wi a Ne ‘ i ‘ 4 bd ' FC Sys | | i Re Sat ‘ Pay! sy rn ~é % f é { ee \ Sy -_ (a

@H PI3B4X Sieh

JSeum

S, listory D ” 2 sity 2 g 7

, CT 06511

Ss —

(Received 2 September 1986) Abstract

Nine skulls of the rare Oryzomys argentatus are compared to 109 skulls of the six races of O. palustris. Mahalanobis distance is greater between O. argentatus and all Floridian forms of O. palustris than the Floridian forms are from each other. In a canonical discriminant analysis, two models grouping O. argentatus with one or both of the insular races of O. palustris (sanibeli and planirostris) were shown by the Roy’s Greatest Root statistic to fit the data less well than a model in which O. argentatus was regarded as distinct. A one-way ANOVA and Duncan's Multiple Range Test on the variation in nasal bone proportions show that there are two significantly different groups of these Oryzomys (p < 0.05): all O. palustris together and O. argentatus alone. We hypothesize O. argentatus originated on the Lower Keys in the late Sangamon and underwent selection for character divergence in sympatry with O. palustris during the Wurm.

© Copyright 1986 by the Peabody Museum of Natural History, Yale University. All rights reserved. No part of this publication, except brief quotations for scholarly purposes, may be reproduced without the written permission of the Director, Peabody Museum of Natural History.

ISSN No. 0-912532-00-9

Postilla Number 198

30 December 1986

Relationships of the Silver Rice Rat Oryzomys argentatus (Rodentia: Muridae)

Numi C. Goodyear and James D. Lazell, Jr.

Key Words

Oryzomys argentatus, Florida Keys, speciation, morphometrics, discriminant analysis, silver rice rat.

Introduction

Oryzomys argentatus, the silver rice rat, was originally described on the basis of two specimens (Spitzer and Lazell 1978). The work was criticized by Humphrey and Barbour (1979) and Barbour and Humphrey (1982) who felt that the use of ratios was not appropriate, and that the species was probably invalid. Because of a lack of data they did not substantiate these claims; they stated that the issue was moot because the taxon was probably extinct. Since, we have found that O. argentatus is extant on at least nine of the Lower Florida Keys (Goodyear, in press). Because the original sample size was too small for a statistically significant analysis of differences between O. argentatus and O. palustris, we present herein additional data corroborating our earlier results. We compare nine silver rice rat skulls with 109 skulls from the six races of O. palustris. The possibility of a close relationship between O. argentatus and the other insular forms, O. p. sanibeli and O. p. planirostris is especially considered. Two skulls are shown in Figure 1.

2 Silver Rice Rat

Postilla 198

Fig. 1

Skulls of male rice rats in dorsal view. Right, Oryzomys argentatus, YPM 4667, Raccoon Key, Monroe Co., Florida. Left, Orzyomys palustris sanibell, YPM 4670, Sanibel Island, Lee Co., Florida. Bar = 1 cm.

Specimens Examined and Related Abbreviations

Abbreviations for museums in which

specimens are housed are as follows:

AMNH American Museum of Natural History, New York

ANSP Academy of Natural Sciences, Philadelphia

FSM Florida State Museum, Gainesville

NCSM North Carolina State Museum, Raleigh

MCZ Museum of Comparative Zoology, Harvard University

USNM US National Museum of Natural History, Smithsonian Institution

YPM Peabody Museum of Natural History, Yale University

Not all specimens could be used for all

characters.

Oryzomys argentatus. YPM 4664-9; USNM 514994-5; AMNH 256405-6; FSM 16366.

3 Silver Rice Rat

O. p. palustris. YPM 4406-7; ANSP 11870, 11875; MCZ 1527-8, 2687, 2689-90, 56111-2, 5114, 5121, 5127, 5886, 6454, 6456; USNM 117384, 286831, 71368; AMNH 91146; NCSM 301-2, 472, 484-3, 491.

O. p. texensis. MCZ 2701-4, 2712-3, 2715-7, 2874, ANSP 14439; USNM 43299-300; AMNH 136499-500.

O. p. coloratus. MCZ 4454-59, 4461-2, 4465-9; USNM 71354-5, 73747-9, 228418, 228420-22, 228425, 228427.

O. p. natator. MCZ 3056-60, 7047, 7049, 7051, 7128, 7133-4, 7241-2; USNM 1090, 2250, 2253, 3872, 6172-4, 6176, 6183, 64061-7, 64069, 64071, 78705, 142693, 142697, 142748, 142811, 163993, 163995.

O. p. sanibeli. YPM 4670-2; USNM 301534.

O. p. planirostris. USNM 301533.

Methods

Nine O. argentatus, 16 O. p. texensis, 25 O. p. palustris, 24 O. p. coloratus, 39 O. p. natator, four O. p. sanibeli, and one O. p. planirostris skulls were used in the analysis. Variables used were condylobasal length, zygomatic breadth, nasal length, and nasal width. In the original description of the species the ratios of nasal length/nasal width and condylobasal length/zygomatic breadth were used to discriminate O. argentatus from all subspecies of O. palustris. Pelage color was not used as a variable because this character is so definitive that it would have unfairly weighted the data. In a canonical discriminant analysis—Statistical Analysis System (SAS) CANDISC procedure—we used only the metric data and no ratios to generate the models. In Model | we determined the Mahalanobis distance between all seven groups—the six races of O. palustris and O. argentatus. In Model Il O. argentatus was lumped with the insular rice rat O. p. sanibel. In Model Ill O. p. sanibeli, O. p. planirostris and O. argentatus, the three insular taxa, were lumped. Roy's Greatest Root, a summary statistic not affected by the number of classification groups, was used to determine which model best fit the data.

Postilla 198

The skull measurements were also examined using the SAS DISCRIM procedure with and without the ratios. Last we performed a one-way ANOVA and a Duncan's Multiple Range test on the ratio of nasal length/nasal width between the seven taxonomic groups.

Results

The Roy's Greatest Root, for which larger values mean better fit, was largest for Model | (0.82872), smaller for Model Il (0.814489), and smallest for Model Ill (0.749208). This indicated that it was better to view O. argentatus as a separate taxon, distinct from the other insular forms.

The Mahalanobis distances between O. argentatus and the races of O. palustris (from Model |) are shown in Table 1. The values show that the O. argentatus centroid is farther from all other groups than they are from each other with the exception of one pair, O. p. sanibeli and O. p. texensis which seem quite different from each other. However, the distance between O. argentatus and O. p. texensis is the greatest of all. Canonical variables 1 and 2 are plotted for Model | in Figure 2. Though the program attempted to maximally separate all seven groups, only the O. argentatus lie distant from the main mixed cloud of O. palustris subspecies. One individual O. argentatus lies within the O. palustris cloud. This was a captive female runt, which died at three months, weighing only 31 g (a normal female weighs 60 to 70 g). It had badly recurved teeth and did not grow at the rate of its litter mates. It died retaining the juvenile skull proportions which caused the misclassification. In the five discriminant analyses performed it was consistently grouped with O. palustris. No O. palustris was ever grouped with O. argentatus.

The case of overlap should be disregarded, as in Vogt and McCoy (1980), as the result of one teratogenic individual. However, cranial characters in combination with coloration do render every O. argentatus unequivocally distinct from every O. palustris.

4 Silver Rice Rat Postilla 198

CANONICAL VARIABLE 2

-4.5 -2.5 m0) 5S 1:5 CANONICAL VARIABLE |!

Fig. 2

Canonical variables one and two (Model |: see text) for seven North American rice rats (Oryzomys) using four cranial characters. Solid circles = O. argentatus; open circles = O. palustris palustris;

X = O. p. texensis; + = O. p. coloratus; open triangles = O. p. natator, solid triangles = O. p. sanibeli, * = O. p. planirostris.

Using the DISCRIM procedure with no sanibeli and O. p. planirostris 0%. Predictably, ratios, the program correctly classified with the ratios used instead of the raw data, individuals as follows: O. argentatus, 79%; O. the program correctly classified more O.

p. natator, 77%; O. p. texensis, 63%; O. p. argentatus, 89% (i.e., all but the runt noted

palustris, 40%; O. p. coloratus, 25%; O. p. above); O. p. natator, 85%; O. p. palustris,

5 Silver Rice Rat

Table 1

Postilla 198

Mahalanobis distance values between seven North American Oryzomys. The distance values are in the up- per triangle, and probability values for greater distance are in the lower triangle. O.p.p. = Oryzomys palus- tris palustris, O.p.t. = O. p. texensis, O.p.c. = O. p. coloratus, O.p.n. = O. p. natator, O.p.s. = O. p. sanibe-

li, O.p.pl. = O. p. planirostris, O.a. = O. argentatus.

SPECIES O.P.P. O.P.T. O.P.C. O.p.p. = 1.09 12 O.p.t. 4.37 = 2.28 O.p.c. 6.61 18.26 —

O.p.n. 8.03 29.24 2.69 O.p.s. 0.66 1.58 0.23 O.p.pl. 0.01 0.03 0.04 O.a. 5.80 8.92 4.28

32%; O. p. coloratus, 21%; O. p. texensis, O. p. sanibeli and O. p. planirostris 0%. The only O. argentatus misclassified was the female runt, YPM 4664.

In the one-way ANOVA on the ratio of nasal length divided by nasal width, there was a highy significant difference between O. argentatus and O. palustris (pb < 0.01). A Duncan's Multiple Range test showed that there were two significantly different groups: one consisted of O. argentatus alone; the other contained all the subspecies of O. palustris (p < 0.05).

Discussion

Since it was originally described (Spitzer and Lazell 1978) the validity of O. argentatus has been repeatedly challenged (Humphrey and Barbour 1979; Barbour and Humphrey 1982). Because the species is so rare (Goodyear, in press) collection of information has been slow. Research on aspects of the natural history and the results of this taxonomic re-evaluation show that the animal is indeed as distinctive as originally thought. Radiotelemetry of individuals (Spitzer 1983) showed that, unlike their closest geographic relative, O. p. coloratus, they are primarily salt marsh inhabitants that can have home ranges 10 to 100 times larger than one might expect for an animal of their size and guild. Their pelage is distinctive: always the silver-grey of the limey Florida Keys’ mud regardless of condition of

O.P.N. O.P.S. O.P.PL. O.A. 1.18 Zier. 1.10 3.16 2.09 2.85 Ho) 3.37 0.68 1.28 1.98 2.68 — 1.47 1.46 255 0.20 = 2.71 2.25 0.01 0.36 = 2.86 2.68 7.68 16.15 —

molt or diet. The coats of rice rats from temperate Florida tend to be grey-brown in juveniles (because of shorter guard hairs) and develop red hues when animals mature. Humphrey et al. (1986) raise the point that pelage is affected by ‘‘chemicals and light in the environment and by fumigants in museum cabinets.’’ This may cause some of the variation they observed in specimens of the Sanibel and Pine Island rice rats. We find that silver rice rats raised in the laboratory on Purina Dog Chow are identical in all respects to those captured wild, and that our museum specimens appear unchanged.

The silver rice rat's distinctive features may be due to an early separation from the mainland stock of Oryzomys and a subsequent period of sympatry when character divergence occurred. Oryzomys is known to be a successful colonizer of oceanic islands: the only rodent to have naturally reached the Galapagos Archipelago over more than a thousand kilometers of ocean (Heller 1904). We suspect that Oryzomys arrived in the Florida Keys as waters of the Sangamon Interglacial receded and exposed the islands (Lazell 1984; Goodyear, in press). We calculate the probable earliest exposure of land in what are now the Keys to be about 75 000 years BP from the curves provided by Morris et al. (1977). This date is corroborated by the climatic curves of Brunner (1982). If they arrived over water, differentiation may have begun before the Wurm reconnected the islands with the mainland. During the transition

6 Silver Rice Rat

from interglacial to glacial, and in the more recent reversal of that transition, Florida Bay was largely freshwater swamp, continuous with the Everglades; the oolitic Lower Keys were connected by mangrove swamp (Hoffmeister 1974). At these times, flanking the 65 000 year glacial maximum, the mainland and the Florida Keys rice rats may have been sympatric. This could explain the degree of character divergence seen in osteology, pelage, behavior, and habitat preferences that persist today after their re-isolation.

While the ecological significance of the elongate nasal bones of O. argentatus is not apparent to us at present, the other three characters seem the very sorts which selection for character divergence might produce. While O. argentatus and the distant forms of O. palustris utilize salt marsh (Spitzer 1973), the proximate form O. p. coloratus seems confined to freshwater areas. We have trapped extensively for coloratus in the Everglades and Upper Keys and have never taken it in brackish or salt habitats. The vast home ranges documented for O. argentatus indicate a very different foraging strategy for this species compared to any subspecies of O. palustris (Goodyear, in press).

Finally, pelage color is far more different between O. argentatus and its nearest neighbor O. p. coloratus than between argentatus and geographically remote forms like nominate palustris and texensis. The Everglades O. p. coloratus is richly patterned in russet and ochraceous tones; O. argentatus is overall chinchilla grey to ash white. The grey-brown nominate palustris and texensis,

Literature Cited

Postilla 198

geographically remote from O. argentatus, most closely resemble it in color. The insular form O. p. sanibeli has a dark grey-brown dorsum shading to warm fawn-brown on the sides (Hamilton 1955; YPM specimens); the ventral hairs of sanibeli are cream-white tipped with plumbeus bases. The ventral hairs of O. argentatus are ashy to the bases.

Pelage color differences are often strong evidence of character divergence in small mammals. For example, two species of mice, Peromyscus leucopus and P. gossypinus, can scarcely be separated on any consistent characters where far removed from each other, but sharp color (and hind foot) distinctions are seen where their ranges overlap (Webster et al. 1985). A similar case is described by these authors involving the shrews Blarina brevicauda and B. carolinensis.

Character divergence between presently allopatric, isolated species resulting from past contact was discussed theoretically by Williams (1969). Lazell (1972, p. 103-104) provided further discussion and an Antillean example.

Acknowledgments

Our field work was supported in part by The Conservation Agency, Friends of the Everglades, U.S. Fish and Wildlife Service, Earthwatch, The Explorers Club, The Nature Conservancy, and State of Florida Department of Parks and Recreation. The University of Rhode Island provided computer time. George Garrett and numerous others assisted us in the field.

Barbour, D. B. and S. R. Humphrey. 1982. Status of the silver rice rat (Oryzomys argentatus). Florida

Sci. 45, 2:106-112.

Brunner, C. A. 1982. Paleoceanography of surface waters in the Gulf of Mexico during the late

Quaternary. Quat. Res. 17:105-119.

Goodyear, N. C. In press (1987). Distribution and habitat of the silver rice rat, Oryzomys argentatus. J.

Mammal. 68.

Hamilton, W. J. 1955. Two new rice rats (genus Oryzomys) from Florida. Proc. Biol. Soc. Washington

68:83-86.

7 Silver Rice Rat Postilla 198

Heller, E. 1904. Mammals of the Galapagos Archipelago, Exclusive of the Cetacea. Proc. Calif. Acad. Sci. So), FZ.

Hoffmeister, J. E. 1974. Land from the Sea: the Geologic Story of South Florida. U. Miami Press, Coral Gables, Florida, 143 p.

Humphrey, S. R. and D. B. Barbour. 1979. Status and habitat of eight kinds of endangered and threatened rodents in Florida. Spec. Sci. Rep. No. 2. Office of Ecological Services, Fla. State Mus., 17 p. Humphrey, S. R., R. W. Repenning and H. W. Setzer. 1986. Status survey of five Florida mammals. Tech. Rep. No. 22, Fla. State Mus., Gainesville, 38 p.

Lazell, J. D., Jr. 1984. A new marsh rabbit (Sylvilagus palustris) from Florida’s Lower Keys. J. Mammal. 65) 126-33:

1972. The anoles (Sauria, Iguanidae) of the Lesser Antilles. Bull. Mus. Comp. Zool. (Harvard) 143, 1:

ao:

Morris, B., J. Barnes, F. Brown and J. Markham. 1977. The Bermuda marine environment. Bermuda Biol. Stn. Spec. Publ. 15. iv + 120 p.

Spitzer, N. C. 1973. Rice rat's world. Man and nature, Mass. Audubon Soc., March:24-26.

1978. Cudjoe Key rice rat, p. 7-8. In J. Layne (ed.), Rare and Endangered Biota of South Florida. Vol. |, Mammals. Gainesville, Univ. Presses Florida, 52 p.

1983. Aspects of the biology of the silver rice rat, Oryzomys argentatus. Unpubl. M.S. thesis, Univ. Rhode Island, Kingston, 100 p.

Spitzer, N. C. and J. D. Lazell, Jr. 1978. A new rice rat (genus Oryzomys) from Florida’s Lower Keys. J. Mammal. 59, 4:787-92.

Vogt, R. C. and C. J. McCoy. 1980. Status of the emydine turtle genera Chrysemys and Pseudemys. Ann. Carnegie Mus. 49, 5:93-102.

Webster, W., J. Parnell and W. Biggs. 1985. Mammals of the Carolinas, Virginia, and Maryland. Univ. North Carolina Press, Chapel Hill. vii + 255 p.

Williams, E. E. 1969. The ecology of colonization as seen in the zoogeography of anoline lizards on small islands. Quart. Rev. Biol. 44, 4:345-89.

The Authors

Numi C. Goodyear. Doctoral candidate, Department of Zoology, University of Rhode Island, Kingston, RI 02881.

James D. Lazell, Jr. Curatorial Affiliate for Recent Vertebrates, Peabody Museum of Natural History, Yale University, and President, The Conservation Agency, 6 Swinburne Street, Jamestown, RI 02835.

TN 3 9088 01206 5181

.

i yee