Introduction

Five species of broad-clawed shrews occur in Mexico and Central America. Two of these, Cryptotis griseoventris and C. goodwini, occur in close proximity to each other and can be extremely difficult to differentiate because they have few cranial or mandibular differences apart from size.

Pelage cannot be used to distinguish C. griseoventris and C. goodwini because tone and color vary geographically within species and because most older museum specimens have foxed over time. Post-cranial skeletal material, especially the humerus, has been used to distinguish shrews, but skeletons typically were not collected in the past.

Previously, size differences of the skulls were sufficient to distinguish C. griseoventris and C. goodwini. However, new specimens are more difficult to identify using skull size alone because they are intermediate between the two species. Dried skins prepared for museum collections have the forefoot bones left in them and can sometimes show distinct morphological differences between shrew species. We were interested to see if similar differences could be used to distinguish C. griseoventris and C. goodwini, two separate populations of C. griseoventris, and classify newly-collected specimens from Guatemala.

Materials and Methods

We used digital x-ray technology to obtain images of the forefoot bones of 13 C. goodwini from the type locality Calel, Guatemala, 8 C. griseoventris from the type locality in San Cristóbal, Mexico, 17 C. griseoventris from Todos Santos, Guatemala, 11 new specimens treated as unknowns, and 5 older specimens from other localities that were also treated as unknowns.

Bones and claws of the right forefoot of Cryptotis goodwini as viewed dorsally. Roman numerals designate individual digits. Abbreviations: ph., phalanx	Measurements used in this study: CL = length of claw, CW = width of claw, DPL = length of distal phalanx, DPW = width of distal phalanx, MPL = length of middle phalanx, MPW = width of middle phalanx, PPL = length of proximal phalanx, PPW = width of proximal phalanx, ML = length of metacarpal, and MW = width metacarpal.
Digital x-ray images of the right manus of Cryptotis. Original negatives were converted to positive images. A: C. goodwini from Calel, Guatemala (type locality); B: C. griseoventris from Todos Santos Cuchumatán, Guatemala; C: C. griseoventris from San Cristóbal de las Casas, Chiapas, Mexico (type locality).

Lengths and widths of phalanges, metacarpals, and claws were measured from digital images and recorded for digits I, III, and V of the forefoot of each specimen.

Mean, standard deviation, and range were calculated for each measurement. These data were analyzed using univariate and bivariate plots in Microsoft Excel and multivariate plots in Systat 11 to determine which measurements were most useful for differentiating known specimens of C. goodwini and C. griseoventris.

We then used these measurements to determine whether the two populations of C. griseoventris could be distinguished and whether they could be used to classify the unknowns.

To enhance existing differences that were otherwise difficult to distinguish with a single digit or bone, we used cumulative lengths and widths of a single digit or the sum of the same bone on three separate digits.

Results

C. goodwini averages longer bone length measurements for all bones in all digits with the exception of digit I, in which C. goodwini has a shorter CL than either population of C. griseoventris. However, most bone length measurements overlap greatly between the two species. Widths of phalanges and metacarpals from C. goodwini are almost always longer than those of C. griseoventris with very little overlap between the two species. PCA analysis showed considerable differences in digit III between species.

Metacarpals and phalanges of C. griseoventris from Todos Santos, Guatemala, are longer and wider for nearly all measurements of digits I, III, and V, than those of C. griseoventris from San Cristóbal, Mexico. Most of the variation between the two populations of C. griseoventris was manifest in CL, DPL, and DPW of digits III and DPL of digit 5.

The most useful measurements for classifying the unknown specimens were the widths of bones for all digits and DPL for digits III and V. Lengths of bones and claws were not as helpful because of the large amount of overlap between measurements. The most useful combination of measurements for distinguishing groups was the sum of DPW, PPW, and MW of digit I. PCA analysis of digit III was also useful in classifying unknown specimens.

Discussion

Identifying species using lengths of bones and claws was difficult because of the amount of overlap between the two species with the exception of DPL of digit 3 and 5. Widths were good identifiers of species because there was little overlap between species. Our analysis showed that though specimens of C. goodwini averaged longer and wider forefoot bones than those of C. griseoventris, C. goodwini tended to have proportionately wider forefoot bones and forefeet.

Specimens from Todos Santos, Guatemala, tended to have longer and wider forefoot bones and forefeet than those from San Cristóbal, Mexico. Measurements that are useful in discriminating between the two populations of C. griseoventris and C. goodwini are not as useful in identifying the unknowns. This may mean that these measurements vary more by population and are not good indicators of species.

The measurements that were very similar for both populations of C. griseoventris were the most useful in identifying unknown specimens of C. griseoventris and C. goodwini and were usually the widths of bones. This may mean that these measurements, usually widths, have the least amount of overlap outside of populations. Combined width measurements are even better identifiers of species because they augment differences between species that were otherwise difficult to see.

Conclusions

Our study investigated the ability of forefoot morphology to identify individuals of C. griseoventris and C. goodwini. We found that forefoot morphology was not only a good identifier of species, but it also could be used to distinguish discreet populations within species.

• Cryptotis goodwini has proportionally wider forefoot bones and forefeet than those of C. griseoventris.
  
  
• Widths of forefoot bones are more useful for distinguishing species than lengths because bone lengths tend to overlap between species.
  
  
• Combinations of measurements are most useful in distinguishing species because they enhance differences that would otherwise be hard to see.
  
  
• Measurements that are useful in distinguishing populations are not as useful in distinguishing species.
  
  
• Cryptotis griseoventris from San Cristóbal, Mexico, tended to have shorter and thinner forefoot bones than those from Todos Santos, Guatemala.

These findings may be able to tell us more about the biology, ecology, and distribution of broad-clawed shrews.

Literature Cited

Woodman, N. and Croft, D. A. 2005. Fossil Shrews from Honduras and their significance for late glacial evolution in body size (Mammalia: Soricidae: Cryptotis). Fieldiana: Geology (new series), 51:1-30.

Woodman, N. and Morgan, J. J. P. 2005. Skeletal morphology of the forefoot in shrews (Mammalia: Sorcidae) of the Genus Cryptotis, as revealed by x-rays. Journal of Morphology, 226:60-73.

Woodman, N. and Timm, RM. 1999 Geographic variation and evolutionary relationships among broad-clawed shrews of the Cryptotis goldmani - group (Mammalia: Insectivora: Sorcidae). Fieldiana: Zoology (new series), 91:1-35.

Acknowledgments

I would like to sincerely thank Neal Woodman for his support and patience throughout this project. I would also like to thank Jim Mead for use of the annex of his office and my fellow RTP interns for putting up with me for 10 weeks. This project was funded by the Bill and Jean Lane Internship Endowment.