coefficients of variation from an analysis of PRL 0.21 0.23 -0.08 sample means, we excluded the most vari- ZAL 0.24 0.01 0.12 able, thereby reducing the number of vari- IFL 0.22 -0.23 0.15 ables to 21 (see table 3). Table 3 gives the CIL 0.24 -0.11 0.06 eigenvectors for each variable on the first three IDB 0.16 0.37 0.68 principal components. PDB 0.22 0.08 0.07 The first principal component is related to PLB 0.22 0.18 0.32 size (fig. 6) in that the larger species, Graomys POB 0.21 -0.17 -0.01 domorum and G. are at AZB 0.22 -0.29 0.10 (M) griseoflavus (G), PZB 0.23 -0.21 0.05 the right; the medium-size species ofAndal- MTB 0.22 0.07 -0.05 galomys (D, 0, P) are in the center; and the ONL 0.24 -0.06 0.02 small species, Calomys callosus (C) and both NLL 0.24 -0.11 0.03 species of Eligmodontia (E), are at the left. ROL 0.23 -0.10 -0.00 The overlapping of G. domorum and G. gris- IOB 0.23 -0.06 -0.22 eoflavus suggests their overall cranial simi- BCB 0.21 0.07 -0.23 larity. Andalgalomys olrogi is well separated BUD 0.15 0.55 -0.39 (by the second principal component) from A. SKD 0.24 0.04 -0.22 p. pearsoni and A. p. dorbignyi. The two spec- ZPW 0.19 -0.31 -0.20 imens of A. p. dorbignyi bridge the scatters Cum. % 77.2 84.0 86.8 of A. p. pearsoni and G. griseoflavus. These a Dimension measured. Abbreviations are spelled out two specimens are different in age, sex, and in the Methods section. 14 AMERICAN MUSEUM NOVITATES NO. 2890

0 0 0 2.0 +

P GP p G E P p D G E p 0.4 p p Dp C. 0.0- p M G a. E E E M M E E E G M M GM G E G M M C E C G M GM C G C -2.0± c G c

n i i -6 -2 2 6 Pc1 Fig. 6. Graph ofthe results of a principal components analysis; Principal Component 1 (PC 1) versus Principal Component 2 (PC2). C = Calomys callosus, D = Andalgalomys pearsoni dorbignyi, E = Eligmodontia typus, G = Graomys griseoflavus, M = G. domorum, 0 = A. olrogi, and P = A. p. pearsoni. size; the older (based on tooth wear) speci- DISCUSSION: Data presented show (1) that men (AMNH 260762) is slightly smaller than the two Bolivian phyllotine specimens (on the younger one (MSB 55245). the basis of which a new subspecies is here We did a canonical variates analysis named) are morphologically (phenetically) (CANDISC procedure ofSAS) using the same different from other phyllotines, (2) that they samples as above (except omitting two spec- are nearer to the two previously recognized imens of Eligmodontia referred to E. sp. B), taxa ofAndalgalomys than to any other phyl- and using the reduced number of variables lotine, and (3) that, withinAndalgalomys, they (21). This analysis uses specified groups (the are nearer to A. pearsoni than to A. olrogi. It species above). In a plot ofCanonical Variate is possible to interpret some ofthe diagnostic 1 versus Canonical Variate 2 (fig. 7), the two characters of Andalgalomys as synapomor- species of Graomys lie close together; A. p. phies and thus develop a cladistic hypothesis dorbignyi is between them and A. p. pearsoni of monophyly. Within the genus it is more while A. olrogi is further away. Eligmodontia difficult to interpret any shared characters of and C. callosus are well removed from the A. pearsoni and the new taxon as synapo- other clusters at the bottom of the graph. In morphies rather than as primitive plesio- the graph of Canonical Variate 1 versus Ca- morphies. The distinctive characters ofA. ol- nonical Variate 3 (fig. 7), there is more sep- rogi may be interpreted as derived from those aration ofthe species of Graomys; A. p. pear- ofA. pearsoni. Within the known range ofA. soni lies between Eligmodontia and A. p. p. pearsoni in Paraguay, no geographic cline dorbignyi, and A. p. dorbignyi lies near C. in size is noticeable in the limited number of callosus. A. olrogi and Eligmodontia overlap. specimens available. 1 987 OLDS, ANDERSON, AND YATES: ANDALGALOMYS 15

0 0 0 0 D 0 0 D G G PP G G PPPp P GGG M CN G G G M MM z p c 0 p G M

ccC -6- E E E C C EE E E i I i i -10 -4 2 8 CAN I

3- K E MM E M M M M E E 0 M MM E M M O 0OE GG 0 P P _ P 0± p G p G G p G p D GG z p 0 G G 0 P P C p -3-f C D C

C C C

-10 -4 2 8 CAN 1 Fig. 7. Graphs ofthe results of a canonical variates analysis: upper graph, Canonical Variate 1 (CAN 1) versus Canonical Variate 2 (CAN 2), and lower graph, Canonical Variate 1 (CAN 1) versus Canonical Variate 3 (CAN 3). Abbreviations as in figure 6. 16 AMERICAN MUSEUM NOVITATES NO. 2890

Generally, the concept ofsubspecies is used Baker, R. J., M. W. Haiduk, L. W. Robbins, A. for well-marked populations with contiguous Cadena, and B. F. Koop geographic ranges that are hypothesized to 1982. Chromosomal studies of South Amer- interbreed with one or more other subspecies ican bats and their systematic implica- of the species. This assumes the "biological tions. In M. A. Mares and H. H. Geno- ways (eds.), Mammalian biology in species concept." There is no consensus on South America, vol. 6, pp. 303-327. how "well marked" the population must be Linesville, Pa.: Spec. Publ. Ser. Pyma- or how narrow or sharp the zone of intergra- tuning Lab. Ecol., xii + 539 pp. dation between subspecies must be. Hershkovitz, P. Recognition of a new subspecies, Andal- 1962. Evolution of Neotropical cricetine ro- galomys pearsoni dorbignyi, for the two Bo- dents (Muridae) with special reference livian specimens is based on the highly prob- to the phyllotine group. Fieldiana, Zool., able distinctness of the population they 46: 1-524. represent and the hypothesis that A. p. dor- Hooper, E. M., and G. G. Musser bignyi and A. p. pearsoni will be found to 1964. The glans penis in Neotropical crice- tines (Muridae) with comments on clas- intergrade somewhere in the area between the sification ofmuroid rodents. Misc. Publ. presently known geographic localities ofthese Mus. Zool. Univ. Michigan, 123: 1-57. subspecies. An alternative hypothesis is that Myers, P. these two taxa will be found sympatrically in 1977. A new phyllotine rodent (genus Gra- this area and maintain their morphological omys) from Paraguay. Occas. Pap. Mus. distinctness. These alternative hypotheses can Zool. Univ. Michigan, 676: 1-7. be tested by further collecting in the inter- Olds, N., and S. Anderson vening area. In press. A diagnosis ofthe tribe Phyllotini (Ro- dentia, Muridae). Florida State Mu- seum. Pearson, 0. P., and J. L. Patton REFERENCES CITED 1976. Relationships among South American phyllotine rodents based on chromo- Anderson, S. some analysis. J. ., 57(2): 339- 1969. Taxonomic status ofthe woodrat, Neo- 350. toma albigula, in southern Chihuahua, SAS Institute Inc. Mexico. Misc. Publ. Univ. Kansas Mus. 1982. SAS user's guide: statistics, 1982 ed. Nat. Hist., 51: 25-50. Cary, N.C.: SAS Institute Inc., 584 pp. 1985. Lista preliminar de mamiferos bolivi- Williams, D. F., and M. A. Mares anos. Cuadernos, Acad. Nac. Cienc. Bo- 1978. A new genus and species of phyllotine livia, vol. 65, Cienc. Naturaleza, no. 6, rodent (Mammalia: Muridae) from Mus. Nac. Hist. Nat., Zoologica, no. 3, northwestern Argentina. Ann. Carnegie pp. 5-16. Mus., 47(9): 193-221.

APPENDIX. SPECIMENS EXAMINED OR REPORTED A. olrogi (6): Argentina: Catamarca; 15 km MVZ 145276 (holotype) (Villa Hayes = S of Highway 62 on Rio Amanao and W of 25006'S, 57034'W). Paraguay: Boqueron; 15.6 Andalgala, CM 44020. Argentina: Catamar- km by road N of Filadelphia, UMMZ ca; "Km marker 10" on Highway 62, 10 km 134386. Paraguay: Nueva Asuncion; 0.5 km W of Andalgala, CM 44021. Argentina: Ca- S ofTeniente Enciso, UCONN 17579, 17580, tamarca; 15 km W (on Route 62) of Andal- Teniente Enciso (cited as 2.5 km S by Myers, gala on W bank Rio Amanao, CM 44022- 1977), UCONN 17562, 17563, 17566, 17568, 44024 (44024 = holotype). Argentina: Ca- 17575. Paraguay: Nueva Asuncion; 1 km SW tamarca; 8 km W of Andalgala on Highway km 620 on the Trans-Chaco Road, AMNH 62, CM 86451. (Andalgala = 1060 m; 27°36'S, 262344-262347, UMMZ 158621, 130037, 66019'W; Amanao = 27°33'S, 66°31'W.) and the following uncataloged specimens at A. pearsoni pearsoni (24): Paraguay: Bo- UMMZ, collected by T. W. Nelson: TWN queron; 410 km by road NW of Villa Hayes, 183, 184, 193, 195, 201, 202, 218, 229, 233. 1987 OLDS, ANDERSON, AND YATES: ANDALGALOMYS 17

A. pearsoni dorbignyi (2): Bolivia: Santa modontia have been referred to E. typus Cruz; 29.5 km W ofRobore, 475 m (18°19'S, (Hershkovitz, 1962); however, our sample 60002'W), AMNH 260762, MSB 55245. includes two species. We are not certain of Calomys callosus (35): Bolivia: Tarija; 8 their correct names. km S and 10 km E Villa Montes, 467 m Graomys domorum (11): Bolivia: Santa (21019'S, 63025'W), AMNH 246740, 246760. Cruz; 5 km by road SE of Comarapa, 1695 Bolivia: Tarija; 2 km S and 10 km E Tiquipa, m (17°58'S, 64°29'W), AMNH 260748, Laguna Palmar, 555 m (20°56'S, 63°21'W), 260751-260754. Bolivia: Cochabamba; AMNH 246745, 246834, 246857, 246860, Cochabamba, 2600 m (17°24'S, 66°09'W), 246862. Paraguay: Chaco; 50 km WSW For- FMNH 50961-50963, 50967-50969. tin Madrejon, AMNH 248449-248452, G. griseoflavus (12): Bolivia: Tarija; 8 km 248455-248459; UMMZ 124232, 124233, S and 10 km E Villa Montes, 467 m (21°l9'S, 124237-124250, 125460-125462. 63025'W), AMNH 246761, 246770, 246771, Calomys lepidus (22): Bolivia: Oruro; 3.5 246773, 248438. Bolivia: Santa Cruz; Tita, km E Huancaroma, 3720 m (17040'S, 300 m (18°25'S, 62°10'W), AMNH 260757. 67027'W), AMNH 260607-260614, 260616, Argentina: Jujuy; Santa Barbara, AMNH 260619-260624. Bolivia: Oruro; 2.5 km NE 185222. Paraguay: Chaco; Guachalla, FMNH Huancaroma, 3720 m (17040'S, 67028'W), 54359-54361, 54363, 54364. MSB 55264, 55266-55271. Bolomys lasiurus (9): Bolivia: Santa Cruz; Calomys sorellus (9): Peru: Huancavelica; 7 km N and 38 km W of Robore, 550 m Lircay, 3278 m, FMNH 95396-95399, (18016'S, 60007'W), AMNH 260552, 260553, 95415-95417, 95548, 95549. 260509-260513, MSB 55250. Bolivia: Santa Calomys tener (32): Brasil: Sao Paulo; Ita- Cruz; 29.5 km W ofRobore, 475 m (18°l9'S, petininga, USNM 460552-460554, 462124- 60002'W); AMNH 260507. 462126, 484409, 484411-484415, 484417- Thrichomys apereoides (1): Bolivia: Santa 484419, 484430, 484432, 484434, 484436- Cruz; 7 km N and 38 km W of Robore, 550 484439, 484443-484445, 484450-484452, m (18016'S, 60°07'W), AMNH 260860. 484470-484472, 543055. Ctenomys minutus (3): Bolivia: Santa Cruz; Eligmodontia sp. A (9): Argentina: Pata- 41 km by road W ofRobore, 550 m (18018'S, gonia; Arroyo Aike, AMNH 25702, 25703. 60002'W), AMNH 260835, 260836, MSB Argentina: Mendoza; 312 km (by El Man- 55367. zano Rd) of Tunuyan, CM 43657, 43671, Chaetophractus vellerosus (1): Bolivia: 43673,43675,43681,43682. Bolivia: Oruro; Santa Cruz; 41 km by road NW of Robore, 12 km S and 1.5 km E Eucaliptus, Lecheria, 500 m (18°18'S, 60°02'W), AMNH 260318. 3359 m (17°42'S, 67030'W), AMNH 246776. Tolypeutes matacus (1): Bolivia: Santa Eligmodontia sp. B (2): Argentina: Men- Cruz; 29.5 km W ofRobore, 475 m (18019'S, doza; La Paz, MCZ 41079, 41100. All Elig- 60002'W), AMNH 260320.

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, New York 10024.