Occasional Papers Museum of Texas Tech University Number262 10November2006 Morphometric Variation in Two Populations of the Cactus Mouse (Peromyscus eremicus) from Trans-Pecos Texas W. D. II, Richard Manning, Franklin Yancey, and Clyde Jones Abstract External and cranial measurements of specimens of the cactus mouse (Peromyscus eremicus) from two localities in the Trans-Pecos region of Texas (Harte Ranch addition to Big Bend National Park; Big Bend Ranch State Park) were analyzed statistically to determine morphomet¬ ric variation in mice from these two localities. Univariate and multivariate statistical analyses were performed on 50 specimens from each locality. Discriminant function analysis of cranial characteristics was used to classify individual specimens to each locality. Results suggest that there is minimal significant difference in mice from the two localities, but not enough to war¬ rant recognition of geographic variation, and there are no significant morphological differences between sexes. Key words: Big Bend National Park, Big Bend Ranch State Park, Peromyscus eremicus, univariate and multivariate statistics Introduction The cactus mouse, Peromyscus eremicus (Baird with rocky outcrops. Often the species can be taken 1858), is a rather common rodent found in desert scrub in man-made structures (Schmidly 2004; Yancey et al. habitats throughout the southwestern United States, 2006). from western Texas to southern California, and from northern and central Mexico to Baja California and some The biology of the cactus mouse (Peromyscus adjacent islands, both in the Pacific Ocean and the Sea eremicus) was summarized by Veal and Caire (1979). of Cortez (Veal and Caire 1979; Hall 1981). In Texas, In their treatment, they stated that considerable varia¬ the cactus mouse is found from Val Verde County on tion exists among the 15 subspecies of P. eremicus. the Edwards Plateau (Goetze 1998), to throughout most They cited the following selected external and cranial of the Trans-Pecos (Yancey 1997; Schmidly 2004). measurements in particular: total length, length of tail, The preferred habitat is desert scrub, especially areas length of hind foot, length of ear, greatest length of 2 Occasional Papers, Museum of Texas Tech University skull, and zygomatic breadth. Dice (1939) also re- this in mind, we undertook a study to examine geo- ported morphological differences amongst the sexes, graphical variation and sexual dimorphism of cactus with female P. eremicus significantly larger than males mice from two areas of the Trans-Pecos in West Texas, in the following measurements: body length, length of The subspecies reported here is P. eremicus eremicus ear, length of mandible, and bullar width of skull. With (Baird) (see Hall 1981; Manning and Jones 1998). Methods and Materials The source areas for our specimens were: 1) the length of hind foot (HF), and length of ear pinnae Harte Ranch addition to the northwestern part of Big (EAR). Cranial measurements included: greatest length Bend National Park (HR), Brewster County, Texas; of skull (GLS), zygomatic breadth (ZB), breadth of and, 2) Big Bend Ranch State Park (BBRSP), Presidio brain case (BBC), postorbital constriction (POC), depth County, Texas. Atotal of 50 adult specimens (25 males of cranium (DC), length of rostrum (LR), breadth of and 25 females) was selected from each locality for rostrum (BR), length of maxillary toothrow (LMAX), analysis. All of the specimens (most of which were breadth across upper first molars (BM1-M1), width collected by the authors) are deposited in the mammal of first upper molar (WUM1), mastoid breadth (MB), collection of the Natural Science Research Laboratory and length of mandible (LMAND). at The Museum of Texas Tech University. Univariate and multivariate statistical analyses External measurements were taken from speci¬ were preformed using S-plus software (1998) on a men labels, whereas cranial measurements were taken Dell computer. Our null hypothesis was that there is from museum material by one of the authors (RWM) no significant difference between external and cranial using digital calipers. Standard external measurements measurements of cactus mice at two localities from included: total length (TL), length of tail vertebrae (TV), Trans-Pecos Texas. Results Results of univariate analyses of characters are different between the two samples (see Table 1). presented, by locality, in Table 1 (external measure¬ Univariate tests suggest that three cranial characters, ments) and Table 2 (cranial measurements). Results depth of braincase (DC), length of maxillary toothrow of multivariarte analysis of variance (MANOVA) for (LMAX), and length of mandibular toothrow external characteristics suggest that there is a signifi¬ (LMAND), were significantly different between the cant difference between sample localities (p = 0.00035) two samples (see Table 2). but there is no significant difference between sexes (p = 0.47021) or sex by locality (p = 0.67538). Re¬ Discriminant function analysis (DFA), using cra¬ sults of multivariarte analysis of variance (MANOVA) nial characteristics, was used to classify individual for cranial characteristics suggest that there is a sig¬ specimens to either locality (or group). Multivariate t- nificant difference between sample localities test indicates a significant difference between groups. (p = 0.00005), but there is no significant differ¬ The analysis correctly classified 38 of 50 individuals ence between sexes (p = 0.46617) or sex by locality (locality 1) and correctly classified 42 of 50 individu¬ (p = 0.74785). als (locality 2). Overall, 80% of the specimens were correctly classified using DFA. A Jack-knife DFA (a Univariate tests (t-test) suggest that one external more robust test) correctly identified 71% of the speci¬ character, length of hind foot (HF), was significantly mens. Similar results, 76% correctly classified, were obtained on DFA using external features. Manning et al.—Peromyscus eremicus from Trans-Pecos Texas 3 Table 1. Descriptive statistics for external measurements of cactus mice from Harte Ranch (HR) and Big Bend Ranch State Park (BBRSP), Texas. Standard external measurements (in mm) included: total length (TL), length of tail vertebrae (TV), length of hind foot (HF), and length of ear pinnae (EAR). Sexes were combined because multivariate statistics between sexes were not significant. Numbers presented indicate arithmetic mean, one standard deviation, and sample size (n). NS = univariate non-signifi¬ cance, p = probability. Harte Ranch Big Bend Ranch State Park univariate significance TL 183.6, 7.5 (49) 183.5,8.6(49) NS TV 95.4, 6.1 (49) 95.8, 5.1 (49) NS HF 19.4, 0.8 (50) 20.0, 0.5 (50) p< 0.001 EAR 18.1, 1.0(50) 18.3,0.7 (50) NS Table 2. Descriptive statistics for cranial measurements (in mm) of cactus mice from Harte Ranch (n = 50) and Big Bend Ranch State Park (n = 50), Texas. Cranial measurements included: greatest length of skull (GLS), zygomatic breadth (ZB), breadth of brain case (BBC), postorbital constriction (POC), depth of cranium (DC), length of rostrum (LR), breadth of rostrum (BR), length of maxillary toothrow (LMAX), breadth across upper first molars (BM1-M1), width of first upper molar (WUM1), mastoid breadth (MB), and length of mandible (LMAND). Sexes were combined because multivariate statistics between sexes were not significant. Numbers presented indicate arithmetic mean and one standard deviation. NS = univariate non-significance, p = probability. Harte Ranch Big Bend Ranch State Park univariate significance GLS 25.0, 0.47 24.9, 0.71 NS ZB 12.6, 0.35 12.5, 0.35 NS BBC 11.7, 0.22 11.6, 0.26 NS POC 3.95, 0.11 3.97, 0.11 NS DC 8.99, 0.21 8.81,0.22 /?<0.001 LR 9.04, 0.27 9.08, 0.38 NS BR 3.70, 0.31 3.62, 0.31 NS LMAX 3.57,0.10 3.65, 0.11 /?<0.001 BM1-M1 4.69, 0.10 4.67, 0.14 NS WUM1 1.09, 0.03 1.09, 0.03 NS MB 10.9, 0.28 10.9, 0.29 NS LMAND 3.59, 0.09 3.66, 0.12 p< 0.001 Discussion, Summary, and Conclusions Although statistical analysis suggests there are mm; and length of mandibular toothrow, HR, 3.59 mm significant univariate differences between our site-spe¬ versus BBRSP, 3.66 mm), we do not feel that these cific samples in one external character (length of hind are strong enough differences to support the hypoth¬ foot, HR sample, 19.4 mm versus BBRSP sample, esis that there is marked dimorphism between speci¬ 20.00 mm) and three cranial characters (depth of brain mens from our two localities. Cactus mice from case, HR, 8.99 mm versus BBRSP, 8.81mm; length of BBRSP have slightly longer maxillary and mandibular maxillary toothrow, HR, 3.57 mm versus BBRSP, 3.65 toothrows and slightly deeper crania when compared 4 Occasional Papers, Museum of Texas Tech University to cactus mice from HR. Perhaps this is the result of Results of our multivariate tests further indicate that selection pressures for a more powerful bite to pro¬ there is no statistically significant difference between cess different, tougher, or larger seeds or other foods. sexes at either locality. There is no demonstrable sig¬ nificant sexual dimorphism between male and female Dice (1939) reported morphological differences cactus mice (P. eremicus eremicus) from our two between the sexes of P. eremicus, with females being Trans-Pecos Texas localities. significantly larger than males in the following characters: body length (which we reject), length of ear Other subspecies of cactus mice, especially some (which we reject), length of mandible (which we sup¬ of the eight insular taxa of P. eremicus (tiburonensis, port), and bullar width of skull (which we reject). We cedrosensis, avius, insulicola, polypolius, cinereus, do not feel this is a strong enough difference to sup¬ collatus, and pullus), are worthy of morphological in¬ port the hypothesis that there is marked sexual dimor¬ vestigation. phism between males and females in these samples. Acknowledgments We wish to thank Meredith Longoria for assistance running the statistical analyses. Literature Cited Dice, L. R. 1939. Variation in the cactus-mouse, S-Plus, 6.1 release. Copyright 1998, 2002. Insightful Cor¬ Peromyscus eremicus. Contributions of the Labo¬ poration. ratory of Vertebrate Genetics, University of Michi¬ Veal, R., and W. Caire. 1979. Peromyscus eremicus. Mam¬ gan 8:1-27. malian Species 118:1-6. Goetze, J. R. 1998. The mammals of the Edwards Plateau, Yancey, II, F. D. 1997. The mammals of Big Bend Ranch Texas. Special Publications, Museum of Texas State Park. Special Publications, Museum of Tech University 41:1-263. Texas Tech University 39:1-210. Hall, E. R. 1981. The mammals of North America. John Yancey, II, F. D., R. W. Manning, and C. Jones. 2006. Wiley and Sons, New York. Mammals of the Harte Ranch Area of Big Bend Manning, R.W., and C. Jones. 1998. Revised checklist of National Park, Brewster County, Texas. Oc¬ Recent land mammals of Texas, 1998. Occasional casional Papers, Museum of Texas Tech Univer¬ Papers, Museum of Texas Tech University 182:1- sity 253:1-15. 19. Schmidly, D. J. 2004. The mammals of Texas, revised. University of Texas Press, Austin. Manning et al.—Peromyscus eremicus from Trans-Pecos Texas 5 Addresses of authors: Richard W. Manning Clyde Jones 107 LBJ Cove Department of Biological Sciences and San Marcos, TX 78666 Natural Science Research Laboratory rmanning7@austin. rr. com The Museum of Texas Tech University Lubbock, TX 79409-3191 Franklin D. Yancey, II cjmajones@aol. com Madera Center of Reedley College 30277 Avenue 12 Madera, CA 93638 frank.yancey@scccd. edu