Nongeographic variation in 23 external and cranial characters was assessed for Microtus mexicanus. Univariate and multivariate statistical analyses were used to assess variation because of age and sex. Discriminant function analysis was used to summarize variation among three age groups. Adults and subadults of both sexes showed considerable overlap whereas juveniles were distinct. Multivariate analysis of variance and one-way analysis of variance for each character indicated that there were slight differences between the sexes and that the extent of sexual dimorphism varied among the three populations examined. The population from the United States had no significant sexual dimorphism whereas there was slight sexual dimorphism in voles from Mexico. Nongeographic variation within populations typically results from the effects of growth and aging as well as differences between the sexes. These phenomena may be studied preparatory to analyses of geographic variation (Choate and Williams, 1978; Wilhelm, 1982) to account for possible variation that is not associated with geography and to determine if discrete subsets of the population can be pooled in statistical analyses, thereby increasing the power of the tests (Zar, 1984). However, information on nongeographic variation is valuable for other reasons as well. For example, growth patterns can be used to estimate environmental stress (Brisbin et al., 1986). The presence or absence of sexual dimorphism can be indicative of the mating system (Kleiman, 1977; Ralls, 1977), niche partitioning between males and females (Schoener, 1967), or increased demands of large neonates on the females during reproduction (Ralls, 1976; Myers, 1978). Several studies have investigated sexual dimorphism in Microtus mexicanus (Wilhelm, 1982; Hamilton and Heidt, 1984; Hoffmeister, 1986). However, these studies utilized only univariate statistics in their analyses. Willig ' Present address: Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131. This content downloaded from 157.55.39.35 on Thu, 01 Sep 2016 05:36:59 UTC All use subject to http://about.jstor.org/terms 98 TRANSACTIONS OF THE KANSAS ACADEMY OF SCIENCE et al. (1986) showed that the use of multivariate techniques is the preferred method for examining differences between sexes and that univariate techniques can lead to erroneous conclusions. The purpose of this study was to use univariate and multivariate techniques to assess nongeographic variation in morphologic characters of Microtus mexicanus and to compare our results with those obtained from previous studies that employed only univariate techniques. MATERIALS AND METHODS Twenty-three morphologic characters were recorded from 290 specimens collected from three populations of Microtus mexicanus (see Appendix I for specimens examined). Nineteen cranial measurements were taken to the nearest 0.1 mm with a dial caliper and were chosen based on their repeatability and utility in describing size and shape of the skull. Measurements of external characters were recorded from specimen tags. These mensural characters were total length (TL), length of tail (LT), length of hind foot (HF), length of ear (LE), width of nasals (WN), rostral breadth (RB), interorbital constriction (IC), zygomatic breadth (ZB), prelambdoidal breadth (PLB), mastoidal breadth (MB), width of foramen magnum (WFM), height of foramen magnum (HFM), width of incisive foramen (WIF), length of incisive foramen (LIF), condylobasilar length (CBL), length of maxillary diastema (LXD), length of rostrum (LR), length of nasal (LN), length of maxillary toothrow (LMT), rostral depth (RD), greatest length of skull (GLS), and length of mandibular diastema (LND). The characters that were included in this study were described by Best (1978), Choate and Williams (1978), DeBlase and Martin (1981), or Hamilton and Heidt (1984) with the following exceptions: rostral breadth, greatest distance across rostrum anterior to zygomatic arches; width of incisive foramen, greatest distance across the incisive foramen; and rostral depth, least height from nasal to premaxillary. The shape of the skull was used for aging specimens because the continual growth of Microtus molars prohibits the use of tooth wear (Carleton, 1985). Specimens were assigned to one of three age categories: juvenile, subadult, and adult. Adults were characterized by a flattened braincase, angular skull, prominent mastoidal crest, prominent temporal ridges and evidence of reproduction as indicated on skin tags. Subadults were characterized by a more rounded braincase and a less angular skull than adults, mastoidal crest not prominent, and temporal ridges not developed. Finally, juveniles were characterized by a distinctly rounded braincase, no angularity to the skull, and mastoidal crest not developed. Analysis of nongeographic variation, including differences among the age groups and between sexes, was accomplished using the computer programs BIOSTAT I (Pimentel and Smith, 1986a) and BIOSTAT II (Pimentel and Smith, 1986b). The program MDA was used to perform one-way analyses This content downloaded from 157.55.39.35 on Thu, 01 Sep 2016 05:36:59 UTC All use subject to http://about.jstor.org/terms VOLUME 93, NUMBERS 3-4 99 Table 1. Nongeographic variation in differences among three age groups of 102 Microtus mexicanus from Ixtlan de Juarez, Mexico. Means for each age group and F-ratios resulting from one-way analysis of variance are presented for each character. Refer to Methods for definitions of characters.