Editorial Type:
Article Category: Research Article
 | 
Online Publication Date: 01 Dec 2001

Geographic Variation and Ontogenetic Change in the Diet of the Mexican Pacific Lowlands Garter Snake, Thamnophis validus

,
, and
Page Range: 1034 – 1042
DOI: 10.1643/0045-8511(2001)001[1034:GVAOCI]2.0.CO;2
Save
Download PDF

Abstract

The Mexican Pacific lowlands garter snake, Thamnophis validus, is found in aquatic habitats on the Pacific coast of México from the states of Sonora to Guerrero and in the Cape Region of the Baja California peninsula. We characterized the diet of Thamnophis validus through examination of stomach contents of museum specimens and from literature records. The diet of T. validus consists almost entirely of fishes and larval and adult anurans. There is strong geographic variation in the diet: snakes from uplands of Baja California contained only anurans, whereas the majority of snakes from mainland México that contained prey had eaten fishes. Mainland snakes show a pronounced ontogenetic shift from feeding primarily on anurans to feeding primarily on fishes. When mainland snakes are divided into three size classes, the intermediate class shows the most generalized diet (i.e., the most even distribution between fishes and anurans). We suggest that this nonmonotonic relationship between predator size and the level of specialization may be relatively common in snakes. Prey mass and the variance in prey mass both increase with increasing predator size in T. validus, as is the case for many carnivorous animals. Thamnophis validus apparently differs from other aquatic garter snakes (T. atratus, T. couchii, T. hammondii, T. melanogaster, and T. rufipunctatus) in feeding heavily on adult anurans and in the common use of aquatic open-mouth searching behavior.

Copyright: The American Society of Ichthyologists and Herpetologists
 Fig. 1. 
 Fig. 1. 

Logistic regression with prey type (fishes or amphibians) as the dependent variable and snake SVL as the independent variable, using data for mainland snakes only. The points represent individual observations (i.e., snakes containing either fishes or amphibians).−2 log (likelihood) for the model = 38.070, χ2 = 9.734, df = 1, P = 0.002. The curve represents the probability that an observation at that SVL is a fish, according to the logistic regression. A model including sex of the snake as an additional independent variable did not significantly improve the fit of the model, and sex was not significantly related to prey type


 Fig. 2. 
 Fig. 2. 

Ln-total prey mass (a) and ln-mean prey mass (b) plotted against ln-snake SVL. The lines are simple linear regressions. For statistical analyses, see Table 4


Received: 05 Sept 2000
Accepted: 28 Mar 2001
  • Download PDF