Michael L. Kennedy

An assessment of indices of relative and absolute abundance for monitoring populations of small mammals

Abstract Knowledge of small-mammal population densities and species richness values are crucial to wildlife conservation and many ecological investigations. The relationship between estimates of absolute abundance and relative abundances of small-mammal populations is unclear. Therefore, from 1999–2001 measures of relative and absolute abundance of small mammals were compared across a terrestrial landscape using 3 vegetation communities (forest, early-successional field, and forest-field edge) in western Tennessee. We used a catch/unit effort index derived from transect sampling to estimate relative abundance and 2 estimates of absolute abundance, Mt+1 (the number of unique individuals captured) and Nest (a population estimate obtained from statistical models). These 2 estimates were derived from mark-recapture sampling on trapping grids. Using both sampling procedures, we conducted trapping with Sherman live traps during autumn, winter, and spring for 2 years. We tested the prediction that the pattern of total captures of small mammals across a landscape from a catch/unit effort index reflects the same pattern as those derived from Mt+1 and Nest. Also, we tested the prediction that the pattern of species richness of small mammals derived from a catch/unit effort index on transects corresponds with the species richness based on Mt+1 determined from grid sampling. We conducted comparisons of patterns using a Spearman rank correlation coefficient. Overall, we captured 988 small mammals, representing 9 genera and 11 species. The pattern of captures from the catch/unit effort index was proportional to those derived from Mt+1 and Nest. (P ≤ 0.001 in each correlation analysis; rs17 = 0.880 and rs17 = 0.810, respectively). The pattern of species richness derived from the catch/unit effort index was similar to those derived from the mark-recapture sampling (P ≤ 0.001; rs17 = 0.861). Capture frequencies across vegetation communities and within seasons from both transect sampling and grid sampling did not differ for 4 of 6 seasons (χ22 = 4.4329, P = 0.1090; χ22 = 18.2224, P = 0.0001; χ22 = 0.2129, P = 0.8990; χ22 = 1.6395, P = 0.4405; χ22 = 3.8688, P = 0.1445; χ22 = 7.1750, P = 0.0277). Results of this study suggest that measures of relative abundance provide patterns of population trends proportional to those derived from estimates of absolute abundance.

Morphologic Variation in the Raccoon, Procyon lotor, and Its Relationship to Genic and Environmental Variation

Morphologic variation in the raccoon, Procyon lotor , from the southeastern United States was assessed with univariate and multivariate analyses. The species was found to be sexually dimorphic with skulls of males generally more massive. This may relate to intraspecific territorial defense rather than intersexual resource partitioning. A matrix of correlation among skull characters was computed, and the first three principal components were extracted. These accounted for 84.2% of the variation in the character set among males and 86.5% among females. Three-dimensional projection of localities onto principal components showed that, for both males and females, large individuals occurred in western and northern states and the smallest animals occurred in the Florida Keys. Size gradations were seen with most raccoons similar to those in nearby geographic areas. No relationship was detected between genic and morphologic variation; however, a relationship was found between patterns of morphologic and environmental variation. This relationship was stronger in males. Canonical correlation suggested that trends in size of raccoons follow Bergmanns rule, but other arguments are against this. It appears that patterns of variation in size in raccoons have developed under a complex array of environmental pressures which cannot be individually dissected. It may be that the raccoon genome, while relatively static, allows for a high degree of morphologic plasticity in the face of environmental variation.

Interspecific Morphologic Variation in Kangaroo Rats (Dipodomys): Degree of Concordance with Genic Variation

Schnell, G. D. (Department of Zoology and Stovall Museum, University of Oklahoma, Norman, Oklahoma 73019), T. L. Best (Natural Sciences Research Institute, Natural History Museum, Eastern New Mexico University, Portales, New Mexico 88130), and M. L. Kennedy (Department of Biology, Memphis State University, Memphis, Tennessee 38152) 1978. Interspecific morphologic variation in kangaroo rats (Dipodomys): degree of concordance with genic variation. Syst. Zool. 27:34-48.-The 24 forms of kangaroo rats (Dipodomys) that have been recognized as species by one or more recent authors were studied using phenetic techniques. A total of 41 morphometric characters (4 skin, 16 skull, and 21 post-cranial) were measured for up to 10 males and 10 females of each species. The standardized average measurements were analyzed and then measurements were divided by principal component I projections based on unstandardized data to reduce the influence of overall size. Phenograms and three-dimensional principal component models were produced and the results compared with the findings of previous authors, including those from allozyme analyses. Several species pairs and groups involving a few species were stable throughout our studies and consistent with groups identified in phylogenetic, specialization, or other classificatory schemes. These include: compactus and ordii; elephantinus and venustus; agilis, paralius, peninsularis, and antiquarius; nitratoides and merriami; and ornatus and phillipsii. However, marked morphologic gaps (other than those based on size) are not present between many of the species groups and attempts to place species into such groups in classifications probably result in an indication of distinctness that goes beyond what is found in nature. In terms of interspecific comparisons, there was no association between phenetic groups based on morphologic data and those constructed using allozyme findings. When comparisons were made among previous classifications and our results, the allozyme data produced the most divergent classifications. Overall, the associations between our morphologic results and previous classifications were relatively weak, reflecting by and large the variance in arrangements of species groups, rather than major differences within groups. Given the relative indistinctness of morphologic groups, it seems unlikely that any one classification can accurately represent interspecific phenetic, cladistic, and/or phylogenetic affinities within the genus. [Phenetics, taxonomic congruence, Dipodomys, kangaroo rats, morphometrics, allozymes.] Significant effort has been expended by a number of investigators studying evolutionary mechanisms to determine the extent of genic variation in local populations, among populations, and among closely related species (see Lewontin, 1974, and Powell, 1975, for general reviews; Selander and Johnson, 1973, summarized data for vertebrates). Workers have attempted to investigate and elucidate a fundamental aspect of evolutionary theory-that race formation and eventually speciation occur by the conversion of genetic variation within populations to variation among populations (Lewontin, 1967). Prakash (1969) emphasized that to understand the speciation process we not only should have a knowledge of genetic differences between closely related species (or populations within a species), but also of the correlation between genic differences and differences based on other characteristics, such as those describing various aspects of morphology, ecological preference, and behavior. Relatively few studies have quantitatively compared interspecific differences in morphology against between species allozyme differences. Sokal (1973) indicated that allozyme variation is an unreliable indicator of classical morphologic (phenetic) variation, citing Sneath and

CLASSIFICATION BIAS IN DISCRIMINANT FUNCTION ANALYSES USED TO EVALUATE PUTATIVELY DIFFERENT TAXA

Abstract Discriminant function analysis has been used to assess morphological distinctiveness of putatively different taxa. We used randomizations of previously published morphological data for 2 subspecies of the coyote, Canis latrans frustror and C. l. thamnos, to quantify a recognized but previously unexamined bias in discriminant-function analyses that use resubstitution classification. This bias results in overestimates of intertaxon distinctiveness and is exacerbated when sample sizes are small. An alternative classification technique, jackknife sampling, is relatively unbiased.

Using Geographic Information Systems To Determine Home Range of the Southern Flying Squirrel (Glaucomys volans)

-Home ranges of 12 southern flying squirrels (Glaucomys volans) were studied in the Ouachita Mountains of west-central Arkansas during the winters of 1990-1991 and 1991-1992. Because traditional planimetric home ranges do not account for differences in topography, a Geographic Information System (GIS) which incorporated topography was also used to determine home range. Mean female and male planimetric and (topographical) home-range estimates were 3.8 ha (3.9 ha) and 7.8 ha (9.0 ha), respectively. Differences in home-range size were not statistically significant between sexes, however, home-range estimates of males were more variable. Topographic estimates were significantly larger than planimetric estimates. The GIS more clearly represents home range with respect to habitat characteristics in the context of topographic features and it should be used in future studies.

Ecologic Relationships of Co-Occurring Populations of Opossums (Didelphis virginiana) and Raccoons (Procyon lotor) in Tennessee

Opossums ( Didelphis virginiana ) and raccoons ( Procyon lotor ) were studied at two sites in Tennessee using mark-recapture techniques. Winter density, relationship of trap success with selected habitat variables, interspecific association, and the degree of mutual habitat utilization were determined. Density of opossums was estimated as one animal/9.9 ha at one site and one animal/44.5 ha at the second location. Abundance of raccoons was estimated as one animal/70.4 ha at one site and one animal/34.5 ha at the other. Number of opossums and raccoons was not significantly correlated with any of the habitat variables measured at either site. Interspecific association values for co-occurring populations were 0.111 and 0.267; these values were not statistically significant and indicated that opossums and raccoons occurred independently of one another in regard to spatial distribution. Both species exhibited a wide range of ecological tolerance for habitat.

GEOGRAPHIC VARIATION IN THE BLACK BEAR (URSUS AMERICANUS) IN THE EASTERN UNITED STATES AND CANADA

The pattern of geographic variation in morphologic characters of the black bear (Ursus americanus) was assessed at 13 sites in the eastern United States and Canada. Thirty measurements from 206 males and 207 females were recorded to the nearest 0.01 mm using digital calipers and subjected to principal components analysis. A matrix of correlations among skull characters was computed, and the first 3 principal components were extracted. These accounted for 90.5% of the variation in the character set for males and 87.1% for females. Three-dimensional projection of localities onto principal components showed that, for males and females, largest individuals occurred in the more southern localities (e.g., males-Louisiana-Mississippi, eastern Texas; females-Louisi- ana-eastern Texas) and the smallest animals occurred in the northernmost locality (Quebec). Gen- erally, bears were similar morphologically to those in nearby geographic areas. For males, correla- tions between morphologic variation and environmental factors indicated a significant relationship between size variation and mean January temperature, mean July temperature, mean annual precip- itation, latitude, and actual evapotranspiration; for females, a significant relationship was observed between morphologic variation and mean annual temperature, mean January temperature, mean July temperature, latitude, and actual evapotranspiration. There was no significant correlation for either sex between environmental factors and projections onto components II and III.

Morphometric variation in kangaroo rats (Genus Dipodomys) and its relationship to selected abiotic variables

Geographic patterns of morphometric variation, as represented by projections of samples on the first principal component, were obtained for 14 species of kangaroo rats (genus Dipodomys ) by using principal components analysis. The amount of variation accounted for by the first principal component ranged in males from 37.5 ( D. nelsoni ) to 87.1% ( D. californicns ) with a mean of 61.2%. These values for females were from 51.1 ( D. phillipsii ) to 83.9% ( D. californicus ) with a mean of 65.3%. Characters that consistently accounted for the most variation for this component were total length of skull, basal length, greatest width of skull, and width of mastoid bullae. Values for the projections of the samples on principal component I (frequently considered a body-size index) were compared to values of latitude, and annual values and seasonal ranges of temperature, precipitation, and actual evapotranspiration by using Pearsons product-moment correlation analyses. Of these abiotic variables, annual temperature was considered a slightly more-appropriate predictor of size variation. Beyond a tendency for size to increase with decreasing annual temperature (in accordance with Bergmanns Rule), there was little general pattern of correlation between morphometric variation and these abiotic features. Had fewer species of kangaroo rats been examined, conclusions would have depended on the taxa analyzed. It appears that abiotic parameters may effect morphometric variation in some portions of the distribution of a species, but may be superseded by different factors in other parts of that taxons range.

Hematology, intestinal parasites, and selected disease antibodies from a population of bobcats (Felis rufus) in central Arkansas.

Eight bobcats with adjoining or overlapping home ranges were examined. Hematological values were within previously reported ranges. Six bobcats demonstrated antibody titers to Toxoplasma gondii. Isospora spp., Taenia taeniaeformis, Spirometra mansoides, Physoloptera rara, Toxocara cati, Strongyloides spp., Trichurus spp., Capillaria spp., and Ancyclostoma spp. were found also in the animals examined. The mean number of parasite species per host was 4.1. All bobcats tested negative for serum antibodies to Rocky Mountain spotted fever (Rickettsia rickettsii). Two bobcats had titers ≤1:20 for tularemia (Francisiella tularensis), and two were positive for leptospirosis (Leptospira spp.).

Morphologic Variation in the Masked Shrew (Sorex cinereus) and the Smoky Shrew (S. fumeus)

-Morphologic variation in cranial characters of the masked shrew (Sorex cinereus) and the smoky shrew (S.fumeus) was assessed using multivariate statistical analyses. Matrices of correlations of characters were computed, and the first three principal components extracted. Three-dimensional projections of localities onto principal components indicated that larger male and female S. cinereus occurred in the northeastern portions of its range, along the coast of the Pacific Northwest in the United States, and in central Canada. In general, smaller body size was characteristic of animals occurring in southern latitudes. Selected environmental variables showed significant relationships with size and shape components of morphologic variation. Larger individuals of S. fumeus generally were found in the northern and central part of the distribution (West Virginia, Maryland, northern Virginia, and southern Pennsylvania). Smallest individuals occurred N of this central region and in the most southern localities. Body size of S. fumeus was correlated with selected environmental variables and interspecific relationships.