Revision of the Tribe Phyllotini

< OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN eiaopv A996 DE \ 6 A5VIT l H' *"» NO VOLS DARD ECONOMY THESIS LEAF ATTACH D Q THIS TITLE j D ECONOMY ALTTH 1ST D COLOR OR TITLE I .D. PLE 20706 ra^gr

)IS

ISSN BIX

IINDING WHEEL SYS. ID. *E CY Stf 1 3

NEW SERIF

Revision of the Tribe Phyllotin (Rodentia: Sigmodontinae), with a Phylogenetic Hypothesis for the Sigmodontinae

Scott J. Steppan

february 28, 1995 Miblication 1464

>I I UBLISHED BY FIELD MUSEUM Ol ' - i for i n tffors Ui

n thcre<

reduces th

ieaiion res.:.

• , uete copi

i copy plus two

.'.id Mos-r.ui Ulmoi:

. t" and

n the nieti

w that of r oi Style (I3ih lvcrsily

; given in '.

index ot A

nford

.. lontai

structure, md floristics. Journal of

ind Si .--..

Bulletin 143, Bui

Guatemala tdiana: 1 b >ns are rei in the text (not as "pU mpanicd

I acceptable . . ! < {!

Dens;, ami and not exceed 11V4 i), may obtained in (preferred)

it ion; and photo

color : figure;

corresponding autii can be made and queries ans

a be made of) the s .

. : ON ACID FREE PAPER. Revision of the Tribe Phyllotini

FIELDIANA

Zoology NEW SERIES, NO. 80

Revision of the Tribe Phyllotini (Rodentia: Sigmodontinae), with a Phylogenetic Hypothesis for the Sigmodontinae

Scott J. Steppan

Division of Mammals Field Museum of Natural History Roosevelt Road at Lake Shore Drive Chicago, Illinois 60605-2496

Committee on Evolutionary Biology The University of Chicago Chicago, Illinois 60637

Accepted July 11, 1994 Published February 28, 1995 Publication 1464

PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY © 1995 Field Museum of Natural History Library of Congress Catalog Card Number: 94-62054 ISSN 0015-0754 PRINTED IN THE UNITED STATES OF AMERICA Table of Contents 8. Lateral view of a generalized Phyllotis cranium 22 9. Variation in incisor grooves among Abstract 1 phyllotines 23 Introduction 1 10. Variation in upper incisor dentine fis- Evolutionary Relationships within Sigmo- sures 23

dontinae 2 1 1 . Variation in ventromedial process of Taxonomic History of the Phyllotines ... 6 mandibular ramus 27 Materials and Methods 7 12. Position of anterior root of zygomata .. 29 Taxa and Characters Examined 7 13. Dorsal views of interorbital region .... 31 Quantitative Character Coding 10 14. Dorsolateral views of posterior cranium Analytical Methods 11 34 Comparative Morphology 13 15. Stapedial spine of auditory bulla 35 Dentition 14 16. Medial views of auditory bulla and in- Cranium and Mandible 26 ternal carotid canal 36 Postcranial Skeleton 40 17. Ventral view of hemal arches and hemal External Morphology 49 processes in Nectomys squamipes 48 Characters of the Phallus and Soft Anatomy 18. Ventral and lateral views of bacular ap- 51 paratus in Phyllotis magister 52 Phylogenetic Relationships within Sigmo- 19. Strict consensus cladogram for the Sig- dontinae 53 modontinae; analysis weighted to favor Results 53 sigmodontine monophyly 54 Discussion 60 20. Strict consensus cladogram for the Sig- Phyllotine Monophyly 62 modontinae; unweighted analysis 56 Phylogenetic Relationships within 21. Majority-rule bootstrap consensus tree Phyllotini 63 for the Sigmodontinae; analysis weight-

Results 63 ed to favor sigmodontine monophyly . . 58 Discussion 70 22. Strict consensus cladogram for the Phyl- Taxonomy 72 lotini 64 Acknowledgments 100 23. Eighty percent majority-rule consensus Literature Cited 101 tree, derived from most-parsimonious Appendix: Specimens Examined 104 trees wherein Punomys is not a phyllotine 65 24. Majority-rule bootstrap consensus tree List of Illustrations for the Phyllotini 66 25. Cranium and mandible of Calomys laucha 73 1 . Albumin immunological dendrogram 26. Upper and lower molars of Calomys 3 (from Sarich, 1985) laucha 74 2. Evolutionary scenario for South Ameri- 27. Cranium and mandible of Eligmodontia can sigmodontines (from Hershkovitz, morgani 76 1962) 5 28. Upper and lower molars of Eligmodon- 3. Hypothesized relationships of South tia morgani and Graomys griseoflavus . 11 American "cricetines," based on phallic 29. Cranium and mandible of Graomys gris- characters (from Hooper & Musser, eoflavus 78 1964) 6 30. Cranium and mandible of dar- 4. Nomenclature for dental elements (from Phyllotis wini 80 Reig, 1980) 9

3 1 . and 5. Phenogram and cladogram from electro- Upper lower molars of Phyllotis

darwini and . . phoretic data (from Spotorno, 1986) .. 10 Loxodontomys micropus 81 6. Dorsal view of a generalized Phyllotis 32. Cranium and mandible of Loxodonto- cranium 18 mys micropus 83 7. Ventral view of a generalized Phyllotis 33. Cranium and mandible of Auliscomys cranium 20 pictus 85 34. Upper and lower molars of Auliscomys List of Tables pictus and Galenomys garleppi 86 35. Cranium and mandible of Galenomys

garleppi 87 1 . Species included in sigmodontine analysis 36. Cranium and mandible of Chinchillula 2 2. sahamae 89 Species included in phyllotine analysis . . 8 37. Upper and lower molars of Chinchillula 3. Data matrix for the phylogenetic analysis sahamae and Andinomys edax 90 of the Sigmodontinae 12 38. Cranium and mandible of Andinomys 4. Data matrix for the phylogenetic analysis edax 92 of the Phyllotini 14 39. Cranium and mandible of Irenomys tar- 5. Vertebral counts among Neotropical sig- salis 93 modontines and selected muroids 42 40. Upper and lower molars of Irenomys 6. Distribution of selected characters among

tarsalis . and Euneomys chinchilloides . . 94 oryzomyine and thomasomyine genera 47 41. Cranium and mandible of Euneomys 7. Consistency and retention indexes for sig- chinchilloides 96 modontine characters 55 42. Cranium and mandible of Neotomys 8. Consistency and retention indexes for ebriosus 97 phyllotine characters 68 43. Upper and lower molars of Neotomys ebriosus and Reithrodon auritus 98 44. Cranium and mandible of Reithrodon auritus 99

VI Revision of the Tribe Phyllotini (Rodentia: Sigmodontinae), with a Phylogenetic Hypothesis for the Sigmodontinae

Scott J. Steppan

Abstract

The phylogenetic relationships of the South American rodents of the tribe Phyllotini are reviewed, considering both the phylogenetic relationships of the phyllotines to the other sigmodontine tribes and the relationships within the phyllotines. Cladistic analysis of 40 morphological characters for 28 sigmodontine taxa provides a working hypothesis of sigmodontine phylogenetics, phyllotine monophyly, and likely sister-groups to the phyllotines. Five Old World and six New World cricetid taxa represent outgroups, and together they root the sigmodontine tree within a paraphyletic thomasomyine group. The analysis corroborates the recent proposal of a monophyletic oryzomyine group that includes the tetralophodont genera Holochilus, Pseu- doryzomys, and Zygodontomys. A supratribal clade is indicated that includes the Akodontini, Phyllotini, Scapteromyini, and Punomys. The distinctiveness and monophyly of the Central American tylomyine group is strongly supported. The taxonomic distribution of and variation in morphological characters of the dentition, skull, skeleton, and soft anatomy are discussed. Apparent biases in the evolutionary polarity of reductive characters are identified in detail from a broad taxonomic survey (174 species) for intra- and interspecific variation in number of vertebrae, as well as from optimization of other characters on the phylogenetic hypotheses. Conflicting results from various phylogenetic studies suggest that Sigmodon be considered Sigmodontinae incertae sedis. Pseudoryzomys and Punomys are removed from the phyllotines, and Phyllotini is diagnosed. A cladistic analysis of 35 phyllotine taxa using 98 morphological characters is presented, and the taxonomy of the phyllotine genera is revised. Species of An- dalgalomys are referred to Graomys. Removal of micropus from Auliscomys to the genus Loxodontomys is supported. The two most species-rich genera, Phyllotis and Calomys, appear to be paraphyletic, but their species relationships are insufficiently resolved to justify modifying their taxonomy at this time.

Introduction inhabiting the puna, an alpine steppe community (Reig, 1986). The phyllotines constitute one of the principal This study presents a cladistic analysis of evo- radiations of the New World muroids. Frequently lutionary relationships among members ofthe tribe the most abundant mammals in their range, phyl- Phyllotini. It then provides a taxonomic revision lotine species are concentrated among the pastoral of the tribe with diagnoses of recognized genera habitats of the Andes, stretching from Ecuador to within this phylogenetic context. An impediment Tierra del Fuego, and from the Pacific coast of to any such cladistic analysis within tribes is that Peru and Chile east through Patagonia to south- intertribal relationships among Neotropical sig- eastern Brazil. Maximum diversity is achieved in modontine rodents, and even tribal monophyly, the altiplano, with 44% of the phyllotine species are poorly resolved. This lack of understanding of

FIELDIANA: ZOOLOGY, N.S., NO. 80, FEBRUARY 28, 1995, PP. 1-112 Table 1 . Species included in sigmodontine analysis. higher-level relationships significantly reduces follows Musser and Carleton with (Taxonomy [1993], confidence in hypotheses of character polarities modifications noted.) and specific membership within tribes. Better estimates ofoutgroups to the phyllotines are needed. Old World "cricetids"" Therefore, this study also presents a cladistic anal- Subfamily Calomyscinae for the Calomyscus baluchi ysis subfamily Sigmodontinae (sensu Reig, Subfamily Cricetinae 1 980) in order to provide a provisional hypothesis Cricetulus migratohus ofoutgroup relationships to be applied to the phyl- Mesocricetus auratus lotine analysis, and a revised diagnosis of Phyl- Phodopus sungorus lotini. These two nested analyses will be referred Subfamily Mystromyinae Mystromys albicaudatus to as the sigmodontine and phyllotine analyses. New World "cricetids" Phyllotine membership and defining characters Subfamily Tylomyinae'' have fluctuated among studies, but nearly all Nyctomys sumichrasti workers have recognized the following taxa as Tylomys nudicaudus Aulisco- Subfamily Neotominae' phyllotines: Andalgalomys, Andinomys, Neotoma jloridana mys, Calomys, Chinchillula, Eligmodontia, Gale- Ochrotomys nuttalli nomys, Graomys, Irenomys, and Phyllotis. Prob- Peromyscus leucopus lematic taxa have included Euneomys, Holochilus, Scotinomys teguina Reithrodon, Subfamily Sigmodontinae Neotomys, Pseudoryzomys, Punomys, Tribe Akodontini"' Sigmodon, and Zygodontomys. "Problematic taxa" Akodon albiventer are those that at various times have been included Akodon boliviensis within the phyllotine group as well as genera hy- Oxymycterus hispidus pothesized to have been derived from a phyllotine Tribe Ichthyomyini Anotomys leander ancestor. The phyllotine analysis examines rep- Ichthyomys hydrobates resentatives of all phyllotine genera, as defined by monticolus Neusticomys the results of the sigmodontine analysis. All for- Tribe Oryzomyini'' mally or informally recognized supergeneric groups Holochilus brasiliensis are represented in the analysis, as Neacomys spinosus sigmodontine Nectomys squamipes are all "problematic taxa" except Euneomys. Oligoryzomys fulvescens Oryzomys capito Oryzomys palustris Evolutionary Relationships Pseudoryzomys simplex Zygodontomys brevicauda within Sigmodontinae Tribe Phyllotini Calomys callosus Native muroid rodents are represented in South Graomys griseoflavus America exclusively by the subfamily Sigmodon- Neotomys ebriosus tinae 1 843. Debate continues as to wheth- Phyllotis darwini Wagner, Reithrodon auritus er this taxon includes the North American "cri- Tribe Scapteromyini Kunsia tomentosus Scapteromys tumidus Tribe Sigmodontini Sigmodon hispidus Informal designation of Old World and New World Tribe Wiedomyini "cricetids" reflects historical usage and serves to distin- Wiedomys pyrrhorhinos guish them from murines and arvicolines, but little sup- Thomasomyine grour/ port has been presented for the monophyly of either Chilomys instans group. h Rhipidomys latimanus Sensu Reig (1984). The distinctiveness of this group Thomasomys aureus and its basal position relative to the North American Thomasomys baeops neotomine-peromyscines and South American sigmo- Thomasomys rhoadsi dontines has also been noted by Carleton (1980). < Sigmodontinae incertae sedis Sensu Reig (1980). d Punomys lemminus Sigmodontine tribes regarded as informal groups by Musser and Carleton ( 1 993) are here recognized in their formal tribal designations, sensu Vorontsov (1959). e Contents per Voss and Carleton (1993). ' Monophyly and tribal status argued against by Voss (1993).

FIELDIANA: ZOOLOGY Tylomys Ototylomys

Zygodontomys

Calomys Phyllotis

Oryzomys Nectomys

Sigmodon

Peromyscus Neotoma

Mesocricetus

40 30 20 10 TIME (MYA)

Fig. 1. Albumin immunological dendrogram of New World muroids (modified from Sarich, 1985).

cetids," the neotomine-peromyscines (Carleton & but the few available molecular or cladistic studies Musser, 1984; Musser & Carleton, 1993), or is are consistent with monophyly (Carleton, 1973; limited to the predominantly South American spe- Catzeflis et al., 1993). cies sensu Reig (1980, 1986). The northern and While Carleton cautioned that assuming mono- southern continental groups have also been char- phyly of the complex-penised sigmodontines acterized as having "simple" or "complex" penis ("South American cricetines") "as presently con- types, respectively (Hershkovitz, 1966b; Hooper stituted" was premature (1980, p. 140), his dis- & Musser, 1964). The subfamily Sigmodontinae tance Wagner tree (1980, Fig. 41) does support is here considered to be limited to the predomi- monophyly of the Neotropical sigmodontines pro- nantly complex-penised, largely Neotropical spe- vided they are not defined as identical with "com- cies in accordance with the taxonomy of Reig plex penis" murids and the Central American Nyc- (1980) and evolutionary scenarios of Hershkovitz tomys is excluded. The distinctiveness ofNyctomys

( 1 962, 1 966a), excluding the Central American ge- from the other "complex penis" forms has been nus Nyctomys. The taxonomy of muroid rodents noted repeatedly for several aspects of the male used in this paper is presented in Table 1 . Crice- reproductive system (Arata, 1964; Hershkovitz, tidae was not recognized by Musser and Carleton 1966b; Hooper & Musser, 1964; Voss & Linzey, (1993), and in this paper I use the term "cricetid" 1981). Sarich (1985) presented an albumin im- for the assemblage of subfamilies sharing the den- munological dendrogram for New World "crice-

tal morphology associated with earlier definitions tids" (Fig. 1 ). The South American sigmodontines of Cricetidae (Simpson, 1 945). as defined in this study were a monophyletic branch Sigmodontinae (as was defined to include the in an unresolved trichotomy with the Central North American neotomine-peromyscines) was American Tylomys (which Carleton [1980] found one of only two muroid subfamilies that Carleton to be most closely related to Nyctomys) and the and Musser (1984, p. 300) were unable to diag- North American Peromyscus and Neotoma. Cat- nose, owing to their "immense heterogeneity." zeflis et al. (1993) reported a DNA hybridization Monophyly of the Neotropical "complex penis" study that clearly distinguishes the North and South sigmodontines has not been clearly demonstrated, American groups as separate lineages and referred

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI them to the subfamilies Neotominae and Sigmo- Chilomys, Aepeomys [Reig, 1980]). Gardner and dontinae, following Reig (1980). Neotominae was Patton (1976) derived all sigmodontine lineages represented by Neotoma and Peromyscus while from an Oryzomys karyotype. Reig (1986, 1987) Sigmodontinae was represented by Sigmodon, considered oryzomyines to be the direct or indirect Oryzomys, Zygodontomys, Akodon, and Phyllotis. descendants [sic] of the ancestral Oryzomys-like Catzeflis et al.'s (1993) analysis and review of pre- sigmodontine. Carleton's (1980, Fig. 41) Wagner vious DNA hybridization studies not only support tree placed Akodon and Oxymycterus at the base the monophyly of the South American Sigmodon- of the South American lineage, while Sarich's tinae relative to the Neotominae, but also relative (1985) immunological tree (Fig. 1) placed Sig- to other "cricetid" groups: cricetines and arvicol- modon at the base of the South American branch, ids. Molecular data sets thus support the definition with Zygodontomys basal to the phyllotines and of Sigmodontinae used in this paper. oryzomyines sensu stricto. Additional support for a monophyletic Sigmo- Hershkovitz (1962) envisioned two lineages dontinae comes from distributions of ectoparasites arising from a pentalophodont (with complete me- and endoparasites. Wenzel and Tipton (1966) solophostyle), Thomasomys-like stock: a thoma- found that mites and lice (as well as the less host- somyine group that gave rise to the oryzomyine specific ticks) found on complex-penised "crice- group, and a tetralophodont lineage that gave rise tids" (sigmodontines) belonged to a radiation of to the paraphyletic akodontine group, from which South American origin. Congruently, Slaughter and radiated the ichthyomyine, phyllotine, and sig- Ubelaker (1984) found members of the nematode modont groups (Fig. 2). Later, he presented the genus Parastrongylus, belonging to a species group scapteromyines as the sister-group to the oxy- largely restricted to Old World "cricetids," to be mycterines, and these together as part of the ako- present in several oryzomyines and Sigmodon, but dont radiation (Hershkovitz, 1966b). Vorontsov not in North American neotomine-peromyscines. (1959) outlined what amounts to a monophyletic The nematodes show strong host specificity and group consisting of his Phyllotini, Euneomys, and are not likely to have distributions strongly af- Sigmodontini (including Reithrodon and Neoto- fected by climate, a criticism Carleton (1 980) made mys). The Wagner tree generated by Carleton of the flea data from Wenzel and Tipton (1966). ( 1 980) placed Oxymycterus as a basal South Amer- Slaughter and Ubelaker (1984) argued that the ican genus and Scapteromys as highly derived, in neotomine-peromyscines had diverged from the contrast to his earlier noncladistic study of stom- lineage that later gave rise to the Old World "cri- ach morphology that hypothesized a sister-group cetids" and the sigmodontines prior to the com- relationship between these two (Carleton, 1973). plex-penised lineage having acquired the parasite. The evolutionary scenario diagrammed by Gard- Few hypotheses of relationships among the sig- ner and Patton (1976) treats the akodontines and modontines have been proposed, and only one has oxymycterines as sister-groups comprising an in- utilized phylogenetic methods (Carleton, 1980), dependent offshoot from Oryzomys. Other inde- wherein the analysis of the South American sig- pendent lineages include the thomasomyines, modontines was peripheral to the principal objec- ichthyomyines, and a group composed of the phyl- tives of the study. More comprehensive attempts lotines, sigmodonts, and scapteromyines. (Gardner &Patton, 1976; Hershkovitz, 1962; Reig, Finally, Reig (1984, 1986) wove a complex bio- 1 986) have lacked the analytical rigor of cladistics. geographic scenario for the South American "cri- Nonetheless, the several scenarios and studies pro- cetids." He considered phyllotines to be most like- vide an important conceptual framework. ly derived from akodontines, though he suggested The group most commonly identified as the bas- that phyllotines could be independent offshoots al member of the sigmodontines has been the spe- from the oryzomyines. He also hypothesized that cies-rich oryzomyines, which have often been por- Zygodontomys was an independent oryzomyine trayed as paraphyletic. The definition of offshoot of undetermined affinity and proposed "oryzomyines" has varied, either referring to ory- the descent of ichthyomyines from a Thomaso- zomyines sensu stricto (Melanomys, Microryzo- mys-like ancestor, scapteromyines from akodon- mys, Neacomys, Nectomys, Nesoryzomys, Oeco- tines, and sigmodonts from the phyllotine Neo- mys, Oligoryzomys, Oryzomys, Scolomys, tomys. Although there is no single point on which Sigmodontomys [Hershkovitz, 1962; Musser & all of these authors agree, a consensus would place Carleton, 1993]) or also including the thomaso- the oryzomyines or thomasomyines at the root of myines (Thomasomys, Rhipidomys, Delomys, the sigmodontines.

FIELDIANA: ZOOLOGY SYLVAN PASTORAL — iv .^p-~-- X . i CHINCHILLULA ;^ , 2 * ^' l-V-+s,v, ANDI NOMYS

PENTALOPHODONT SYLVAN CRICETINES

Fig. 2. Evolutionary scenario for the South American sigmodontines (from Hershkovitz, 1962).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI Oxymycterus Oryzomys Akodon

Holochilus Eligmodontla

Neacomys, Notiomys Sigmomys

Sigmodon

Rheomys

Nyctomys

Fig. 3. Hypothesized relationships of South American "cricetines," based on phallic characters (from Hooper & Musser, 1964).

Taxonomic History of the chilus was best placed with the oryzomyines. Phyllotines Reithrodon was placed as a basal phyllotine; Neo- tomys and Pseudoryzomys were not examined. In- The following is a summary of the more recent explicably, Geoxus {"Notiomys") was dia- taxonomic history of the phyllotines. Additional grammed as part of a phyllotine lineage, though details, particularly of the period before 1962, can in the text it was described as allied to akodontines, be found in Olds and Anderson (1 989) and in Tate Oxymycterus, and phyllotines. (1932a,b,c). Phyllotines have commonly been viewed as a Hershkovitz (1962, and Fig. 2) portrayed the paraphyletic group. Gardner and Patton (1976) phyllotines as a monophyletic group derived from diagrammed their view of evolutionary relation- akodont stock. In his detailed revision of the phyl- ships among Neotropical "cricetids" based pri- lotines and discussion of sigmodontine morpho- marily on karyotypic data. They showed sigmo- logical evolution, Hershkovitz included Zygodon- donts and scapteromyines as derived from tomys (whose southern forms have since been primitive phyllotines. Pearson and Patton (1976) removed to Bolomys) and Pseudoryzomys but ex- and Gardner and Patton (1976) agreed on the included Reithrodon and Neotomys (which he conclusion of Andinomys, Auliscomys, Calomys, sidered to be sigmodonts along with Sigmodon Chinchillula, Eligmodontia, Neotomys, Phyllotis and Holochilus), as well as Euneomys, Irenomys, (including Graomys), and Reithrodon as phyllo- and Punomys. tines. Their analysis relied on similarity in number The glans penis of Neotropical "cricetids" was and form of unhanded chromosomes. They ex- first systematically examined by Hooper and Mus- plicitly excluded Zygodontomys but did not ex- ser (1964), who inferred evolutionary relation- amine the genera A ndalgalomys (first described in ships among 1 9 genera based on estimates of over- 1978 as a Graomys), Euneomys, Galenomys, Ir- all similarity (Fig. 3). They found no diagnostic enomys, Pseudoryzomys, or Punomys. trait among the diverse phalli of five phyllotine Reig (1980, 1986) viewed all the major sigmo- genera. They seem to have excluded Zygodonto- dontine tribes as paraphyletic and stated that the mys (diagrammed near the base of the sigmodon- phyllotines most likely evolved "directly from the tine radiation; Fig. 3), although their discussion oryzomyines through the akodontines" (Reig, indicates that it could also be placed at the base 1 986, p. 426). Reig ( 1 986) also suggested that both of the phyllotines. The cited similarity between sigmodont genera, Holochilus and Sigmodon, were Eligmodontia and Akodon could lead to the in- (independently?) derived from a Neotomys-like terpretation of Eligmodontia as either a basal phyl- ancestor in Peru. lotine or akodontine. They suggested that Holo- Spotorno (1986) explored the akodontine and

FIELDIANA: ZOOLOGY phyllotine radiations (which he viewed as sister- Olds and Anderson (1989) also identified and groups) using banded karyotypes, electrophoresis, diagnosed a distinct "Reithrodon-group" that in- glans penis and bacular morphology, and cranial cluded Euneomys and Neotomys. They alluded to morphometries. He argued that the phyllotines a relationship of this group to the remaining sig- were monophyletic, citing simplification and pla- modonts but left this relationship unspecified. nation of their molars, differentiation of the distal Braun (1993) generally agreed with Olds and baculum, and a poorly developed base of the prox- Anderson (1989) on the composition of Phyllotini imal baculum as characteristic features. Like but additionally included Pseudoryzomys as the Hershkovitz (1962) and Reig (1986), Spotorno most basal phyllotine. She did not recognize a considered the phyllotines to be derived from an Reithrodon group but instead found support for a akodontine ancestor. Though he drew no definite generic group that included Reithrodon, Euneo- conclusions about phylogenetic relationships be- mys, Neotomys, and Auliscomys along with An- tween genera, his concept of the phyllotines in- dinomys, Chinchillula, Galenomys, Irenomys, and cluded Andinomys, Auliscomys, Calomys, Chin- Punomys. Braun also elevated Auliscomys boli- chillula, Eligmodontia, Euneomys, Graomys, viensis to Maresomys, reinstated Paralomys for Irenomys, Phyllotis, and Reithrodon. Spotorno did Phyllotis gerbillus, additionally including Phyllotis not explain why he considered Reithrodon a phyl- amicus in the reinstated genus, and resurrected lotine but Neotomys a sigmodont. Punomys was Loxodontomys for micropus, which she removed listed as Sigmodontinae incertae sedis and not an- from Auliscomys. alyzed. Pseudoryzomys and Zygodontomys were In an earlier version ofthis study (Steppan, 1 993), omitted. using nearly the same phyllotine data set (see Ma- Olds and Anderson (1989) presented the first terials and Methods), I excluded Punomys from formal diagnosis of Phyllotini and the first im- the phyllotines. Phyllotis was found to be para- plicitly cladistic treatment of the group, providing phyletic, but the internal nodes were very poorly a foundation for this examination of phyllotine resolved. Calomys was also paraphyletic, with C. monophyly and tribal relationships. They includ- sorellus as the sister-species to all remaining phyl- ed Punomys and excluded Pseudoryzomys and lotines. Confirmation was found for the Reithro- Zygodontomys. In their survey of 33 sigmodontine don group, which was most closely related to Au- genera ( 1 4 phyllotine and 1 9 nonphyllotine), they liscomys, Galenomys, and the resurrected could not find any unique synapomorphies for the Loxodontomys. Graomys appeared paraphyletic phyllotines. All phyllotines were found to have the with respect to Andalgalomys, but character sup- following combination of characters: port was not strong. Eligmodontia was most closely related to Graomys, and this group appeared to hairy heel, ears moderate to large, palate long (except be derived from Phyllotis. in Irenomys), incisive foramina long, parapterygoid fossa relatively broader than mesopterygoid fossa (except in Punomys), sphenopalatine vacuities large, supraorbital region never evenly curved in cross section, interparietal well developed, zygomatic notch deeply Materials and Methods excised (less so in Irenomys), teeth tetralophodont, M3 more than half the length of M2. (Olds & An- Taxa and Characters Examined derson, 1989, p. 63.)

Determining whether these characters are actually Two separate but nested phylogenetic analyses synapomorphies for the phyllotines requires a were conducted: one on the Sigmodontinae and phylogenetic hypothesis for the subfamily. Olds one on the Phyllotini. The sigmodontine analysis and Anderson (1989) incorporated into their di- included 29 ingroup species representing all named agnosis characters that "may be synapomorphic" tribes or major generic-groups. Eleven outgroup and recognized the difficulty of diagnosing the taxa included Nyctomys and Tylomys (Central phyllotines given the current knowledge of tribal American genera of uncertain affinities to the sig- relationships by describing the diagnosis as a "hy- modontines and neotomine-peromyscines), the pothesis for future testing and elaboration" (p. 63). terminal neotomine-peromyscines Neotoma and I will test their hypotheses of phyllotine mono- Peromyscus, the basal neotomine-peromyscines phyly and associated synapomorphies by con- Ochrotomys and Scotinomys (Carleton, 1 980), and ducting a more broadly based cladistic analysis for the Old World "cricetids" Calomyscus, Cricetulus, the subfamily. Mesocricetus, Phodopus, and Mystromys. Rela-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 2. included in 1 tribal Table Species phyllotine analysis. tromyinae (see Table , with subfamily and classification). Estimates of the number of phyl- Thomasomyine group lotine species vary with group limits and specific Thomasomys baeops status of taxa, with most estimates between 40 and Tribe Oryzomyini 45. The phyllotine analysis included 35 phyllotine Holochilus brasiliensis OTUs 33 in 14 Nectomys squamipes representing putative species phyl- Pseudoryzomys simplex lotine genera, in addition to 1 2 species belonging brevicauda Zygodontomys to 1 1 outgroup genera (Table 2). Tribe Ichthyomyini Character assessments were made from direct Ichthyomys hydrobates examination of museum Muse- Tribe Akodontini specimens (Field Akodon albi venter um of Natural History, Chicago, fmnh; Museum Akodon boliviensis of Vertebrate Zoology, University of California, andinus Chroeomys Berkeley, mvz; National Museum of Natural His- Oxymycterus hispidus tory, Smithsonian Institution, Washington, D.C., Tribe Scapteromyini of of Zool- Scapteromys tumidus usnm; University Michigan Museum Sigmodontinae incertae sedis ogy, Ann Arbor, ummz; Museo Nacional de His- lemminus Punomys toria Natural, Santiago, Chile, mnhn; The Muse- Tribe Phyllotini um, Michigan State University, East Lansing, msu; Andalgalomys pearsoni examined listed in the Phal- Andinomys edax specimens Appendix). A uliscomys boliviensis lic measurements for some species were measured Auliscomys pictus from published illustrations (Hooper & Musser, sublimis Auliscomys 1964; Spotorno, 1986). Evidence of two pairs of Calomys callosus preputial glands was gathered from the literature Calomys hummelincki Calomys laucha (Voss & Linzey, 1981) for some species. Stomach Calomys lepidus and some hyoid data are from Carleton (1980), Calomys sorellus gallbladder data are from Voss (1991), and mam- Chinchillula sahamae mae number is from Gyldenstolpe (1932), Hersh- Eligmodontia morgani and Olds Euneomys chinchilloides kovitz(1955, 1959, 1962, 1966b), (1988). Euneomys petersoni Dental nomenclature follows Reig (1977, and Fig. Galenomys garleppi 4). domorum Graomys A broad survey of characters from varied ana- Graomys griseoflavus tomical systems was conducted, resulting in 40 Irenomys tarsalis for the and 98 Loxodontomys micropus characters sigmodontine analysis Neotomys ebriosus characters in the phyllotine analysis covering den- amicus Phyllotis tal, cranial, postcranial, external, gastrointestinal, andium Phyllotis and male reproductive tract systems. Seventeen Phyllotis caprinus characters were shared by the two analyses, but 5 Phyllotis darwini 1 in Phyllotis definitus of the 7 were coded differently each. Many of Phyllotis gerbillus those 1 7 were included in the phyllotine analysis Phyllotis haggardi to help define outgroup relationships. Previous Phyllotis magister surveys have found little variation in soft anatomy Phyllotis osilae that was not evidenced Phyllotis wolffsohni among phyllotines already Phyllotis xanthopygus rupestris in the skeleton (Carleton, 1973; Voss & Linzey, Phyllotis xanthopygus xanthopygus 1981; Voss, 1991). The 40 sigmodontine charac- Reithrodon auritus evae ters represent 1 1 4 character states and a minimum Reithrodon auritus pachycephalus of 74 character state transitions. The 98 phyllotine Reithrodon typicus characters represent 265 character states and a " Removal of micropus from Auliscomys to Loxodon- minimum of 167 character state transitions. Char- tomys recommended by Braun (1993) and Steppan ( 1 993). acter state descriptions were defined so as to be more objective or quantitative than they have been tionships among the Old World "cricetids" are in the past. Ambiguous terms such as "relatively unclear, but in a recent treatment (Musser & Carle- broad," "large," or "well developed" were gen- ton, 1993) these five taxa represented the murid erally but not entirely avoided. Quantitative char- subfamilies Calomyscinae, Cricetinae, and Mys- acters or those with quantitative components were

FIELDIANA: ZOOLOGY Anterolabial conule Anterohngual conule Anterof lexus

Protostyle Anteroloph Parastyle Paraflexus Protophule Paroloph PARACONE Paralophule Mesoflexus Mesostyle Mesoloph Metaflexus Metalophule METACONE Metoloph

Posteroflex js

Posterostyle Posteroloph

Anterolabial conulid Anterolingual conulid Anteroflexid

Protoflexid Anterolophid Metastylid Protostyhd Metaflexid Metalophid Anterolabia cingulum METACONID Anterior murid Protolophulid PROTOCONID Metalophulid Mesof lexid Hypoflexid Mesolophid

Mesostylid Ectoslylid Entoflexid

Ectolophid Entolophulid ENTOCONID Mesoconid Entolophid Posteroflexid Median muri d Hypolophulid HYPOCONID Posterostylid Posterolophid

Fig. 4. Master plan of the occlusal surfaces of idealized first upper and lower molars of a cricetid rodent. All possible elements are shown with their corresponding names (from Reig, 1 980).

measured using a digital caliper precise to ± 0.005 siomorphic character states are not always desig- mm and values were rounded to the nearest 0. 1 nated "0." Characters were treated as ordered un- mm for coding. External measurements were re- less otherwise noted. corded from specimen tags. Character polarities Outgroup taxa in the phyllotine analysis were were determined by outgroup rooting within the selected to include representatives of each of the parsimony analysis rather than a priori, so ple- sigmodontine tribes and major generic groups (ex-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI Reithrodon Auliscomys Phyllotis t Eligmodontia Reithrodontomys Peromyscus MDH-' EAP* And -cE Irenomys ACON •* 'Abe Andinomys c Euneomys Abrothrix SI, Oligoryzomys Oryzomys b _JPEPj-C Sigmodon Scotinomys PEP CI* PEP B2 C P.r Neotoma •Mm Microtis Mui Mesocricetus MDH 1* PEP Dl" £ Mus Rattus

Fig. 5. A. UPGMA dendrogram ofelectrophoretic similarities (from Spotorno, 1 986). B. Seventy percent majority- rule consensus tree of 1 4 1 equally most-parsimonious trees derived from the original allele data. In cladistic reanalysis, individual alleles were treated as character states, proteins as characters. Character states were unordered in a cladistic analysis using PAUP. A strict consensus of the most-parsimonious trees is completely unresolved.

cept the monotypic Wiedomyini). Both analyses this data set (see Discussion under Phylogenetic used the preferred method ofMaddison et al. ( 1 984) Relationships within Sigmodontinae). when outgroup relationships are not well resolved, by simultaneously resolving ingroup and outgroup Quantitative Character Coding relationships under global parsimony. The resulting sigmodontine network was rooted by desig- Quantitative characters, in this case the four ra- nating the Old World "cricetids" as outgroups. tio characters (M3 length/M2 length, M2 width/ The phyllotine network was rooted in accordance M2 length, ear/body, interparietal/parietal), were with the results from the sigmodontine phylogeny coded using a minor variation on segment-coding and consistent with the common estimate of basal (Chappill, 1989). In segment-coding, itself a vari- sigmodontines (Hershkovitz, 1962; Reig, 1980, ation ofrange-coding (Colless, 1 980), the total range 1986; Voss, 1993; Voss & Carleton, 1993). Sig- of mean values for the taxa is divided into a num- modon was not included in the final phyllotine ber of equal-length segments. Segment-coding cat- analysis because previous molecular and morpho- egorizes an ordered series of OTUs into discrete logical phylogenies were highly discordant on its character states by creating a discriminant crite- position among sigmodontines. DNA hybridiza- rion that is a multiple of the pooled within-group tion has been reported to show Sigmodon to be standard deviation. Thus, the extent to which the basal member of a Neotropical group, outside OTUs are grouped together is objectively deter- a group that included the oryzomyines, Akodon, mined by the actual observed variability within and Phyllotis (Catzeflis et al., 1993), though the each of the OTUs. This characteristic of segment- data were not presented. Likewise, albumin im- coding and similar techniques, such as generalized munological distances placed Sigmodon outside gap-coding (Archie, 1985), is justified by the ar- a clade that included oryzomyines, akodontines, gument that the ease with which an evolutionary phyllotines, and ichthyomyines (Sarich, 1985). unit can evolve from one character state to another

Sigmodon was clustered with the North American (e.g., response to selection) is a function of the neotomine-peromyscines in phenetic (Spotorno, amount of genetic variance present for that trait 1986, and Fig. 5 A) and cladistic (Fig. 5B; reanaly- (Farris, 1966; Kluge & Farris, 1969; Archie, 1985). sis of data in Spotorno, 1986) analyses of electro- Segment-coding proceeds by first calculating the phoretic data, and its inclusion resulted in three pooled within-group standard deviation (sp ) for discordant and unconventional tree topologies with the set of taxa and then choosing a value for the

10 FIELDIANA: ZOOLOGY multiplier (c). This size of the multiplier deter- the segments, whose lengths were calculated a mines the percentage of overlap between the dis- priori, to shift as a group so that segment bound- = tributions of two OTUs (e.g., lsp 31% overlap aries could fall within the largest available gaps. = between two populations, 3sp 7% overlap) (Ar- The segments were not allowed to shift more than chie, 1985). Use of larger multipliers represents a one-half a segment. The objective of this shift was more conservative estimate of the number of bi- to avoid arbitrarily splitting two taxa with very ologically significant state transitions. Characters similar values and placing them into different that vary little between taxa but show high intra- character states. The multiplier used for all four specific variability would be subdivided less than quantitative characters was 4, chosen to yield an characters with relatively little intraspecific vari- overlap between OTUs of less than 5% (Archie, ation. The OTUs are then ordered (usually as- 1985). This is larger than that recommended by cending) by the magnitude of their means. Starting Chappill (1989), who preferred 'A or lh (yielding at one end of the series, all those OTUs that fall overlaps of 45% and 40%, respectively), but those within a group bounded by csp are joined in a small multiplier values would result in dozens of "segment." This step is repeated for each subse- character states for each character, raising the like- quent segment. The process is repeated until the lihood of excessive influence by the quantitative last OTU in the series has been joined into a seg- characters on the phylogenetic analysis. A multi- ment. These subsets are then converted to codes plier of 4 is commonly used with generalized gap- by increasing the code value by 1 for each segment coding and its related techniques. transition. The size of the segments is determined a priori as a multiple, c, of the pooled within-group standard deviation, sp . In this way, the number of character states is determined by the amount of Analytical Methods infraspecific variation relative to interspecific variation. Phylogenetic hypotheses were generated under One drawback of generalized gap-coding and its the principle of Wagner parsimony using the com- related techniques is that the position of one end puter program PAUP, version 3.1.1 (Swofford, of a subset is strongly influenced by the distribu- 1993). Heuristic tree-search algorithms were em- tion of OTU values at the other end. Nearly iden- ployed rather than the exact methods of exhaus- tical OTUs can be categorized into two different tive search or branch-and-bound, which required states because of the specific value of the OTU at prohibitively long computer runs with the many the other end of the subset. In other words, taxon taxa included in this study. Minimum-length trees sampling can significantly affect a generalized gap- were accumulated from multiple replicate analy- coding scheme and the addition of even a single ses, each starting with a different random tree. taxon can require a recoding of all others. Tradi- Experience with these data sets demonstrated that tional gap-coding techniques place state transi- with this many taxa (>40), most single replicates tions at large gaps but have a series of other short- will not find trees of the minimum length. Con- comings. The number of character states increases sensus trees were produced from the accumulated with the number of OTUs in generalized gap-cod- minimum-length trees. The sensitivity of the re- ing, independent of the range and multiplier value sulting topologies was tested by multiple runs in (the number often decreases in gap-coding). With which particularly interesting or pivotal taxa or a large number of OTUs, there will be a large characters were excluded. Additionally, 100 (sig- number of character states. If the magnitude of the modontine) and 200 (phyllotine) replicate boot- multiplier is increased to compensate for this ef- strap analyses were performed on the data sets to fect, the result is to concentrate the transitions provide nonparametric estimates for the confi- toward the extremes of the series, decreasing the dence to be placed in each node of the trees. Boot- phylogenetic information content. Standardizing strapping randomly resamples the characters in segment (group) lengths a priori minimizes or the data set with replacement (Felsenstein, 1985). eliminates these shortcomings. For more detailed The tree-search algorithm of PAUP can be con- discussions and critiques of the various quanti- strained so that it retains only those trees con- tative coding techniques, the reader is referred to forming to an a priori tree topology. The difference prior studies (Archie, 1985;Chappill, 1989; Mick- in tree length between the most-parsimonious trees evich & Farris, 1981) and references therein. overall and the constrained tree provides addi- The modification used in this study is to allow tional information in evaluating alternative phy-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 11 Table 3. Data matrix for sigmodontine analysis. Table 3. Extended.

17 Table 4. Data matrix for phyllotine analysis.

1 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Thomasomys baeops Ichthyomys hydrobates Holochilus brasiliensis Nectomys squamipes Pseudoryzomys simplex Zygodontomys brevicauda Akodon albi venter Akodon boliviensis Chroeomys andinus Oxymycterus hispidus Scapteromys tumidus Punomys lemminus Calomys callosus Calomys hummelincki Calomys laucha Calomys lepidus Calomys sorellus Andalgalomys pearsoni Graomys domorum Graomys griseoflavus Eligmodontia morgani Galenomys garleppi Auliscomys boliviensis Auliscomys pictus Auliscomys sublimis Euneomys chinchilloides Euneomys petersoni Reithrodon auritus evae Reithrodon auritus pachycephalus Reithrodon typicus Neotomys ebriosus Loxodontomys micropus Irenomys tarsalis Andinomys edax Chinchillula sahamae Phyllotis amicus Phyllotis andium Phyllotis caprinus Phyllotis darwini Phyllotis definitus Phyllotis gerbillus Phyllotis haggardi Phyllotis magister Phyllotis osilae Phyllotis wolffsohni Phyllotis xanthopygus rupestris Phyllotis xanthopygus xanthopygus Table 4. Extended.

21 Table 4. Continued. Table 4. Continued.

71 max

jug

srz

sqm

soc

FIELDIANA: ZOOLOGY but this observation depends on how much vari- 00 = absent ation is subsumed within the orthodont category. 10 = present, small, set medially = Hershkovitz (1962) described A. boliviensis and 20 present, medium to large, set laterally = Galenomys as having proodont incisors, but my ?1 2 lateral roots observations of many of the same specimens that he examined led me to categorize them as ortho- The character states and transition series match dont, as are A. sublimis and A. pictus. The lower Carleton (1980) with the exception that Carleton incisors of Galenomys are also highly proodont, hypothesized that two lateral roots were derived and Hershkovitz (1962) reported that Galenomys directly from roots absent, while I allow two roots is more pronounced in this regard than any other to be derived from any of the states in a single "cricetine." step. Carleton hypothesized the absence of labial Most of the remaining phyllotines have opis- roots to be plesiomorphic for neotomine-pero- thodont incisors. Hyper-opisthodont incisors are myscines, but a large lateral root is the widespread limited to C. callosus and C. laucha, Eligmodon- and possibly plesiomorphic condition among tia, Reithrodon, Neotomys, Irenomys, and Loxo- phyllotines. Auliscomys and Andinomys have a dontomys. small medial root, while Euneomys, Neotomys, and Irenomys lack it altogether. The condition in 3 P. Upper Incisor Dentine Fissure - 3 states Loxodontomys is unclear because it possesses a (Fig. 10). second root along the lateral border, which may = long straight slit be a modification of the primitive condition. = 1 short, not quite linear slit, "comma"-shaped Molar Roots of M2— Not coded. The wide- = 2 tripartite, "Y"-shaped spread condition among phyllotines is a single large lingual root in addition to the anterior and pos- In most phyllotines and other sigmodontines, terior roots. Reithrodon auritus pachycephalus and the dentine of the incisor is cleaved anteroposteri- possibly Euneomys chinchilloides have a partially orly into a long, straight slit (Fig. 1 0C). Members bifurcated lingual root. The condition in other spe- of the Reithrodon, Auliscomys, and Andinomys cies of those genera is not known. Chinchillula generic-groups are characterized by modifications lacks the lingual root entirely, consistent with the of this condition. The genera Auliscomys, Chin- general reduction in the number of molar roots in chillula, Galenomys, and Irenomys show a shorter that genus. slit that becomes rounded or "comma"-shaped at 6P. Molar Roots of M3 — 3 states. the anterior end (Fig. 10B). This condition can also be found in P. definitus and in some speci- = 3 roots mens of P. wolffsohni. In the third condition, the 1=2 roots = anterior end splits in two, becoming tripartite or 2 1 root "Y"-shaped (Fig. 10A). This condition is found in the Reithrodon generic-group, in Loxodonto- mys, and in northern Andinomys edax. The tri- Phyllotines show one, two, or three roots in the partite condition is best developed in Loxodon- third upper molar, with three roots the common tomys and has not been observed by me outside condition. Carleton (1980) considered three roots the phyllotines. This trait often shows some vari- to be plesiomorphic for the neotomine-peromys- ability within species (most notably in Andinomys, cines. Reduced numbers occur in Andinomys, which exhibits both states "0" and "2"), and wear Chinchillula, Loxodontomys, A. boliviensis, Phyl- patterns can make it difficult to determine whether lotis gerbillus, P. darwini, and P. xanthopygus. All the straight or "comma"-shaped condition is pres- these examples have two roots except for Chin- ent. chillula, which has the most highly derived condition of a single root. 4P 5P. Labial Root of Ml •4 states, 2 sub- characters. 7P. Labial Root of m! — 2 states.

Fig. 6. Dorsal view of a generalized Phyllotis cranium, ab, antorbital bridge; fr, frontal; ip, interparietal; jug, jugal; lc, lachrymal; max, maxillary; mrz, maxillary root of zygomatic arch; nas, nasal, nlf, nasolacrimal foramen; par, parietal; pre, premaxillary; soc, supraoccipital; sqm, squamosal; srz, squamosal root of zygomatic arch; zn, zygomatic notch.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 19 20 FIELDIANA: ZOOLOGY = absent transformation series. In some outgroup taxa that = 1 present superficially appear to lack the anteromedian flexus (e.g., Pseudoryzomys, Zygodontomys), the rem-

Carleton ( 1 980) considered absence of the labial nant enamel island from a fully infolded and cutoff root to be plesiomorphic for the neotomine-per- flexus can be seen in relatively unworn teeth. Cut- omyscines, but the presence of a labial root is the off flexi can also be clearly seen at all ages in many widespread condition among phyllotines. Only oryzomyines. There is thus the distinct possibility Euneomys, Neotomys, and Irenomys lack this root. that the absence of an anteromedian flexus could These are the same species that lack the labial root be a secondary loss from a derived, infolded con- on M 1 . The intermediate condition found in some dition rather than from reduction of the flexus Mis (presence of a small, medially positioned root) depth. Additionally, enamel islands, sometimes is not observed in the ml of any phyllotine. connected to the flexus, can be seen in slightly worn teeth of some phyllotine species. The con- 8 P. Molar Roots of m2— 2 states. dition for species lacking this ontogenetic infor- = 2 roots mation is conservatively coded as unknown, "?".

1 = 3 roots A distinct anteromedian flexus is found in An- dalgalomys, Calomys, Galenomys, and Aulisco- Three roots is the widespread condition, again mys pictus. in apparent contrast to Carleton's (1980) hypothesis that two roots was plesiomorphic for the neo- IS. Mesoloph(-id)— 3 states. tomine-peromyscines. The derived reduced state = mesoloph(-id) joined with mesostyle(-id): is again found in Euneomys and Irenomys, but not (pentalophodont) in Neotomys. Two roots are also found in Andi- = 1 small mesoloph or mesolophid present, does nomys, Auliscomys, and Loxodontomys as well as not join with mesostyle(-id), or partially fused P. magister, P. haggardi, and P. wolffsohni. How- with paracone ever, sample sizes are usually one or two, so in- = 2 absent: (tetralophodont) dividual variation is difficult to p^sess. Hershkovitz 1 out the for — ( 993) pointed potential 9P. Molar Roots of m3 2 states. mistaking a paralophule (arising from the para- with a from the = 2 roots cone) mesoloph (arising mure). Mesoloph is weakly developed in some akodon- 1 = 3 roots tines and Anotomys and usually partially fused to Again, the widespread state among the phyllo- paracone when present; there may be either a well- tines is for the full complement of three roots. Two developed mesoloph or paralophule in Scaptero- roots were found in Andalgalomys, Eligmodontia, mys (a partially fused mesoloph seems most likely). Galenomys, and Neotomys. It is unknown if three Among sigmodontines, complete mesolophostyles roots are found in the unexamined species of An- (mesoloph fused with the mesostyle) are found dalgalomys and Eligmodontia. only in oryzomyines and thomasomyines. The pentalophodont condition is conventionally hy- 10P. Anteromedian Flexus Ml— 4 states. pothesized to be plesiomorphic (e.g., Hershkovitz, = absent or limited to shallow groove 1962, 1966b, 1993; Carleton, 1980). However, = 1 distinct or prominent placement of the root to the sigmodontine tree can = 2 infolded to form lake strongly affect this polarity assignment. Mesolophs = 3 loss from state "2," with reduction of lake are entirely absent in phyllotines and most ichthyomyines. Character states "0" and "3" are difficult to distinguish because both represent absence of the an- Enteroloph, Ectolophid— Not analyzed. The teromedian flexus but are at opposite ends of the absence of an enteroloph or ectolophid was listed

Fig. 7. Ventral view of a generalized Phyllotis cranium, als, alisphenoid; boc, basioccipital; bsp, basisphenoid; cc, carotid canal; ect, ectotympanic part of auditory bulla; fm, foramen magnum; fo, foramen ovale; if, incisive foramen; max, maxillary; mlf, middle lacerate foramen; mpf, mesopterygoid fossa; mpi, medial process of incisive foramen; ms, mastoidal capsule; occ, occipital condyle; pal, palatines; pet, petrosal part of auditory bulla; pi, posterolateral palatal pit; ppf, parapterygoid fossa; ppp, parapterygoid process; pre, premaxillary; psh, presphenoid; sf, stapedial foramen; spv, sphenopalatine vacuity; sqm, squamosal; sts, stapedial spine of auditory bulla.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 21 Fig. 8. Lateral view of a generalized Phyllotis cranium, ab, antorbital bridge; als, alisphenoid; ect, ectotympanic part of auditory bulla; fo, foramen ovale; fr, frontal; hp, hamular process of squamosal; ip, interparietal; jug, jugal; max, maxillary; ms, mastoidal capsule; nas, nasal; nlf, nasolacrimal foramen; of, optic foramen; par, parietal; pfg, postglenoid foramen; ppp, parapterygoid process; pre, premaxillary; soc, supraoccipital; spf, sphenopalatine foramen; sqm, squamosal; ssf, subsquamosal foramen; sts, stapedial spine of auditory bulla; tt, tegmen tympani; zp, zygomatic plate; zs, zygomatic spine.

by Olds and Anderson (1989) as a possible phyl- 13 P. Flexus Penetration Ml— 3 states. lotine synapomorphy, but they chose not to in- = flexi from opposite sides do not reach each clude it in the diagnosis. Olds and Anderson other implied that it shows a parallel pattern to the = 1 enamel overlaps, or flexi meet at midline mesoloph(-id) and cited Carleton's (1980) general 2 = flexi cross beyond each other concept of the derived and concerted simplification of various enamel structures. However, while This character varies strongly with age, and the most sigmodontines have a mesoloph(-id), among characterizations here are for adults with moder- surveyed species only Punomys has a distinct en- ately worn teeth. Wear reduces the apparent pen- teroloph and ectolophid. It is therefore phyloge- etration, so that in many taxa (e.g., Phyllotis), the netically uninformative within the context of this flexi of well-worn molars do not overlap. Taxa study. with highly involuted molars, the "sigmodont"

1 IP. Mesostyle Ml —2 states. condition of Hershkovitz (1955), include Euneo- mys, Reithrodon, and Neotomys. Less strongly in- = absent = voluted but still overlapping flexi are found in 1 present Graomys and Loxodontomys. Nonoverlapping flexi A small mesostyle on the labial border, uncon- occur sporadically among the phyllotines and can nected to a mesoloph, is found in Calomys and be found in species of Calomys, Phyllotis, Andal- Chinchillula. All other phyllotines lack the me- galomys, Eligmodontia, and Auliscomys in addi- sostyle. tion to some of the monotypic genera.

1 2P. Parastyle/Anteroflexus M 1 — 3 states. 1 4P. Anterolabial Cingulum m 1 — 3 states.

= absent = anterolabial cingulum absent = = 1 present, indistinct 1 anterolabial cingulum weakly developed, lost = 2 present, distinct with wear 2 = anterolabial cingulum distinct An indistinct parastyle and shallow anteroflexus are found in Calomys and Loxodontomys and may A distinct anterolabial cingulum that is present be plesiomorphic for the phyllotines. Both are ab- at all ages is the common condition in phyllotines. sent in C. sorellus and all remaining phyllotines. There is considerable variation in length and width The problematic Punomys has a well-developed that is discussed under protoflexid. Variation is parastyle. also associated with the size, shape, and overall

22 FIELDIANA: ZOOLOGY I /

Fig. 9. Variation in incisor grooves among phyllotines; pointers identify grooves. A, Auliscomys sublimis (fmnh 10771 1); B, Auliscomys pictus (fmnh 64344); C, Euneomys chinchilloides (fmnh 50600); D, Irenomys tarsalis (fmnh 124057); E, Reithrodon auritus (fmnh 134228); F, Neotomys ebriosus (fmnh 24777).

Fig. 10. Upper incisor dentine lake. A, tripartite, Loxodontomys micropus (fmnh 23237); B, curved, Irenomys tarsalis (fmnh 133164); C, straight, Graomys griseoflavus (fmnh 50923).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 23 complexity of the procingulum. Reduction of the the same uncertainties as I do; they did not hy- anterolabial cingulum occurs in Andalgalomys, pothesize an ancestral condition. Carleton (1980) Eligmodontia, Auliscomys, Reithrodon, and Neo- considered the absence of the flexid (undivided tomys, while it is completely absent in Euneomys. anterocone) to be primitive, but at a much more inclusive taxonomic level. The taxonomically 15P. Protoflexid Ml— 2 states. scattered occurrence of a faint remnant of the lake = short anterolabial cingulum, which may curl in young individuals of some Graomys, Phyllotis, toward protoconid; simple protoflexid and Auliscomys raises the likelihood that the ob- = 1 long anterolabial cingulum, fusing with pro- served absence in most Phyllotis and Auliscomys toconid and leaving protoflexid as lake is a secondary loss. Many of the species, particularly most Phyllotis, have been coded as unknown A medium to short anterolabial that cingulum to reflect that uncertainty. Prominent anterome- curve toward the but that does may protoconid dian flexids are found in most Calomys, Andi- not fuse is and widespread among sigmodontines nomys, and P. gerbillus. An infolded lake is found thus to be for the appears plesiomorphic phyllo- in Euneomys, Reithrodon, and Irenomys. tine. In the derived condition, which was not ob- 18P. Procingulum Separation 2 states. served in the outgroups, the anterolabial cingulum Ml— is and fuses with the at least very long protoconid = procingulum attached by anterior mure basally, leaving the protoflexid as a lake. The de- = 1 procingulum separated, mure cut by oppos- rived condition is found in all C. sorellus, Phyllotis, ing flexids Chinchillula, Galenomys, A. sublimis, and Gra- in is enomys. The protostylid may be made distinct by a The procingulum Euneomys (Fig. 40D) pinching of the cingulum proximally or may be tirely separate from the primary cusps by the ab- unrecognizable within the gradual thinning of the sence ofa connecting mure. This condition persists anterolabial cingulum. The protoflexid lake is most even in highly worn teeth. This trait is unique to the and all pronounced in almost all species of Phyllotis. Euneomys among phyllotines among examined sigmodontines. The separation of the 1 6P. Cusp Arrangement m 1 — 3 states. procingulum in the fossil Bothriomys was presum- = primary cusps opposite in position ably important to Hershkovitz's decision to syn- = it with 1 primary cusps intermediate onymize Euneomys (Hershkovitz, 1962). 2 = primary cusps alternate However, the procingulum is distinctly triangular in Bothriomys, similar to Reithrodon and Neoto- Primary cusps can be positioned so that the la- mys, rather than the equally distinct round pro- bial and lingual pairs are each opposite each other, cingulum in Euneomys. with no anterior or posterior shift. In the alternate 1 9P. 1 — 3 states. arrangement, the metacone is situated across from Posterolophid/Stylid m the a shift of hypoflexid, produced by posterior = absent the lingual conids relative to the labial conids. = 1 intermediate, posteroflexid present as groove, Carleton (1980) tentatively hypothesized opposite or obvious in juvenile, absent with strong cusps as plesiomorphic, but the intermediate con- wear dition is most the phyllotines. widespread among 2 = distinct at all ages Opposite cusps are found in Euneomys, Irenomys, and Loxodontomys. Close attention to ontogenetic The posterolophid is never highly developed and in but from a series is needed to verify cusp homology among large phyllotines ranges prominent the "sigmodont" genera in order to score this char- lophid to absence. The intermediate condition is acter. common, where the posteroflexid may be prominent in juvenile teeth but then lost with strong 17P. Anteromedian Flexid Ml — 3 states. wear. The widespread condition among phyllo- = absent or limited to shallow groove tines is for a prominent lophid to be present at all = are absent in the 1 prominent ages. Posterolophids entirely only = molars of and Chinchillula. 2 infolded to form lake, which may be lost with prismatic Irenomys wear The intermediate condition is found in Andalga- lomys, Galenomys, and all Phyllotis except P. ger-

Variation in this trait parallels that in M 1 . Ad- billus, P. amicus, and P. caprinus. dressing the oryzomyines where the lake is highly 20P. Posterolophid/Stylid m3 — 3 states. developed, Voss and Carleton (1993) described the lake as an "internal enameled pit" and cited = absent

24 FIELDIANA: ZOOLOGY = 1 intermediate, posteroflexid present as groove, Andalgalomys, Auliscomys boliviensis, P. osilae, or obvious in juvenile, absent with strong Euneomys, and Neotomys. The more developed wear but still bud-like procingulum appears in Ireno- = 2 distinct at all ages mys and Andinomys.

23P. Hypoflexus Reduction M3— 3 states. The range of variation among sigmodontines in = this structure is largely the same as for ml, al- no reduction relative to M2 = though in general the third molar is less complex 1 reduced relative to M2 = than the first molar. Reflecting this, no postero- 2 highly reduced relative to M2, to absent lophid is found in most phyllotines. The only ex- This character and the several ceptions are bud-like posterolophids in Euneomys following (24-30) describe the but subtle variation in and the occasional remnant of a posteroflexid in complex the of the third molars. General of Neotomys. shape descriptions shape (e.g., "S"-shape, "C"-shape) were tried ini- 2 IP. Procingulum M2— 4 states. tially, drawing from the descriptions in Carleton's character 5. = absent (1980) These generalities, however, = did not adequately describe the independent vari- 1 anteroflexus appears as groove = ation among the dental elements, other 2 protoflexus may appear also; if so, procin- among state determinations gulum poorly developed as broad, shallow things making problematic. Therefore, shape was broken down into its projection with concave anterior edge; if not, specific elements so that could be more de- then distinct antero- or paraflexus they precisely = fined. Although this runs the risk of 3 procingulum distinct, well developed overinflating the taxonomic importance of potentially interre- lated characters, I did not find objective grounds The procingulum on the second molar is phys- for reducing the weight of these characters in the ically constrained by the first molar and is much analysis. For each multistate character in this group, less complex than the procingulum on the first character states were treated as ordered. molar. For example, anterior conules are always absent. In phyllotines, it usually appears com- 24P. Reduction of Mesoflexus M3 — 3 states. pressed against the first molar. Nonetheless, de- = no reduction relative to M2 velopment of the procingulum on M2 exhibits rel- 1 = reduced relative to M2 atively high variability between individuals of the = 2 highly reduced relative to M2, or absent same species. Complete loss of the anteroflexus and procingulum is found in Irenomys, Chinchil- 25P. Posterior Shift of Mesoflexus M3— 2 lula, P. osilae, and P. haggardi. A well-developed states. procingulum is found in Reithrodon, Neotomys, = no shift relative to M2 Loxodontomys, and Andinomys. The comparative = 1 posterior shift relative to M2 presence or absence of accessory anterior or posterior lophs is diagnostic among the "sigmodont" 26P. Hypoflexus Lake M3 — 2 states. genera. Among them, only Neotomys has a pos- = hypoflexus present, no lake teroloph (on M3), and Euneomys lacks the pro- = 1 hypoflexus pinched to form lake cingulum on both M2 and M3 that is seen in Reith- rodon and Neotomys. In some phyllotines, the hypoflexus appears to 22P. Procingulum m2 — 3 states. have been pinched off to form a lake. The hypo- = absent flexus lake is sometimes extended anteroposteri- = 1 protoflexid appears as groove; if pronounced orly orthogonal to the general orientation of the in juvenile, then wears away with age hypoflexus, as in P. osilae. A hypoflexus lake is = 2 procingulum well developed found in some Calomys, Euneomys, Graomys domorum, and all Phyllotis except P. wolffsohni. The Development of the procingulum on m2 also widespread condition among sigmodontines and exhibits relatively high variability among individ- most phyllotine genera is an intact hypoflexus. uals of the same species. The common condition 27P. Rotation of Flexus Axes M3 — 2 states. for phyllotines is a weakly developed procingulum = in which the protoflexid appears as a groove and, no rotation of hypoflexus and mesoflexus axes if pronounced in juveniles, wears away with age. relative to M2 = Absence of a procingulum at all ages is found in 1 axes rotated relative to M2

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 25 28P. Mesoflexid Reduction m3 — 3 states. Anderson (1989) considered an M3 more than half the length of M2 to be diagnostic of the phyllo- = no reduction relative to m2 = tines. My observations of the taxonomic distri- 1 reduced relative to m2 bution of this character disagree with that reported 2 = highly reduced relative to m2, or absent by Olds and Anderson. I find that virtually all 29P. Anterior Shift of Mesoflexid m3 — 2 sigmodontines (26 of 28 surveyed) possess this states. condition. The source of this systematic discrep- ancy is unknown but may be due to measurement = no shift relative to m2 = criteria. They stated that "phyllotines seem broad- 1 anterior shift relative to m2 ly plesiomorphic in this regard" (p. 62), yet in- 30P. Posterior Shift of Hypoflexid m3 — 2 cluded it in the diagnosis. states. 3S. Shape M2 — 2 states. = no shift relative to m2 = = width < 0.91 length 1 posterior shift relative to m2 = 1 width > 0.91 length 3 1 P. Fusion of Opposing Flexi in M3 — 2 states.

at the = flexi do not meet Width was measured widest point, rather = than the occlusal to avoid variation due 1 flexi meet, median mure cut surface, to wear. Character state values were determined The widespread condition among both out- using segment coding with a multiplier of 4. Gen- groups (sigmodontines) and phyllotines is for the eralized gap-coding with a criterion variable of 4 opposing flexi not to meet at the midline. In Ir- results in the same state definitions. Olds and An- enomys, they meet and the enamel from opposing derson (1989) reported that the width in phyllo- 2 flexi fuse but remain intact; the flexi are not con- tines at the posterior half is /3 or more of the tinuous (Fig. 40). This is the condition for all pairs greatest length of the tooth, in contrast to the nar- of flexi and flexids in Irenomys. In three species, rower teeth of oxymycterines, akodontines, scap- Andalgalomys pearsoni and Graomys griseoflavus teromyines, and ichthyomyines. They suggested as well as Andalgalomys olrogi (according to pub- that phyllotines probably possess the plesiomor- lished photos [Olds et al., 1987]), the opposing phic condition. As with the ratio of M3/M2, my flexi in M3 do in fact fuse and become continuous. survey disagrees: no sigmodontine examined has 2 Graomys domorum is variable for this trait. Most a tooth narrower than /3 proportion. Width of the but not all individuals of Graomys griseoflavus occlusal surface, as was measured by Olds and exhibit fused flexi. In Andalgalomys, the flexi are Anderson (1989), is strongly affected by wear, and fused in M2 as well. these differences in definition and age criteria may account for the differences in coding. 32P. Ratio of M3 Length to Alveolar Length of Molar Tooth Row— 3 states.

= < 0.205 Cranium and Mandible

1 = 0.205-0.25 2 = > 0.25 33P. Capsular Projection of Mandible— 2 states. The occlusal length of M3 was compared to the = distinct capsule, or elevation of superior alveolar length of the maxillary tooth row. Among masseteric ridge, usually ventral to the coro- those taxa with enlarged molars, two classes are noid process recognized: most have moderate molars (< 0.25 = 1 indistinct or absent tooth row length) while large molars (> 0.25 tooth row length) are found in Irenomys and Neotomys. In many murids, a distinct capsule forms around the root of the lower incisor; this capsule then 2S. Length M3 — 3 states. projects out from the body of the mandible. This = < 0.63 length M2 character can vary among individuals, so mod- = 1 0.63 to 0.96 length M2 erate sample sizes are needed for coding. = 2 > 0.96 length M2 34P. Height of Coronoid Process— 3 states. = Character state values were determined using above maximum height of mandibular con- segment-coding with a multiplier of 4. Olds and dyle

26 FIELDIANA: ZOOLOGY Fig. 1 1 . Ventromedial process of the mandibular ramus (vmr). A, process distinct, Neotomys ebriosus (fmnh 24775); B, process weakly present, ramus sharply angled, Loxodontomys micropus (fmnh 124393); C, process absent, ramus rounded, Phyllotis andium (fmnh 19468).

= 1 subequal well posterior to the ventral curvature of the di- 2 = below mandibular condyle astema. The knob just reaches the dorsal edge of the mandible in C. callosus, C. laucha, Andalga- Coronoid that rise above the maxi- processes lomys, Graomys, most Phyllotis, Galenomys, and the relative mum height of mandibular condyle, R. auritus. The most extreme condition, which to the basal of the are plane mandible, common appears derived in reference to its absence among and were found all sur- among phyllotines among other sigmodontines, has the knob exceeding the veyed outgroups. dorsal edge of the mandible and is found only in and P. 3 5 P. Anterior Masseteric Ridge Position— 4 Eligmodontia gerbillus. states. 36P. Medioventral Process of Mandibular

Ramus— 3 states 1 = anterior edge not formed into a knob, well (Fig. 1). ventral to dip of diastema = ramus rounded when viewed = process absent, 1 knob slightly below dorsal edge of mandible or not = ventrally sharply angled 2 knob just reaches dorsal of mandible = edge 1 process weakly present, or ramus sharply an- 3 = knob exceeds dorsal edge gled, near 90° 2 = process distinct The anteriormost extent of the masseteric ridge is usually enlarged into a small knob or swelling, At the posterior terminus of the symphysis, the and the ridge varies between taxa in its dorsal/ ventral surface of the ramus curves so that the two ventral position as well as in anterior/posterior halves of the mandible diverge. The radius of cur- position relative to the mental foramen and the vature at this point is variable among taxa. It rang- ventral curvature of the diastema. This character es from smoothly curved, to sharply angled (« is surprisingly stable within species. A knob just 90°), to the presence of a distinct pair of processes below the dorsal edge of the mandible at the di- nearly rejoining at the midline (Fig. 1 1). Greater astema is common among the phyllotines. In Ir- development of this trait is generally associated enomys and Andinomys, the knob is below and with robustness of the jaw. Intraspecific variation

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 27 can sometimes cross over the categorical boundary tocone in Irenomys, Loxodontomys, Auliscomys between smoothly and sharply rounded. The more sublimis, and A. pictus. Even longer foramina that widespread condition among sigmodontines is extend to the hypocone and metacone are diag- smoothly rounded (Fig. 1 1 C). The sharply angled nostic of Reithrodon. The fossil Proreithrodon condition (Fig. 1 1 B) is found in G. domorum, the (Roverto, 1914), which was synonymized by remaining Phyllotis, and all other genera except Hershkovitz (1955) with Reithrodon, appears to Neotomys. The condition in Neotomys is further have much less extensive incisive foramina than modified into distinct processes (Fig. 1 1 A), diag- any living Reithrodon, raising questions about its nostic for the genus, although some individuals of proper taxonomic placement. Andinomys have moderately developed processes. 39P. Maxillary Septum of Incisive For- 37P. Premaxillary Protrusion— 3 states. amina— 3 states. = = < premaxillaries terminating behind the ante- length '/2 incisive foramina = rior plane of the incisors 1 length Vi-% incisive foramina = = 1 premaxillaries terminating at or slightly an- 2 length > % incisive foramina terior to incisive plane The length of the maxillary septum, which is 2 = premaxillaries produced well anterior to in- visible in the incisive foramina, can vary inde- cisive plane pendently of the posterior extent of the incisive This character is defined by the point at which foramina. The phyllotines show little variation for the anterior edges of the premaxillaries terminate this character, with all species surveyed having relative to the most anterior margin of the upper septa that extend between lh and % the length of incisors. This character is structurally associated the incisive foramina. Calomys hummelincki has with incisor procumbency, but most variation ap- septa that may be less than lh the foramina length. pears to be independent of that dental character. Greater variation in this trait occurs outside the phyllotines, where most akodontines have very 4S 38P. Posterior Extent of Incisive For- short septa (< Vi) and most oryzomyines have amina— 4 states. septa that exceed % the incisive foramina length. = not reaching anterior margin of anterolabial 40P. Orientation of Incisive Foramina— 2 and anterolingual conules of M 1 = states. 1 level with anterolabial and anterolingual = conules separation of anterior apexes < 80% sepa- 2 = extending to level of paracone and protocone ration of posterior apexes = = 3 extending to level ofhypocone and metacone 1 separation of anterior apexes 80-100% of posterior apexes This character is often imprecisely but conve- niently designated as "length" of the incisive fora- Among sigmodontines, the posterior apexes of mina. Among sigmodontines, foramina that do the incisive foramina are typically more widely not reach the first molars are characteristic of the set apart than the anterior apexes. Among some ichthyomyines, many oryzomyines, and some tho- phyllotines, the anterior region is more robustly masomyines. Most other sigmodontines exhibit excavated. In Reithrodon, Irenomys, and Andi- state "1" where the incisive foramina reach the nomys, the anterior apexes are separated by a dis- anterior conules. The widespread presence of state tance between 80 and 100% that separating the "0" ("short foramina") among the outgroups sur- posterior apexes. In Neotomys, the anterior apexes veyed in this study suggests that condition as the are most robust, being as broadly separated as the plesiomorphic condition for sigmodontines, but posterior apexes. Voss and Carleton (1993) considered the plesio- 5S41P. DORSOVENTRAL POSITION OF ANTERIOR morphic state as unknown. The most extensive Root of Zygomata— 3 states (Fig. 12). foramina are found among some akodontines and = phyllotines. antorbital bridge lying well below dorsal sur- The widespread condition among phyllotines is face ofrostrum ('A-Vi less than rostrum height, for the incisive foramina to terminate approxi- as measured from the midpoint between mately medial to the anterior conules of M 1 . Fora- height of zygomatic spine and anteriormost mina that do not reach the anterior alveoli of the border of orbit) = molars are found among phyllotines only in Neo- 1 antorbital bridge below rostrum (displaced tomys. Foramina extend to the paracone and pro- < '/» rostrum height)

28 FIELDIANA: ZOOLOGY 2 = insertion high, close on dorsal surface ros-

trum (< Vs) or posterior surface of bridge joins at dorsal level of surface

In sigmodontines, the anterior root of the zygomatic arch rests upon the zygomatic plate, which leans out from the rostrum. The structure connecting the dorsal surface of the zygomatic plate with the rostrum is the antorbital bridge. The dorsoventral position of the antorbital bridge was measured at the midpoint between the zygomatic spine and the anteriormost margin of the orbit (Fig. 12). The widespread condition among sigmodontines and the likely plesiomorphic condition among phyllotines is for the antorbital bridge to lie well below the dorsal surface of the rostrum, between Vi and Va the depth of the rostrum below the dorsal surface (Fig. 1 2A). This condition characterizes Calomys and Andalgalomys, although some individuals off. amicus have low-lying antorbital bridges. Most phyllotines have the condition where the antorbital bridge is displaced between Va and V$ the rostral depth below the dorsal surface. A still more extreme condition, where the antorbital bridge lies close on the dorsal surface of the rostrum ( < V» the rostral depth), is found in Chinchillula, Euneomys, Neotomys, Reithro- don, and Auliscomys boliviensis. In Euneomys, the bridge actually inserts onto the dorsal surface of the rostrum (Fig. 1 2B).

6S 42P. Posterior Margin of Zygomatic Plate— 2 states. = anterior to M 1 alveolus = 1 or anterior to alveolus subequal Fig. 12. Position of the anterior root of the zygomata, with measurement positions indicated; pms, pre- This character was the position of the judged by maxillo-maxillary suture. A, Vi-'/i below dorsal surface posterior margin of the zygomatic plate (equiva- of rostrum, Calomys laucha (fmnh 29246); B, inserting lent to the anterior margin ofthe orbit when viewed high on the rostrum, Euneomys chinchilloides (fmnh ventrally) relative to the anterior alveolus of the 133088). first upper molar. Both states are present among "cricetid" outgroups, but state "0" appears to be ly plesiomorphic (Voss, 1988). The only other tax- the more common. on identified with a masseteric tubercle is the peromyscine Megadontomys (Voss, 1988). 7S. Masseteric Tubercle— 2 states. 8S. Zygomatic Notch— 3 states. = absent = 1 present = absent = 1 shallow, depth < Vi width of notch, and notch The tendon of the superficial masseter attaches less than half length of zygomatic plate at the bases of the inferior zygomatic root. Among 2 = deep, well developed ichthyomyines and to a lesser extent Scotinomys, the point of insertion is on a distinct bony spur Carleton (1980) concluded that an absent or (Voss, 1988, Fig. 12) projecting out from the zy- barely described zygomatic spine was the primi- gomatic root, rather than being marked by a patch tive condition for North American neotomine- of rugose bone. Absence of the masseteric tubercle peromyscines and other nonsigmodontine muroid is widespread among New World murids and like- rodents, and Olds and Anderson (1989) cited

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 29 = Carleton (1980) as justification for their polariza- < 20° (in frontal view) = tion of "no notch" as primitive in phyllotines. 1 > 20° However, Carleton's observation is weakly informative for phyllotine systematics without a strong The lateral inclination of the zygomatic plate phylogenetic hypothesis. Carleton's character 25 out from the rostrum varies moderately among actually referred to both the zygomatic notch and phyllotines (~ 10—40°) but also shows significant zygomatic spine, with emphasis on the spine, not variation within species. Zygomatic plates that are the notch. He viewed these two aspects of the max- inclined greater than 20° from the vertical are illary root of the zygomata as highly correlated, in widespread among sigmodontines and most likely a data set dominated by neotomine-peromys- plesiomorphic for the phyllotines. cines. However, within the phyllotines, as an ex- 45P. Premaxillo-Maxillary Suture Orienta- ample, the spine (which defines the anterior portion— 2 states. tion of the notch) can vary greatly between = relatively closely related species with little varia- 90-135° angle formed relative to palatine tion in the posterior development of the notch. plane by the suture on the lateral surface of The development of the spine is primarily con- rostrum and below antorbital foramen = trolled by the degree of excavation of the ventral 1 suture nearly horizontal at ventral end, body of the zygomatic plate, in addition to the sharply angled {> 90°) in middle of rostrum overall breadth of the plate. These two aspects of In most phyllotines and all other sigmodontines, the zygomatic plate need to be distinguished when the premaxillo-maxillary suture on the side of the discussing the sigmodontines. Deep notches are rostrum is oriented essentially dorsoventrally and found in most sigmodontines except ichthyomy- forms a 90-1 35° angle with the ventral edge of the ines, many thomasomyines, and some oryzo- rostrum as it passes around to the ventral side. myines. Euneomys, Neotomys, and Reithrodon are the only 43P. of Zygomatic Spine— 4 Development sigmodontine genera possessing a suture that makes states. a sharp bend (> 90°) in the middle of the rostrum and is horizontal when it around to = absent, anterior border of zygomatic plate nearly passes the ventral side 1 The extinct Proreithro- rounded or receding dorsally (Fig. 2). = don (Ameghino, 1 908) does not appear to possess 1 absent, anterior border nearly flat, vertical this derived condition. 2 = moderate, anterior border weakly curved clearly — 3 strongly developed, pronounced concavity 9S 46P. Posterior Extension of Nasals — 2

states (Fig. 1 3). The of variation for this character in the range = posteriormost terminus lies clearly posterior phyllotines equals or exceeds that found among to the dorsal maxillary-frontal suture at its all other sigmodontines. The length of the spine contact with the is influenced the width of the dorsal of lachrymal by region = 1 terminus lies subequal with or anterior to the the zygomatic plate and the degree of excavation maxillary-frontal suture in the ventral body of its anterior border (see Fig.

8). Voss and Carleton (1993) considered a non- The extension of the nasals was evaluated against spinous plate to be primitive for sigmodontines, the point at which the maxillary-frontal suture, and the common condition among sigmodontines running along the dorsal surface of the rostrum, surveyed in this study is for the anterior border of reaches the edge of the rostrum and contacts the the zygomatic plate to be flat and oriented nearly lachrymal. Nasals often extend further in adults vertically. Flat margins of the zygomatic plate or than in juveniles, but ontogenetic variation among weakly developed spines are the widespread con- adults does not seem significant. The nasals do not ditions among phyllotines. Zygomatic spines are extend posterior to the suture contact (Fig. 13A) most developed in Reithrodon, where their length in most "cricetid" outgroups, except some Old is greater than the width of the zygomatic plate at World "cricetids." Within the sigmodontines, oc- the plate's narrowest point. Auliscomys, Chin- currence of "long" nasals (Fig. 1 3B) is scattered, chillula, and Galenomys lack spines and have a with both conditions being found within several reduced dorsal body of the zygomatic plate, pro- tribes. ducing a convexly rounded anterior border that "Long" nasals are widespread among phyllo- recedes dorsally (e.g., Fig. 34). tines, with shorter nasals limited to Graomys and 44P. Inclination of Zygomatic Plate— 2 states. Reithrodon.

30 FIELDIANA: ZOOLOGY m and was considered plesiomorphic for neotomine- omys, and (weakly) P. wolffsohni. The overhanging peromyscines (Carleton, 1 980) and sigmodontines supraorbital is also characteristic of many oryzo- (Voss, 1 993). Two modifications of the plesiomor- myines. Because it is unclear which states might phic condition can be easily recognized. In one be adjacent to overhanging supraorbitals in a condition commonly seen among other muroid transformation series, this character was conser- rodents, the maximum constriction is anterior and vatively treated as unordered. the supraorbital region diverges posteriorly. An- 5 IP. Supraorbital Ridge— 2 states. dalgalomys, Graomys, Eligmodontia, Calomys (C = callosus, C. laucha, C. hummelincki), P. gerbillus, absent or directed laterally = and P. amicus share this condition, which is as- 1 lateral edges of supraorbital ridged and di- sociated with ridged and overhanging supraorbit- rected dorsally als. Another derived condition, where the maximum constriction is situated posteriorly and the This character distinguishes those supraorbital supraorbital region diverges anteriorly, is found in edges that are raised into dorsally directed ridges Andinomys and two species of Auliscomys, A. bo- from those with no or overhanging ridges. The liviensis, and A. sublimis. condition, widespread among sigmodontines and

Carleton ( 1 980) combined this and the following likely plesiomorphic for the phyllotines, is to lack character into a single character, "supraorbital any dorsally directed ridges. Dorsally directed shape and temporal ridges." Among species that ridges are found in Euneomys, Reithrodon, and I surveyed, the two characters are associated in Chinchillula. that posteriorly divergent supraorbitals are always 52P. Supraorbital Knobs— 2 states. overhanging and anteriorly divergent supraorbitals are not, but multiple combinations were ob- = absent = served. 1 small swellings or knobs on anterior supraorbital region, just posterior to lachrymal 1 IS 50P. Supraorbital Edge— 4 states.

= rounded when supraorbital region smoothly In some phyllotines that otherwise show a slight viewed in cross-section medial trough along the midline of the supraor- = 1 at of dorsal surface, or weakly angled edge bital region (i.e., similar to weak ridging of the for half the of the angled length supraorbital supraorbital edges), bony processes of the frontals region are found along the supraorbital margins just pos- 2 = but distinctly angled (~ 90°), not overhang- terior to the lachrymal bones. Supraorbital knobs ing are unrelated to the inflated frontals found in oxy- 3 = into sharply angled overhanging shelf, at least mycterines. The widespread condition is for no posteriorly such swellings. Supraorbital knobs are found in P. darwini, Andinomys, Euneomys, Neotomys, and Reithrodon. The variety of supraorbital morphologies ex-

hibited it by sigmodontines makes difficult to pro- 53P. Mediodorsal Fusion of Frontals — 3 vide states. precise definitions of character Olds states. and Anderson followed Hershkovitz (1989) (1962) = in describing the diagnostic condition for phyllo- complete 1 = or vascularized tines as "supraorbital region never evenly curved partially open 2 = distinct and consistent fontanelle in cross section." The edge can vary from smoothly rounded, to being distinctly but obtusely angled, to acutely angled into an overhanging shelf. The The widespread condition among sigmodon- widespread condition for phyllotines is for the su- tines and clearly the plesiomorphic condition for praorbital edges to be angled along nearly the en- phyllotines is for the frontals to be completely fused tire supraorbital region. A possibly derived con- along the midline. In Irenomys, Neotomys, and dition where only the posterior region is weakly Andinomys, that fusion is not complete for most angled is found in Loxodontomys, A. sublimis, and individuals. In Irenomys and Neotomys, the gap P. osilae. Care must be taken in scoring the char- usually is associated with what appears to be a acter because the ridge becomes more sharply de- sinus emerging from within the frontals. In some fined with age. The sharply ridged, overhanging individuals, channels in the bone indicate that a condition is found in some Calomys (C callosus, vessel loops up from the frontals and back down C. laucha, C. hummelincki), Andalgalomys, Gra- into the frontals or cranium. A distinct fontanelle

32 FIELDIANA: ZOOLOGY is present in Andinomys as noted by previous Character state values were determined using workers (e.g., Hershkovitz, 1962). More than 90% segment-coding with a criterion variable of 4. De- of Andinomys individuals examined (« 50) pos- scriptions of interparietal size or shape are often sess a fontanelle, and several have a second fon- imprecise (e.g., "well developed, at least trans- tanelle in the posterior region of the frontals as versely" [Olds & Anderson, 1989], "small, trian- well. gular, or irregularly oblong" [Voss, 1988]) due partially to high intraspecific variability in both 12S 54P. Shape of Frontoparietal Suture— 2 attributes. I chose to define one aspect of size and states (Fig. 13). shape as the length of the interparietal relative to = rounded, edge of frontal convex the parietal, measured along the midsagittal line. = 1 straight or slightly sigmoidal to concave Because this character is variable within most of the recognized tribes as well as in the outgroups, polarity is difficult to determine and easily affected The shape of the frontoparietal suture was eval- by tree topology. This character appears to be more uated for the body of the suture, excluding the informative within tribes rather than between. Olds lateral margins, which usually curve anteriorly to and Anderson (1 989) included a "well developed" "horn"-shaped extensions of the parietal: the horns interparietal in their differential diagnosis of the are present in taxa exhibiting either condition. phyllotines. They described a well-developed in- Convex, rounded sutures (Fig. 1 3 A) are found in terparietal as plesiomorphic for the phyllotines in all surveyed outgroups, are widespread among sig- their character review and thus did not list it with modontines, and thus are likely the plesiomorphic the likely synapomorphies. sigmodontine condition. Ichthyomys (Fig. 13B) possesses the more extreme condition where the 56P. Medial Length of Interparietal/Parie- frontal edge of the suture is concave. tal— 3 states. Convex sutures are likewise widespread among - < 0.33 phyllotines, with the presumptively derived con- 1 = 0.33-0.45 dition limited to Andinomys, Chinchillula, Ireno- 2 = > 0.45 mys, and Reithrodon. This coding scheme does not apply well to Neotomys, whose frontals are A separate coding scheme was used for the phyl- very rounded and sometimes "heart"-shaped, and lotine analysis because the range of variation was so Neotomys was coded as unknown. less than among the sigmodontines, but taxonom- 55P. Angle of Frontoparietal Suture — 2 ically significant variation could still be recog-

states (Fig. 1 3). nized. As with character 1 IS, lengths of the inter- = parietal and parietal were measured along the obtuse angle = midline. The common condition among phyllo- 1 acute or right angle tines is for a long interparietal (state "2"), but the plesiomorphic condition is unclear with all in- The angle was evaluated by judging the orien- stances of short interparietals among other sig- tation along the body of the sutures. In cases with modontines. A moderate interparietal (between 33 strong curvature, particularly near the medial apex, and 45% the medial length of the parietal) is found the condition was evaluated by extending lines in most Calomys except C. sorellus, which, like joining the medial apex to the triple junctures of most other phyllotines, has a larger interparietal the frontals, parietals, and squamosals on both that is greater than 45% of the parietal length. sides. In the majority of cases without strong cur- Moderate interparietals are also found in Aulis- vature, these two criteria are congruent. Obtusely comys and Neotomys. Although not coded in this angled sutures (Fig. 13 A) are widespread among study because of the difficulty in coding the com- sigmodontines and phyllotines and are thus likely plex variation observed, shape of the interparietal plesiomorphic for the phyllotines. This character may be a useful character at other taxonomic lev- was not phylogenetically informative among the els. For example, my observations indicate that nonphyllotines surveyed for this study. despite substantial individual variation, particular 1 3S. Ratio of Interparietal/Parietal shapes characterize and distinguish species, sub- Length— 3 states. species, and even populations within the Phyllotis darwini complex. = < 0.43 = 1 0.43-0.70 1 4S. Parietal/Occipital Contact— 2 states (Fig. = 2 > 0.70 14).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 33 Fig. 14. Dorsolateral views of posterior cranium. A, parietal and occipital in contact, Nectomys squamipes (fmnh 141633); B, interparietal and squamosal in contact, Nyctomys sumichrasti (fmnh 35186).

= parietal and occipital in contact with each inflated: sloping posterolaterally), but because this other, interparietal and squamosal not in condition is so restricted among phyllotines, trans- contact verse (moderately inflated) may be a phyllotine = 1 interparietal and squamosal in contact with synapomorphy. An oblique bulla is found in C each other, parietal and occipital not in con- callosus, Andalgalomys, and some Andinomys. A tact rounded anterior border to the bulla is found in C. lepidus and The widespread condition seen in sigmodon- many Phyllotis. tines and outgroups is for the posterolateral mar- 15S 5 8 P. Tegmen Tympani— 2 states. gins of the parietal to contact the anterolateral = tegmen tympani contacts posterior suspen- corners of the occipital (Fig. 14A), even in species sory process of squamosal across middle lac- with well-developed interparietals. This parietal/ erate foramen little more a milli- occipital contact may be than = 1 tegmen tympani does not contact squamosal meter in length but precludes any contact between the lateral margins of the interparietal and the pos- In most sigmodontines, the tegmen tympani, teromedial corners of the squamosal. Only tylo- also known as the periotic portion of the petrosal, myines have interparietals that are broadly in con- crosses the middle lacerate foramen to contact the tact with the squamosals (Fig. 14B), precluding posterior suspensory process of the squamosal contact between parietals and occipitals. Neoto- (Voss, 1993, Fig. 8). Voss (1993) and Voss and mys is polymorphic for this character, with all four Carleton (1993) considered this to be the plesio- bones nearly meeting at a single vertex. morphic condition. Voss and Carleton (1993) also 57P. Orientation of Anterior Border of proposed that the absence of this contact, and in Auditory Bulla— 3 states. particular the absence of the suspensory process, = is a synapomorphy of the tribe Oryzomyini, but oblique (viewed ventrally) = their sample of sigmodontine genera was limited. 1 transverse Absence of the contact was found in all surveyed 2 = rounded oryzomyines, including the tetralophodont Hol- This character largely describes the degree of ochilus, Pseudoryzomys, and Zygodontomys, in the anterior inflation in the bulla. The widespread phyllotine Reithwdon, and in the Old World cri- condition among sigmodontines is oblique (un- cetine Mystromys.

34 FIELDIANA: ZOOLOGY 59P. Shape of Stapedial Spine of Auditory

Bulla— 2 states (Fig. 15).

= circular to ovoid in cross-section = 1 laterally appressed against auditory bulla, not smoothly rounded in cross-section

The lateral compression may be a consequence of the high degree of inflation of the auditory bulla in Andalgalomys and Graomys, the only taxa in which this character has been observed (Fig. 1 5B). The stapedial spine is bounded medially and/or laterally by the stapedial artery or its subsidiary branches, the supraorbital and infraorbital.

16S. Squamosal Fold— 2 states.

= absent = 1 present

The squamosal fold is a thickening along the anterior border of the postglenoid foramen, with the dorsal edge nearly folded over the anterior margin, strongly obscuring a large tegmen tympani and small postglenoid foramen. A squamosal fold has only been observed in Kunsia, Tylomys, and Nyctomys.

17S. Subsquamosal Foramen— 3 states. = present, well developed = 1 reduced to slit, little or no exposure to occipital 2 = absent

Carleton (1980) considered the absence of the subsquamosal foramen to be plesiomorphic for New World murids, despite the more widespread occurrence of state "0." The various conditions are dispersed throughout the sigmodontines, but all phyllotines have a well-developed subsquamosal foramen.

60P. Thickness of Hamular Process of Squamosal— 4 states. = process wholly absent (i.e., subsquamosal foramen absent) = 1 broad along entire length, subsquamosal foramen often reduced = 2 bridge reduced in thickness, posterior terminus appears flattened = 3 posterior end reduced as well, not greatly thicker than bridge

The presence and thickness of the hamular process of the squamosal is dependent on the presence of the subsquamosal foramen. The most frequent condition among phyllotines is for a thin, fragile lincki, Euneomys, Neotomys, Andinomys, and Chinchillula.

62P. Internal Carotid Canal— 2 states (Fig. 16). = bounded by both basioccipital and the ectotympanic portion of auditory bulla = 1 bounded entirely (or nearly so) by petrosal and ectotympanic portions of auditory bulla

The widespread condition among sigmodontines and probably the plesiomorphic condition for phyllotines is for the carotid canal to be bounded by both the auditory bulla and the basioccipital bone as the internal carotid artery passes between them to enter the braincase (Fig. 16 A). The alternate condition differs in having a flange of the petrosal extend between the basioccipital and the internal carotid artery, thus along with the ectotympanic portion of the bulla entirely or almost

entirely forming the carotid canal (Fig. 1 6B). This condition is found in Andinomys, Chinchillula, Euneomys, Irenomys, and Neotomys.

63P. Extension of Eustachian Tube— 3 states. = tube does not reach posterior lobe of pterygoid process Fig. 16. Medial views of auditory bulla and internal = 1 tube subequal to posterior lobe of pterygoid carotid canal. A, internal carotid bounded by both ba- does not extend anterior to the base sioccipital and ectotympanic, Phyllotis andium (fmnh process, 8 1 249); B, internal carotid bounded by petrosal and ecto- of process = tympanic portions of auditory bulla, Neotomys ebriosus 2 tube extends anteriorly past base ofpterygoid carotid (fmnh 24775). boc, basioccipital; cc, canal; ect, process ectotympanic part of auditory bulla; pet, petrosal part of auditory bulla; sf, stapedial foramen. The widespread condition among phyllotines and sigmodontines, and most likely plesiomorphic for the phyllotines, is for the anterior flange of the = eustachian tube to be subequal with the pterygoid 1 postglenoid foramen distinctly anterior to process, reaching the posterior lobe of the process, subsquamosal foramen but not extending anterior to the base of the process. This is the intermediate state in the transition This character is a simplified description of a series represented among the phyllotines. The eu- complex trait that depends on the orientation and stachian tubes of many Andinomys have lanciolate thickness of the hamular process of the squamosal, projections that can extend up to 2 mm anterior as well as the size, shape, orientation, and relative to the bases of the pterygoid processes. positions of the temporal vacuities: the subsqua- 1 8S 64P. Relative Width of Mesopterygoid mosal foramen and the postglenoid foramen. The Fossa— 3 states. anterior extent of each is particularly influential = in determining state assignments. The common distinctly broader than adjacent paraptery- condition among phyllotines is for the vacuities goid fossae = to be positioned essentially dorsoventrally. The 1 subequal = common condition among the outgroups is for the 2 distinctly narrower than adjacent parapter- vacuities to be positioned distinctly anteroposte- ygoid fossae riorly, with the postglenoid foramen more anterior than ventral to the subsquamosal foramen. This Previous studies have not always specified a cri- condition is found in C. callosus and C. humme- terion for evaluating the size of the two fossae.

36 FIELDIANA: ZOOLOGY This is particularly important because both fossae 2 = > 2.4 times anterior width may converge or diverge posteriorly to different and thus the relative will degrees, proportions vary. The categorical values are the same as for the Olds and Anderson evaluated (1989) mesoptery- parapterygoid fossa. Here the common condition goid breadth at the posterior of the zygo- margin in phyllotines is a relatively parallel-sided meso- matic aperture. I chose to make the comparison pterygoid fossa ("0")- Moderate posterior diver- at the basisphenoid-presphenoid suture, believing gence (1.5-2.4 anterior breadth) is found in Cal- that this would be a more stable reference land- omys lepidus, Andalgalomys, Eligmodontia, and mark. The different criteria may be the explana- P. definitus. The most posteriorly divergent con- tion for the differences in Holochilus, coding (i.e., dition is found in Reithrodon. Pseudotyzomys). Olds and Anderson (1989) con- 67P. Parapterygoid Fossa Depth— 3 states. sidered that a parapterygoid fossa relatively broader than the mesopterygoid fossa was diagnostic for = flat, even with bony palate the phyllotines. = 1 slightly to moderately excavated above level The condition is variable sigmodontines, among of bony palate but within the phyllotines the widespread condi- = 2 deeply excavated above level of bony palate tion is for the mesopterygoid fossa to be distinctly narrower than the fossa. The two parapterygoid The widespread condition among sigmodon- fossae are subequal in breadth in Chinchillula, P. tines and phyllotines is for a shallow parapterygoid wolffsohni, Irenomys, and Andinomys. Broad fossa, essentially even with the bony palate. A mesopterygoid fossae are found in several out- parapterygoid fossa excavated slightly beyond the group taxa (Holochilus, Nectomys, Oxymycterus, level of the bony palate is found in Euneomys and and Zygodontomys). Andinomys. Reithrodon and Neotomys share a well- 65P. Parapterygoid Shape— 3 states. excavated parapterygoid fossa. Depth and breadth ofthe parapterygoid fossa may be functionally cor- = posterior width < 1.5 times anterior width related with hypsodont grinding molars because 1 = 1.5-2.4 times anterior width the internal pterygoids originate in the parapter- 2 = > 2.4 times anterior width ygoid fossae (Kesner, 1980; Rinker, 1954), and

Carleton 1 considered a fos- This character describes the degree of posterior ( 980) accordingly deep sa divergence in the parapterygoid fossa. The com- as derived. mon condition and among surveyed sigmodontines 20S 68P. Sphenopalatine Vacuities— 5 states. the widespread condition among phyllotines is for = the posterior breadth to be 1.5-2.4 times the an- absent, roof of mesopterygoid fossa wholly terior breadth. The probably derived condition ossified = where the posterior breadth is less than 1.5 times 1 narrow slit encompassing presphenoid-ba- visible within the anterior breadth is found in Euneomys, Lox- sisphenoid juncture, wholly fossa odontomys, Neotomys, and Reithrodon. The para- mesopterygoid 2 = distinct but orbital pterygoid fossa diverges more strongly only in P. vacuity constricted, wings osilae. of presphenoid not fully separated posterior to medial pterygoid processes 19S. Shape of Mesopterygoid Fossa— 3 states. = 3 vacuity large, medial pterygoid processes ful- = anterior to orbital of posteriorly convergent, "horseshoe"-shaped ly wings presphenoid, = not visible in no 1 parallel sided, "U"-shaped mesopterygoid fossa, large = of lateral into the va- 2 posteriorly divergent, "V"-shaped projections margins cuity of this = The coding character for the two analyses 4 vacuity very large, orbital wings of presphe- differs to reflect the greater range of variation noid filamentous, very large optic foramen among sigmodontines than among phyllotines. Most outgroups are either convergent or parallel- The ancestral condition for this character is dif- sided. All phyllotines are divergent, although some ficult to determine. All states except the most open are nearly parallel-sided. vacuities seem to be common among other "cricetids." Carleton (1980) considered a wholly os- 66P. Shape of Mesopterygoid Fossa— 3 states. sified roof of the mesopterygoid fossa to be prim- = posterior width < 1.5 times anterior width itive for neotomine-peromyscines, and Olds and = 1 1.5-2.4 times anterior width Anderson (1989) considered "large" sphenopala-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 37 tine vacuities (following the description in Carle- The position of the mesopterygoid fossa was ton [1980]) to be a synapomorphy for phyllotines. evaluated relative to the line connecting the pos- With the additional character states used in this terior borders ofthe upper third molars. This char- study, very open sphenopalatine vacuities (states acter has sometimes been referred to as "palate "3" and "4") are only found in the phyllotines and length." I used the same criterion for coding "short" in Sigmodon. Voss (1991) reported significant but palates as did Olds and Anderson (1989), namely, discrete variation for this trait among Zygodon- that the mesopterygoid fossa extend anteriorly be- tomys brevicauda. The open sphenopalatine va- yond the posterior edge of M3, but my observa- cuity that he illustrated for some populations would tions disagree with their coding ("long" palates) be categorized as state "2" under my coding scheme for Ichthyomys, Holochilus, and Neotomys. rather than state "3," as characterizes the phyl- Hershkovitz (1962) regarded a "short palate" as lotines. primitive for the sigmodontines. Variation in the extent of excavation in the The widespread condition among phyllotines is sphenopalatine vacuities is much greater among for the anterior border of the mesopterygoid fossa sigmodontines than within the phyllotines. All to be within Vi to 1 tooth-length (M3) posterior to phyllotines have large sphenopalatine vacuities that the posterior alveoli. A shorter palate, with the are fully exposed within the mesopterygoid fossa. mesopterygoid fossa within Vi of a tooth-length, is Reithrodon is characterized by especially large found in some Reithrodon auritus and Chinchil- sphenopalatine vacuities in addition to large optic lula. The mesopterygoid actually reaches the pos- and anterior lacerate foramina. The convergence terior plane of the third molars in Irenomys, Neo- of the foramina and vacuities in Reithrodon has tomys, and Andinomys. resulted in filamentous orbital wings of the pre- 7 IP. Medial Process of Posterior Palate— 2 sphenoid. State "4" is subsumed in state "3" in states. the sigmodontine analysis because it is represented only in Reithrodon, making it uninformative in = absent the = analysis. 1 present

69P. Position of Orbital Wings of Presphe- The common condition among sigmodontines noid— 2 states. and the phyllotines is for there to be no medial process from the posterior margin of the palate. = wings anterior to a distinct constriction of the presphenoid 72P. Posterior Palatine Ridge— 2 states. = 1 wings posterior to maximum constriction = absent or indistinct = 1 present, a longitudinal ridge formed by con- The widespread and almost certainly plesiomor- vergence of parallel channels arising from phic condition in phyllotines is for the wings to palatine foramina join the presphenoid anterior to the sharp constriction that occurs just anterior to the basisphe- The posterior palatine ridge is a distinct ridge noid. Among phyllotines, the derived condition, running down the midline of the palate, most pro- where the wings join equal with or posterior to the nounced posteriorly. It is usually, but not always maximum constriction, is found only in Andal- (e.g., Punomys), continuous with a median process galomys and Graomys. The derived condition is of the posterior palate. The ridge appears to be largely due to the gradual thinning of the presphe- formed principally by the near coalescence of two noid in those two genera, rather than the abrupt channels running along the palate. Among phyl- constriction present in the other phyllotines. lotines, the posterior palatine ridge is found only in Reithrodon and Neotomys. 2 IS 70P. Position of Anterior Border of Mesopterygoid Fossa— 4 states. 73P. Posterolateral Palatal Pits— 2 states. = = lying > 1 M3 tooth-length posterior to M3 anterior to mesopterygoid fossa = x = of 1 lying between h and 1 tooth-length posterior 1 posterior to anterior border mesoptery- to M3 goid fossa 2 = lying between and < Vi tooth-length posterior to M3 A pair of pits are always found among phyllo- = 3 reaching posterior plane ofpaired M3s ("short tines in the posterior region of the palate framing palate") the anterior border of the mesopterygoid fossa.

38 FIELDIANA: ZOOLOGY The widespread condition is for these pits to be Coding largely follows Carleton ( 1 980), who fol- just anterior to the mesopterygoid fossa. Alter- lowed Bugge's (1970) polarity. Readers are di- natively, the pits are displaced posteriorly into the rected to Bugge ( 1 970) for discussion of the carotid anterior region of the parapterygoid fossae. This system in muroid rodents and to Brylski (Brylski, condition is found in Irenomys, Andinomys, and 1990; geomyoids), Carleton (1980; New World P. osilae. Of these three species, only P. osilae has "cricetines"), Voss (1988; ichthyomyines), Carle- a "long" palate; that is, two of three phyllotines ton and Musser (1989; oryzomyines), and Voss with pits posterior to the edge of the mesoptery- and Carleton (1993; oryzomyines) for more de- goid fossa have a mesopterygoid fossa that extends tailed descriptions of this character in New World far anteriorly. However, the two traits are not as muroids. Each of these workers considered a com- highly correlated as that observation suggests. plete stapedial system (state "0") to be primitive Among outgroup taxa surveyed, four out of seven for the group they were discussing. In the primitive with pits posterior to the mesopterygoid fossa in condition, the carotid artery splits into the internal fact have "long" palates where the mesopterygoid carotid and the stapedial artery. The stapedial ar-

x is greater than h ofa tooth-length from the molars. tery passes into the auditory bulla, through the stapes, and into the cranium. After a of the 74P. Orientation of Maxillary Tooth Rows— split stapedial, the supraorbital branch passes along the 3 states. internal surface of the squamosal and alisphenoid = posteriorly divergent bones, leaving a groove as evidence of its passage, = 1 parallel and eventually emerges through the sphenofrontal = 2 convergent foramen. The "primitive" condition, with functional stapedial foramen, sphenofrontal foramen, and all is found in this The orientation of the maxillary tooth rows is squamosal groove present, in basal some estimated by lines passing through the apexes of survey ichthyomyines (Voss, 1988), some the alveoli. Akodontines generally have conver- thomasomyines, Wiedomys, oryzomyines, and vir- gent tooth rows, oryzomyines generally parallel, akodontines, scapteromyines, Punomys, all Also but Punomys has divergent tooth rows. Thus, the tually phyllotines. common among sigmodontines in this are those conditions plesiomorphic condition for phyllotines is unclear. survey considered to be derived. These derived condi- 7 5 P. Sphenopalatine Foramen— 3 states. tions are characterized by loss of the sphenofrontal = foramen and loss of the the absent or nearly ossified supraorbital branch of = Further loss of the infraorbital 1 present, small to moderate size stapedial artery. = branch of the results in the re- 2 present, large stapedial artery duction of the stapedial foramen. These derived conditions are found in oryzomyines, some The common condition among sigmodontines many thomasomyines, terminal ichthyomyines (accord- is for the sphenopalatine foramen to be absent or ing to Voss, 1988), Sigmodon, and constricted, but the widespread condition among among phyllotines in Reithrodon and Voss phyllotines is for a distinct and moderate-sized Neotomys. (1991) reported that brevicauda shows foramen to be present. The constricted condition Zygodontomys well-defined discrete variation for this also occurs in Auliscomys sublimis and Loxodon- geographic trait, with one set of populations a com- tomys. A particularly large sphenopalatine fora- having plete stapedial circulation and the other set men is found in Reithrodon, consistent with the lacking a complete stapedial circulation. It is general high degree of fenestration in its basal cra- interesting nium. that this species shows geographic variation for two traits that are otherwise conserved at the ge- 22S 76P. Carotid Circulation— 3 states. neric or even tribal level. = both stapedial and sphenofrontal foramina 77P. Squamosal Fenestra— 2 states. present, squamosal-alisphenoid groove = squamosal fenestra present where mastica- present = tory-buccinator nerve passes over the squa- 1 stapedial foramen present, but sphenofrontal mosal-alisphenoid groove foramen absent, no squamosal-alisphenoid = 1 squamosal fenestra absent groove = 2 both stapedial and sphenofrontal foramina Where the squamosal-alisphenoid groove cross- absent, no squamosal-alisphenoid groove es the trough formed on the exterior of the alisphe-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 39 = noid by the masticatory -buccinator nerve, a fe- 2 entoglossal process absent, basihyal straight, nestra often appears. Squamosal fenestrae were thyrohyal short only observed in individuals with a squamosal— alisphenoid groove (i.e.. complete stapedial sys- Transformation series and state descriptions tem), which is likely a structural necessity for the follow Carleton (1980. Fig. 1 1), from which most bone to be thin enough for the fenestra to occur. of the data were gathered, but coding order was Under this argument, a fenestra cannot be present modified to allow linear ordering of the character in animals with a derived stapedial system. This states for the analysis. Carleton considered the in- character was therefore coded as unknown in taxa termediate condition, with entoglossal process with the derived stapedial system, which still al- small, to be plesiomorphic for most muroid groups lows the potential for incongruent ancestral state he surveyed. .All sigmodontines surveyed by Carle- assignments under parsimony, but this should be ton ( 1 980) or myself lack an entoglossal process. unimportant to the analysis if it should happen. Polarity of this character is uncertain. Most phyllotines have the squamosal fenestra, although there Posteranial Skeleton is significant individual variation. Taxa with the complete supraorbital circulation that generally 25S. .Articulation of First Rib— 2 states. lack the fenestra include Galenomys. Euneomys. = first rib articulates with first thoracic verte- Irenomys. southern Andinomys. most Phylloi is. and bra A reduced circu- only Auliscomys pictiis. supraorbital = 1 first rib articulates with transverse process of lation, where the stapedial foramen in the auditory seventh cervical vertebra in addition to first bulla is present but constricted, and both the thoracic vertebra sphenofrontal foramen and squamosal-alisphenoid groove are absent, is found among phyllotines only in Xeoiomys. Reithrodon is the only .Articulation with the seventh cervical vertebra phyllotine without a stapedial artery. is recognized by the presence of a faceted articulation surface on the v ertebra's transverse process, 23S ~SP. Alisphenoid Strut— 3 states. in addition to apparent contact (Carleton. 1980, = absent or filamentous 15). Carleton (1980) considered dual articu- = Fig. 1 consistent dorsal process, but does not fully lation as derived within neotomine-peromyscines. cross foramen ovale The presence of both states among nonsigmodon- = 2 present and bony tines and the universal dual articulation among sigmodontines suggests that dual articulation may The absence of an alisphenoid strut is wide- be a synapomorphy for the sigmodontines. spread among the phyllotines. but some individual 26S. Number of Thoracic and Lumbar Ver- variation is apparent. V'oss (1993) and Voss and tebrae— 3 states. Carleton (1993. Fig. 10) considered presence of the alisphenoid strut to be plesiomorphic for sig- 0=14 thoracic and 5 or 6 lumbar = modontines. Among phyllotines. a complete and 1 13 thoracic and 6 lumbar consistent alisphenoid strut is found in Andalga- 2=12 thoracic and 7 lumbar lomys. Eligmodontia, Graomys. Irenomys. and Reithrodon. In an intermediate condition, at the Thoracic v ertebrae w ere defined as having com- usual position of the strut, a process of the alisphe- plete pairs of fully formed and articulating thoracic noid extends into the foramen ovale from the dor- ribs. Supernumerary ribs are easily distinguished sal side. This condition is found in Euneomys and from complete ribs because they never articulate Chinchillula. Because there is no direct indication with the preceding v ertebra (as do complete ribs), that a dorsal process is the intermediate state in are always shorter and thinner, may diverge at odd transformations between absence and presence, the angles, and often exhibit enlarged and deformed character states are treated as unordered. heads. Supernumerary ribs that lack these conditions were considered to be associated with lumbar 24S. Hvoid— 3 states. vertebrae. Carleton (1980) considered 13 thoracic = entoglossal process long and attenuate, ba- and 6 lumbar vertebrae to be plesiomorphic for sihyal arched, thyrohyal long the New World muroids. The number of vertebral = 1 entoglossal process a small knob, basihyal elements was surveyed across various taxa in ad- arched. thvTohyal long dition to those included in the phylogenetic anal-

40 FIELDIANA: ZOOLOGY ysis (Table 5, 179 species total). The Central off; one appeared to have 1 2 ribs, but two addi- American tylomyines, of uncertain affinities to the tional disarticulated ribs were found with this sigmodontines or neotomine-peromyscines, gen- skeleton; and only one had a complete thirteenth erally have more thoracic vertebrae (13, 14, or 1 5) pair. Of the two skeletons examined by Carleton

while maintaining the six lumbar vertebrae. Thir- ( 1 980), one was too damaged for an accurate count teen thoracic and six lumbar vertebrae appear to by me, and the other was not found. In G. do- be the widespread conditions among sigmodon- morum, not previously reported, seven skeletons tines and are found in thomasomyines, ichthyo- clearly had the derived condition, one had an myines, akodontines, phyllotines, and scaptero- extra thoracic pair (13T [thoracic] + 7L [lum-

myines. The apparent consistency ofthis character bar]), one was missing a lumbar vertebra ( 1 2T + among generic groups suggests that it may be high- 6L), and three had the plesiomorphic condition ly informative regarding oryzomyine relationships (13T + 6L). Ten of 13 skeletons of Sigmodon (Table 6). This trait appears more labile among hispidus examined had the derived 12 thoracic the phyllotines than among the other sigmodon- and 7 lumbar rib pairs, as was observed by Voss tines (Steppan, 1993). (1992) in his revision of the South American spe- A consistent pattern is observed among species cies of Sigmodon. Three of the five skeletons ex- that are polymorphic for the number of thoracic amined by Carleton (1980) were reexamined by rib pairs and vertebral number. Of the 57 species me; one had 13 rib pairs, one had a supernu- found to be polymorphic for rib pairs, only 13 merary pair, and one had a single supernumerary have minority variants with fewer than the modal rib. This variation does not appear to be a prep- number of thoracic ribs. In all other species, the aration artifact. None of the specimens in species minority variant has an additional pair of either with moderate to large series and characterized

complete ribs (3 1 species ) or incompletely formed as having 1 3 rib pairs were observed to have lost supernumerary ribs (24 species) (the numbers of the last pair without leaving some evidence on species just listed do not add up to 57 because the vertebrae. Additionally, the second lumbar some species have three or four conditions pres- vertebra can be recognized by an enlarged trans- ent). Only 6.3% of the 978 individuals examined verse process relative to those on the thoracics. differ from their species' modal count for true ribs. This enlarged process would have obstructed an additional rib, if a rib had been present. 79P. Number of Thoracic Rib Pairs— 2 states. 27P. Number of Caudal Vertebrae— 6 states. 0=13 thoracic ribs

1 = 12 thoracic ribs - > 40

1 = 36^0 = The coding for the phyllotine differs 2 30-35 analysis = from the sigmodontine analysis in not including 3 24-29 4 = more than 1 3 ribs. The widespread and probably 20-23 5 = 11-19 plesiomorphic condition among phyllotines is 1 3 ribs. Twelve ribs are found in Andalgalomys, Gra- omys, Reithrodon, and some Calomys. Coding was derived from a histogram of all sur-

Some of these observations differ from pub- veyed sigmodontine species (Table 5). Vertebral lished counts and deserve further discussion. Both counts show limited variation in well-preserved Carleton (1980) and Olds (1988) reported that C. specimens, typically with 80% ofspecimens within callosus has the plesiomorphic 1 3 ribs. Nineteen a range of two, with a few notable exceptions (Ho- of the 20 skeletons of callosus examined in this lochilus brasiliensis, 25-34; Microryzomys minu- study clearly had only 1 2 ribs and one had a thin, tus, 27-38; Oecomys concolor, 32-38; Oligoryzo- short thirteenth pair that did not articulate with mys microtis, 31-38; Macrotarsomys bastardi, the twelfth thoracic vertebra. Of the 1 1 skeletons 30-40). This variation within species may reflect examined by Carleton (1980), I could find only currently unrecognized taxonomic diversity or one individual that had more than 12 ribs (and misidentification of specimens. Carleton (1980) the degree of articulation was unclear). Carleton was unable to reliably estimate polarity for this (1980) also reported that both Graomys griseo- character. Nearly the entire range of variation is Jlavus and Sigmodon hispidus possess the plesio- found among the outgroups, from state "1" in morphic condition. Of the 13 skeletons of G. Nyctomys to state "5" in some "cricetids," but the griseqflavus examined, 1 1 had 1 2 ribs, with no extremes are likely to be derived in both sigmo- evidence that the thirteenth pair had been broken dontines and outgroups.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 41 Table 5. Vertebral counts among Neotropical sigmodontines and selected muroids.

Taxon a Thoracic^ Lumbar CaudaL

Tribe Akodontini

Akodon yAbrothrix) longipilis 10

A. (Abrothrix) sanborni A. (Akodon) aerosus A. albiventer A. azarae

A. boliviensis

A. cursor A. mollis A. neocenus A. olivaceus

A. puer

A. puer lutescens A. subfuscus A. torques

A. urichi A. xanthorhinus

A. (Deltamys) kempi A. (Microxus) bogotensis

A. (A/.) mimus

A. (Thaptomys) nigrita Bolomys amoenus

B. lasiurus

B. obscurus B. temchuki Chelemys macronyx Chroeomys andinus C. jelskii Geoxus valdivianus Notiomys edwardsii Oxymycterus delator O. inca

O. platensis O. rufus

Thalpomys lasiotis Table 5. Continued.

Taxona N" Thoracic Lumbar'' Caudal'

/. tweedii 4 Neusticomys monticolus N. venezuelaef Rheomys mexicanusf

R. raptor R. thomasi

R. underwoodi Table 5. Continued.

Taxon" N* Thoracic^ Lumbar" Caudal'

O. eliurus 1 O. jlavescens O. fulvescens

O. longicaudatus

O. magellanicus O. microtis

O. microtis fornesi

O. nigripes

Oryzomys albigularis

O. alfaroi O. bolivaris O. buccinatus O. capito

O. chapmani O. couesi O. intermedins O. keaysi O. melanotis O. nitidus O. palustris

O. polius

O. ratticeps O. subflavus O. talamancae

Pseudoryzomys simplex

Sigmodontomys alfari Zygodontomys brevicauda Table 5. Continued.

Taxon" V Thoracic^ Lumbar' CaudaF

R. venezuelae 12 39^12 12 39-40 R. venustus 13 Thomasomys aureus 13 41-43 13 41 T. baeops 13 37^1 T. cinereus 13 14 42 T. daphne 13 41 T. gracilis 13 38 T. hylophilus 13 41 T. oreas 13 38 T. paramorum 13 35-39 (38) T. pyrrhonoius 13 > 38 13 + 1 39 T. rhoadsi 13 Wiedomys pyrrorhinos 12 39

Tribe Phyllotini Andalgalomys pearsoni 12 31 Andinomys edax 13 28-30 (30) Auliscomys boliviensis 6 13 25-27 A. pictus 4 13 22-26 1 14 27 A. sublimis 4 13 24-25 Calomys bolivae 4 12 * 25 C. callosus 19 12 22-25 (23-24) 1 12 + 1 2 12

1 13 25 C. hummelincki 5 12 21-23 2 12 21-22

1 13 C. laucha 9 13 21-23

1 12 + 1 C. lepidus 4 13 20-22 C. musculinus 7 13 27-28

1 13 24-26 2 12

1 ? + 1

1 14 29 C. sorellus 11 13 27-28 2 13 30-31

1 14 29

C. tener 1 12 24

C. venustus 1 12 22

1 13 25

Chinchillula sahamae 1 13 > 21 Eligmodontia morgani 6 13 25-26 E. puerulus 1 13 26

Euneomys chinchilloides 10 13 22-24 (22) Galenomys garleppi 6 13 17-19 Graomys domorum 6 12 31-32, 35 3 13 32-34

1 13

1 12 G. griseojlavus 12 12 32-35 1 13 35

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 45 Table 5. Continued.

Taxon" V : Thoracic Lumbar" Caudal'

Irenomys tarsalis 8

Loxodontomys micropus

Phyllotis amicus P. andium P. caprinus P. darwini P. gerbillus

P. haggardi P. osilae P. wolffsohni P. xanthopygus rupestris

P. xanthopygus xanthopygus

Reithrodon auritus Table 5. Continued.

Taxona N* Thoracic^ Lumbar CaudaP

Eliurus myoxinux 1 Gymnuromys roberti Macrotarsomys bastardi

Nesomys audeberti N. rufus Fig. 17. Ventral view of caudal vertebrae 1-5 in Nectomys squamipes (fmnh 14163), indicating the hemal arches (ha) and hemal processes (hep).

80P. Number of Caudal Vertebrae— 3 states. condition among phyllotines is in Reithrodon. In 16 skeletons of Reithrodon auritus pachycephalia, = < 25 one had a longer spine on the third thoracic ver- 1 = 25-30 tebra and five had second and third spines of ap- 2 = > 30 proximately equal length.

Tail length— Not analyzed. Tail length is highly 82P. Height of Neural Spine of Second Cervi- correlated with number of caudal vertebrae and cal Vertebra— 3 states. provides the more common definition of this gen- = not significantly enlarged eral trait. Olds and Anderson (1989) described tail = 1 enlarged, distinct knob length as variable within Phyllotini, but they none- = 2 very enlarged into distinct keel, "plow"- theless chose to include it in their data matrix. shaped, may overlap C3 While indeed being quite variable within the phyllotines (ranging from 0.33 to 1.4 times head and The widespread condition among phyllotines is body length), tail length seems to be more con- for the neural spine on the second cervical vertebra served and therefore informative outside the phyl- to be enlarged into a distinct knob, usually wider lotines. Most sigmodontines (oryzomyines, tho- than long. In the likely derived condition, the knob masomyines, ichthyomyines) have moderate to is further enlarged into a distinct plow-shaped keel long tails (> 0.85 head and body length). Olds and that may overlap the third cervical vertebra. This Anderson (1989) agreed with Hershkovitz (1962, condition is found in Andinomys and Neotomys. p. 54) that short and long tails are derived from The least developed condition was found by this moderate-length tails. It is unclear in both works survey only in Akodon. whether this polarity only refers to sigmodontines 8 3 P. Length of Neural Spine of Second Cervi- in general or is also a statement about the ancestral cal Vertebra— 2 states. condition in phyllotines. = does not overlap C3 8 IP. Neural Spine of Second Thoracic Ver- = 1 does overlap C3, excluding situation where tebra— 2 states. height is very enlarged = longest spine present on T2 (at least twice as Among phyllotines without a greatly enlarged long as nearby spines) = neural spine on the second cervical vertebra, the 1 short on T2, instead longest on T3 widespread condition is for that spine to not overlap the third cervical vertebra. Alternatively the All sigmodontines except the ichthyomyines knob is angled posteriorly and overlaps the third have a greatly enlarged neural spine on the second cervical. This condition is found in Andalgalomys, vertebra that acts as the site of attachment for the Euneomys, and some Graomys griseoflavus. nuchal ligament (Voss, 1988). Voss reported that in ichthyomyines the spine of the third thoracic is 28S. Hemal Arch— 3 states (Fig. 17). both enlarged and the attachment site for the lig- = absent ament. An enlarged spine on the second thoracic = 1 present, with simple posterior border vertebra is plesiomorphic and widespread for the = 2 present, with spinous posterior border phyllotines just as in the sigmodontines. However, Euneomys also has an enlarged spine on the third In the majority of sigmodontines, the median thoracic. The only other occurrence ofthis derived coccygeal artery, which passes along the ventral

48 FIELDIANA: ZOOLOGY = side of the tail, is bounded by hemal processes at absent = the vertebral junctions, usually starting between 1 present the second and third, becoming most pronounced Polarity of this character is difficult to deter- between the fourth and fifth, and then diminishing mine, but most outgroup taxa lack it. All sigmo- along the next five to ten vertebrae. The hemal dontines have a supertrochlear foramen except arch is a complete bony ring enclosing the artery Rhipidomys latimanus. Carleton (1980) did not and is typically located at the joint between the report on this character. second and third caudal vertebrae (Fig. 1 7). Hemal arches are commonly found among other mam- 3 1 S. Proximal Extent of Fifth Metatarsal— malian orders. In species with additional arches 2 states. between the first and second and/or the third and = peroneal process of fifth metatarsal equal with fourth caudal vertebrae, the arch between the sec- or proximal to the distal edge of calcaneum ond and third is usually the most complex or most (articular surface with the cuboid) developed. Among surveyed pentalophodont gen- = 1 fifth metatarsal not proximal to cuboid/cal- era with 1 2 thoracic vertebrae (Table 6), all appear caneum articulation to possess a "ring"-shaped hemal arch (except Wiedomys). The posterior edge of the arch is Taxonomic coverage is limited, but a "long" extended into a distinct spinous process in Nec- peroneal process (state "0") was found only among tomys, Pseudoryzomys, Zygodontomys, most Ory- the outgroups. zomys, and possibly Microryzomys. An unmodi- 32S. Trochlear Process of Calcaneum — 2 fied hemal arch is found in Oecomys, Neacomys, states. Oryzomys albigularis, O. capito, and possibly O. = xanthaeolus, as well as sporadically among some level with posterior articular facet, process thomasomyines, but sample sizes are small and broad and shelf-like = arches are often damaged. I observed a hemal arch 1 gap between proximal edge of process and in only two tetralophodont genera, Andinomys and posterior articular facet, process shorter and Sigmodon (in both cases, the arch lacked a spinous less shelf-like process). Coding for this character largely follows Carle- 84P. Position of Deltoid Tuberosity— 2 states. ton (1980), except that his state "2" was not ob- = served among species I surveyed and is not in- < 59%, measured from condyle of humerus cluded. Carleton (1980) considered the most to notch of deltoid tuberosity relative to total proximal position of the process (state "0") to be length plesiomorphic for the neotomine-peromyscines, 1 = > 59% but that is unlikely to be true also for the sigmodontines, which I did not observe state "0." The position of the deltoid tuberosity on the among Both character states were found the out- humerus is measured as the percentage of total among groups. humerus length from the condyle to the notch of the tuberosity. The widespread condition for phyllotines is for the tuberosity to be less than 59% of External the total length from the condyle. Morphology

29S. Entepicondylar 2 states. Foramen— 85P. Ventral Surface of Claws (Manus)— 3 states. = present

1 = absent = open basally = 1 closed basally, without strongly developed The presence of an entepicondylar foramen, lo- keel = cated above the medial epicondyle of the humerus 2 fused, forming distinct keel and next to the supertrochlear foramen (Carleton, The widespread condition among phyllotines is 1 980, Fig. 1 3), was considered by Carleton ( 1 980) for claws to be closed basally, without developing to be primitive for neotomine-peromyscines. It is a distinct keel. In Eligmodontia, Chinchillula, Au- present in tylomyines (sensu Carleton, 1 980), most liscomys pictus, and A. sublimis the claw is strongly neotomine-peromyscines and Old World "crice- fused basally, forming a distinct keel. Phyllotis tids," but absent in all sigmodontines examined. wolffsohni, alone among phyllotines, has a claw 30S. Supertrochlear Foramen— 2 states. that is not closed basally.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 49 86P. Length of Dl Relative to D5 (Pes)— 2 34S. Ear (Pinna) Size— 4 states. states. = < 0.068 combined head and body length = = Dl distinctly shorter than D5 1 0.069-0. 108 combined head and body length = = 1 D 1 and D5 subequal in length, not extending 2 0.1 08-0. 150 combined head and body length = past first interphalangeal joint of D2-4 3 > 0. 1 50 combined head and body length

Character state values were determined The widespread condition among phyllotines is using the on for digit 1 to be distinctly shorter than digit 5. segment-coding technique log-transformed ratios as described in the Materials and Methods Reduction in the length of D5 so that it is subequal section. are minor differences between Olds in length with Dl and does not extend past the There Anderson and this in the size bases of D2—4 is present in Reithrodon and Elig- and (1989) study estimation and of several These modontia. Outside the phyllotines, such short out- coding species. differences be a of Four of er digits are also seen in genera with varying lo- may product sampling. the five with the ears sam- comotor modes: Oxymycterus, Sigmodon hispidus, species relatively largest Kunsia, Bolomys, Akodon (Thaptomys), Holochi- pled belong to phyllotines. lus, and Scapteromys. 89P. Countershading of Tail— 3 states. 87P. Position of Hypothenar Pad— 2 states. = distinctly bicolored = = hypothenar extending distally beyond prox- 1 indistinctly bicolored = imal base of the first interdigital pad 2 monocolored = 1 intermediate to first interdigital and thenar Both distinctly and indistinctly bicolored tails pads are common among phyllotines and sigmodontines, polarity equivocal. Indistinctly bi- The widespread condition among phyllotines is making colored tails have a weak but noticeable contrast for the hypothenar to extend distally beyond the between the darker dorsum and lighter underside. base of the first interdigital pad. This is in contrast bicolored tails have a contrast to the widespread condition among surveyed out- Distinctly sharp between the dorsal and ventral sides. Monocol- groups, where the hypothenar is intermediate be- ored tails are found scattered among the phyllo- tween the first interdigital and thenar pads without tines. significantly overlapping the position of either. The is absent in the derived hindfeet hypothenar highly 90P. Furring of Tail Dorsum— 3 states. of Eligmodontia. = sparsely furred, scales evident 33S 8 8 P. Furring of Soles of Feet (Pes)— 3 = 1 furred, scales visible but indistinct states. 2 = densely furred, scales scarcely visible = hair on heels sparse only furred tail is the most common = Moderately 1 heels furred, distal pads naked character state among phyllotines. 2 = heel and distal pads furred 9 IP. Body Pelage Pattern— 3 states. Olds and Anderson (1989) listed this character = countershaded among the possible phyllotine synapomorphies and distinctly 1 = or not countershaded included it in the diagnosis but did not include it indistinctly in their data matrix Table Their dis- (1989, 1). The widespread condition among phyllotines is cussion that it is a condition implied widespread for the dorsum and sides of the body to be a con- while it among sigmodontines among phyllotines trasting color and tone from the undersides. No is characterized species-specific adaptations to by phyllotine is considered to be monocolored, but local environments. Furred heels are found in tho- Euneomys and Neotomys are indistinctly bicol- masomyines, akodontines, and some phyllotines. ored. Furring ofthe soles ofthe hindfeet in phyllotines varies from sparse lateral fur that extends toward 92P. Pectoral Streaks— 2 states. the sole, to more extensive furring around and absent onto the sole of the heel, to furring among the present distal pads. The third condition is easily distinguished from the first two, which have no hair at Only a minority of phyllotine species possess all in the broad distal portion of the sole. pectoral streaks, and those few are concentrated

50 FIELDIANA: ZOOLOGY in the genus Phyllotis. A pectoral streak is usually coding scheme. Spotorno (1992) considered re- an ocher coloration to the fur along the midline duced or absent lateral mounds to be derived for in the pectoral region, approximately circular to New World murids and linked with a suite of traits longitudinal in shape, often presenting a subtle partially controlled by hormonal levels. contrast to the predominantly neutral color of the 93P. Distal/Proximal Bacular Length— 3 undersides and usually limited to a streak rather states (Fig. 18). than spreading laterally toward the shoulders. Pec- toral streaks are found in Neotomys, Loxodonto- = < 0.63 = mys, P. osilae, P. magister, P. definitus, P. wolff- 1 0.63-0.77 = sohni, and at small to moderate frequencies in 2 > 77 some populations of P. xanthopygus. The landmarks chosen to define this character 35S. Mammae Number— 4 states. are from the distal tip of the medial digit of the = 4 distal baculum to the distal tip of the proximal

1 = 6 and from the tip of the proximal to the line con- 2 = 8 necting the widest points ofthe base. The common 3 = 10 or more condition for phyllotines is for an intermediate- sized distal baculum, 63-77% the length of the baculum. A distal Four pairs of pectoral, postaxial, abdominal, and proximal large baculum, ranging from 79 to 105% the character- inguinal mammae constitute the widespread con- proximal length, izes most members of the P. dition among the Sigmodontinae. Six mammae xanthopygus species P. darwini, P. (postaxial, abdominal, and inguinal) are found in group (P. xanthopygus, magister, and P. caprinus; Steppan, 1993). No sigmodon- the thomasomyines (e.g., Thomasomys, some De- tines outside have such lomys, Rhipidomys), all ichthyomyines, some ory- Phyllotis relatively large distal bacula (Hooper, 1962; Hooper & Musser, zomyines (just Neacomys), Wiedomys, and the this "problematic" Rhagomys (Reig, 1980). Voss 1964; Hershkovitz, 1966b; Spotorno, 1986; (1993) considered six mammae to be plesio- study). morphic for the sigmodontines. Mammae number 94P. Length of Lateral Mounds Relative to may be intraspecifically variable in Delomys dor- Medial Mound— 2 states. salis but (6-8; Voss, 1993), sharp geographic seg- = > % regation of the mammary counts suggests that two 1 = < % species may be present. More than eight mammae are found in some and only Calomys, Sigmodon, The widespread and likely plesiomorphic con- Holochilus. mammae is Thus, eight likely plesio- dition among phyllotines is for the lateral mounds for the but the of morphic phyllotines, presence to be greater than % the length of the medial bacu- six versus mammae be infor- eight may highly lar mound. Auliscomys sublimis and A. pictus share mative for pentalophodont sigmodontines (Table the derived condition where the lateral mounds 6). are less than % the medial length. Andinomys is unusual in having lateral mounds that are longer than the medial (Spotorno, 1986). Characters of the Phallus and

Soft Anatomy 9 5 P. Hooks on Lateral Mounds— 2 states (Fig. 18). 36S. Bacular Complexity— 2 states. = absent = = lateral bacular mounds present 1 present, pointing basally

1 = lateral bacular mounds absent The widespread condition among sigmodon- The complex penis with lateral bacular mounds tines and phyllotines, and thus likely the plesio- present has conventionally been considered prim- morphic condition among phyllotines, is for the itive for most muroids (Hooper & Musser, 1 964; lateral mounds of the distal baculum to be simple Hershkovitz, 1966b). Carleton (1980) hypothe- and unmodified. The presence of basally directed sized that the simple penis was primitive for the hooks projecting from the tips ofthe lateral mounds neotomine-peromyscines. I coded taxa with high- is unique to the Phyllotis xanthopygus species- ly reduced and peculiar lateral mounds (Nyctomys, group (Fig. 18). The condition in P. definitus is Scapteromys) as unknown given this simplified unclear; while Spotorno (1986, Fig. 5.9) indicates

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 51 Fig. 18. Ventral and lateral views of bacular apparatus in Phyllotis magister (fmnh 107469). dn, dorsal knob of lateral mounds; ldb, length of distal baculum; lh, lateral hooks; lpb; length of proximal baculum.

the presence of small hooks, species similarly drawn = absent = in that publication do not possess hooks among 1 one pair present FMNH specimens examined by me. Preparation 2 = two pairs present differences are probably the cause of the differing Data and character coding are adapted from Voss observations. and Linzey (1981). Male accessory glands show 9 6 P. Knob on Dorsal Surface of Lateral little phylogenetically informative variation among Mounds— 2 states (Fig. 18). the sigmodontines (Voss & Linzey, 1981). The exceptions are ventral prostrates, which vary only = absent = among akodontines, and the number of preputial 1 present glands. At least one species in each major generic- group possesses a single pair, with the exception Small knobs project from the middorsal surface of the oxymycterines and Sigmodon. A single pair of the lateral mounds in the species Phyllotis xan- or no preputials are both widespread among other thopygus, P. caprinus, and P. magister (Fig. 1 8). muroids (Carleton, 1980), but absence of prepu- Suitably prepared phalli of P. darwini were not tials is found only in several oryzomyines and tho- available for examination. All other examined masomyines (4 of 1 7 species surveyed). Two pairs species lacked the dorsal knobs. were not reported by Carleton (1980) or Voss and 37S. Preputial Glands— 3 states. Linzey (1981) outside the South American Sig-

52 FIELDIANA: ZOOLOGY = modontinae. The plesiomorphic state for sigmo- absent = dontines therefore appears to be one pair. Two 1 present pairs of preputials occur in Chroeomys jelskii (my Carleton ( 1 980) and Voss (1993) considered the observations suggest small third or even fourth presence of a gallbladder to be plesiomorphic pairs), oxymycterines (sometimes raised to tribal among New World "cricetids." Voss (1991) ex- status, otherwise treated as a subgroup within the amined 93 species of sigmodontine rodents for the akodontines), Sigmodon, and most phyllotines. presence of a gallbladder. This character shows 97P. Preputial Glands— 3 states. little variation within tribes or major generic = groups. The only exceptions seem to be charac- single large lateral pair = teristic ofunrelated scattered genera. A gallbladder 1 single large lateral pair with very small ( < 1 is present in all surveyed akodontines except Len- mm) medial pair = oxus and Akodon cursor, all ichthyomyines except 2 single large lateral pair with medium length Ichthyomys, all thomasomyines except Rhipido- {2-A mm) medial pair mys, all phyllotines, the scapteromyines, and Sig- modon (Punomys was not It is absent Two sources of data on preputial glands were examined). in all oryzomyines, including available for this study: Voss and Linzey's (1981) Pseudoryzomys, Zygodontomys, and Holochilus. All 16 study on male accessory glands in New World mu- surveyed phyllotines have a gallbladder (Voss, 1991). roids and my observations of partially cleared and stained phalli. The coding was adjusted to reflect 39S. Gastric Epithelium 1—4 states. the range of variation observed in phyllotines as = hemiglandular compared to sigmodontines. Voss and Linzey = 1 intermediate, reduction in glandular zone (198 1) found one pair (» 10 mm) in Calomys (C around antrum callosus and C. laucha), while the rest of the sur- = 2 discoglandular veyed phyllotines (Andalgalomys pearsoni, Elig- 3 = pouched modontia typus, Graomys griseqflavus, P. darwini, and P. osilae) had a second smaller ventral pair 40S. Gastric Epithelium II—2 states. (2.5-3.5 mm). The material examined by Voss and = hemiglandular Linzey (1981) consisted of phalli dissected from = 1 intermediate, reduction in glandular zone fluid-preserved carcasses stored in 70% alcohol. I along greater curvature examined phalli that had been partially cleared and stored in With exami- glycerin. microscopic Data and coding are slightly modified from nation of dissected and backlit I found that phalli, Carleton (1973). Character 40S is treated as a sep- detail could be observed than with greater un- arate character to distinguish the intermediate cleared alcohol-preserved specimens. My obser- condition found in ichthyomyines from that found vations of of Calo- species Akodon, Chroeomys, in Thomasomys and Scapteromys. The large ma- mys, Phyllotis, Auliscomys, Irenomys, and jority of sigmodontines possess the hemiglandular Neotomys all support the observations ofVoss and condition. Linzey (1981). However, C. sorellus was found to have a very small ventral pair of glands in exactly the same position and with the same texture and as the ventral in shape pair the other phyllotines. Phylogenetic Relationships within This ventral pair is so small, 0.5-0.8 mm, that it Sigmodontinae is possible that Voss and Linzey would not have recognized it using their methods had they ex- Results amined C. sorellus. Because I did not have access to cleared of C. callosus and appropriately phalli When the analysis includes the dummy variable C. laucha to confirm the absence of such a small to favor sigmodontine monophyly, 28 equally ventral pair of these two were cod- glands, species most-parsimonious trees result, the strict consen- ed as unknown All nine of P. ("?"). specimens sus of which is presented in Figure 19. Each of the chilensis examined me have an xanthopygus by most-parsimonious trees (including outgroups) is additional third 1 mm and situated be- pair, long 287 steps long with a CI of 0.26 and RI (Farris, tween the lateral and medial pairs. 1989) of 0.61. When only the sigmodontines are the values are 195 CI 38S 98P. Gallbladder— 2 states. considered, respective steps,

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 53 og <2-2 3 § IS It 5 5 IS 8 §"9 c

54 FIELDIANA: ZOOLOGY Table 7. Consistency and retention indexes for sigmodontine characters.

Number Character name CI* RI"

1 . Mesoloph(-id) 0.333 2. Length M3 3. Shape M2 4. Posterior extent of incisive foramina 5. Dorsoventral position of anterior root of zygomata 6. Posterior margin of zygomatic plate 7. Masseteric tubercle 8. Zygomatic notch 9. Posterior extension of nasals 10. Posterior extension of premaxillaries

1 1 . Supraorbital edge 1 2. Shape of frontoparietal suture 1 3. Ratio of interparietal/parietal length 14. Parietal/occipital contact 15. Tegmen tympani 16. Squamosal fold 1 7. Subsquamosal foramen 18. Relative width of mesopterygoid fossa 1 9. Shape of mesopterygoid fossa 20. Sphenopalatine vacuities

2 1 . Position of anterior border of mesopterygoid fossa 22. Carotid circulation 23. Alisphenoid strut 24. Hyoid 25. Articulation of first rib 26. Number of thoracic and lumbar vertebrae 27. Number of caudal vertebrae 28. Hemal arch 29. Entepicondylar foramen 30. Supertrochlear foramen

3 1 . Proximal extent of fifth metatarsal 32. Trochlear process of calcaneum 33. Furring of soles of feet (pes) 34. Ear (pinna) size 35. Mammae number 36. Bacular complexity 37. Preputial glands 38. Gallbladder 39. Gastric epithelium I 40. Gastric epithelium II • Oryzomys capito

, Oryzomys palustris Nectomys squamipes Zygodontomys brevicauda Holochilus brasiliensis Pseudoryzomys simplex Oligoryzomys fulvescens Neacomys spinosus Chilomys instans Anotomys leander Ichthyomys hydrobates tE Neusticomys monticolus w^Akodon boliviensis 1— Akodon albiventer Kunsia tomentosus Sigmodon hispidus Calomys callosus Graomys griseoflavus Reithrodon auritus Mystromys albicaudatus Phyllotis darwini Neotomys ebriosus Punomys lemminus Scapteromys tumidus Oxymycterus hispidus Thomasomys rhoadsi Rhipidomys latimanus Thomasomys aureus Thomasomys baeops ^ Scotinomvs teauina Ochrotomvs nuttalli Peromyscus leucoous Wiedomys pyrrhorhinos Calomyscus baluchi Nyctomvs sumichrasti Tylomys nudicaudus Neotoma floridana Cricetulus migratorius Mesocricetus auratus Phodopus sunaorus

56 FIELDIANA: ZOOLOGY modontines and place Wiedomys among them, but tomys (Fig. 19). Monophyly of the ichthyomyines all four steps are needed to remove Mystromys. has already been established (Voss, 1 988), but this The seemingly unlikely placement of Mystromys, study more directly tests monophyly by placing both on conventional systematic and biogeograph- the ichthyomyines in a broader cladistic analysis. ic grounds, leads me to prefer the weighted tree In the bootstrap analysis weighted to favor sig- with the monophyletic Sigmodontinae. Characters modontine monophyly (Fig. 21), Ichthyomyini that unequivocally support a monophyletic Sig- is found in 97% of replicates. Trees five steps long- modontinae given the outgroup topology in Figure er than the most-parsimonious trees (with or with- 1 9 {Scotinomys as the sister-group to Sigmodon- out a dummy variable) must be examined before tinae) include complex baculum and entoglossal a nonmonophyletic Ichthyomyini is found. Ich- process of hyoid absent. Other possible sigmo- thyomyine monophyly is supported by nasals that dontine synapomorphies given different outgroup extend posterior to lachrymal, masseteric tuber- topologies (e.g., monophyletic neotomine-pero- cles, small or very small pinnae, and reduced gas- myscines) would include dual articulation of the tric glandular epithelium. Sensitivity analyses first rib with the transverse processes of the sev- demonstrate that, although the ichthyomyines are enth cervical and first thoracic vertebrae, fifth most commonly placed in a clade with other te- metatarsal not posterior to cuboid/calcaneum ar- tralophodont groups, the akodontines and phyllo- ticulation, entepicondylar foramen ofhumerus ab- tines, the sister-group to the ichthyomyines is sent, supertrochlear foramen of humerus present, sometimes the oryzomyines or a thomasomyine. and a gap between trochlear process and articular It appears that the ichthyomyines occupy a rela- facet of the calcaneum. tively basal position in the sigmodontine tree. Neither the weighted nor the unweighted anal- The oryzomyine taxa, exclusive of the thoma- yses show a monophyletic Neotominae, but they somyines, form the second major clade. Support- do strongly support a monophyletic tylomyine ing the oryzomyines sensu stricto are the putative group sensu Carleton ( 1 980) and Reig ( 1 984), Nyc- synapomorphies of nasals extending posterior to tomys and Tylomys. This pairing is found in 92% lachrymal (except in some Oryzomys), alisphen- of bootstrap replicates (Fig. 21) and is supported oid strut absent, long palate (except in Holochilus), by the following putative synapomorphies: no pa- eight or more mammae, 1 2 thoracic vertebrae (ex- rietal/occipital contact, squamosal fold, preputial cept in Neacomys), fewer than 36 caudal verte- glands present, overhanging supraorbital, subs- brae, hemal arch present, and gallbladder absent. quamosal foramen absent, stapedial branch of the The distribution of some of these potentially di- carotid artery reduced or absent, more than 30 agnostic characters is explored in greater taxonom- caudal vertebrae, and gallbladder absent. All these ic detail in Table 6. The most-parsimonious trees character states are found in Ototylomys as well given sigmodontine monophyly (Fig. 19) place except a squamosal fold; the presence of preputials Wiedomys as the sister-group to the oryzomyines. and gallbladder in Ototylomys is unknown. Oto- In this hypothesis, absence of the alisphenoid strut, nyctomys was not examined. Absence of subsqua- 1 2 ribs, and possibly presence of the hemal arch mosal foramen and reduction of carotid circula- (weakly developed in Wiedomys) would be synap- tion can also be found in Neotoma. Four steps are omorphies of the more inclusive clade. However, required to collapse the branch joining the tylo- the bootstrap consensus places Chilomys as the myines and forming a trichotomy with Neotoma. sister taxon to Oryzomyini, indicating that the sis- The phylogenetic analysis identifies two well- ter-group to Oryzomyini is unresolved. Two ad- defined branches within the Sigmodontinae tree, ditional steps are needed to draw the putatively ichthyomyines and the oryzomyines sensu Voss basal oryzomyines (Oligoryzomys or Neacomys) and Carleton (1993), including the tetralophodont out of the oryzomyines. Three additional steps are genera Holochilus, Pseudoryzomys, and Zygodon- needed to disrupt oryzomyine monophyly when

Fig. 20. Strict consensus cladogram of the 28 equally most-parsimonious trees of the South American Sigmo- dontinae, without a dummy variable. Overall length for each tree is 282 steps. The South African-endemic Mystromys is placed among the phyllotines, and the neotomine-peromyscine Scotinomys is placed near the base of the sigmodontines. Outgroup taxa are underlined.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 57 Oryzomys capito Oryzomys palustris 46 30- Nectomys squamipes 39H Zygodontomys brevicauda l l— Holochilus brasiliensis r46- 17- £ Pseudoryzomys simplex fulvescens 1-13- Oligoryzomys Neacomys spinosus Chilomys instans Akodon boliviensis 83 1— Akodon albiventer 11 Oxymycterus hispidus { 16 Scapteromys turn idus { Kunsia tomentosus Sigmodon hispidus 27 Reithrodon auritus r-9- { Graomys griseoflavus Calomys callosus 31-ft Phyllotis darwini 22. 36 r m Neotomys ebriosus Punomys lemminus Anotomys leander Ichthyomys hydrobates »{ 52 r57- 22- Neusticomys monticolus Thomasomys rhoadsi Wiedomys pyrrhorhinos 38H Rhipidomys latimanus 42 aureus 52 Thomasomys { 28- Thomasomys baeops Scotinomys teguina Ochrotomys nuttalli 41 l 31- 1 Peromyscus leucopus Tylomys nudicaudus 92 sumichrasti 83-I Nyctomys Neotoma floridana Cricetulus migratorius 54 j 1— Mesocricetus auratus • 4e\ 19 Phodopus sungorus Calomyscus baluchi Mystromys albicaudatus

58 FIELDIANA: ZOOLOGY Oligoryzomys and Neacomys are excluded from zygomatic notch, reduction to fewer than 30 cau- the analysis. dal vertebrae (with hypothesized subsequent in- The root ofthe sigmodontine tree is placed with- crease), and eight or more mammae. If the ich- in the thomasomyines, which appear to be highly thyomyines are not the immediate sister-group to paraphyletic. However, the bootstrap percentages the tetralophodont tribal-group, then reduction or are lowest for this region of the tree (Fig. 21). This partial fusion of the mesolophostyle would also be analysis suggests that Thomasomys is also para- a supporting character. The tetralophodont tribal- phyletic and its members occupy basal positions group is also generally characterized by the ple- on the tree. Four additional steps are needed to siomorphic presence of a gallbladder, a stapedial join T. rhoadsi with its congeners in a monophy- artery imparting a squamosal groove, moderately letic Thomasomys. However, placement ofthe sig- haired heels, nonoverhanging supraorbital surface, modontine root is unstable because most char- 1 3 thoracic rib pairs (with subsequent reduction), acters that are variable within the sigmodontines no hemal arch, and relatively small M3 (with en- are also variable among the outgroups. The result largement among the phyllotines). Bootstrap per- of this is that character polarities determined by centages for this clade are very low (Fig. 21). outgroup criteria are only as accurate as the out- Within the tetralophodont tribal-group, Scap- group topology. Sensitivity analyses (not illustrat- teromyini (Hershkovitz, 1966), which includes ed) demonstrated that shifts in outgroup relation- Scapteromys and Kunsia, appears polyphyletic, and ships had significant impact on the placement of an Akodontini that includes the oxymycterines the sigmodontine root. Among these various anal- (sensu Reig, 1987) without the scapteromyines may yses, the topological relationships within the sig- be either paraphyletic (bootstrap consensus tree) modontine portion of the unrooted network were or polyphyletic (most-parsimonious trees). How- relatively stable. It was the placement of the root ever, only one additional step is needed to be con- within that network that was most unstable. Thus, sistent with a monophyletic Akodontini (Akodon if the neotomine-peromyscines and tylomyines plus Oxymycterus). Two additional steps are need- were sister-groups and the Old World "cricetids" ed for a monophyletic Scapteromyini. Punomys is were paraphyletic, then the morphology of the an- placed as the sister taxon to the phyllotines plus cestral sigmodontine would most closely resemble Sigmodon. the Old World "cricetids," and the sigmodontine Monophyly of the phyllotines as currently de- root would be placed just outside the phyllotine fined is not directly supported. This analysis places group. If the tylomyines were the sister-group to Sigmodon within the phyllotine radiation. Three the sigmodontines, then the root would be placed additional steps are needed to bring Sigmodon near the base of the oryzomyines. In most rooting down to the immediate sister taxon to the phyl- alternatives, though, the thomasomyines were still lotines. In that topology, Punomys is placed as the paraphyletic and generally basal. sister taxon to Oxymycterus in an akodontine clade. The remaining branch with the most character In the most-parsimonious trees (Fig. 19), Sigo- support is that consisting of akodontines, scapter- modon, Reithrodon, and Neotomys have the lon- omyines, phyllotines, Sigmodon, and Punomys, gest branch lengths among the terminal sigmo- and is the sister-group to the ichthyomyines in dontine taxa. Characters that support a phyllotine Figure 19. I will refer to it by the shorter name of plus Sigmodon clade (or a phyllotine clade with the tetralophodont tribal-group because it is com- Sigmodon excluded) include complete loss of the prised by most of the taxa hypothesized by Hersh- mesoloph, premaxillaries subequal in extent with kovitz to be derived from tetralophodont stock nasals, large sphenopalatine vacuity, mesoptery- (i.e., lacking a complete mesolophostyle). The is- goid narrower than parapterygoid, and large pin- sue of Sigmodon is addressed separately below. nae (except Neotomys). The clade ofPunomys plus This clade is supported by the following putative the phyllotines is supported by a zygomatic arch synapomorphies, as hypothesized from the map- inserting at a moderate to high position on the ping of character transformations on the tree: in- rostrum and a moderately long interparietal. cisive foramina extending to the molars, deep Traditional notions of a Sigmodontini (Reig,

Fig. 2 1 . Majority-rule bootstrap consensus tree ( 1 00 replicates) for the Sigmodontinae, including dummy variable weighted to favor sigmodontine monophyly. Numbers indicate percentage of replicates containing the specified clades.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 59 1980) are not supported. The shortest trees that sigmodontine radiation based on a few dental place Sigmodon and Holochilus as sister taxa are characters seems prone to mislead. Additionally, six steps longer than the most-parsimonious trees. most of the Bensonomys fossils cannot be a Cal- Notably, those constrained trees move Sigmodon omys as the genus is currently defined (Steppan, into the oryzomyines while leaving the rest of the 1993); they possess mesolophs that are entirely tree nearly unchanged. The shortest trees that in- absent among all extant Calomys and in the sister- clude a Sigmodontini sensu Hershkovitz (1955; group to Calomys, the remaining phyllotines. This Sigmodon, Holochilus, Reithrodon, Neotomys) are argument does not preclude the possibility that 1 1 steps longer. In those alternative trees, the weight prephyllotine ancestors or their relatives may have of Reithrodon and Neotomys draws Holochilus into possessed reduced mesolophs. a terminal phyllotine branch. The placement of Chilomys is unstable, and in the past it has been placed either with the thomasomyines (Hershkovitz, 1966a; Voss, 1991) or with the oryzomyines sensu stricto (Reig, 1980, Discussion 1986; Voss & Linzey, 1981). The proper status of the Wiedomyini is difficult to assess from this Three principal conclusions can be drawn from analysis, but Wiedomys does not appear to be far the sigmodontine analysis. First, the Oryzomyini derived from a basal "thomasomyine" grade. (Voss & Carleton, 1993), which includes the te- Voss (1993) argued that formally maintaining tralophodont genera Holochilus, Pseudoryzomys, genera in the tribe Thomasomyini was unjustifi- and Zygodontomys, but which excludes the tho- able owing to their collective lack of unifying, apo- masomyines, is confirmed. Second, the phyllotines morphic characters and the inability to assign in- are members of a tetralophodont tribal-group that dividual "thomasomyine" genera to other includes the akodontines, scapteromyines, and demonstrable monophyletic groups. As an alter- Punomys. Third, the root of the sigmodontines is native, he recommended referring to thomaso- placed within the thomasomyine group. Mono- myines formally as a "plesion," citing Wiley (1 98 1) phyly of the Sigmodontinae is indicated given res- for comparison, to emphasize that they share ervations regarding Mystromys and Scotinomys. primitive attributes. Wiley (1981, p. 219) explic- Additionally, the monophyly of the Ichthyomyini itly reserved the nonranked category "plesion" for and Tylomyinae is strongly supported. fossil taxa, allowing them to be classified with Re- A problem for any sigmodontine phylogeny giv- cent taxa without revising established classifica- en the current state of muroid systematics is the tions for either the fossil or Recent taxa. rooting of the tree. Many workers have taken the The study strongly supports a tylomyine group position that the oryzomyines or thomasomyines that includes Nyctomys, Tylomys, and Ototylomys are the basal sigmodontines (Gardner & Patton, (and presumably Otonyctomys, which was not ex- 1976; Hershkovitz, 1962, 1993; Reig, 1980, 1986), amined). While this study groups the tylomyines partially because oryzomyines sensu lato are char- with Neotoma, other studies indicate that the ty- acterized by many traits considered widespread lomyines are basal members of a New World ra- and potentially plesiomorphic among other cri- diation with no clear affinities to either the sig- cetine or murid rodents. Phyllotines have been modontines or neotomine-peromyscines considered by neontologists to be highly derived (Carleton, 1980;Sarich, 1985; Haiduketal., 1988; (Hershkovitz, 1962; Reig, 1986). However, Jacobs Catzeflis et al., 1993). Using G-banded chromo- and Lindsay ( 1 984) concluded that phyllotines were somes, Haiduk et al. (1988) suggested that Nyc- primitive among sigmodontines, based almost en- tomys was a basal member of a neotomine radi- tirely on the identification ofBensonomys, the old- ation after its separation from the South American est putatively sigmodontine fossils, as a subgenus sigmodontines. However, because the sigmodon- of Calomys. Their hypotheses of primitive traits tines available for comparison constituted a tax- among sigmodontines are merely those of Ben- onomically biased and incomplete sample, in- sonomys (Jacobs & Lindsay, 1 984, Table2) but are cluding only oryzomyines and Sigmodon, suggested limited to molar and lower mandible characters, relationships to the sigmodontines as a whole must the only characters observable from the fossils. In be tentative at best. The monophyly and basal contrast, this study found few dental characters phylogenetic position of the tylomyines should be (only three were included) that were informative recognized formally. Tribal status as the Tylo- at the tribal to subfamily level, and rooting of the myini (Carleton, 1980) would be most in keeping

60 FIELDIANA: ZOOLOGY with the results of this study (tylomyines as sister- (Spotorno, 1986) (Fig. 5 A) and cladistically (Fig. group to Neotoma), but the available molecular 5B, reanalysis of data in Spotorno, 1986) place studies suggest that elevating the tylomyines to an Sigmodon among the North American neoto- equal status with the sigmodontines and neotom- mine-peromyscines. Second, when Sigmodon is ine-peromyscines as the subfamily Tylomyinae included in the phyllotine data set (results not (sensu Reig, 1984) would be more generally con- shown), three discordant alternative placements sistent. for Sigmodon are represented among the most- This analysis is generally at odds with the re- parsimonious trees. Sixty percent of most-parsi- lationships and evolutionary tendencies in chro- monious trees include Sigmodon, with 27% placing mosomal evolution proposed by Gardner and Pat- Sigmodon as the sister-taxon to the phyllotines ton (1976). They hypothesized that chromosome plus akodontines and 13% placing Sigmodon as change was unidirectional, dominated by Robert- the basal oryzomyine. The last two alternatives sonian fusions that reduced diploid number. Par- push Graomys to a basal position among the phyl- tially as a consequence of their model, the genus lotines, in contrast to its typical terminal position Oryzomys was proposed to be the stem group from (Steppan, 1993). The placement of Sigmodon which all sigmodontines arose. This analysis clear- among the phyllotines in the sigmodontine anal- ly shows that Oryzomys is a member of a derived ysis may be due to the taxa sampled and the char- clade separated from all other sigmodontine tribes. acters included. For example, three of the five Nor is Oryzomys the basal member of the ory- phyllotines in this sigmodontine analysis possess zomyine clade, although the genus as currently 1 2 ribs (Calomys, Graomys, Reithrodon), which is constituted is probably not monophyletic. Gard- most-parsimoniously hypothesized as a synapo- ner and Patton (1976) also hypothesized that Ho- morphy joining them with Sigmodon. However, lochilus was derived from a phyllotine stock; it is this results from the improbable sampling of three clearly an oryzomyine. Akodontines were dia- independent rib losses from the both common and grammed as independently derived from Oryzo- plesiomorphic phyllotine condition (Steppan, mys (Gardner & Patton, 1976, Fig. 10), while this 1993). Likewise, Sigmodon, Graomys, and Reith- study indicates that the akodontines are related to rodon are associated in this analysis by the pres- the phyllotines and scapteromyines. Even a cur- ence of an alisphenoid strut, but the strut appears sory mapping ofdiploid numbers on the trees from to be independently evolved in Graomys and this study indicates that Robertsonian fusion can- Reithrodon (Steppan, 1993). Sigmodon possesses not be used as the exclusive model ofchromosom- the posteriorly divergent, ledged supraorbital re- al evolution in sigmodontines. Gardner and Pat- gion found in some oryzomyines, Andalgalomys, ton's model (which has also been proposed by and Graomys, but that is very different from the Bianchi et al. [1971] and Pearson and Patton narrow and vertically ridged supraorbitals of [1976]) has been salutary, but it has unfortunately Reithrodon and Neotomys. Third, Sigmodon lacks sometimes been evoked as a phylogenetic axiom an angled premaxillo-maxillary suture, a syna- to polarize phylogenetic relationships (e.g., Vitullo pomorphy that joins Reithrodon, Neotomys, and etal., 1990). Euneomys and is unique to them (Steppan, 1993), The previous discussion has referred to Sig- and also lacks grooved incisors (except S. alstoni), modon as an entity separate from the phyllotines another putative synapomorphy ofthe Reithrodon despite its placement in Figure 19. There are three group. reasons for believing that the results of the phy- Sigmodon and some Holochilus are apparently logenetic study shown in Figure 1 9 are misleading the only sigmodontines other than Calomys with in regard to Sigmodon. First, three independent more than eight mammae. This would seem to molecular data sets place Sigmodon outside any support the Sigmodontini sensu Reig ( 1 980). How- clade formed by the remaining sigmodontines. ever, the remaining characters do not support the DNA hybridization places Sigmodon outside a Sigmodontini, and this analysis places Sigmodon clade that includes Oryzomys, Zygodontomys, and Holochilus far apart on the tree (Fig. 1 9) and Akodon, and Phyllotis (Catzeflis et al., 1993). Pro- requires an additional six steps to unite them. Ad- tein immunological distances place Sigmodon as ditionally, albumin immunological cross-reac- the outgroup to all the other sigmodontines (Sar- tions between Sigmodon and Holochilus are much ich, 1985), while the rest of the immunological lower than in other intratribal comparisons (Sar- tree (Fig. 1 ) is in general agreement with this study. ich, 1985). Electrophoretic allele data analyzed phenetically On the basis of the results here, and consistent

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 61 with the frequently suggested but rarely acted upon Similarly, ribs show six losses (given phyllotine observations of previous workers (Gardner & Pat- phylogeny of Steppan [1993] and data in Table 5; ton, 1976; Hooper & Musser, 1964; Sarich, 1985; Sigmodon, Calomys, Graomys, Reithrodon, Rhip- Voss & Myers, 1991), Holochilus should be re- idomys, and oryzomyines plus Wiedomys) versus moved from the Sigmodontini, leaving only Sig- two gains (Anotomys and Neacomys). modon. These various analyses disagree so strong- In contrast, mammae number shows at least five ly on the position of Sigmodon that it seems best increases (sigmodontines, tetralophodont tribal- to draw no conclusions at this time on its rela- group, oryzomyines, Calomys, Holochilus, Sig- tionships. Thus, if Sigmodontini is retained as a modon) versus one to two reductions (Scotinomys, separate tribe, it should be considered Sigmodon- tylomyines plus Neotoma). Given this tree topol- tinae incertae sedis, or even Muridae incertae sed- ogy, alisphenoid strut (four gains, four losses) and is. caudal vertebrae (approximately five gains, five An interesting series of patterns is developed losses) do not show a bias. Somewhat surprisingly, with characters that involve gain or loss, and in the quantitative character pinna size shows seven particular gain or loss of discrete structures, as to eight reductions versus one (phyllotines) to three opposed to topological conformation of parts. (Neusticomys, Thomasomys baeops) enlarge- Among characters that show significant homopla- ments, depending on resolution of polytomies. The sy (i.e., multiple observations of evolutionary preponderance of pinna reduction over enlarge- transformations), structures that are lost generally ment may in part be an artifact of the particular are not regained. Gallbladders, stapedial branch selection of character states. of the carotid artery, mesoloph(-id), and thoracic ribs all show a bias toward loss (or bias against reacquisition) based on character mapping and op- Phyllotine Monophyly timization (accelerated transformation, ACCT- RAN, and delayed transformation, DELTRAN, Because of the placement of Sigmodon within options were both used). Unless otherwise noted, the phyllotines, this study does not provide strong character transformations discussed below were support for the monophyly of the tribe Phyllotini. hypothesized using both optimizations or by However, if the molecular studies are correct in ACCTRAN, which disfavors the observation of placing Sigmodon in a basal position among the directional bias in character transformations. sigmodontines, then its placement among the Across the entire tree (and other records; Voss, phyllotines in this study may be due to conver- 1991), the gallbladder is lost nine times (oryzo- gence and the absence of any taxa along the Sig- myines, Ichthyomys, Rhipidomys, Ochrotomys, modon lineage to provide evidence for character tylomyines, Cricetulus, Akodon cursor, Geoxus transformations, thus leaving Sigmodon with a very valdivianus) but never regained. Mesolophs are re- long branch length. Very long branches due to in- duced at least two times (tetralophodont tribal sufficient taxonomic sampling can lead to erro- group plus ichthyomyines, ancestor of Holochilus, neous groupings (Felsenstein, 1978; Huelsenbeck Pseudoryzomys, and Zygodontomys) and entirely & Hillis, 1 993). Forced removal ofSigmodon does lost three to four times (phyllotines, some ichthyo- not affect the remaining hypothesized relation- myines, Sigmodon, Holochilus), but (depending ships. Therefore, I define the Phyllotini as the on phylogenetic resolution) only gained to a poorly common ancestor of the following genera and all developed condition in Anotomys, maybe some its descendants: Andalgalomys, Andinomys, Au- akodontines not included here, and originally in liscomys, Calomys, Chinchillula, Eligmodontia, an ancestor to the sigmodontines. The stapedial Euneomys, Galenomys, Graomys, Irenomys, Neo- branch of the carotid artery shows at least seven tomys, Phyllotis, and Reithrodon. Punomys does reductions or losses among sigmodontines (ory- not appear to be a phyllotine, but in a taxonom- zomyines, Thomasomys/Rhipidomys clade, some ically more focused study (Steppan, 1993) Puno- ichthyomyines, Reithrodon, Sigmodon, Neoto- mys is equivocally associated with some phyllo- mys, Chilomys), but, depending on relationships tine taxa. I therefore agree with the conclusions of within the oryzomyines and under ACCTRAN Olds and Anderson (1989) on the content of Phyl- optimization, only three reversals (Neacomys, lotini, with the single exception of not including Neusticomys, Zygodontomys). Under DELTRAN Punomys. Diagnostic synapomorphies for the optimization, there are nine reductions in carotid Phyllotini are moderate or large pinnae (> 0.15 circulation and only one reversal {Zygodontomys). head and body length), parapterygoid fossa rela-

62 FIELDIANA: ZOOLOGY tively broader than mesopterygoid fossa, very open (Steppan, 1993), using nearly the same data set sphenopalatine vacuities, complete loss of the me- (see Materials and Methods). While the most-par- soloph, posterior extensions of premaxillaries and simonious trees place Punomys in a clade with nasals subequal, and (except Calomys) two pairs Andinomys and Irenomys (Fig. 22, strict consen- of preputial glands. In contrast, the following char- sus), trees only one step longer place Punomys acter states hypothesized by Olds and Anderson outside the phyllotines (Fig. 23, 80% majority rule (1989) to be phyllotine synapomorphies instead consensus). In Steppan (1993), these alternative appear to be plesiomorphic: hairy heel, palate long, topologies were equally parsimonious, and the incisive foramina long, supraorbital region never nonphyllotine hypothesis for Punomys was pre- evenly curved in cross-section, interparietal well ferred. The 200 replicate bootstrap analysis with developed, zygomatic notch deeply excised, teeth this data set (Fig. 24) also places Punomys as the tetralophodont (which includes vestigial meso- sister taxon to the phyllotines, the same as in the lophs [Olds & Anderson, 1989]), and M3 more sigmodontine analysis. The root was placed be- than half the length of M2. A formal diagnosis is tween Thomasomys and the oryzomyines by des- presented in the Taxonomy section. ignating Thomasomys as the outgroup. Although most recent studies have excluded Details of character support for specific nodes

Pseudoryzomys from the phyllotines (e.g., Olds & can be found in Steppan (1993). Differences be- Anderson, 1989; Voss & Myers, 1991; Voss & tween the two sets of analyses and principal areas Carleton, 1993), others have included it (Braun, of congruence will be highlighted here. Selected 1993; Reig, 1986). I agree with Voss and Carleton nodes are labeled on Figure 23 for references in (1993) on moving Pseudoryzomys to the Oryzo- the text. myini. Other oryzomyines (e.g., Microakodonto- This analysis results in 22 equally most-parsi- mys, Oligoryzomys sp., Hershkovitz, 1993) exhib- monious trees, each 763 steps long, with a CI of it reduction and loss of the mesoloph(-id) in 0.22 and an RI of 0.55 (Fig. 22, strict consensus). association with the transition from forest to grass- Excluding the outgroups, the trees are 525 steps = = land and scrub communities (Hershkovitz, 1993). long, CI 0.29, RI 0.57. Both sets of CIs are Braun (1993) concluded from a cladistic analysis at the mean of the observed range of values for that Pseudoryzomys is either the basal phyllotine this many taxa (Archie, 1 989) or somewhat below or the sister taxon to the phyllotines. Her study the "expected" value of 0.34 for 47 taxa (Sander- included two oryzomyine and one thomasomyine son & Donoghue, 1 989). In each of the 1 2 1 equally species in the outgroups, but these were used to most-parsimonious trees that do not include Pu- define ancestral states and were not included in nomys within the phyllotines (Fig. 23, 80% ma- the actual numerical analysis. Thus, intentionally jority rule consensus), the respective values are = = or not, Pseudoryzomys was assumed a priori to be 764 steps long, CI 0.22, RI 0.55 overall, and = = a member of a clade that included the phyllotines 498 steps long, CI 0.30, RI 0.58 excluding and the other outgroups, two species of Akodon. outgroups. Both sets of trees are of equal length With only two closely related Akodon remaining when Punomys is pruned from the trees. Based on as outgroups to the phyllotines, the analysis could the arguments from the sigmodontine analysis and not test the tribal relationships of Pseudoryzomys in Steppan (1993), the alternative hypothesis of a or the monophyly of the phyllotines. This analysis nonphyllotine Punomys will form the basis for the demonstrates that Pseudoryzomys is unrelated to remaining analyses and discussion, as well as cal- the phyllotines: at least 10 additional steps are culating CIs and RIs for individual characters (Ta- required to place it as a basal phyllotine or among ble 8). the tetralophodont tribal-group. The differences between the preferred hypothesis from this analysis and Steppan (1993, Fig. 1) are minor and limited to several collapsed branches due to the greater number of trees in this anal- Phylogenetic Relationships within ysis and the exclusion of two taxa {species nova, Phyllotini northern populations of Andinomys edax) included in the earlier study. The major difference be- Results tween the two studies is greater character support in this analysis as reflected in higher average boot- As should be expected, the results of this anal- strap percentages (46% in Fig. 24 versus 41% in ysis conform closely to those produced earlier Steppan [1993, Fig 3]).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 63 Calomys callosus Calomys hummelincki Calomys laucha Calomys lepidus Calomys sorellus Andalgalomys pearsoni Graomys griseoflavus Graomys domorum Eligmodontia morgani Phyllotis gerbillus Phyllotis amicus Phyllotis darwini Phyllotis xanthopygus xanthopygus Phyllotis xanthopygus rupestris Phyllotis caprinus Phyllotis magister Phyllotis osilae Phyllotis haggard! Phyllotis definitus Phyllotis andium Phyllotis wolffsohni Chinchillula sahamae Galenomys garleppi Auliscomys boliviensis Auliscomys sublimis Auliscomys pictus Euneomys chinchilloides Euneomys petersoni Reithrodon typicus Reithrodon auritus pachycephalus Reithrodon auritus evae Neotomys ebriosus Loxodontomys micropus Irenomys tarsalis Andinomys edax Punomys lemminus Akodon albiventer Akodon boliviensis Chroeomys an din us Oxymycterus hispidus Ichthyomys hydrobates Scapteromys turn idus Holochilus brasiliensis Pseudoryzomys simplex Zygodontomys brevicauda Nectomys squamipes Thomasomys baeops

Fig. 22. Strict consensus cladogram of 22 equally most-parsimonious trees for the Phyllotini (763 steps long = = overall, CI 0.22, RI 0.55). Thomasomys and the four oryzomyine taxa (Holochilus, Nectomys, Pseudoryzomys, Zygodontomys) were designated as outgroups. Punomys is placed among the phyllotines with Andinomys and Ir- enomys.

64 FIELDIANA: ZOOLOGY Calomys callosus Calomys hummelincki Calomys laucha Calomys lepidus Calomys sorellus Andalgalomys pearsoni Graomys griseoflavus Graomys domorum Eligmodontia morgan! Phyllotis gerbillus Phyllotis amicus Phyllotis darwini Phyllotis xanthopygus xanthopygus Phyllotis xanthopygus rupestris Phyllotis caprinus Phyllotis magister Phyllotis osilae Phyllotis haggard! Phyllotis definitus Phyllotis andium Galenomys garleppi Auliscomys boliviensis Auliscomys sublimis Auliscomys pictus Euneomys chinchilloides Euneomys petersoni Reithrodon typicus Reithrodon auritus pachycephalus Reithrodon auritus evae Neotomys ebriosus Loxodontomys micropus Irenomys tarsalis Andinomys edax Chinchillula sahamae Phyllotis wolffsohni Punomys lemminus Akodon albiventer Akodon boliviensis Chroeomys andinus Oxymycterus hispidus Scapteromys turn idus Ichthyomys hydrobates Holochilus brasiliensis Pseudoryzomys simplex Zygodontomys brevicauda Nectomys squamipes Thomasomys baeops

Fig. 23. Eighty percent majority-rule consensus tree of 121 equally most-parsimonious trees wherein Punomys is not a phyllotine. Each tree is one step longer than the most-parsimonious overall (764 steps long overall; excluding = = outgroups, each tree is 498 steps long, CI 0.30, RI 0.58). This topology represents the preferred hypothesis of phylogenetic relationships among the phyllotines. Labeled nodes are referred to in the text.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 65 Calomys callosus r—74 43 Calomys hummelincki — 32-1 t_ Calomys laucha | 26- I— Calomys lepidus Calomys sorellus Andalgalomys pearsoni 84J Graomys griseoflavus Graomys domorum 38 r"C Eligmodontia morgani Phyllotis amicus ,—24. 57 Phyllotis gerbillus

Phyllotis andium

Phyllotis darwini 41 2 Phyllotis caprinus •— 21- 1—20'

I .— r C— Phyllotis xanthopygus xanthopygus 68-^ Phyllotis xanthopygus rupestris definitus i&T Phyllotis I 16 — Phyllotis magister

Phyllotis osilae Phyllotis haggard! Galenomys garleppi Auliscomys boIMensis ,£84-i Auliscomys sublimis 94- | Auliscomys pictus 1—12- Euneomys chinchilloides 1—39- 9ft ] Euneomys petersoni

84. Reithrodon typicus -10O| Reithrodon auritus pachycephalus 83H I—76 1—35' Reithrodon auritus evae Neotomys ebriosus Loxodontomys micropus 5- Irenomys tarsalis 30 i— 31 Andinomys edax r£ Chinchillula sahamae Phyllotis wolffsohni Punomys lemminus Akodon albiventer 1—53- 64- boliviensis 35 Akodon 2 c Oxymycterus hispidus 20-r tI— Chroeomys andinus Scapteromys tumidus Ichthyomys hydrobates Nectomys squamipes .ssT Holochilus brasiliensis •c Pseudoryzomys simplex 5{ Zygodontomys brevicauda Thomasomys baeops

Fig. 24. Majority-rule bootstrap consensus tree (200 replicates) for the Phyllotini. Numbers indicate percentage of replicates containing the specified clades.

66 FIELDIANA: ZOOLOGY Calomys is paraphyletic in the majority of the dicate that despite low bootstrap percentages, most-parsimonious trees with C. sorellus as the Phyllotis sensu stricto (node D) is moderately sta- sister taxon to the remaining phyllotines. The stable. Phyllotis sensu stricto is supported by a mod- tus of Calomys sorellus and C. lepidus is unre- erate to large distal baculum (93) and a series of solved. Support for paraphyly (or monophyly) of homoplasious characters. Calomys is weak: the bootstrap tree (Fig. 24) in- The highest bootstrap percentage for any Phyl- dicates a monophyletic Calomys but in only 26% lotis node that is also found in Figure 23 is 68%, of bootstrap replicates. The placement of C. so- for the clade consisting of P. darwini, P. caprinus, rellus with the remaining phyllotines is supported and the two subspecies of P. xanthopygus. Char- by a ventral pair of preputial glands (97), loss of acter support for this xanthopygus species-group the parastyle/anteroflexus Ml (12), more than 25 is principally provided by three phallic characters: caudal vertebrae (80), and long interparietal (56). hooks on the lateral mounds (95), dorsal knobs on The basal nodes of the remaining phyllotines the lateral mounds (96), and a large distal baculum are unresolved in the preferred hypothesis (Fig. relative to proximal baculum (93). The bootstrap 23) but are resolved in the analysis with Punomys percentage is 48% for the clade of P. magister and in the ingroup (Fig. 22) and in the previous anal- P. definitus, two very restricted and geographically ysis (Steppan, 1993). All analyses reveal four clades distant taxa that had been considered conspecific among the taxa terminal from Calomys. These by Pearson (1958), but P. magister is placed as the include a PhyllotislGraomys clade (node B), an sister-species to the xanthopygus species-group in Auliscomys group (node E), a clade including the most-parsimonious trees (Figs. 22, 23). Spe- Reithrodon and Loxodontomys (node F), and the cific character support for sister-species status is sister taxa Andinomys and Irenomys. The last three weak but includes nasals slightly broader than = clades form part of the sister-group to the Phyl- minimum interorbital distance (48, CI 0.17 lotisl Graomys clade in Figure 22. This clade, in- overall, but character state unique within the Phyl-

is lotis streaks 1 = 0. 1 cluding Reithrodon and Auliscomys, much more clade) and pectoral (9 , CI 7). highly differentiated, as reflected in the greater ge- The Graomys/Andalgalomys clade is placed as neric diversity as currently recognized (nine genera the sister-group to Eligmodontia (node C), though versus four). this grouping is not as well supported as the Gra- The bootstrap consensus tree joins Phyllotis sen- omys/Andalgalomys clade in the bootstrap tree. su stricto with the clade containing Graomys and Phyllotis gerbillus and P. amicus next join succes- Eligmodontia, but this node is not fully resolved sively to this group in the shortest trees (Fig. 23) in the consensus of the most-parsimonious trees or as a sister clade in the bootstrap consensus tree (Fig. 23, above node B). In any case, Phyllotis is (Fig. 24). This more inclusive clade is supported shown to be polyphyletic. In addition to the core by a posteriorly divergent supraorbital (50) and Phyllotis clade (node D), other species currently premaxillaries protruded well anterior to the in- included in Phyllotis are basal members of clades cisive plane (37). with Eligmodontia and Graomys (P. amicus, P. The clade consisting of Graomys and Andal- gerbillus), Auliscomys and Reithrodon groups (P. galomys is supported by orbital wings of the pre- wolffsohni), and a more inclusive clade (node B, sphenoid posterior to maximum constriction of P. andium). Bootstrap percentages for all these the presphenoid (69), a small but distinct zygo- groupings are low, however, as Phyllotis occupies matic spine (43), a sharply ridged, overhanging the most poorly resolved region of the tree. Eight supraorbital region (50), a laterally apressed sta- additional steps are needed for a monophyletic pedial spine (59), and 12 ribs (79). Paraphyly of Phyllotis. Furthermore, five additional steps are Graomys with respect to Andalgalomys is strongly needed just to bring andium and wolffsohni into indicated as G. griseoflavus and A. pearsoni appear a Phyllotis clade while still leaving amicus and as sister-species. This pairing is found in 83% of gerbillus at the base of the Graomys clade. Char- bootstrap replicates and is supported by no hy- acters supporting the inclusion of P. wolffsohni at poflexus reduction M3 (23), no posterior shift of the base of the Reithrodon and Auliscomys groups mesoflexus M3 (25), fusion of opposing flexi M3 include a nonlinear, "Y"- or "comma"-shaped fis- (31), and flattening of hamular process (60). Four sure in the upper incisors (3), premaxillaries ter- additional steps are needed for a monophyletic minating behind the anterior edge of the incisors Graomys, as the sister taxon to Andalgalomys. (37), and subequal mesopterygoid and parapter- Auliscomys pictus, A. sublimis, A. boliviensis, and ygoid fossae widths (64). Sensitivity analyses in- Galenomys together compose the Auliscomys group

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 67 Table 8. Consistency and retention indexes for phyllotine characters.

Number Character name CF RP

1. Incisor grooves 0.444 2. Incisor procumbency 3 Upper incisor dentine fissure 4. Labial root of M 1

5. Labial root of M 1 6. Molar roots of M3

7. Labial root of m 1 8. Molar roots of m2 9. Molar roots of m3 10. Anteromedian flexus Ml

11. Mesostyle M 1 12. Parastyle/anteroflexus M 1 13. Flexus penetration M 1 14. Anterolabial cingulum m 1 15. Protoflexid m 1 16. Cusp arrangement m 1 17. Anteromedian flexid ml 18. Procingulum separation ml 19. Posterolophid/stylid m 1 20. Posterolophid/stylid m3 21. Procingulum M2 22. Procingulum m2 23. Hypoflexus reduction M3 24. Reduction of mesoflexus M3 25. Posterior shift of mesoflexus M3 26. Hypoflexus lake M3 27. Rotation of flexus axes M3 28. Mesoflexid reduction m3 29. Anterior shift of mesoflexid m3 30. Posterior shift of hypoflexid m3 31. Fusion of opposing flexi in M3 32. Ratio of M3 length to alveolar length of molar tooth row 33. Capsular projection of mandible 34. Height of coronoid process 35. Anterior masseteric ridge position 36. Medioventral process of mandibular ramus 37. Premaxillary protrusion 38. Posterior extent of incisive foramina 39. Maxillary septum of incisive foramina 40. Orientation of incisive foramina 41. Dorsoventral position of anterior root of zygomata 42. Posterior margin of zygomatic plate 43. Development of zygomatic spine 44. Inclination of zygomatic plate 45. Premaxillo-maxillary suture orientation 46. Posterior extension of nasals 47. Posterior extension of premaxillaries 48. Nasal width 49. Interorbital shape 50. Supraorbital edge 51. Supraorbital ridge 52. Supraorbital knobs 53. Mediodorsal fusion of frontals 54. Shape of frontoparietal suture 55. Angle of frontoparietal suture 56. Medial length of interparietal/parietal 57. Orientation of anterior border of auditory bulla 58. Tegmen tympani 59. Shape of stapedial spine of auditory bulla 60. Thickness of hamular process of squamosal 61. Positions of temporal vacuities Table 8. Continued.

Number Character name CI" RP

62. Internal carotid canal 0.250 63. Extension of eustachian tube 64. Relative width of mesopterygoid fossa 65. Parapterygoid shape 66. Shape of mesopterygoid fossa 67. Parapterygoid fossa depth 68. Sphenopalatine vacuities 69. Position of orbital wings of presphenoid 70. Anterior border of mesopterygoid fossa 71. Medial process of posterior palate 72. Posterior palatine ridge 73. Posterolateral palatal pits 74. Orientation of maxillary tooth rows 75. Sphenopalatine foramen 76. Carotid circulation 77. Squamosal fenestra 78. Alisphenoid strut 79. Number of thoracic rib pairs 80. Number of caudal vertebrae 81. Neural spine of T2 82. Height of neural spine of C2 83. Length of neural spine of C2 84. Position of deltoid tuberosity 85. Ventral surface of claws (manus) 86. Length of Dl relative to D5 (pes) 87. Position of hypothenar pad 88. Furring of soles of feet (pes) 89. Countershading of tail 90. Furring of tail dorsum 91. Body pelage pattern 92. Pectoral streaks 93. Distal/proximal bacular length 94. Length of lateral mounds relative to medial mound 95. Hooks on lateral mounds 96. Knob on dorsal surface of lateral mounds 97. Preputial glands 98. Gallbladder Less stable is the position of Chinchillula. Sup- ment regarding suprageneric relationships, and that port for its placement at the base of the Auliscomys discordance may be due in part to the strictly cla- group (Fig. 22) comes from the anterior border of distic approach of this study in contrast to the the zygomatic plate rounded or receding dorsally evolutionary systematics of most prior studies. (43) and premaxillaries terminating behind the an- Some of the earlier studies (e.g., Hooper & Musser, terior plane of the incisors (37). In the alternative 1964) make pairwise statements of similarity that hypothesis found in some of the most-parsimo- are difficult to translate into a hierarchical phy- nious trees comprising Figure 23 and in the boot- logenetic hypothesis, and others implicitly rely on strap consensus tree (Fig. 24), Chinchillula is basal paraphyletic groups. In his revision of the genus to the Andinomys and Irenomys clade. Phyllotis, Pearson (1958) found consensus with The best supported of the generic-groups is the Ellerman (1941) and Osgood (1947) and recog- Reithrodon group (node F), previously defined by nized four subgenera — Graomys, Auliscomys, Olds and Anderson (1989), consisting of Reith- Loxodontomys, and Phyllotis— and removed P. rodon, Neotomys, and Euneomys. Inclusion of gerbillus to the related genus Paralomys. The phy- Loxodontomys with the Reithrodon group is sup- logenetic relationships implied by placing these ported by a relatively parallel-sided parapterygoid subgenera under Phyllotis is consistent with this fossa (65), a tripartite fissure in the upper incisors study in regard to Graomys being closely related

(3), and a narrow mesopterygoid fossa (64) and is to Phyllotis but is incongruent in regard to Aulis- found in 35% of bootstrap replicates. comys and Loxodontomys, which this study shows The Reithrodon group is supported by a sharply to be more closely related to other genera. Pearson angled premaxillo-maxillary suture (45, unique (1958) also did not recognize Eligmodontia as part within the Sigmodontinae), sigmoidal molars, sen- of a Phyllotis group. su Hershkovitz (1955) (represented here by mul- Hershkovitz (1962) revised the phyllotines and tiple characters), no anterior shift by the meso- recognized a Calomys section, which could be a flexid m3 (29), distinctly grooved incisors (1), clade or a grade, and a Phyllotis section, which anterior root of the zygomata inserting high, close should translate as a clade. The Calomys section to dorsal surface of the rostrum (41), absence of was primarily distinguished from the Phyllotis sec- labial root ml (7), and supraorbital knobs (52). tion by crested (bunodont) rather than flat or ter- The close relationship of Reithrodon with Neoto- raced molars. Zygodontomys from his Calomys mys is supported by loss of supraorbital branch of section and Pseudoryzomys from his Phyllotis sec- the stapedial artery (76), a deeply channeled pos- tion have since been removed from the phyllo- terior palate with distinct median ridge (72), an- tines. The remainder of his Calomys section con- terior apexes of incisive foramina well separated sists of Calomys and Eligmodontia. Like Pearson (40), and deeply excavated parapterygoid fossae (1958), he included Auliscomys and Graomys (67). within the genus Phyllotis and indicated that Eu- All the analyses place two genera not generally neomys along with the sigmodonts (Reithrodon, recognized by previous workers as closely related: Neotomys, Holochilus, and Sigmodori) might be Andinomys and Irenomys. Their grouping together considered the sister-groups to the phyllotines. is supported by an anterior masseteric ridge below Species-group assignments are more similar be- = and well posterior to the diastema (35, CI 0.30 tween Hershkovitz ( 1 962) and this study; for ex- overall, but character state unique among the phyl- ample, Hershkovitz's P. darwini complex is al- lotines), anterior apexes of incisive foramina rel- most identical to the P. xanthopygus-magister atively widely separated (40), frontals incomplete- species-group seen in this study.

ly fused or apparently vascularized along the Pearson and Patton ( 1 976) and Spotorno ( 1 986) midline (53), and posterolateral palatal pits in the have diagrammed hypotheses of evolutionary re- anterior parapterygoid fossa (73). The bootstrap lationships based on karyotypic data. Species that consensus tree (Fig. 24) groups the taxa in 30% of share the same diploid and fundamental numbers replicates. are generally also found by this analysis to be closely related— for example, A. pictus with A. sublimis, Discussion and P. xanthopygus and P. darwini with P. caprinus. However, P. amicus and P. magister also share Comparisons between the results of this study the same karyotypic formula but are morpholog- and previous studies show only moderate agree- ically quite distinct. Similarly, P. haggardi and P.

70 FIELDIANA: ZOOLOGY gerbillus share their karyotypic formulas with the Braun (1993) recently reported results of phe- xanthopygus species-group. Higher-order rela- netic and cladistic analyses of the phyllotines based tionships show less comparability across the two on 36 craniodental and 10 external characters, re- data sets. For example, Spotorno (1986) placed corded as 39 qualitative and 7 quantitative char- Andinomys at the base of the phyllotine radiation, acters. Qualitative and quantitative characters were while the karyotypes of Reithrodon, Euneomys, equally influential in the analyses because of the and Neotomys are as diverse as those of the phyl- large number of character states in all the quan- lotines as a whole and give no indication of close titative characters (8-10 each). Her cladogram relationship. In fact, Spotorno (1986; p. 22) ex- shows some similarities to mine, although the ro- plicitly acknowledged that gross karyotype is a poor bustness of her cladistic results are unknown due estimator of homology, concluding from G-band- to software limitations and the procedures used ing patterns that the close similarity of the P. xan- and because confidence estimates (e.g., bootstrap thopygus and Euneomys karyotypes "represent[s] values or additional steps required to break up independent acquisitions within each taxon." Spo- clades) were not reported. Character support for torno (1986) also screened electrophoretic alleles. clades was not generally reported either. A prin- His UPGMA dendrogram (Fig. 5 A) separates An- cipal conclusion was that Pseudoryzomys was the dinomys, Irenomys, and Euneomys from Reith- sister taxon to the phyllotines and, thus, may be rodon and Loxodontomys by placing them near the basal phyllotine. By only including the phyl- the base of the tree. A cladistic reanalysis of the lotines, two akodonts, and Pseudoryzomys in the same data set (Fig. 5B) provides little resolution. actual numerical analysis without any oryzo- The electrophoretic data set in Spotorno (1986) myines, the tribal status of Pseudoryzomys could does not seem to be highly informative for the not be tested. The results of this study indicate phyllotines. that Pseudoryzomys is not a phyllotine, nor is it Reig (1986) presented a biogeographic scenario within a clade that includes the phyllotines and for the diversification of the phyllotines and other akodontines. sigmodontine groups. His scenario drew upon mo- From the results of these analyses, I propose lar morphology and its dietary correlates, ecology, several generic groups to provide an informal ref- karyology, biogeography, and the limited fossil ev- erence structure to communicate some of the bet- idence. Paraphrasing in cladistic terminology, Reig ter supported hypotheses. The informal nomen- (1986) visualized Calomys as the most basal and clature reflects the uncertainty regarding key nodes, generalized phyllotine genus, with C. sorellus as some more inclusive than the generic-groups I pro- the most basal member of either a Calomys or pose, that would allow a proper cladistic allocation phyllotines-minus-Ca/omysclade. His view of the of monophyletic groups to formal subtribes. The lowland Calomys (e.g., C callosus, C. laucha) as groups I recognize are the Graomys group (node derived or terminal species is consistent with this C) including Andalgalomys, Graomys, and Elig- study. Phyllotis and a Neotomys-Sigmodon-Holo- modontia; the Auliscomys group (node E) includ- chilus complex constitute the basal members ing Auliscomys and Galenomys; the Reithrodon among the remaining phyllotines and evolved in group (node F) including Euneomys, Neotomys, the central and southern altiplano. Auliscomys, and Reithrodon; and the Andinomys group in- Galenomys, and the sister taxa Chinchillula and cluding Andinomys and Irenomys. The species Andinomys are then hypothesized to be indepen- amicus and gerbillus, currently assigned to Phyl- dently evolved from a highly paraphyletic Phyl- lotis, are most likely basal members of the Gra- lotis. In sharp contrast to the results of this study, omys group and should be assigned to it provided Reig (1986) hypothesized that Graomys and future studies confirm that hypothesis. Loxodon- Auliscomys are sister taxa. Andalgalomys, Pseu- tomys may well be a basal member of the Reith- doryzomys, and Eligmodontia would be indepen- rodon group, but this study does not definitively dently derived from Calomys. Finally, Loxodonto- demonstrate it. Chinchillula currently stands as mys and Euneomys are independent southern Phyllotini sedis mutablis for nearly equally par- offshoots ofa paraphyletic Auliscomys. Thus, Gra- simonious placements at the base of the Aulisco- omys would be closely related to Euneomys and mys or Andinomys groups. Phyllotis as currently unrelated to Andalgalomys, while Reithrodon, recognized is almost certainly paraphyletic, but the Neotomys, and Euneomys would be unrelated to relevant region of the phylogeny is not sufficiently each other. resolved that I can confidently propose an alter-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 71 native hypothesis with which to formally revise er than mesopterygoid fossa, very open spheno- the taxonomy. Finally, more confidence can now palatine vacuities, complete loss of the mesoloph, be placed in my earlier suggestion (Steppan, 1993) posterior extensions of premaxillaries and nasals that, due to the paraphyly of Graomys, Andalga- subequal, and (except in some Calomys) two pairs lomys should be subsumed within it. of preputial glands. I will also suggest, but not diagnose, a more formal taxonomy, primarily as a hypothesis to be tested by further studies and as a possible aid to Calomys Waterhouse description of the phylogenetic hypotheses pre- (Figs. 25, 26) sented here. The most appropriate taxonomic category between genus and tribe would be the sub- Calomys Waterhouse, 1837. Proc. Zool. Soc. Lond., tribe. To maximize information content and allow 1837:21. in the use of both subtribes and generic groups Hesperomys Waterhouse, 1839. Zoology Voy. "Bea- future taxonomies, recognized subtribes should be gle," 2:75. more inclusive than the generic groups that I just listed. In this scenario, the clade comprised min- Type Species— Mus (Calomys) bimaculatus imally of Phyllotis sensu stricto and the Graomys (Waterhouse), 1837, by original designation. group would be subtribe Phyllotina (node B). The Included Species— The specific taxonomy of sister-group to Phyllotina would be Reithrodonina Calomys is less stable than the other phyllotine and is expected to include the Auliscomys, Andi- genera. Species have undergone numerous reas- nomys, and Reithrodon groups. Finally, the sister- sessments in recent years, principally due to the group to the clade combining Phyllotina and input of karyotypic data. The nomenclature used Reithrodonina would be Calomyina, including only in the phylogenetic analysis largely followed Calomys, unless C. sorellus were in fact the sister- Hershkovitz (1962), and nomenclatural uncer- group to Phyllotina-plus-Reithrodonina, in which tainties of certain specimens are indicated in the case sorellus would have to receive its own genus Appendix. In her unpublished revision of the ge- and possibly subtribe. nus, Olds (1988) recognized 10 species: bimaculatus Waterhouse, 1837, callosus Rengger, 1830, hummelincki Husson, 1960, laucha Olfers, 1818, lepidus Thomas, 1884, murillus Thomas, 1916, Taxonomy musculinus Thomas, 1913, sorellus Thomas, 1 900, tener Winge, 1888, and venustus Thomas, 1894.

Numbers in parentheses listed in the diagnoses Vitullo et al. ( 1 990) did not recognize bimaculatus, refer to the character number from the phyloge- murillus, or tener but did recognize fecundus netic analysis of the phyllotines (Table 4). All di- Thomas, 1926, and callidus Thomas, 1916. In agnoses are phylogenetic in nature, in that they contrast to Olds (1988), Musser and Carleton list character states hypothesized to be derived (1993) recognized nine species, among them cal- relative to a named, more inclusive taxon. Asterisk lidus. They treated fecundus as a subjective syn- (*) signifies autapomorphy relative to the other onym of boliviae Thomas, 1901, and synonymized phyllotines. bimaculatus and murillus with laucha and musculinus, respectively. Whatever their rank, these taxa are properly included in Calomys, with res- Tribe Phyllotini Vorontsov, 1959 ervations regarding sorellus as discussed under Comments. Type Genus— Phyllotis, by tautonomy. Diagnosis— Members ofthe tribe Phyllotini de- Included TAXA—Andinomys, Calomys, Chin- scended from a common ancestor with the follow- chillula, Eligmodontia, Euneomys, Galenomys, ing traits: no reduction in the mesoflexus M3 rel- Graomys, Irenomys, Loxodontomys, Neotomys, ative to M2 (24), posterior margin of zygomatic Phyllotis, and Reithrodon. plate anterior to M 1 (42), mesopterygoid fossa ly- Diagnosis— Members of the Neotropical sub- ing more than 1 tooth-length posterior to M3 (70), family Sigmodontinae (family Muridae) descend- and length of distal baculum 63-77% of proximal ed from a common ancestor with the following length (93). traits: moderate or large pinnae (> 0. 1 5 head and Distribution— From central Peru at high ele- body length), parapterygoid fossa relatively broad- vations, through Bolivia, northern and central Ar-

72 FIELDIANA: ZOOLOGY = Fig. 25. Cranium and mandible of Calomys laucha (fnmh 23405). Scale bar 10 mm.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 73 Fig. 26. Upper and lower molars of Calomys laucha (fmnh 29246).

= gentina, far northern Chile, Paraguay, Uruguay, to FN 64). Group II consists of a Robertsonian = = southwestern and eastern Brazil. Calomys hum- series from venustus (2N 54-56, FN 66) to = = = melincki has a poorly known distribution in north- fecundus (2N 50, FN 66), and callidus (2N = ern Venezuela, eastern Colombia, and several off- 48, FN 66). Group III consists of callosus (2N = = = = shore Caribbean islands. 36, FN 48), lepidus (2N 36, FN 68), and = = Comments— Phylogenetic relationships within musculinus (2N 38, FN 56). Number of ribs Calomys are poorly known. Olds (1988) revised is consistent with Olds's scheme, as specimens as- the genus and recognized five species-groups, two signed by her to callosus, venustus, and tener all of which are monospecific, of unknown interre- share the derived condition of 1 2 ribs and seven lationship. Her groupings were based on tail length, lumbar vertebrae. Calomys laucha and lepidus have diploid number, and multivariate size and shape the primitive 13 ribs and six lumbar vertebrae. of skulls. The five groups (with modal mammae This observation contradicts the statement that counts) are as follows: callosus (10), venustus (10- "C callosus, C. laucha, and C. lepidus generally

14), and tener (10); laucha (8), bimaculatus (10), [have] ... 1 3 thoracic, with 1 3 pairs of ribs, 6-7 and hummelincki (8); murillus (10-12) and mus- lumbar" (Olds, 1988, p. 50). Many of the same culinus ( 1 2-14); lepidus (8); and sorellus (8). These specimens were examined in both studies. This groupings bear little resemblance to the three rec- observation also contradicts the coding by Carle- ognized by Vitullo et al. (1990) based on karyo- ton (1980), who coded callosus as having 13 ribs. types. Group I is closely related to the ancestral Nineteen of the 20 skeletons of callosus examined type hypothesized by Pearson and Patton (1976) in this study clearly had only 1 2 ribs, and one had = = and includes sorellus (2N 64, FN 68), laucha a thin, short thirteenth pair that did not articulate = = = (2N 64, FN 68), and hummelincki (2N 60, with the twelfth thoracic vertebra.

74 FIELDIANA: ZOOLOGY The most-parsimonious trees from the phylo- Included Species— At least three species have genetic analysis (Figs. 22, 23; Steppan, 1993) in- been demonstrated from karyotypic data: typus,

dicate that sorellus is not a member of a Calomys puerulus Philippi, 1 896, and morgani Allen, 1 90 1 . clade but may instead be the sister taxon to all Diagnosis— Members of the Graomys group other phyllotines. The analysis above is equivocal, descended from a common ancestor with the fol-

but the hypothesis that sorellus shares a more re- lowing traits: hyper-opisthodont upper incisors (2),

cent ancestor with other phyllotines than it does anterolabial cingulum ml indistinct ( 1 4), posterior with Calomys is supported by the presence of a shift of hypoflexid m3 relative to m2 (30), knob ventral pair ofpreputials (97), more than 25 caudal of anterior masseteric ridge exceeds dorsal edge of vertebrae (80), loss of the parastyle/anteroflexus diastema (35), supraorbital edges angled (50), eu- Ml (12), a long interparietal (56), and two roots stachian tubes do not reach posterior edge of pter- on m3 (9). ygoid processes (63), parapterygoid fossa 1.5-2.5 Necromys Ameghino (1889) was synonymized times mesopterygoid fossa width (64), maxillary with C. callosus by Hershkovitz (1962), but the tooth rows posteriorly convergent (74), deltoid tu- type specimen was considered a senior synonym berosity greater than 59% of humerus length from of Bolomys by Massoia and Pardifias (1993). condyle (84), ventral surface of claws forming distinct keel (85), Dl and D5 ofpes subequal in length (86), sole of pes furred (88), *hind feet elongated, Graomys Group hypothenar pad absent, *fused plantar pads D2- 4, and hair in pectoral region entirely white from Included Taxa— Eligmodontia and Graomys. base to tip. Diagnosis— Members ofthe tribe Phyllotini de- Distribution— Found throughout southern Ar- scended from a common ancestor with the follow- gentina, far eastern Chile in the south, along the ing traits: no reduction in the mesoflexus of M3 eastern slopes of the Andes in western and north- relative to M2 (24), hypoflexus of M3 intact, not ern Argentina, and in the altiplano of northern pinched to form lake (26), premaxillaries pro- Chile, southwestern Bolivia, and extreme southern duced well anterior to incisors (37), supraorbital Peru. region posteriorly divergent (49), supraorbital edg- Comments— Three species are now recognized, es ridged, overhanging (except Eligmodontia) (50), based mostly on karyotypes: typus, puerulus, and maxillary tooth rows parallel (except Eligmodon- morgani. The specimens examined from Neuquen tia) (74), squamosal fenestra present between mas- Province in western Argentina and from southern ticatory-buccinator trough and squamosal-ali- Chile are tentatively referred to the species mor- sphenoid groove (77), alisphenoid strut present gani, the name offered by Kelt et al. (1991) for (78), and tail dorsum sparsely furred (except G. specimens with a karyotype of 2N = 32-33, FN pearsoni) (90). = 32. Kelt et al. (1991) and Ortells et al. (1989) Comments— This generic group may well in- also recognize two other species of Eligmodontia: clude Phyllotis amicus and P. gerbillus, but as the = puerulus (2N 50, FN = 48) from Bolivia, Peru, current analysis cannot confidently resolve the is- and northern Chile and typus (2N = 43—44, FN = sue, I prefer not to make formal nomenclatural 44) from central and eastern Argentina. Zambelli changes. Should further studies confirm the inclu- et al. ( 1 992) found typus to be sympatric with mor- sion of amicus and gerbillus, then they should be gani at two localities in Neuquen Province near included in the Graomys group as well. where some of the specimens examined in this study were collected. The recent chromosomal Eligmodontia F. Cuvier studies do not include morphological data that would allow more confident (Figs. 27, 28) species assignments for the specimens examined in this study.

Eligmodontia F. Cuvier, 1837. Ann. Sci. Nat. (Paris), Graomys Thomas ser. 2, 7:168. Eligmodon Wagner, 1841. Arch. Naturg., 1:125. (Figs. 28, 29) Heligmodontia Agasiz 1 846. Nomencl. Zool. Mamm., Addenda, 5:136, 175. Graomys Thomas, 1916. Ann. Mag. Nat. Hist., ser. 8, 17:141. Type Species— 1 Eligmodontia typus Cuvier, 837, Andalgalomys Williams and Mares, 1978. Ann. Car- by original designation. negie Mus., 47:197.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 75 Fig. 27. Cranium and mandible of Eligmodontia morgani (fmnh 133070).

76 FIELDIANA: ZOOLOGY A

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 77 Fig. 29. Cranium and mandible of Graomys griseoflavus (fmnh 28423).

78 FIELDIANA: ZOOLOGY Type Species— Mus (Phyllotis) griseo-flavus Diagnosis— See discussion under Comments. Waterhouse, 1837, by original designation. Distribution— As currently defined, the genus Included Species— Includes griseoflavus (Wa- is distributed throughout nonforest habitats in the terhouse, 1837), domorum (Thomas, 1902), olrogi Andes from Ecuador to the Straits of Magellan, (Williams & Mares, 1978), and pearsoni Myers, from near sea level along the Pacific coast, to the 1977. The status of the geographically restricted altiplano. edithae Thomas, 1 9 1 9, is less clear, and specimens Comments— Woodman (1993) has recently rec- were not examined. ommended that species names for all genera end- Diagnosis— Members of the Graomys group ing in the feminine -otis and derived from the descended from a common ancestor with the fol- Greek oroa, for ear, should be feminized. Four lowing traits: distinct and large anterolabial cin- species of Phyllotis would be affected by this no- gulumml (14, 15), noshiftofmesoflexid m3 (29), menclatural change: arnica from amicus, caprina moderately developed zygomatic spine, anterior from caprinus, definita from definitus, and xan- border of zygomatic plate concave (43), supraor- thopyga from xanthopygus. However, Pritchard bital sharply ridged, overhanging (50), *laterally (1994) argued that mammalian generic names apressed stapedial spine (59), *orbital wings of the ending in -otis are actually Latin derivations of presphenoid posterior from maximum constric- the Greek, and Latin words ending in -is are of tion (69), maxillary tooth rows parallel (74), 12 the 3rd declension and may be masculine, femi- thoracic and seven lumbar vertebrae (79), and nine, or neuter, depending on priority of usage moderately sized distal baculum (93). when grouped with specific names. The first usage Distribution— Found at moderate to low ele- of Phyllotis with a specific epithet requiring gender vations in southern Bolivia, western Paraguay, and agreement was with the masculine xanthopygus northern and central Argentina, and reported from (Waterhouse, 1837). At issue is whether the -otis southwestern Brazil. ending was a proper feminine form of otoo or was

Comments— Synonymy of A ndalgalomys is re- a latinization of the Greek. The Greek otis (otict) quired by the strong support for G. (Andalgalo- means bustard and, while feminine, refers to a bird mys) pearsoni and G. griseoflavus sharing a com- that lacks external pinnae, and thus may not have mon ancestor more recently than either does with originally been derived from "feminine-eared G. domorum. creature," a literal translation. It certainly serves as a poor reference for names relating to modifications of ears. I retain the historical usage for but resolution Phyllotis Waterhouse Phyllotis species names, proper may await further of the of the (Figs. 30, 31) investigation etymology original Greek forms. My analysis of morphological characters strong- Phyllotis Waterhouse, 1837. Proc. Zool. Soc. Lond., 1837:27. ly indicates that Phyllotis is either polyphyletic or Paralomys Thomas, 1926. Ann. Mag. Nat. Hist., ser. paraphyletic, but no single alternative hypothesis 17:315. 9, exists with which to confidently revise existing taxonomy. Therefore, the genus is left unrevised and Type Species: Mus darwini (Waterhouse), 1837, undiagnosed. It seems likely that amicus and ger- by subsequent designation (Thomas, 1 884, p. 449). billus should be removed; the name Paralomys Included Species: Minimally, Phyllotis sensu Thomas, 1926, is available for a genus including stricto includes caprinus Pearson, 1958, chilensis gerbillus. In fact, Braun ( 1 993) recently placed them Mann, 1 945, definitus Osgood, 1915, darwini (Wa- in Paralomys, although her analysis indicates that terhouse, 1837), haggardi Thomas, 1898, magister it is paraphyletic. Should amicus and gerbillus not Thomas, 1912, osgoodi Mann, 1945, osilaei. A. form a monophyletic group, then Paralomys would

1 1 Allen, 90 , and xanthopygus (Waterhouse, 1837). apply only to gerbillus. Phyllotis andium falls out- Additional species included pending further re- side the main Phyllotis clade and may well be a visionary studies (see Comments): andium Tho- basal member that requires generic status in order mas, 1912, wolffsohni Thomas, 1 902, amicus Tho- to maintain a strictly monophyletic Phyllotis. The mas, 1900, and gerbillus Thomas, 1900. Musser more highly derived wolffsohni seems to be a basal and Carleton (1993) also recognized bonariensis member of the clade including Auliscomys, An- Crespo, 1 964, which along with osgoodi was not dinomys, and Reithrodon. After excluding these examined for the cladistic analyses. four species, nine remain to form a monophyletic

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 79 Fig. 30. Cranium (fmnh 22325) and mandible (fmnh 22328) of Phyllotis darwini.

80 FIELDIANA: ZOOLOGY 5 o c

>> 5

a D

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 81 group in the most-parsimonious tree: caprinus, Distribution— Southern Andes of Chile and chilensis, darwini, definitus, haggardi, magister, Argentina from about 36°S to the Straits of Ma- osgoodi, osilae, and xanthopygus. Though not in- gellan. cluded in the cladistic analysis, chilensis and os- Diagnosis— Members of the tribe Phyllotini and goodi are clearly members of Phyllot is sensu stric- allied to the Reithrodon group, descended from a to, virtually identical to xanthopygus rupestris for common ancestor with the following traits: un-

the characters used in the cladistic analysis. grooved upper incisors (1), tripartite dentine lake in This study follows the recommendation of the upper incisors (3), two labial roots Ml (5),

Walker et al. (1984) to restrict the specific name two roots M3 (6), labial root ml present (7), two

darwini to coastal Chilean populations and to in- roots m2 (8), indistinct parastyle/anteroflexus M 1

clude the north Andean subspecies (e.g., rupestris) (12), anterolabial cingulum distinct (14), proto- within xanthopygus. Musser and Carleton (1993) flexid m2 present as groove (22), hypoflexus M3 cited Walker et al. (1984) in aligning the north reduced relative to M2 (23), mesoflexid m3 re- Andean subspecies under darwini and restricted duced and shifted anteriorly relative to m2 (28, xanthopygus to the south Andean subspecies (e.g., 29), length of M3 greater than 20% alveolar length vaccarum), but Walker et al. (1984) were quite of tooth row (32), posterior border of mandibular clear about the restriction of darwini to coastal symphysis sharply angled (36), premaxillaries ter- Chilean populations. The specimens referred to minating at or slightly anterior to incisive plane darwini in Braun (1993) were instead xanthopy- (37), incisive foramina terminating at level of gus. paracone and protocone (38), antorbital bridge

l My recognition of chilensis as a distinct species Vs- A below dorsal surface of rostrum (41), nasals differs from all recent treatments (Hershkovitz, broader than interorbital constriction (48), supra- 1962; Mann, 1978; Pearson, 1958). Unpublished orbital edges angled for Vi of length (50), supra- morphometric, anatomical, and molecular data orbital ridges and knobs absent (51, 52), internal collected by me indicate the specific separation of carotid bounded by both auditory bulla and oc- chilensis from P. x. rupestris in southern Peru. The cipital (62), mesopterygoid fossa distinctly nar- two taxa form clearly bimodal clusters in principal rower than parapterygoid fossa (64), posterior component space with little overlap. Where good width of parapterygoid less than 1.5 times anterior series exist along transects, the transition between width (65), parapterygoid fossa recessed slightly the taxa is sharp, identifiable to within a kilometer. above palate (67), maxillary tooth rows posteriorly, Morphological intermediates (outliers) are not as- divergent (74), sphenopalatine foramen absent or sociated with proximity to the transition zone. nearly ossified (75), squamosal fenestra present Phyllotis chilensis is autapomorphic in being the between squamosal groove and masticatory-buc- only phyllotine examined with three pairs of pre- cinator trough (77), tail sparsely furred (90), pec- putial glands. Additionally, Peruvian rupestris toral streaks present (92), and distal baculum less possesses the narrowest upper incisors in relation than 63% length of proximal baculum (93). to incisor depth of any Phyllotis. Spotorno and Comments— Osgood (1947) described Loxo- Walker (1983) also found that chilensis was elec- dontomys as a subgenus of Phyllotis, and Braun

trophoretically more closely related to darwini than ( 1 993) elevated it to a genus for micropus. Generic to populations of true xanthopygus. It is not en- status was supported by Steppan (1993). Inclusion tirely clear whether chilensis or rupestris shows within Auliscomys is strongly argued against by evidence of introgression with the other subspecies this study, but karyotypic data would seem to sug- of xanthopygus, but current DNA sequence and gest a close association (Walker & Spotorno, 1992). revisionary studies should resolve the question. However, the karyotypic analysis assumed rather than tested a monophyletic Auliscomys sensu lato Loxodontomys Osgood by not including outgroups in the analysis. The (Figs. 31, 32) basal position of Loxodontomys relative to the divergence of the Auliscomys and Reithrodon groups allows the possibility that Loxodontomys Loxodontomys Osgood, 1947. J. Mamm., 28:172. and Auliscomys have retained relatively primitive Type Species— Mus micropus Waterhouse, 1 837, karyotypes. Simonetti and Spotorno (1980) moved by original designation and monotypy. micropus from Phyllotis to Auliscomys because of Included Species— Includes the single species its similar karyotype and proximity to Auliscomys micropus. species in an ordination analysis. The karyotypes

82 FIELDIANA: ZOOLOGY Fig. 32. Cranium and mandible of Loxodontomys micropus (fmnh 23287).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 83 are indeed similar, but the multivariate analysis Type Species— Reithrodon pictus Thomas, 1 884, was based on only 4 external and 1 1 partially re- by original designation. dundant molar measurements. Additionally, mi- Included species—pictus (Thomas, 1884), bo- cropus was compared XoAuliscomys, Phyllotis, and liviensis (Waterhouse, 1 846), and sublimis (Thom- Andinomys, but to none of those taxa that this as, 1900). analysis indicates that it is related to. Their mul- Diagnosis— Members of the Auliscomys group tivariate analysis does not conflict with the results descended from a common ancestor with the fol- of this study. lowing traits: upper incisors with shallow grooves Osgood (1943) had included micropus in Aulis- or striae in most specimens (1), incisors orthodont

of but root 1 set comys, a subgenus Phyllotis, only supported (2), small labial of M , medially (4), pos- this alignment by the development of the parastyle terolophid ml distinct at all ages (19), procingul- M2 (which he described as variable among groups um M2 moderately to well developed (21), highly allied to Phyllotis) and the "somewhat more oblique reduced mesoflexid on m3 (28), posterior shift of pattern" (Osgood, 1 943, p. 2 1 3) of the molar lophs. hypoflexid on m3 (30), supraorbital region ante- Otherwise, he found micropus "somewhat anom- riorly divergent, narrowest point posterior (except alous" in association with Auliscomys. pictus) (49), interparietal/parietal length between Pine et al. (1979) proposed that Loxodontomys 0.33 and 0.45 (56), internal carotid canal bounded Osgood was technically a nomen nudum because by both auditory bulla and occipital (62), and max- in Osgood's description, "it" referred to micropus illary tooth rows parallel (74). rather than to Loxodontomys. However, in defin- Distribution— Altiplano from central Peru to ing Loxodontomys as a monotypic taxon, "it" re- southwestern Bolivia and far northern Argentina ferred to both the species and subgenus, and "it" and Chile. was compared to other genus-level taxa, not to Comments— Braun (1993) erected Maresomys individual species. to contain boliviensis primarily because it was less similar to sublimis and pictus than they were to Auliscomys Group each other and because her cladistic analysis placed it as the sister-species to Galenomys garleppi. No Included Taxa—Auliscomys and Galenomys. putative synapomorphies for this topology or es- Diagnosis— Members of the tribe Phyllotini al- timates of confidence were reported. My study does lied to Chinchillula, Loxodontomys, and the An- not support sister taxon status between boliviensis dinomys and Reithrodon groups, descended from and Galenomys but, instead, strongly supports a a common ancestor with the following traits: curved monophyletic Auliscomys (84% of bootstrap rep- dentine lake in the upper incisors (3), length of licates, hypothesized synapomorphies in diagno- M3 greater than 20% alveolar length of tooth row sis). Phenetic analyses in Braun (1993) and Spo- (32), capsular projection of mandible indistinct or torno (1986) indicate that dissimilarities among absent (33), medioventral process of mandibular boliviensis, sublimis, and pictus are comparable to ramus absent, the ramus not sharply angled (36), that seen within Calomys and Phyllotis. The cur- premaxillaries terminating behind the incisive rent taxonomy in regard to boliviensis, sublimis, plane (37), *zygomatic spine absent, zygomatic and pictus appears to fully satisfy the objectives plate convex and receding dorsally (43), zygomatic of both cladistic and phenetic concepts of system- plate inclined less than 20° (44), posterior terminus atics. of premaxillaries anterior to nasals (47), nasals The species micropus is removed to Loxodon- broader than interorbital constriction (48), meso- tomys (see Comments therein). pterygoid fossa distinctly narrower than parapterygoid fossa (64), and tails usually densely furred and short to less than head (90) very short, always Galenomys Thomas and body length. (Figs. 34, 35)

Galenomys Thomas, 1916. Ann. Mag. Nat. Hist., ser. Auliscomys Osgood 8, 17:143. (Figs 33, 34) Type Species— Phyllotis garleppi Thomas, 1 898, original designation. Auliscomys Osgood, 1915. Field Mus. Nat. Hist. Publ. by Zool. Ser., 10:190. Included Species— Includes the single species Braun 1993:40. Maresomys , garleppi (Thomas, 1916).

84 FIELDIANA: ZOOLOGY Fig. 33. Cranium and mandible of Auliscomys pictus (fmnh 64344).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 85 a D

86 FIELDIANA: ZOOLOGY Fig. 35. Cranium (amnh 246947) and mandible (fmnh 53845) of Galenomys garleppi.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 87 Diagnosis— Member of the Auliscomys group length of the molar tooth row (32); zygomatic plate descended from a common ancestor with the fol- rounded, spine absent (43); anterior root of zy- lowing traits: incisors orthodont (2); lower incisors gomata inserting near dorsal surface of rostrum highly procumbent; three roots on m3 (9); distinct (41); supraorbital ridges raised dorsally (51); fron- anteromedian flexus on Ml (10); reduced hypo- toparietal suture straight, forming a right or acute flexus on M3 (24); highly reduced mesoflexid on angle (54, 55); postglenoid foramen anterior to m3 (28); knob of anterior masseteric ridge reaches subsquamosal foramen (61); internal carotid not dorsal edge of mandible (35); *skull strongly arched bounded by basioccipital (62); anterior border of upward in region of supraorbital to rostrum, ros- mesopterygoid fossa lying between and 0.33 trum flexed downward; internal carotid canal tooth-length behind M3s (70); ventral surface of bounded by both auditory bulla and occipital (62); claws on manus fused, forming distinct keel (85); maxillary tooth rows posteriorly convergent (74), highly distinctive coloration, with white hip *outwardly bowed; no fenestra where masticato- patches contrasting with dark brown to black side ry-buccinator nerve passes over squamosal-ali- patches, tawny over dark gray back, and white sphenoid groove (77); deltoid tuberosity greater postauricular patches. than 59% along humerus from the condyle (84); Distribution— Altiplano of southern Peru, far soles of hindfeet furred (88); tail densely furred western Bolivia, and far northern Chile. (90); and *tail very short, less than 40% head and Comments— I regard Chinchillula as Phyllotini body length. sedis mutabilis to reflect the uncertainty as to the Distribution— Restricted distribution in the two near equally parsimonious positions it can provinces or departments of Puno in Peru, La Paz occupy: the basal member of the Auliscomys group and Oruro in Bolivia, and Tarapaca in Chile. or the basal member of the A ndinomys group. Nei- Comments— Braun (1993) erected Maresomys ther topology is strongly supported, but Chinchil- for boliviensis based on its sister-group status with lula's basal position near the divergence of these Galenomys, creating an otherwise paraphyletic two groups seems highly likely. Auliscomys. The strength of a Galenomys/Mare- somys clade relative to other topologies cannot be Andinomys Group evaluated because no synapomorphies are reported. This study indicates robust support for a Included Taxa—Andinomys and Irenomys. to which is monophyletic Auliscomys, Galenomys Diagnosis— Members of the tribe Phyllotini and the sister-group. allied to the Auliscomys and Reithrodon generic groups, descended from a common ancestor with the following traits: fissure of upper incisor curved

Chinchillula Thomas or tripartite (3); labial root M 1 small or absent (4); (Figs. 36, 37) two roots m2 (8); well-developed procingulum m2 (22); no reduction hypoflexus M3 (23); length of of tooth row ChinchillulaThomas, 1898. Ann. Mag. Nat. Hist., ser. M3 greater than 25% alveolar length 7, 1:280. (32); anterior masseteric ridge well ventral and posterior to dip in diastema (35); posterior border Type Species— Chinchillula sahamae Thomas, of mandibular ramus sharply angled (36); pre- 1898, by original designation. maxillaries at or slightly anterior to incisive plane Included Species— Includes the single species (37); separation of anterior apexes of incisive fo- sahamae. ramina greater than 80% that of posterior apexes Diagnosis— Member of the tribe Phyllotini and (40); nasals broader than interorbital constriction allied to the Auliscomys and Andinomys generic (48); incomplete fusion or vascularization of fron- groups, descended from a common ancestor with tals (53); frontoparietal sutures straight, forming the following traits: characterized by *very large acute or right angle (54, 55); internal carotid not body size; curved fissures in upper incisors (3); bounded by basioccipital (62); mesopterygoid fos- one root on M3 (6); *lingual root of M2 absent; sa and parapterygoid fossa subequally broad (64); three roots on m2 (8); small mesostyles present mesopterygoid fossa reaching M3s (70); posterior (11); opposing flexi on Ml do not overlap (13); palatal pits located subequal or posterior to an- posterolophid ml absent (19); procingulum M2 terior border of mesopterygoid fossa (73); and absent (21); length of M3 less than 0.205 times the maxillary tooth rows posteriorly divergent (74).

88 FIELDIANA: ZOOLOGY Fig. 36. Cranium (fmnh 52479) and mandible (fmnh 52478) of Chinchillula sahamae.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 89 n. 2

90 FIELDIANA: ZOOLOGY Andinomys Thomas Diagnosis— Members of the Andinomys group descended from a common ancestor with the fol- (Figs. 37, 38) lowing traits: incisors deeply grooved (1); labial Andinomys Thomas, 1902. Proc. Zool. Soc. Lond., root Ml absent (4); three roots on M3 (6); labial 1902(1):116. root of ml absent (7); "opposing flexi/flexids of all molars meet at midline, nearly severing mures/ Type Species—Andinomys edax Thomas, 1 902, murids (13); cusps opposite (16); posterolophid of by original designation. ml absent (19); procingulum of M2 absent (21); Included Species— Includes the single species height of the coronoid process subequal with man- edax. dibular condyle (34); incisive foramina extending Diagnosis— Members of the Andinomys group to protocone (38); hamular process of the squa- descended from a common ancestor with the fol- mosal reduced in thickness along entire length (60); lowing traits: molars opisthodont (2); labial root medial process of the posterior palate absent (71); of Ml present, small, set medially (4); primary alisphenoid strut present as a consistent dorsal cusps alternate (16); lophs sharply angled; anter- process that does not cross the foramen ovale (78); omedian Ilex id m 1 prominent ( 1 7); mesoflexid m3 more than 30 caudal vertebrae (80); deltoid tu- highly reduced (28); moderate zygomatic spine, berosity greater than 59% of the length of the hu- anterior border ofzygomatic plate weakly concave merus measured from the condyle (84); heels of (43); supraorbital region anteriorly divergent (49); the pes only sparsely furred (88); and tail mono- supraorbital swellings present (52); *fontanelle colored (89). formed by incomplete fusion of frontals (53); post- Distribution— Low to moderate elevations glenoid foramen anterior to subsquamosal fora- from Nuble Province in south central Chile to the

1 to base Guaitecas men (6 ); eustachian tubes extend anteriad Straits of Magellan; Chiloe Island and ofpterygoid processes (63); parapterygoid recessed Islands. slightly above level of bony palate (67); second Comments— Characters, distribution, and ecol- cervical neural spine enlarged into "plow"-shaped, ogy reviewed by Kelt (1993). distinct keel (82); hypothenar pad intermediate to first interdigital and thenar pads (87); and tail bicolored and densely furred (89, 90). Reithrodon Group Distribution— Found in the highlands of southern Peru, far northern Chile, southwestern Included Taxa— Includes the highly differen- Bolivia, and northern Argentina. tiated genera Euneomys, Neotomys, and Reithro- Comments— Previously, I had treated northern don. This group was first formally identified by and southern populations of A. edax edax as sep- Olds and Anderson (1989). arate OTUs for the cladistic analysis (Steppan, Diagnosis— Members of the tribe Phyllotini al- 1993) because of geographic differentiation for lied to Loxodontomys and the Auliscomys and An- qualitative characters. Subsequent examination of dinomys groups, descended from a common an- additional material has indicated that, although cestor with the following traits: distinct grooves the variation is significant, it is also complex, and on the upper incisors (1); tripartite dentine lake in the recognition of separate taxa would be pre- the upper incisors (3); *sigmoidal molars; three mature until a thorough generic revision has been roots on M3 (6); no labial root on m 1 (7); indistinct conducted. or weakly developed anterolabial cingulum ml (14); procingulum on m2 absent (22); no reduction or shift of m3 mesoflexid (28, 29); length of M3 than 20% alveolar of tooth row (32); Irenomys Thomas greater length height of the coronoid process subequal with the (Figs. 39, 40) mandibular condyle (34); posterior border ofman- dibular sharply angled or with distinct Irenomys Thomas, 1919. Ann. Mag. Nat. Hist., ser. symphysis anterior root 9, 3:201. process (36); strong zygomatic plate; of zygomata inserting near or on dorsal surface of Type Species— Reithrodon longicaudatus Phi- rostrum (41); *premaxillo-maxillary suture with it lies horizon- lippi, 1900, by original designation. acutely angled bend, so that nearly rostrum Included Species— Irenomys tarsalis (Philippi, tal as it passes under ventral surface of 1900). (45); interorbital region narrow; supraorbital

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 91 Fig. 38. Cranium (fmnh 29157) and mandible (fmnh 23435) of Andinomys edax.

92 FIELDIANA: ZOOLOGY Fig. 39. Cranium and mandible of Irenomys tarsalis (fmnh 133164).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 93 I 1 as:

94 FIELDIANA: ZOOLOGY ridges raised dorsally (51); supraorbital swellings most similar to that found in Reithrodon and to a present (52); temporal vacuities positioned anter- lesser extent Neotomys. Likewise, the obtuse ori- odorsally (61); mesopterygoid fossa distinctly nar- entation of the opposing lophids resembles Reith- rower than parapterygoid fossa (64); posterior rodon and Neotomys more than the parallel lo- width parapterygoid less than 1.5 times anterior phids in Euneomys. In fact, Pardinas (in press) width (65); medium to deep parapterygoid fossa reports that the type of Bothriomys is an immature (67); maxillary tooth rows posteriorly divergent individual, and synonymizes it with Graomys. (74); and fewer than 25 caudal vertebrae (80). Pearson and Christie (1991) provided compel- ling morphologic evidence for specific distinction Euneomys Coues between chinchilloides and mordax, fully consis- (Figs. 40, 41) tent with data from karyotypes and ordination of morphometric data (Reise & Gallardo, 1990). Euneomys Coues, 1874. Proc. Acad. Nat. Sci. Phila- Chelemyscus is based on a mismatched skin and 26:185. delphia, skeleton whose associated locality is also suspect IBothriomys Ameghino, 1889. Act. Acad. Nat. Cienc. (Hershkovitz, 1962; Pearson & Christie, 1991). Rep. Argentina, 6:1 18. The skull is the offossor, but it is unclear at Chelemyscus Thomas, 1925. Ann. Mag. Nat. Hist., type ser. 9, 15:584-585. this time to which species of Euneomys it belongs (Pearson & Christie, 1991). Musser and Carleton Type Species— Reithrodon chinchilloides Wa- (1 993) recognized petersoni as a species, in contrast terhouse, 1839, by original designation. to Hershkovitz (1962), Reise and Gallardo (1990), Included Species— Includes Euneomys chin- and Pearson and Christie (1991). However, Reise chilloides (Waterhouse, 1839), E. mordax Thom- and Gallardo's (1 990) results show petersoni to be as, 1 9 1 2, and E. petersoni J. A. Allen, 1 903. Status largely distinct from chinchilloides in multivariate of the fossil Bothriomys catenatus (Ameghino, analyses, with some minimal overlap. Musser and 1889) is uncertain (see Comments). Carleton (1993) correctly call for a rigorous revi- Diagnosis— Members of the Reithrodon group sion. descended from a common ancestor with the following traits: anterior incisor grooves (1); anterolabial cingulum absent (14); *procingulum sepa- Neotomys Thomas rated from anterior murid in ml fusion of by (Figs. 42, 43) metaflexid and protoflexid (18); posteroflexid in m3 present as shallow groove in juveniles, absent in adults anteroflexus shallow (20); M2 (21); hy- Neotomys Thomas, 1894. Ann. Mag. Nat. Hist., ser. poflexus pinched off to form dentine lake in M3 6, 14:346. (26); *dorsal surface ofzygomatic plate rising above dorsal surface of rostrum, anterior root of zygo- Type Species— Neotomys ebriosus Thomas, mata inserting at dorsal surface (41); zygomatic 1894, by original designation. plate inclined less than 20° from vertical (44); ali- Included Species— Includes the single species sphenoid strut present (78); *highly elongated neu- ebriosus. ral spine of third thoracic vertebra (81); and spine Diagnosis— Members of the Reithrodon group of second cervical vertebra overlaps third cervical descended from a common ancestor with the fol- vertebra (83). lowing traits: *anterior surface of upper incisors Distribution— Andes from central Chile and slightly concave; *upper incisor grooves forming Argentina to Tierra del Fuego. distinctly pinched fold on lateral margin (1); lower Comments— Hershkovitz (1962) lists Bothrio- incisors steeply angled; three roots on lower m2 mys, the type of which is the Pleistocene mandible (8); three roots on m3 (9); primary cusps alternate designated catenatus (Ameghino, 1889), as a syn- (16); anteromedian flexid lost (17); reduction of onym ofEuneomys. While the specimen does share mesoflexid in m3 (28); very large M3s, greater than with Eunomys the number and depth of the loph- 'A tooth row length (32); *mandible very deep and ids, as well as the diagnostic separation of the pro- robust; indistinct capsular projection of the man-

1 cingulum on m , the angles of the labial lophids dible (33); coronoid process below level of man- and the triangular shape of the procingulum are dibular condyle (34); *distinct ventromedial pro- very different from those found in extant Euneo- cess of mandibular ramus (36); premaxillaries mys. The triangular shape of the procingulum is protruded well anterior of incisive plane (37); *in-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 95 Fig. 41. Cranium and mandible of Euneomys chinchilloides (fmnh 133088).

96 FIELDIANA: ZOOLOGY Fig. 42. Cranium and mandible of Neotomys ebriosus (fmnh 107842).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 97 98 FIELDIANA: ZOOLOGY Fig. 44. Cranium and mandible of Reithrodon auritus (fmnh 134225).

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 99 cisive foramina not reaching anterior conules of cavated (67); *optic foramen very large, orbital Ml (38); anterior apexes of incisive foramina as wings of the presphenoid filamentous (68); distinct broadly separated as posterior apexes (40); well- posterior palatine ridge (72); *large sphenopalatine developed zygomatic spine (43); supraorbital ridg- foramen (75); *stapedial foramen, sphenofrontal es weak (51); frontals incompletely fused or vas- foramen, and squamosal-alisphenoid groove all cularized (53); interparietal/parietal length be- absent (76); alisphenoid process present (78); 12

tween 0.33 and 0.45 (56); eustachian tubes extend thoracic and 7 lumbar vertebrae (79); nuchal lig- past base of pterygoid processes (63); paraptery- ament sometimes attaches to third thoracic ver- goid deeply excavated (67); mesopterygoid fossa tebra (81); and first and fifth digits of pes short reaches M3s (70); distinct posterior palatine ridge and subequal in length (86). (72); stapedial foramen present, sphenofrontal fo- Distribution— Central and southern Argenti- ramen and squamosal-alisphenoid groove absent na, southern Chile, Uruguay, and southernmost (76); alisphenoid strut absent (78); neural spine on Brazil. second cervical vertebra enlarged into "plow"- Comments— At least 13 species-group names shaped keel (82); tail bicolored and densely furred have been proposed for Reithrodon. Its system- (89, 90); and pectoral streak present (92). atics have recently been reviewed by Ortells et al. Distribution— Altiplano and highlands of Peru, (1988) in the context of new karyotype data. They southwestern Bolivia, far northern Chile, to high- recognized two species: typicus for the Uruguayan- = lands of northern Argentina. Brazilian form with 2N 28 and auritus Fischer, 1814 (= physodes Olfers, 1818) for the pampean and central Argentinean form with 2N = 34. The Reithrodon Waterhouse fossils described as Reithrodon chapalmalense (Figs. 43, 44) Ameghino may not belong to Reithrodon. Based on the drawing in Roverto (1914), one anterior Reithrodon Waterhouse, 1837. Proc. Zool. Soc. Lond., cranium lacks two diagnostic characters: the in- 1837:29. cisive foramina do not extend posterior to the an- 1889. Act. Acad. Nat. Cienc. Ptyssophorus Ameghino, terior conules, and the premaxillo-maxillary su- Rep. Argentina, PI. IV. ture does not make an acute-angled bend Tretomys Ameghino, 1889. Act. Acad. Nat. Cienc. the Rep. Argentina, PL IV. (diagnostic of generic group). These two traits IProreithrodon Ameghino, 1908. Anal. Mus. Nac. cannot be assessed in the Pleistocene fossils at- Buenos Aires, 10:424. tributed to Ptyssophorus (left mandible) and Tre- tomys (left upper molars and anterior root of zy- Type Species— Reithrodon typicus Waterhouse, gomata), which otherwise appear referable to 1837, by original designation. Reithrodon. Included Species— Includes auritus (Fischer, Of 16 skeletons examined, one shows a longer 1814), and typicus Waterhouse, 1837. neural spine on the third thoracic vertebra than Diagnosis— Members of the Reithrodon group on the second thoracic vertebra; this is the con- descended from a ancestor with fol- common the dition otherwise found only in Euneomys. Five traits: an additional shallow lowing groove at mid- other skeletons show spines on both T2 and T3 line of the incisors upper (1); upper incisors hyper- equally elongated, but no alcohol specimens were la- opisthodont (2); distinct, laterally positioned, dissected to verify the insertion of the nuchal lig- bial root of *two Ml (4); lingual roots on M2; ament, which in ichthyomyines is coincident with of a labial root in 1 three roots presence m (7); on the elongated spine (Voss, 1988). The phallus of m2 (8); primary cusps alternate (16); protoflexid Reithrodon has been characterized by Hooper on m2 as short *incisive fora- present groove (22); (1962). Additional taxonomic history can be found mina to extending posterior hypocone (38); *deep- in Osgood (1943) and Tate (1932b). ly excised zygomatic plate and long zygomatic spine (43); vaulted cranium; *tegmen tympani does not contact suspensory process of squamosal (58); hamular process of the squamosal reduced along en- Acknowledgments tire length (60); internal carotid canal bounded by both auditory bulla and occipital (62); posterior Sincere thanks go to Bruce D. Patterson for his width of mesopterygoid fossa greater than 2.5 times contributions in innumerable discussions, his time, the anterior width (66); parapterygoids deeply ex- and his expertise. Philip Hershkovitz shared many

100 FIELDIANA: ZOOLOGY insights into Neotropical mammals and was always Archie, J. W. 1985. Methods for coding variable mor- features for numerical taxonomic available for discussions. Robert Voss generously phological analysis. Systematic Zoology, 34: 326-345. examined specimens and added many insights. . 1989. A randomization test for phylogenetic Thanks go to Sydney Anderson, James L. Patton, information in systematic data. Systematic Zoology, and Philip Myers for making specimens available 38: 239-252. for examination on to Jose loan, Angel Spotorno, Bianchi, N. O., O. A. Reig, O. J. Molina, and F. N. Yanez, and Chris Carmichael for access to spec- Dulout. 1971. Cytogenetics of the South American imens in their care, and to Bill Stanley, Barbara akodont rodents (Cricetidae). I. A progress report of Argentinian and Venezuelan forms. Evolution, 25: 724- Stein, Juan Carlos Torres-Mura, and Laura Abrac- 736. zinskas for assisting with access to specimens. Bruce Braun, J. K. 1992. Systematic relationships of the Patterson and Philip Hershkovitz provided help- tribe Phyllotini (Muridae: Sigmodontinae) of South ful comments on earlier drafts of the manuscript. America. Unpublished Ph.D. diss., University of Michael Carleton, Philip Myers, and three anon- Oklahoma, Norman. ymous reviewers provided detailed comments that . 1993. Systematic Relationships of the Tribe (Muridae: Sigmodontinae) of South Amer- significantly improved the organization and con- Phyllotini ica. Oklahoma Museum of Natural History, Norman, tent ofthis manuscript. Barry Chernoff, John Flynn, 50 pp. Joel Brown, Jack Fooden, Lawrence Heaney, Ol- Brylski, P. 1 990. Development and evolution of the iver and Pearson, Angel Spotorno provided help- carotid circulation in geomyoid rodents in relationship ful comments in discussions. I thank Betty Strack to their craniomorphology. Journal of Morphology, for her instruction and assistance with the scan- 204: 33-45. ning electron microscope. Volunteer Jessica G. Bugge, J. 1970. The contribution of the stapedial artery to the cephalic arterial supply in muroid rodents. Shull drew the illustrations in Figures 6-8, 10-12, Acta Anatomica, 76: 313-336. 17, and 18. Vincent Ferguson printed the photo- Carleton, M. D. 1973. A survey of gross stomach A. Kawamoto assisted with the graphs. Mayumi morphology in New World Cricetinae (Rodentia, Mu- figure preparation and drew the illustrations in roidea), with comments of functional interpretations. Figures 9 and 15. Financial support was provided Miscellaneous Publications of the Museum of Zool- of 146: 1-43. principally by the Rowley Graduate Fellowship of ogy, University Michigan,

. 1 980. of neotom- the Field Museum ofNatural History and the Searle Phylogenetic relationships ine-peromyscine rodents (Muroidea) and a reappraisal Fellowship of the University of Chicago, as well of the dichotomy within New World Cricetinae. Mis- as a Collections Grant from the Amer- by Study cellaneous Publications Museum of Zoology, Univer- ican Museum of Natural History, a Grant-in-Aid sity of Michigan, 157: 1-146. of Research from the National of Sci- Academy Carleton, M. D., and G. G. Musser. 1984. Muroid ences, through Sigma Xi, the Scientific Research rodents, pp. 289-379. In Anderson, S., and J. K. Jones, Orders and Families of Recent Mammals of Society, a Center for Latin American Studies Travel Jr., eds., the World. John Wiley and Sons, New York. Grant, and a Grant-in-Aid of Research from the

. 1989. studies of ro- American Society ofMammalogists. The Marshall Systematic oryzomyine dents (Muridae, Sigmodontinae): A synopsis of Mi- Field III Fund provided financial support for the croryzomys. Bulletin of the American Museum of acquisition of some of the examined. specimens Natural History, 191: 1-83.

Catzeflis, F. M., A. W. Dickerman, J. Michaux, and J. A. W. Kirsch. 1993. DNA hybridization and rodent phylogeny, pp. 159-172. In Szalay, F. S., M. J. Novacek, and M. C. McKenna, eds., Mammalian Phy- logeny: Placentals. Springer-Verlag, New York. Literature Cited Chappill, J. A. 1989. Quantitative characters in phylogenetic analysis. Cladistics, 5: 217-234. Ameghino, F. 1889. Contribution al conocimiento de Colless, D. H. 1980. Congruence between morpho- los mamiferos fosiles de la Republica Argentina. Actas metric and data for Menidia A re- de la Academia National de Ciencias de Cordoba, 6: allozyme species: 29: 288-299. 1-1028 and Atlas. appraisal. Systematic Zoology, E. 1874. of the muridae of North . 1908. Las formaciones sedementarias de la Coues, Synopsis America. of the of Natural Sci- region litoral de Mar del Plata y Chapadmalal. Anales Proceedings Academy 26: 173-196. del Museo National de Buenos Aires, 10: 343-428. ences, Philadelphia, F. et Arata, A. 1964. The anatomy and taxonomic signif- Cuvier, 1837. Du genere Eligmodonte de FElig- icance of the male accessory reproductive glands of modonte de Buenos-Ayres, Eligmodontia typus. An- muroid rodents. Bulletin ofthe Florida State Museum, nates des Sciences Naturelles, 7: 168-171. 9: 1-42. Biological Sciences, Ellerman, J. R. 1941. The Families and Genera of

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 101 Living Rodents. Vol. 2, Family Muridae. British Mu- Kelt, D. A. 1 993. Irenomys tarsalis. Mammalian Spe- seum of Natural History, London, 690 pp. cies, 447: 1-3.

Farris, J. S. 1 966. Estimation of conservation of char- Kelt, D. A., R. E. Palma, M. H. Gallardo, and J. A. acters by constancy within biological populations. Cook. 1991. Chromosomal multiformity in Elig- Evolution 20: 587-591. modontia (Muridae, Sigmodontinae), and verification of the status of E. Zeitschrift fur . 1989. The retention index and the rescaled morgani. Saugetier- 56: 352-358. consistency index. Cladistics, 5: 4 1 7—4 1 9. kunde, M. H. 1980. Functional of the Felsenstein, J. 1978. Cases in which parsimony or Kesner, morphology musculature of the rodent Mi- compatibility methods will be positively misleading. masticatory subfamily crotinae. of 165: 205-222. Systematic Zoology, 27: 401^*10. Journal Morphology, Kluge, A. G., and J. S. Farrjs. 1969. . 1985. Confidence limits on phylogenies: An Quantitative and the evolution of anurans. Zo- approach using the bootstrap. Evolution, 39: 783-79 1 . phyletics Systematic ology, 18: 1-32. Gardner, A. L., and J. L. Patton. 1976. Karyotypic W. M. J. AND D. R. Mad- variation in oryzomyine rodents (Cricetinae) with Maddison, P., Donoghue, comments on chromosomal evolution in the Neo- dison. 1 984. Outgroup analysis and parsimony. Sys- tematic 33: 83-103. tropical cricetine complex. Occasional Papers of the Zoology, Museum of Zoology, Louisiana State University, 49: Mann, G. F. 1978. Los pequenos de mamiferos de 1^*8. Chile (marsupiales, quiropteros, edentados y roe- 40: 1-342. Gyldenstolpe, N. 1932. A manual of Neotropical sig- dores). Guyana (Zoologia), modont rodents. Kungliga Svenska Vetenskapsaka- Massoia, E., and U. F. Pardinas. 1993. El estado demiens Handlingar, 3: 1-164. sistematico de algunos muroideos estudiados por Haiduk, M. W., C. Sanchez-Hernandez, and R. J. Ameghino en 1 889, revalidacion del genero Necromys Baker. 1988. Phylogenetic relationships of Nycto- (Mammalia, Rodentia, Cricetidae). Ameghiniana, 30: mys and Xenomys to other cricetine genera based on 407^*18. data from G-banded chromosomes. The Southwestern Mickevich, M. F., and J. S. Farris. 1981. The im- 33: 397-403. Naturalist, plications of congruence in Menidia. Systematic Zo- Hershkovitz, P. 1955. South American marsh rats, ology, 30: 351-370. Holochilus, with a of ro- genus summary sigmodont Musser, G. M., and M. D. Carleton. 1993. Family dents. Fieldiana: Zoology, 37: 639-673. Muridae, pp. 501-756. In Wilson, D. E., and D. M. . 1959. Two new genera of South American ro- Reeder, eds., Mammal Species of the World: A Tax- dents (Cricetinae). Proceedings of the Biological So- onomic and Geographic Reference. Smithsonian In- ciety, Washington, 72: 5-9. stitution, Washington. 1962. Evolution of cricetine ro- Neotropical Olds, N. 1988. A revision of the genus Calomys (Ro- dents (Muridae) with special reference to the phyllo- dentia: Muridae). Unpublished Ph.D. diss., City Uni- tine Fieldiana: 46: 1-524. group. Zoology, versity of New York, New York. 1 966a. Mice, land bridges and Latin American Olds, N., and S. Anderson. 1989. A diagnosis of the faunal interchange, pp. 725-751. In Wenzel, R. L., tribe Phyllotini (Rodentia, Muridae), pp. 55-74. In and V. J. Tipton, eds., Ectoparasites of Panama. Field Redford, K. H., and J. F. Eisenberg, eds.. Advances Museum of Natural History, Chicago. in Neotropical Mammalogy. Sandhill Crane Press, 1966b. South fossorial American swamp and Gainesville, Fla. rats of the Scapteromyine group (Cricetinae, Murinae) Olds, N., S. Anderson, and T. L. Yates. 1 987. Notes with comments on the glans penis in murid taxonomy. on Bolivian mammals 3: A revised diagnosis of An- Zeitschrift fur Saugetierkunde, 31: 81-149. dalgalomys (Rodentia, Muridae) and the description -. 1993. A new central Brazilian genus and spe- of a new subspecies. American Museum Novitates, cies of rodent sigmodontine (Sigmodontinae) transi- 2890: 1-17. tional between akodonts and oryzomyines, with a dis- M. O. A. cussion of murid molar morphology and evolution. Ortells, O., Reig, N. Brum-Zorrilla, and O. A. Scaglia. 1 988. and Fieldiana: Zoology, n.s., 75: 1-18. Cytogenetics karyosystematics of phyllotine rodents (Cricetidae, Sigmodontin- Hooper, E. T. 1962. The glans penis in Sigmodon, ae): I. Chromosomal multiformity and gonosomal- Sigmomys, and Reithrodon (Rodentia, Cricetinae). autosomal translocation in Reithrodon. Genetica 77: Occasional Papers of the Museum of Uni- Zoology, 53-63. versity of Michigan, 625: 1-11. Ortells, M. O., O. A. Reig, R. L. Wainberg, G. E. Hooper, E. T., and G. G. Musser. 1964. The glans HURTADO DE AND T. M. L. GENTILE DE penis in Neotropical cricetines (Muridae) with com- CATALFO, Fronza. ments on classification of muroid rodents. Miscella- 1989. Cytogenetics and karyosystematics of rodents II. neous Publications Museum of Zoology, University phyllotine (Cricetidae, Sigmodontinae). Chromosomal and autosomal of Michigan, 123: 1-57. multiformity polymor- phism in Eligmodontia. Zeitschrift fur Saugetier- HUELSENBECK, J. P., AND D. M. HlLLIS. 1993. SuCCCSS kunde, 54: 129-140. of phylogenetic methods in the four-taxon case. Sys- tematic Biology, 42: 247-264. Osgood, W. H. 1915. New mammals from Brazil and Peru. Field Museum of Natural Pub- Jacobs, L. L., and E. H. Lindsay. 1984. Holarctic History, Zoology lications, X: 187-198. radiation of Neogene muroid rodents and the origin

of South American cricetids. Journal of Vertebrate . 1943. The mammals of Chile. Field Museum Paleontology, 4: 265-272. of Natural History, Zoological Series, 30: 1-268.

102 FIELDIANA: ZOOLOGY . 1947. Cricetine rodents allied to Phyllotis. tidae). Annates del Museo de Historia Natural Val- Journal of Mammalogy, 28: 165-174. paraiso, 13: 285-297. Pardinas, U. F. J. In press. Sobre las vicisitudes de Simpson, G. G. 1945. The principles of classification los generas Bothriomys Ameghino, 1889, Euneomys and a classification ofmammals. Bulletin ofthe Amer- Coues, 1874, y Graomys Thomas, 1916 (Mammalia, ican Museum of Natural History, 85: 1-350. Rodentia, Cricetidae). Ameghiniana. Slaughter, B. H., and J. E. Ubelaker. 1984. Rela- Pearson, O. P. 1958. A taxonomic revision of the tionships of South American cricetines to rodents of rodent genus Phyllotis. University of California Pub- North America and the Old World. Journal of Ver- lications in Zoology, 56: 391-477. tebrate Paleontology, 42: 255-264. Pearson, O. P., and M. I. Christie. 1991. Sympatric Spotorno, A. E. 1992. Parallel evolution and ontogeny species of Euneomys (Rodentia, Cricetidae). Studies of simple penis among New World cricetid rodents. on Neotropical Fauna and Environment, 26: 121-1 27. Journal of Mammalogy, 73: 504-514.

Pearson, O. P., and J. L. Patton. 1976. Relationships Spotorno, A. E., and L. I. Walker. 1983. Analisis among South American phyllotine rodents based on electroforetico de dos especies de Phyllotis en Chile chromosome analysis. Journal of Mammalogy, 57: Central y sus hibridos experimentales. Revista Chilena 339-350. de Historia Natural, 56: 5 1-59. Pine, R. H., S. D. Miller, and M. L. Schamberger. Spotorno, A. O. 1986. Systematics and evolutionary 1979. Contributions to the mammalogy of Chile. relationships ofAndean phyllotine and akodontine ro- Mammalia, 43: 339-376. dents. Unpublished Ph.D. diss., University of Cali- Pritchard, P. C. H. 1994. Comment on gender and fornia, Berkeley. declension of generic names. Journal of Mammalogy, Steppan, S. J. 1 993. Phylogenetic relationships among 75: 549-550. the Phyllotini (Rodentia: Sigmodontinae) using mor- characters. of Evolu- Reig, O. A. 1977. A proposed unified nomenclature phological Journal Mammalian for the enamelled components of the molar teeth of tion, 1: 187-213. the Cricetidae (Rodentia). Journal of Zoology (Lon- Swofford, D. L. 1993. PAUP: Phylogenetic Analysis don), 181:227-241. Using Parsimony. 3.1.1. Illinois Natural History Sur- vey, . 1980. A new fossil genus of South American Champaign. cricetid rodents allied to Wiedomys, with an assess- Tate, G. H. H. 1932a. The taxonomic history of cer- ment of the Sigmodontinae. Journal of Zoology (Lon- tain South and Central American cricetid rodentia: don), 192:257-281. Neotomys, with remarks upon its relationships; the cotton rats and and the "fish- 1 984. Distribuicao geographica e historia ev- {Sigmodon Sigmomys); rats Neus- olutiva dos roedores muroides sulamericanos (Crice- eating" (Ichthyomys, Anotomys, Rheomys, and American Novi- tidae: Sigmodontinae). Revista Brasileira de Genetica, ticomys, Daptomys). Museum 7: 333-365. tates, 583: 1-10.

. 1932b. The taxonomic history of the . 1986. Diversity patterns and differentiation of genus Reithrodon Waterhouse American Mu- high Andean rodents, pp. 404—439. In Vuilleumier, (Cricetidae). seum 529: 1—4. F., and M. Monasterio, eds., High Altitude Tropical Novitates, Biogeography. Oxford University Press, London. . 1932c. The taxonomic history of the South American cricetid Eu- 1987. An assessment of the systematics and genera Euneomys (subgenera and evolution of the Akodontini, with the description of neomys Galenomys), Auliscomys, Chelemyscus, Chinchillula, new fossil species of Akodon (Cricetidae, Sigmodon- Phyllotis, Paralomys, Graomys, Elig- modontia and American Museum Novi- tinae). Fieldiana: Zoology, 39: 347-399. Hesperomys. tates 541: 1-21. Reise, D., and M. H. Gallardo. 1990. A taxonomic O. 1884. On a collection of Muridae from study of the South American genus Euneomys (Cri- Thomas, central Peru. of the cetidae, Rodentia). Revista Chilena de Historia Nat- Proceedings Zoological Society, 1884: 447-458. ural, 63: 73-82. London,

. 1 894. of Rtnker, G. C. 1954. The comparative myology of the Descriptions some new Neotropical Muridae. Annals & of Natural 6: mammalian genera Sigmodon, Oryzomys, Neotoma, Magazine History, 346-366. and Peromyscus (Cricetinae), with remarks on their intergeneric relationships. Miscellaneous Publications . 1 898. On some new mammals from the neigh- of the Museum of Zoology, University of Michigan, borhood of Mount Sahama, Bolivia. Annals & Mag- 83: 1-124. azine of Natural History, 7: 277-283.

Roverto, C. 1914. Los estratos araucanos y sus fosiles. 1 902. On two new genera of rodents from the Anales del Museo Nacional de Historia Natural de highlands of Bolivia. Proceedings of the Zoological Buenos Aires, 25: 1-247. Society of London, 1902(1): 1 16.

Sanderson, M. J., and M. J. Donoghue. 1989. Pat- 1916. On the groupings of some South Amer- terns of variation in levels of homoplasy. Evolution ican Muridae that have been referred to Phyllotis, Eu- 43: 1781-1795. neomys, and Eligmodontia. Annals & Magazine of Natural 8: 139-143. Sarich, V. 1985. Rodent macromolecular systematics, History, pp. 423-452. In Luckett, W. P., and J.-L. Hartenber- 1919. On small mammals collected by Sr. E. ger, eds., Evolutionary Relationships among Rodents: Budin in north-western Patagonia. Annals & Maga- A Multidisciplinary Analysis. Springer-Verlag, Berlin. zine of Natural History, 9: 199-212.

Simonetti, J. Z., and A. O. Spotorno. 1 980. Posicion VlTULLO, A. D., M. B. Espinosa, AND M. S. Merani. taxonomica de Phyllotis micropus (Rodentia: Crice- 1 990. Cytogenetics of vesper mice, Calomys (Roden-

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 103 tia; Cricetidae): Robertsonian variation between Calo- Woodman, N. 1993. The correct gender of mamma- mys callidus and Calomys venustus. Zeitschrift fiir lian generic names ending in -Otis. Journal of Mam- Saugetierkunde, 55: 99-105. malogy, 74: 544-546. VoronTsov, N. N. 1959. The system of hamster (Cri- Zambelli, A., F. Dyzenchauz, A. Ramos, N. de Rosa, cetinae) in the sphere of the world fauna and their E. Wainberg, and O. A. Reig. 1 992. Cytogenetics phylogenetic relations. Biuletin Moskovkogo Ob- and karyosystematics of phyllotine rodents (Criceti- shtschestva Ispitately Prirody, Otdel Biologia, 64: 1 34- dae, Sigmodontinae). III. New data on the distribution 137 [in Russian]. and variability of karyomorphs of the genus Elig- modontia. Zeitschrift fiir 57: 155-162. Voss, R. S. 1988. Systematics and ecology of ichthyo- Saugetierkunde, myine rodents (Muroidea): Patterns of morphological evolution in a small adaptive radiation. Bulletin of the American Museum of Natural History, 188: 259-493.

. 1 99 1 . An introduction to the Neotropical rodent genus Zygodontomys. Bulletin of the American Museum of Natural History, 210: 1-113. Examined 1 992. A revision ofthe South American species Appendix: Specimens of Sigmodon (Mammalia: Muridae) with notes on their natural and American Museum history biogeography. Specimens are from the Field Museum of Nat- Novitates, 3050: 1-56. ural History unless otherwise designated. The ab- . 1993. A revision of the Brazilian muroid ro- breviations and those dent genus Delomys with remarks on "thomaso- skel(s) phal(s) designate myine" characters. American Museum Novitates, 3073: specimens from which skeletons or phalli were 1-46. examined.

Voss, R. S., and M. D. Carleton. 1993. A new genus for Hesperomys molitor Winge and Holochilus magnus Hershkovitz, with comments on phylogenetic relationships and oryzomyine monophyly. American Mu- seum Novitates, 3085: 1-39. Tribe Phyllotini Voss, R. S., and A. V. LrNZEY. 1981. Comparative gross morphology of male accessory glands among Neotropical Muridae (Mammalia: Rodentia) with Calomys callosus— ARGENTINA. Catamarca: comments on implications. Miscellaneous systematic La Merced (msu 19235, 1 9237- skels). Jujuy: Cal- Publications of the Museum of Zoology, University ilegua (22235 — skel, 23373; regarded as venustus of Michigan, 159: 1-39. by Olds, 1988). Tucuman: Conception (30167- Voss, R. S., and P. Myers. 1 99 1 . Pseudoryzomys sim- as venustus BO- plex (Rodentia: Muridae) and the significance of Lund's 30173; regarded by Olds, 1988). collections from the caves of Lagoa Santa, Brazil. Bul- LIVIA. El Beni: San Juan (1 18807, 1 18808); San letin of the American Museum of Natural History, Joaquin (117123); Yuatre (118296); Yutiole 206:414-^132. (118605, 118610, 118294). Santa Cruz: San Mi- Walker, L. I, and A. E. Spotorno. 1992. Tandem guel Rincon (amnh 260686-260690-skels). Tari- and centric fusions in the chromosomal evolution of ja: Villa Montes (34238 -skel). PARAGUAY. the South American phyllotines of the genus Aulis- comys (Rodentia, Cricetidae). Cytogenetics and Cell Chaco: Fortin Madrejon, WNW (ummz 125466, Genetics, 61: 135-140. 125468-125477— skels). Presidente Hayes: Juan

Walker, L. I., A. E. Spotorno, and J. Arrau. 1984. de Zalazar, 8 km NE (ummz 133915-133922- and reproductive studies of two nominal Cytogenetic skels). Total: 35. subspecies of Phyllot is darwini and their experimental hummelincki— COLOMBIA. La Gua- hybrids. Journal of Mammalogy, 65: 220-230. Calomys Pto. E Maicao Waterhouse, G. R. 1837. Characters of new species jira: W Lopez, (usnm 483982). of Mus, from the collection of Mr. Darwin. Proceed- VENEZUELA. Maru lab colony, originally from ings of the Zoological Society of London, 27-29. Monagas: 45 km S Maturin, close to Rio Tigre

Wenzel, R. L., and V. J. Tipton. 1966. Some rela- 1 47 (usnm 460437^6044 , 460447). Monagas: km tionships between mammal hosts and their ectopar- SSE Maturin, Puente Tigre (usnm 388 104, 388 105, asites, pp. 677-723. In Wenzel, R. L., and V. J. Tipton, 388107, 3881 10, 3881 13). Total: 12. eds., Ectoparasties of Panama. Field Museum of Nat- ural History, Chicago. Calomys laucha— ARGENTINA. Buenos Ai- Wiley, E. O. 1981. Phylogenetics: The theory and prac- res: Dorrego (50939, 50940, 50942-50945; re- tice of John New York, phylogenetic systematics. Wiley, garded as murillus by Olds, 1 988); near Henderson 439 pp. (23395); Partida Balcarce (msu 16815, 16816, Williams, D. F., and M. A. Mares. 1978. A new genus 16818, 168 19 -skels); Urdampilleta (23405; re- and species ofphyllotine rodent (Mammalia: Muridae) as murillus Rio from northwestern Argentina. Annals of the Carnegie garded by Olds, 1988). Negro: Museum, 47: 193-221. Chimpay (50932, 50937; regarded as murillus by

104 FIELDIANA: ZOOLOGY Olds, 1988). BOLIVIA. Tarija: Villa Montes, 10 Phyllotis amicus— PERU. Arequipa: Caraveli, km E (amnh 246668, 246674, 246841, 246849, Atiquipa (107389— phal). La Libertad: Menocu- 246867-skels). PARAGUAY. Presidente Hayes: cho (19258-19263). Lima: Chos (107347, Juan de Zalazar, 8 km NE (ummz 133928— skel). 107352-phals). Total: 9. 20. Total: Phyllotis andium— ECUADOR. Azuay: Valle Calomys lepidus— BOLIVIA. La Paz: Ulla-ulla de Yunguilla (43311). PERU. Ancash: Macate (ummz 121081— skel). PERU. Arequipa: Callalli, (20914, 20915, 20923, 20938, 21 145); Rio Mosna 15 km S (mvz 174019— skel); Laguna Salinas (129248, 129249 -skels). Libertad: Hac. Llagueda (49752-skel). Ayacucho: San Miguel (75419). (19464). Lima: Lima (107361 -phal). Total: 10. Picchu Huancav- Cuzco: Machu (107823— phal). Phyllotis caprinus— ARGENTINA. Jujuy: Mai- elica: Santa Ines (75420). Junin: Carhuamayo mara (85847); Sierra de Zenta, La Laguna (85848, Puno: Hac. Collacachi (54743). (49555, 49749- 85849); Sierra de Zenta (41287). Total: 4. skel, 51429). Total: 10. Phyllotis darwini— CHILE. Aconcagua: Papudo Calomys sorellus— PERU. Ancash: Hda. Catoc (22679-22684); Pte. Los Molles (119507- Nevado (81288, 81289); Quincayhuanca (81276, 1 19509— skels). Coquimbo: Parque Nac. Fray Car- 8 1287). Arequipa: Arequipa (107795— phal); Jorge (133874, 133875, 133879-133881, 133894; aveli (107399-phal); Chivay (107688, 107689, Romero — 133896-skel); (22325-22329). Santiago: 107732 phals); Ayacucho: Chunyacc (ummz Cerro Manquehue (119491-119497; 119500— 120286 120291 — -skel); Jawaymachay (ummz skel). Valparaiso: Olmue (22346). Total: 30. skel); Pacaicasu (ummz 120288, 120289 -skels); Phyllotis definitus —PERU . Ancach: Macate Tambo, San Miguel (75388, 75389); Tucumachay (21126-21128— topotypes). Total: 3. (ummz 120290— skel). Cuzco: Huancarani (mvz Phyllotis gerbillus— PERU. Lambayeque: Lam- 171549-171554— skels). La Libertad: mountains bayeque, 1 6 km NW (mvz 1 4 1 847); Morrope (mvz near Otuzco (19209, 19210). Total: 24. 138148, 138 149 -skels). Piura: Piura (21916, Andalgalomys pearsoni— BOLIVIA. Santa Cruz: 81265-81273). Total: 13. Robore, 29.5 km W (amnh 260762-skel). PAR- ECUADOR. Con- AGUAY. Nueva Asuncion: km 620, Trans-Chaco Phyllotis haggardi— Azuay: trayerbas (amnh 61856— skel). Chimborazo: Mt. road (ummz, uncataloged [T. W. Nelson field num- Chimborazo (53306-53308). Pichincha: Mt. Pi- bers 184, 193, 201, 202, 233]). Total: 6. 1 1 1 1 1 Sa- chincha (443 , 443 3, 443 7, 53305, 920 2); Eligmodontia morgani— ARGENTINA. Neu- loya (53309). Total: 10. quen: Chos-Malal (29 1 53); Las Lajas (29 1 55). Rio PERU . Negro: Choele Choel (41293); Pilcaniyeu (29152). Phyllotis magister— Arequipa: Arequipa Santa Cruz: Piedra Clavada (35351). CHILE. (35360, 35361); Cailloma, Chivay (107690, 3 km N To- Aisen: Chile Chico (133010, 133018, 133022, 107691 —phals). Moquegua: Ilubaya, 133025-skels; 133027; 133068-skel); Coih- rata (107417); Mariscal Nieto (107469 -phal). Tacna:Tarata aique Alto (133005 -skel); Pto. Ibanez (133070). (107561, 107611-107613, 107616, Magallanes: Lake Sarmiento (50582). Total: 14. 107620, 107622-phal, 107623, 107625, 107629; mvz 143749). Total: 17. Graomys domorum— BOLIVIA. Cochabamba: Chucuito Aiquile (50961-50963); Parotani (21525, 21526); Phyllotis osilae— PERU. Puno: (5 1285, 1078 59 35 km S Pena Blanca (amnh 255966 — skel); Tin-Tin 51287; 107843, -phals); Have, (50967-50969, 51920, 51921). Santa Cruz: Com- (107860, 107870-107872, 107874, 107881, 107894- apara (amnh 260750-260754 — skels); Florida 107885, 107887, 107888, 107891, 6 Santa (72884). Total: 17. 107896); Pucara, km S (mvz 173165); Rosa, 12 km S (mvz 173162); Yunguyo (51269, 51270, Graomys griseojlavus — ARGENTINA. Cata- 51274, 51278). Total: 23. marca: Belen (28423); Pta. Tinogasta (29163). Rio Negro: Chimpay (50920, 50923-50928). BOLIV- Phyllotis wolffsohni— BOLIVIA. Chuquisaca: IA. Santa Cruz: Cordillera Guanacos (21431, Padilla, 9 km N (amnh 263693, 2639 1 2, 2639 13- Rio 263914— Cocha- 21432). Tarija: Tablada (29165, 29166); Tiquipa, skels); Limon (amnh skel). bamba: Liriuni Pocona Laguna Palmar (amnh 246777, 246778— skels); (140814); (461 13); Tapa- Villa Montes (amnh 246773, 246779 -skels). cari (mvz 1 20 1 80); Taquina (50957-50960, 51918, PARAGUAY. Boqueron: Colonia Fernheim 51919). Santa Cruz: Comarapa, 21 km W (amnh (54359, 54360); La Urbana (34235). Alto Para- 263914-skel). Total: 14. guay: Puerto Casado (54407). Total: 21. Phyllotis xanthopygus rupestris— BOLIVIA. La

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 105 Paz: Esperanza, Pacajes (53607, 53612). CHILE. Sicuani (83475). Puno: Asillo, 3 mi W, Hac. Po- Antofagasta: E of San Pedro (22303, 22304, 22307, socani (51254, 51255); Picotani (52478-52482); 22308, 22311). Coquimbo: Paiguano (22251, Pto. Arturo (53156). Total: 16. 72 22271-22273, 22277, 22285). Tarapaca: Arica, Galenomys garleppi— BOLIVIA. La Paz: Pa- PERU. km E (133830, 133832, 133835, 133836). cajes Province, Esperanza (53845). Oruro: Eucal- Yura Arequipa: (49744, 49763, 49766). Moque- iptus (amnh 24694 1-246947 -skels). Total: 8. gua: Torata (107405, 107407, 107415-phals). Andinomys edax— ARGENTINA. Jujuy: W of Tacna: Tarata (107617, 1 97642 -phals). Total: 26. Yala (23434, 23435); Tilcara (mvz 120222, 120223, xanthopygus— ARGEN- Phyllotis xanthopygus 141617). BOLIVIA. Cochabamba: Ayopaca Prov- TINA. Santa Cruz: Rio Ecker (124384-124388, ince, El Choro (74869). Potosi: Potosi, 20 mi S 124436-124439). CHILE. Aisen: Chile Chico (mvz 120224-120226); Yuruma (29156, 29157). (133940, 133943, 133944, 133947, 133958, Tarija: Camataqui, 25 mi SSE (mvz 120227- 133973, 133979, 1 33982). Magallanes: Ultima Es- 120232). CHILE. Tarapaca: Arica, 72 km E Lazo (50542). Total: 50. peranza, Laguna (132647, 132648, 132651 -skels). PERU. Puno: Loxodontomys micropus— ARGENTINA. Rio Yunguyo, 6 km S (51279-51283). Tacna: Tarata, Negro: San Carlos de Bariloche (23840-23842). 1.5 mi N (mvz 139480, 139481). Total: 27.

Santa Cruz: Rio Ecker ( 1 24393, 1 24394), ( 1 24397, Irenomys tarsalis— CHILE. Aisen: La Junta 124435). CHILE. Aisen: Pto. Ibafiez (132705, (133164). Chiloe: Rio Inio (22528, 22529, 22531- 132706 — skels); Reserva Nac. Coihaique 22535). Llanquihue: Peulla (50558, 50559-skels; (132874-phal). Rio Nireguao (22229-skel; 50563, 50588, 58589). Osorno: Osorno, 84 km SE 23283-23285, 23287). Llanquihue (50647). Ma- (124056-124058); Osorno, 53 km SSE (133137, gallanes: Pto. Natales (506 1 4); Pta. Arenas (506 1 5- 133138, 133 142- skels); Osorno, 44 km SSE 50622; 127337-phal). Malleco: Cerro Nahuel- (133136, 133139, 133140); Osorno, 32 km SSE buta (50643-50645). Osorno: Refugio, Valle de la (133131, 133155-skel); Maicolpue, 65 km W Picada (127717-127720-skels). Total: 33. Osorno (133 133, 133136, 1 331 39); Rufugio, Valle boliviensis— BOLIVIA. La Paz: Pa- Auliscomys de la Picada (127732-skel). Total: 28. cajes Province, Esperanza (53582, 53583, 53586, Euneomys chinchilloides— ARGENTINA. Ti- 53589-53599). CHILE. Tarapaca: Choquelimpie erra del Fuego: Lago Fagnano (50736). CHILE. (22690, 22691). PERU. Arequipa: Cailloma Aisen: Pto. Ibafiez (133088, 133089-skels: (49765, 49768-49771, 49774-skels); Puno: Pas- 134027; 134181, 134182-skels; 134183, 134184, to Grande, 30 mi W Mazo Cruz (mvz 114719); 1 34 1 86, 1 34233). Magallanes: Pta. Arenas (50600, San Antonio de Esquilache (49574-49577, 49579- 50601). Total: 12. 49581). Total: 30. Euneomys petersoni — CHILE. Aisen: Coih- Auliscomys pictus— PERU. Arequipa: Cailloma aique Alto, 4.5 km E (133082, 133083-skel, (49775_skel; 107678, 107716-phals); Cuzco: 133085-skel, 133086). Magallanes: Ultima Es- Machu Picchu, 20 km E (107804-phal, 107806, peranza, Laguna Lazo (50584-50586, 50588- 107819). Junin: Carhuamayo (54734-54742); Ju- 50590, 50593-skel, 50595-50599); Lago Sar- nin (21132, 21133, 21 135-21 142- topotypes); miento (50583). Total: 18. Pachacayo (20060); Tarma (64344). Puno: Puno, ebriosus— Si- Hac. Collacachi (49751— skel); Santa Rosa, 6 km Neotomys ARGENTINA. Jujuy: W (107918, 107920, 107922, 107925-skel, erra de Zenta (41282). PERU. Pasco: Chigrin 107926; 107968, 107975 -skels). Total: 34. (24776-24778); La Quinua (24775). Ancash: Re- Cuzco: Auliscomys sublimis— PERU. Arequipa: Cail- cuay Ticapampa (81283). Marcapata, Ccolini (75580). Junin: Paccha (64345). Puno: Hac. loma, Chivay (107696, 10771 1 -phals); Laguna Collacachi (49708); Have (107824, 107842 -skel); Salinas (49542, 49543, 49546, 49547); Sumbay Total: 13. (49536, 49537, 49539, 49540, 49544, 49545). Yunguyo (51261, 51263). Puno: Yunguyo, 6 km S (51260); Huacullani Reithrodon auritus evae— ARGENTINA. Ne- (52669-52671); Have, 35 km S (107873); Laguna uquen: Estancia Alicura (mvz 151033); Lago Na- de Loriscota, 5 mi N (mvz 145613— skel); Puno huel Huapi, 11 km NNE (mvz 165853, 169013). (amnh 213596, 213597, 213601). Total: 20. Rio Negro: San Carlos de Bariloche, 18 km SE

(mvz 1 62272); Comallo, 8 km WSW (mvz 1 6403 1 ). Chinchillula sahamae— PERU . Arequipa: Cail- Total: 5. loma (49406, 49407 -skel, 49421, 49422); Sum- bay (49401, 49417, 49418). Cuzco: Cordillera de Reithrodon auritus pachycephalus — CHILE.

106 FIELDIANA: ZOOLOGY Aisen: Chile Chico (134178; 134188, 134189, Geronimo (35182-35186). GUATEMALA. 134225, 134226, 134228, 134229, 134231-skels; Esuintla: Finoa St. Christina (7350 1 -skel). HON- 134232, 134235); Coihaique Alto (134187-skel, DURAS (43308 -skel). MEXICO. Chiapas: 134192; 134196, 1 34202 -skels; 134204, 134205, Ocosingo (64186). PANAMA. Canal Zone: Rod- 134207, 134210, 134213, 134222, 134224). Ma- man Ammo Depot (usnm 396407). Chiriqui: Cer- gallanes: Pta. Arenas (124426— holotype); Rio ro Punta Casa Tillfy (usnm 323881). Darien: La Verde (50570-50576). Total: 29. Laguna (usnm 338266), El Aguacate (usnm Bocas del Toro Cerro Reithrodon typicus typicus— URUGUAY. Lav- 503722), (usnm 323880), Atul Total: 16. alleja: Minas, Arroyo Polanco (27707-27709). (usnm 306974). Rocha: Castillos (27704). San Jose: Puerto Arazati Tylomys nudicaudus— GUATEMALA. Chi- (27653). Trienta y Tres: Quebrada de los Cuervos maltenango: Yerocapa, Finca Recreo (64568— skel, (27705, 27706). Total: 7. 64569). MEXICO. Chiapas: Palenque (66949). Oaxaca: San Gabriel Mixtapec, 23 mi N (ummz 114186, 114187— skels). Veracruz: Cerro Balza- Other Specimens Examined for pote (127151). msu lab colony (ummz 159340, Phylogenetic Analyses 159341— skels). Zoo specimens (60029, 60030, 60767, 1048 15 -skels). US National Zoo (usnm Old World "Cricetids" 398072, 398073, 520882, 520883). Total: 16.

Subfamily Calomyscinae Subfamily Neotominae Calomyscus baluchi— AFGHANISTAN. Bam- ian: 7.5 mi W Shibar Pass (102956, 102957, Neotomafloridana— USA. Arkansas: Logan Co., 102959); S. Atallah (102960-102962, 102964). Magazine Mtn. (67699); Scott Co., Fourche La PAKISTAN. Malakand: Swat, Karakar Pass Fave River (67700); Stone Co., Marcella (63994, (140403-1 40407 -skels). Total: 12. 63996). Illinios: Union Co., Pine Hills (64333- skel, 64335). Total: 6. Ochrotomys nuttalli— USA. Illinois: Alexander Cricetinae Subfamily Co., Olive Branch (2 1 1 98 — skel). Louisiana: Cald- well Co., Columbia (16486-16488, 16490). North Cricetulus migratorius— IRAN. Azerbaijan: Re- Carolina: Buncombe Co., Weaverville (5265). To- zaiyeh (57893, 57900, 96957-96961). Tehran: tal: 6. DoAb Total: 8. (96956). Peromyscus leucopus— USA. Illinois: Johnson Mesocricetus auratus— Asia: captive (122237— Co., Ozark (1 5745-1 5747, 1 5752-skel); Pope Co., skel); domesticated (571 14— skel). Total 2. Golconda (15708, 15709); Union Co., Cobden Phodopus sungorus— Asia: zoo specimens (139895 -skel). Total: 7. (58804, 58983, 58984, 134492 -skels); domesti- Scotinomys teguina— COSTA RICA. Puntar- cated (msu 34350, 35465— skels). Total 6. enas: Monteverde (128560-128563 — skels, 128564-128566). Total: 7.

Subfamily Mystromyinae Tribe Akodontini Mystromys albicaudatus -SOUTH AFRICA. Transvaal: Johannesburg (38147). Total: 1. Akodon albiventer— BOLIVIA. Potosi: Uyumi (mvz 120233— skel). CHILE. Tarapaca: Arica 72 New World "Cricetids" km E (129981-129983-skels, 129986); Putre (1 29992-1 29993-skels). PERU. Moquegua: To- Subfamily Tylomyinae quepala (mvz 145543 — skel). Puno: Yunguyo (51300). Tacna: Tarata (107578-107581, 107600, Nyctomys sumichrasti— COSTA RICA. Puntar- 107618, 107621, 107644). Total: 17. enas: Monteverde (usnm 559055, 556461); San Akodon boliviensis—PERU. Puno: Have, 5 km

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 107 W (107869, 107882, 107886, 107889, 107892); skel). Pasco: Puerto Victoria (5 1358— skel). Puno: Santa Rosa, 6 km W (107917, 107928, 107976- Bella Pampa (mvz 1 16665 — skel); Sagrario skels), 12 km S (mvz 171615, 171617, 171618, (52495 -skel); Rio Cayumba, Hac. Exito (24761- 1 7 1 620, 1 7 1 62 1 - skels); Yunguyo (5 1 294-5 1299). 24763). Total: 9. Total: 19. Nectomys squamipes—ROlAVlA. La Paz: Al- Chroeomys andinus— CHILE. Antofagasta: Po- coche (117119). BRAZIL. Sao Paulo: Barra do Rio cos (mvz 1 19554— skel). Tarapaca: Parque Nac. Juquia (93046-93048); Ilha do Cardoso (141630- Lauca, Parinacota (130001, 130002-skel, 141633,141636— skels); Primeiro Morro (94393- 130004-skel). PERU. Moquegua: Torata 94379); Ribeirao Fundo (94395); Rocha (94400, (107511, 107516, 107522, 107523, 107525, 94402 -skels); Sao Sebastiao (18200). COLOM- 107526, 107540, 107541). Total: 12. BIA. Antioquia: Bellavista (701 10,70111, 70113— Total: 20. Chroeomysjelskii— PERU. Arequipa: Cailloma skel). (49767-skel; mvz 174281, 174283, 174287- Oligoryzomysfulvescens— COSTA RICA. Pun- 174289 -skels). Puno: Macusani (mvz 173249, tarenas: Finca Helechales (usnm 547949). GUA- 173251, 173255). Santa Rosa (107919, 107921, TEMALA. Chiquimula: Esquipulas (73546- 107931, 107932 -skels). Total: 13. 73548 -skels). MEXICO. Oaxaca: San Gabriel 190328 — Oxymycterus hispidus — ARGENTINA. Mi- (amnh skel). Puebla: Huachinango 6 siones: Rio Parana, Caraguaytay (26753-26757; (6 1833, 1834). Veracruz: Achotal( 14 105-14 108, Cerro 26841, 26856, 26857 -topotypes); Pto. Aguirre 15882X). PANAMA. Azul (usnm 305677). Total: (23843). Total: 9. 11. Oryzomys capito— ECUADOR. Napo, San Jose dePayamino( 125052, 125058, 125059, 125063- skels). PERU. Cuzco: Quincemil (75222, 75255, Tribe Ichthyomyini 75260, 75270). Madre de Dios: Manu (139864, 139865, 139868, 139869 -skels). Total: 12. Anotomys leander— ECUADOR. Napo: Papal- Oryzomys palustris— USA. Florida: Highlands 6.9 km 1 1 55598-1 55630- lacta, W (ummz 26926, Co., Lake Istikpogo (amnh 2425 1 7, 2425 1 8), Lake skels). Pichincha: Chinchin Cocha (53367). Total: 8. Placid (amnh 242519, 242521, 242524). Missis- Ichthyomys hydrobates— COLOMBIA. Cauca: sippi: Copian Co., Burnell (48450, 48452-48455). Chisquio (90293). VENEZUELA. Merida: La Mu- New Jersey: Salem Co. (amnh 232365). Texas: — cuy (ummz 156375 skel). Total: 2. Jefferson Co., Hildebrandt Acres ( 1 34427, 1 34428). Neusticomys monticolus— COLOMBIA. Antio- Total: 14. quia: Santa Barbara (71219-71223); Urrao Pseudoryzomys simplex— BOLIVIA. El Beni: (7 1 2 1 8 - skel). Huila: San Agustin (7 1 224, 71225). San Joaquin (1 18810). PARAGUAY. Chaco: Ta- ECUADOR. Napo: Papallacta (ummz 155604- caagle (34236). Presidente Hayes: Villa Hayes, 24 155606— skels). Pichincha: Old Santo Domingo km NW (ummz 133913). Total: 3. Trail (ummz 126299, 155789, 155790, 155793- Zygodontomys brevicanda— COLOMBIA. Bo- Total: 15. skels). livar: Socorre, upper Rio Sinu (69152— skel). PANAMA. Santa Rita de le Charrero (msu 20669 -skel). SURINAM. Brokopondo: Kwak- oegron (95688, 95784-skels; 95788). TRINI- Tribe Oryzomyini DAD. Cuara (5348). Princetown (5349-5351- topotypes). VENEZUELA. Zulia: Empalado Sa- Holochilus brasiliensis— Beni: BOLIVIA. mouth bana (18740- skel). Total: 10. Rio Baures (amnh 2 1021 8-2 10223 -skels). PAR- AGUAY. Presidente Hayes: Chaco, 15 km NNW (ummz 125997-126003, 126005 -skels). PERU. Tribe Loreto: Rio Amazonas (88913-88917); Yarina- Scapteromyini cocha (55471, 55476, 62089). Total: 22. Neacomys spinosus— COLOMBIA. Putumayo: Kunsia tomentosus—BOlAYlA. Beni: San Joa- Total: 2. Mecaya (7 1 784— skel). PERU. Cuzco: Paucartam- quin (122710, 122711). bo, 72 km NE (ummz 160544— skel). Madre de Scapteromys tumidus— BRAZIL. Rio Grande Dios: Rio Alto Madre de Dios (ummz 160545 — do Sul (amnh 235430-235432-skels). PARA-

108 FIELDIANA: ZOOLOGY GUAY. Cordillera: Tobati(uMMZ 125954, 125956, Thomasomys baeops— COLOMBIA. Huila: San 137071-skels). URUGUAY. Canelones: Bal- Agustin (71466, 71467, 71472, 71476). ECUA- neario, Salinas (122712, 122713). San Jose: Rio DOR. Azuay (amnh 47677, 6 1 925, 6 1 953); El Oro Santa Lucia (122714; amnh 188783, 188784- (amnh 47699); Loja (amnh 61363). Total: 9. skels). Taecuarembo: Rio Negro (amnh 188782). Thomasomys rhoadsi — ECUADOR. Pichin- Total: 13. cha: Cerro Antisana (43246-43250); (amnh 213548-skel). Total: 6.

Tribe Sigmodontini Sigmodontinae incertae sedis Sigmodon hispidus- BRITISH HONDURAS. Cayo, 12 km S (ummz 62985 -skel). MEXICO. Punomys lemminus— PERU. Arequipa: Huay- Chiapas: Bochil (ummz 92598— skel). Veracruz: larco (mvz 116036). Puno: Abra Aricoma (mvz Tenochtitlan (ummz 116335, 116336, 116338— 139588, 139589); Limbani, 8 mi. SSW (mvz skels). USA. Florida: Alachua Co., Gainesville 114757, 114758, 116190-116194); San Antonio (7955-7963). Georgia: Camoen Co., St. Marys de Esquilache (49710— holotype). Tacna: Tarata, (7953); Lanier Co., Oldfield (135121, 135122- 20 km NE (mvz 1 15948). Total: 12. skels); Lowndes Co., 1-75 and GA-31 (135123- 135 125 -skels). Louisiana: Hackley (16383- 16386, 16388). New Mexico: Otero Co., near Tularosa (125371, 125373-skels). Total: 27. Other Specimens Examined for

Vertebral Counts (Table 5) Tribe Wiedomyini

VENEZUELA. Merida Wiedomys pyrrhorhinos — BRAZIL. Ceara: Aepeomys lugens— Total: 1. Ibiapaba (25249). Pernambuco: Exu (136941); (usnm 387955). Garanhuns (136942). Locality unknown (usnm Akodon {Abrothrix) longipilis— ARGENTINA. 538306, 538314, 538382, 538386-538388). To- Rio Negro (mvz 155725, 155726, 155728, 155729, tal: 9. 1 63364); Santa Cruz (ummz 1571 54). CHILE. Val- paraiso (130905, 130907-130910). Total: 11. Akodon {Abrothrix) sanborni— CHILE. Chiloe Thomasomyine Group (127565, 127566, 127568, 127569, 127572); Osorao (mvz 154128, 154130). Total: 7. Akodon aerosus— PERU. Cuzco (mvz 166777- Chilomys instans— COLOMBIA. Huila: Las 166779, 166784); Puno (mvz 173172, 173174, Bardas (71493); San Agustin (71499, 71609). EC- 173175, 173180). Total: 8. UADOR. Cafiar: Chical (amnh 62922 -skel). Pi- Akodon azarae— ARGENTINA. Buenos Aires chincha: Saloya (53403); Volcan Pichincha (22233); Corrientes (mvz 166106— skel); La Pam- (53405). VENEZUELA. Merida: Paramo Tambor pa (mvz 173730— skel). Total: 3. (22 172 -skel). Total: 7. Akodon cursor — BRAZIL. Rio de Janeiro Rhipidomys latimanus— COLOMBIA. Antio- (26626); Pernambuco (123060); Sao Paulo quia: San Jeronimo (70235, 70237, 70238-skels, Total: 11. Huila: EC- (141606-141614). 70241, 70242). Pitalito(71710-skel). — UADOR. Imbabura NE Penahevva (ummz 77245, Akodon mollis PERU. Ancash (129212, 77247 -skels). Total: 8. 129213, 129215, 129216). Total: 4. Thomasomys aureus— BOLIVIA. Cochabamba Akodon neocenus— ARGENTINA. La Pampa Total: 3. (amnh 260422). La Paz: Yerbani (ummz 1 55942— (mvz 173726, 173732, 182025). skel). COLOMBIA. Antioquia: SE Medellin Akodon olivaceus — ARGENTINA. Neuquen (70330-skel); Ventanas (70321 -skel). Caldas: (mvz 163455, 166063); Rio Negro (mvz 155752- Termales (71263, 71264). PERU. Cuzco: Limac- 155754). CHILE. Aisen (22230, 22237, 22238); punco (75228, 75230-75235); Paucartambo, 72 Chiloe (132169-132173, 132275, 132278, km NE (mvz 16671 1 -skel; ummz 160575-skel). 132279); Valdivia (mvz 154125-154127). Total: Total: 15. 19.

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 109 Akodon puer— PERU. Arequipa (mvz 174009, gro (mvz 174385-174387). CHILE. Aisen 174100); Puno (mvz 173238, 173240, 173241, (132957-132959, 132962, 132963); Magallanes 173244, 173245). Total: 7. (50530). Total: 9. — Akodon subfuscus PERU . Ayacucho (mvz Delomys dorsalis— BRAZIL. Sao Paulo (145370, 174246-174248); Puno (mvz 173229, 173234- 145371). Total: 2. 173236). Total: 7. Delomys sublineatus — BRAZIL. Sao Paulo Akodon nr/cAi— TRINIDAD AND TOBAGO. (141628, 141629). Total: 2. VENEZUELA. Tobago (usnm 540711). Aragua Eliurus myoxinus— MALAGASY. Ampitambe (usnm 517596, 517599, 517600); Bolivar (usnm (amnh 31801). Total: 1. 448568). Total: 5. Geoxus valdivianus— ARGENTINA. Rio Ne- Akodon torques— PERU. Cuzco (mvz 166744- gro (mvz 155817, 163382, 172206); Santa Cruz 166751, 166760, 166762). Total: 10. (usnm 49522). CHILE. Osorno (127726, 133090, — Rio Akodon xanthorhinus ARGENTINA. 133099-133101). Total: 6. Negro (mvz 158401); Tierra del Fuego (usnm Gymnuromys roberti — MALAGASY (usnm 482127, 482129, 482130, 482132-482134). 49670, 49671). Total: 2. CHILE. Magallanes (127301-127305). Total: 12. Holochilus brasiliensis vulpinus— ARGENTI- Akodon (Deltamys) kempi— URUGUAY. Ro- NA. Entre Rios (ummz 166268, 166269, 166399, cha (amnh 206139, 206140, 206142). Total: 3. 166692). URUGUAY. Canelones (amnh 206362); Akodon (Microxus) bogotensis— VENEZUELA. Sori (amnh 206372). Total: 6. Merida (usnm 3746 1 2); Tachira ( 1 8678). Total: 2. Holochilus chacarius— PARAGUAY'. Alto Par- Akodon {Microxus) mimus— BOLIVIA. Cocha- aguay (ummz 166255, 166256, 166259, 166267, bamba (amnh 260429, 260586, 260587, 260592, 166314, 166438, 166450). Total : 8. 260593, 260595-260599); La Paz (ummz 126779, Ichthyomys tweedii — EUCADOR. Pichincha 1 55946-1 55948). PERU. Puno (mvz 1 7 1 750). To- Total: 4. tal: 15. (ummz 126300, 155782, 155786, 155787). molitor— URUGUAY. Trienta Akodon (Thaptomys) nigrita— PARAGUAY. Lundomys y Tres Total: 2. Itapua(uMMZ 125959-125963, 125968). Total: 6. (amnh 206392, 296393). — Baiomys musculus — MEXICO. Veracruz Macrotarsomys bastardi MALAGASY. Fi- narantsoa Toliara (56145). Total: 1. (usnm 328800, 328806); (usnm 578716-578718). Total: 5. Bolomys amoenus— PERU. Puno (49750; mvz 173191-173193, 173198-173201). Total: 8. Melanomys caliginosus— COLOMBIA. Meta (amnh 1 5497). Santa Marta (usnm 280606). COS- Bolomys lasiurus — PARAGUAY. Paraguari TA RICA. Heredia (128471, 128472, 128476, (ummz 133947, 137083-137087). Total: 6. 128477, 128484, 128485, 128488, 128489). To- obscurus— ARGENTINA. Buenos Ai- Bolomys tal: 10. res (msu 17387). Total: 1. Microryzomys altissimus— ECUADOR. Cariar Brachytarsomys albicauda —MALAGASY (amnh 63052); Imbabura (125043); Pichincha (amnh 100690-100692). Total: 3. (amnh 213549). Total: 3. Brachyuromys ramirohitra —MALAGASY Microryzomys minutus — BOLIVIA. Cocha- (5623). Total: 1. bamba (amnh 260419). ECUADOR. Tungurahua Calomys musculinus— ARGENTINA. Buenos (47597). VENEZUELA. Merida (usnm 374373, Aires (msu 1 9398); Rio Negro (msu 1 9 1 86, 1 9 1 87, 374380, 374443); Sucre (amnh 69894, 69896). 19189). Total: 4. Total: 7. Cricetulus barabensis —Locality unknown Mystromys albicaudatus — Location unknown (USNM 521109). Total: 1. (usnm 396240). Total: 1. — Cricetulus longicaudatus CHINA. Qinghai Neacomys guianae — BRAZIL. Para (usnm (usnm 449101, 4491 15—4491 18). Total: 5. 549553). FRENCH GUYANA. Paracou (amnh Cricetus cricetus— GERMANY, (amnh 31814). 266555-266558). Total: 5. POLAND (usnm 494 1 3). US National Zoo (usnm Neacomys spinosus — BOLIVIA. Santa Cruz 294370, 294371). Total: 4. (amnh 261987, 261989-261991, 263815). EC- Chelemys macronyx— ARGENTINA. Rio Ne- UADOR. Napo (usnm 534372, 574567). Total: 7.

110 FIELDIANA: ZOOLOGY Neacomys tenuipes— COLOMBIA. Antioquia Oligoryzomys longicaudatus — CHILE. Co- (usnm 499556, 499557, 499559). VENEZUELA. quimbo (133215, 133217, 133218; 133596- Aragua (usnm 517585). Total: 4. 133599). Total: 7. Nesomys audeberti— MALAGASY. Finarant- Oligoryzomys magellanicus — ARGENTINA. soa (usnm 448946, 449232). Total: 2. Tierra del Fuego (usnm 482125, 482126). Total: 2. Nesomys rufus — MALAGASY. Finarantsoa Oligoryzomys microtis— BOLIVIA. Beni (amnh (usnm 448954, 448962^148964). Total: 4. 266947-255950, 255952, 255953). Total: 6. Nesoryzomys narboroughi— ECUADOR. Ga- Oligoryzomys microtis fornesi— PARAGUAY. lapagos (30844; usnm 364938, 364939). Total: 3. Canendiyu (ummz 126013, 126082, 133826, Nesoryzomys indefessus— ECUADOR. Gala- 137018, 137028). Total: 5. pagos (30853). Total: 1. Oligoryzomys nigripes— ECUADOR. El Oro Notiomys edwardsii— ARGENTINA. Rio Ne- (amnh 47744, 47745); Pasaje (amnh 61313). gro (mvz 163065). Total: 1. PARAGUAY. Paraguari (ummz 133880, 133882, 137041, Total: 7. Oecomys bicolor— BOLIVIA, La Paz (ummz 137042). COSTA RICA. Heredia 155945); Santa Cruz (amnh 246808, 262009- Oryzomys albigularis— Puntarenas 2620 11,26381 6). COLOMBIA. Caqueta (72093). (128164, 128462). (128468, 128470, ECUADOR. Napo (125044, 125045, 125047). 128573; usnm 559053). ECUADOR. Caiiar (amnh GUYANA. Muzaruni-Potaro (amnh 64130). 63330); Pichincha (94979). PANAMA. Darien PANAMA. San Bias (usnm 335532, 335533). (usnm 338207). Total: 9. PERU. Madre de Dios (ummz 160550). Total: 14. Oryzomys alfaroi— PANAMA. Darien (usnm Oecomys concolor — VENEZUELA. Aragua 383247, 383256, 383258). GUATEMALA. Es- (usnm 399535, 517572, 517573). Total: 3. quintla (usnm 275364). Total: 4. Oecomys mamorae — BOLIVIA. Santa Cruz Oryzomys bombycinus— PANAMA. Cerro Azul Total: 2. (amnh 262013). Total: 1. (usnm 305649, 305653). Oecomys paricola— GUYANA. Mazaruni-Po- Oryzomys buccinatus— PARAGUAY. Canen- Cordillera taro (amnh 48142). Total: 1. diyu (ummz 133798); (ummz 126005, Total: 3. Oecomys roberti — BOLIVIA. Pando (amnh 126006). 248996, 262825). BRAZIL. Para (usnm 549537, Oryzomys capito— BOLIVIA. Beni (amnh 549539, 547540). Total: 5. 209961, 209968, 210016-210018). Total: 5. Oecomys superans — COLOMBIA. Caqueta Oryzomys chapmani— MEXICO. Oaxaca (amnh (125064). PERU. Pasco (amnh 213540, 232140). 254698-254700). Total: 3. Total: 3. Oryzomys couesi— GUATEMALA. Esquintla Oecomys trinitatis— PANAMA. Darien (usnm (usnm 275365). NICARAGUA. Managua (amnh 305708, 310549, 310550). VENEZUELA. Merida 178028, 178030). Total: 3. (21823). Total: 4. intermedins— BRAZIL. Sao Paulo — Oryzomys Oligoryzomys andinus BOLIVIA. Oruro (94549, 94550, 141637-141641). Total: 7. (amnh 260405); Potosi (amnh 255946). PERU. Oryzomys keaysi — BOLOVIA. Cochabamba Ancash (fmnh 129240). Total: 3. (amnh 260346-260349, 260351). Total: 5. Oligorzoymys chacoensis — BOLIVIA. Chuquis- Oryzomys melanotis— MEXICO. San Luis Po- aca (amnh 262126, 126127, 262129-262131). tosi (amnh 254713-254716). Total: 4. PARAGUAY. Chaco (ummz 125539-125542). «///#«£— Total: 9. Oryzomys PARAGUAY. Itapua (ummz 126008). Total: 1. Oligoryzomys rfe///co/fl -URUGUAY. Duraz- no (amnh 205955, 205957, 205959, 205960). To- Oryzomys polius— PERU. Amazonas (129242, tal: 4. 129243, 129245). Total: 3. PARAGUAY. Oligoryzomys destructor— PERU'. La Libertad Oryzomys ratticeps— Canendiyu (31702). Total: 1. (ummz 1 33794); Itapua (ummz 1 36007); Missiones 125458, Total: 4. Oligoryzomys eliurus— BRAZIL. Mato Grosso (ummz 125459). (amnh 134899). Total: 1. Oryzomys subjlavus—BOlAYlA. Beni (amnh Total: 4. Oligoryzomys flavescens— URUGUAY. Rocha 210024-210027). (amnh 205597-205601). Total: 5. Oryzomys talamancae— COLOMBIA. Choco

STEPPAN: REVISION OF THE TRIBE PHYLLOTINI 111 (72408). EUCADOR. Loja (amnh 61335). VEN- Rhipidomys venezuelae— VENEZUELA. Ara-

5 1 5 1 1 1 EZUELA. Zulia (usnm 448597, 448605, 448607- gua (USNM 7589, 759 , 5 7592). Zulia (usnm 448609). Total: 7. 448629). Total: 4. Oryzomys xantheolus — PERU. La Libertad Rhipidomys venustus— VENEZUELA. Merida (44433). Total: 1. (usnm 387908). Total: 1. Ototylomys phyllotis— BELIZE. Orange Walk Sigmodon alleni— MEXICO. Guerrero (ummz District (58546). GUATEMALA. Alta Verapaz 109664); Michoacan (ummz 109660, 119602, (64564, 64565); Peten (ummz 63556); San Miguel 119604). Total: 4. (ummz 110907). MEXICO. Yucatan (amnh Sigmodon fulviventer— USA. Arizona (ummz 91222). Total: 6. 85455, 85456). Total 2. Ototylomys phyllotis fumeus— NICARAGUA. Sigmodon leucotis— MEXICO. Oaxaca (ummz Chinandega (ummz 1 15519). Total: 1. 96298; 127152). Total: 2. Oxymycterus delator— PARAGUAY. Canen- Sigmodon ochrognathus— USA. Texas (ummz diyu (ummz 126079, 126085-126087), (amnh 79140, 79153-79155). Total: 4. 248437). Total: 5. Sigmodontomys alfari — COLOMBIA. Pal- Oxymycterus inca —BOLIVIA. Santa Cruz mares del Pacifico (usmn 483981). ECUADOR. 263349- (amnh 260604, 260605, 263344-263347, Camolima (ummz 7724 1). PANAMA. Tacaracuna 263353). Total: 11. (usnm 310588, 310589, 310590). Total: 5. Oxymycterus paramensis— ARGENTINA. Ju- Thalpomys lasiotis— BRAZIL. Distrito Federal Entre Rios juy (22234). BOLIVIA. (usnm 271399). (128327). Total: 1. PERU. Puno Total: 3. (52476). Thomasomys cinereus— ECUADOR. El Oro Oxymycterus rufus— ARGENTINA. Buenos (amnh 47659^17661, 47667). Total: 4. Aires BOLIVIA. La Paz (usnm 236296). (amnh Thomasomys daphne— BOLIVIA. La Paz (ummz 249003-249005); Tarija (amnh 262968). BRA- 155894). PERU. Cuzco (ummz 160580). Total: 2. ZIL. Santa Catarina (usnm 236672). URUGUAY. Thomasomys gracilis— PERU. Cuzco (ummz Canelones (amnh 206168, 206169); Cerro Largo 160584). Total: 1. (amnh 206177). Total: 9. Thomasomys hylophilus— COLOMBIA. Norte Rheomys raptor — COSTA RICA. San Jose de Santander (18586). VENEZUELA. Tachira (UMMZ 1 1 1985, 1 1 1986, 1 12300, 1 12301). To- (usnm 442305). Total: 2. tal: 4. Thomasomys oreas — PERU. Cuzco (ummz Rheomys thomasi— EL SALVADOR. Chala- 160587). Total: 1. tenanio (mvz 98811, 98813, 98815); San Miguel — ECUADOR. Pi- (MVZ 98799-98801, 99805, 131998). GUATA- Thomasomys paramorum chincha (amnh 213545-213547; fmnh 47595, MALA. Huehuetenango (ummz 1 18235). Total: 9. 47596, 125061). Total: 6. Rheomys underwoodi— COSTA RICA. Alajuela ECUADOR. Cafi- (ummz 115389, 115460). Total: 2. Thomasomys pyrrhonotus— ar (amnh 63316, 63326). Total: 2. Rhipidomys couesi — BOLIVIA. Chuquisaca Tscherskiatriton— KOREA. Kumhwa (amnh 263919); La Paz (amnh 262991, 262992); NORTH Total: 1. Tarija (amnh 246827). BRAZIL. Mata Grosso (usnm 294636). (amnh 134522). TRINIDAD AND TOBAGO. Tylomysfulviventer — PANAMA. Darien (usmn Trinidad (amnh 212140, 212141, 235071). 310600, 310602-310605). Total: 5. Total: 8. Tylomys mirae— COLOMBIA. Boyaca (71216); Rhipidomys fulviventer— COLOMBIA. Antio- Caldas (71215). Total: 2. quia (71720). Total: 1. Tylomys panamensis— PANAMA. Canal Zone Rhipidomys macconnelli— VENEZUELA. Bo- (usnm 396409); Cerro Azul (usnm 306972); (usnm livar (mvz 160083). Total: 1. 503720, 503721). Total: 4. Rhipidomys mastacalis — BRAZIL. Goyaz Tylomys watsoni— COSTA RICA. Cartago (amnh 1 34524); Para (usnm 549554). VENEZUE- (usnm 566460). Total: 1. LA. Bolivar (usnm 448620, 448621). Total: 3. Rhipidomys scandens — PANAMA. Darien (usnm 338265). Total: 1.

112 FIELDIANA: ZOOLOGY

ted Listing i »f ZotilOH

ictotyp Lawrenc

) 1990. 2 tables.

Panicuh

me Macaques ;k Foode Zoology, n.s.,

Museum, Stockholm .

' • Pattei :.r. , ''let .

' ; le ime: u ta >h linifo tti (<

lolotsv, ii.s.. noin 67 199 ! • •

; tatus, Relative Abundant and Behavior i n.s no. 3 HHus., ilivia. Bv Susan E. Davis Fieldiana Zool T)\ 1,1993 pages

• ; , . ice list. \tl ord( i rnusl 3 oui ; jcation number and »

'• ; > icck: rder i in able in its add current destination p< ;

k. Prices and terms sxii i he 1 subsidiary

>s all reque: ts

- '>. vision ;. brar Pubi cations

' : Ro< elt •' ad Lake Shore Dj

( hi< y [Uii • is 6060^-2498, I

CKMAN

)ERY INC. i i NOV 96