We used Principle Components Analysis (PCA), a multivariate method, to analyze the cranial size and shape of eight Neotropical cat species. Preliminary patterns of sexual dimorphism and geographic variation were considered qualitatively by interpreting scatterplots generated using a PCA on SYSTAT.

Size variation accounted for at least 80% of the variation between individual specimens except in Leopardus tigrinus (71.6%). In all species except L. tigrinus and L. pardalis, most non-size variation was due to canine diameter.

Leopardus wiedii, Puma concolor, Panthera onca, and Oncifelis geoffroyi males tended to have larger canines than females. L. wiedii from Mexico were smaller in general, while Panamanian individuals were larger. L. pardalis from Panama were larger and had greater breadth across canines and canine diameter than individuals from Texas, USA. P. onca from Argentina tend to be larger with larger canines, whereas individuals from Guyana were smaller with smaller canines. P. concolor from Argentina were larger than others. Among the small species (group 1), L. tigrinus was readily differentiated from the other three by size and shape. The remaining three species are remarkably similar in skull size and shape. L. pardalis and Puma yaguarondi are distinct in both size and shape. Finally, we found P. concolor and P. onca to be extremely similar in patterns of variation of size and shape of the skull.

INTRODUCTION

One of the most controversial, yet important disciplines of biology is community ecology. The relationships studied in this field are critical in the understanding of interspecific evolution. However, the complexity of the causes and effects of interspecific and intraspecific interactions found in nature make it extremely difficult to understand the ever-dynamic relationships behind the structure of the community. Therefore, universal theories and explanations are elusive and contentious.

According to the theory of competition, two species utilizing the same resources cannot coexist in the same place at the same time. Therefore, they must somehow differentiate or divide the spatial, temporal, or trophic resources they utilize. Brown and Wilson (1956) first coined the term character displacement, based on previous experiments by Vaurie (1950), which examined two species of nuthatches (Sitta neumayer and S. tephronota) (Brown and Wilson 1956). They described it as a situation in which species that are similar in allopatry are differentiated in sympatry.

Similarity is shown using a morphological, behavioral, or ecological character. Characters must have an effect on competition with other species in the community. In this case, character displacement, or lack there of, will be observed in the variance of a given morphological trait related to competition for food resources across a guild.

This portion of our study addresses patterns of variation between species and groups of species in the Neotropical cat guild. The results are preliminary, but provide necessary information to address some of the above-mentioned ecological questions.

METHODS

Our six measurements were: condylobasal length (CBL), zygomatic breath (ZB), maxillary tooth row (MTR), breath across canines (BAC), greatest canine diameter (GCD), and mandibular height (MANHT).

The eight species span four genera, Leopardus, Puma, Panthera, and Oncifelis, and can be divided into three size groups.

Leopardus tigrinus (Figure 1), Leopardus colocolo (Figure 2), Leopardus wiedii (Figure 3), and Oncifelis geoffroyi (Figure 4) form the smallest group (380-880mm (HB), 1.4 - 7.8 kg) (Emmons, 1997; Redford and Eisenberg, 1992; Reid, 1997; Wilson in litt.).

Puma yaguarondi (Figure 5) and Leopardus pardalis (Figure 6) fall into a medium-sized group (488-1000mm (HB), 9-11.5 kg) (Emmons, 1997; Redford and Eisenberg, 1992; Reid, 1997; Wilson in litt.).

Puma concolor (Figure 7) and Panthera onca (Figure 8) compose the last size group (860-1850mm, 85-100 kg) (Emmons, 1997; Redford and Eisenberg, 1992; Reid, 1997; Wilson in litt.).

Measurements were made to the nearest 0.01 mm using Mitutoyo dial calipers. Panthera onca and Puma concolor CBL measurements were made to the nearest millimeter using an anthropomometer. We found the amount of variance due to each factor based on the eigenvalue.

We then used component loading values to determine the characters controlling the size, shape, and variance of the skull. Invariably, the first Principle Component (PC1) far outweighed the second Principle Component (PC2) (Table 1), and accounts for size-related variation. PC2 then provides information on the next main source of non-size, or shape, variation. Sexual dimorphism and geographic variation patterns were also considered qualitatively by interpreting scatterplots generated using a PCA on SYSTAT.

RESULTS

The first principle component accounted for 72% of the variation in Leopardus tigrinus. That percentage was higher in all of the other species, ranging up to 91% in Puma yaguarondi. Individual character loadings tended to be uniformly high on PC1, indicating that this first principle component is a reflection of overall size variation in each of the samples.

Values for the second principle component ranged from 4% (P. yaguarondi)to 15% (L. tigrinus). Combined, the first two components accounted for 87-97% of the total variation in each species (Table 1).

The character loadings on PC2 are not only lower, as expected, but the variation in individual character loadings is much greater. Greatest canine diameter contributed the most to the second principle component in six of the species. Breadth across canines contributed the most to PC2 in the two exceptions, L. tigrinus and L. pardalis (Table 1).

In some cases a second character contributed heavily to PC2, in addition to one of the canine-related characters. This was the case for L. colocolo, for example, where zygomatic breadth contributed significantly to the second component.

In the case of L. pardalis, maxillary toothrow length and mandible height also loaded heavily on PC2, in addition to the two canine characters (Table 1).

Males tend to be larger than females in some species. For instance, the mean condylobasal length (an indicator of overall size) of male L. wiedii is 93.81, but the mean of CBL for L. wiedii females is 88.192.

We analyzed geographic variation by combining specimens from each country within the overall range of each species. In general, animals from southern localities tended to be larger than those from more northern ones. For example, L. wiedii individuals have a mean CBL of 100.514, but Mexican L. wiedii have a mean CBL of 86.99. Also, L. pardalis from Panama are larger and have larger canines than L. pardalis from USA, Texas, which are smaller and have smaller canines. (Panamanian CBL mean= =131.03, USA CD mean =43.02; Texas CBL mean=118.053, CD mean=38.441).

A PCA of the first size group (L. wiedii, O. geoffroyi, L. tigrinus, L. colocolo) showed that O. geoffroyi, L. colocolo, and L. wiedii are very similar to each other in the measured characteristics. In fact, they are not distinguishable by these characteristics. However, L. tigrinus was very easily distinguished both by size (78% variance), and shape (18%).

The characters most prevalent in factor 2 are condylobasal length, mandible height, maxillary tooth row, and greatest canine diameter (In order of decreasing values). The individuals in the second size group (P. jaguarondi and L. pardalis) are easily distinguished by size, canine diameter, and mandible height.

DISCUSSION AND CONCLUSIONS

So far, we have identified the most important characters for each species and group of species. We have also identified the degree of sexual dimorphism, and potentially interesting geographic populations. In addition, males and females seem to vary in canine diameter suggesting that the sexes may act as separate morphospecies (Dayan et al., 1990).

Slightly different canine sizes might help the cats avoid intraspecific competition, because they would target different sizes of prey. Canine size is primarily responsible for the high PC2 canine diameter score. In addition, canine size is likely to be acted on by competition, because it is directly related to an animal's ability to kill a prey item. The canine delivers the lethal bite to prey by separating the cervical and dorsal vertebrae (Leyhausen 1979). However, the difference seen in male and female canine diameter may also be due to display functions during mating (Dayan et al. 1990)

Next, we plan to use multivariate and traditional approaches to further investigate outlying populations and the factors controlling variance. This study provides useful preliminary information on the geographic variation of these cats. It serves as a baseline for longer-term community-wide character displacement and niche expansion studies.

Acknowledgements

The authors would like to sincerely thank National Science Foundation for supporting this work, Mary Sangrey for directing the RTP program, Linda Gordan and Bob Randall for their help during during data collection, and Ralph Chapman for his help with statistics.

References

Brown, W. L., Jr., and E. O. Wilson. 1956. Character displacement. Systematic Zoology 7:49-64

Dayan, T., D. Simberloff, E. Tchernov, and Y. Yom-Tov. 1990. Feline Canines: Community-
wide character displacement among the small cats of Israel. The American Naturalist
136:39-60.

Emmons, L.H. 1997. Neotropical Rainforest Animals. The University of Chicago Press,
Chicago, IL.

Redford, K. H. and Eisenberg, J. F. 1992. Mammals of the Neotropics, the Southern Cone. The
University of Chicago Press, Chicago, IL.

Reid, F. A. 1997. Mammals of Central America and Southeast Mexico. University of Oxford
Press, N.Y., N.Y.

Vaurie, C. 1950. Notes on Asiatic nuthatches and creepers. American Museum Novitates, No.
1472:1-39.

ENTOMOLOGY

Kathie John
"An Inventory of Leafroller Species of Moths of the Great Smoky Mountains National Park"

Abstract / Summary

VERTEBRATE ZOOLOGY - FISHES

Murilo Carvalho
"Revision of Hypostomus species of the streams in the upper Rio Parana basin on the State of Sao Paulo, Brazil (Siluriformes: Loricariidae)"

Abstract / Summary