Although the landscape matrix is increasingly being incorporated into spatial-ecological population studies, little consideration is being given to the likely possibility that patch quality is confounded with the composition of the matrix that surrounds a patch. For example, the nutritional quality of host-plant patches to an herbivore may be highly correlated with matrix composition, consequently obfuscating the importance of the matrix to interpatch dispersal. From a literature survey of the effects of the matrix on herbivore movement among host-plant patches, we found that 58% of the studies (7/12) failed to experimentally or statistically isolate the effects of the matrix from potential patch-quality effects on dispersal. Most studies consisted of mark-recapture experiments in natural landscapes where the matrix was not manipulated and patch quality was ignored. Of the few studies that evaluated the relationship between matrix composition and patch quality, all of them (4/4) found that these two landscape factors covaried. These data suggest that in most matrix studies, effects of the matrix on dispersal may be wholly, or in part, due to underlying differences in plant quality. We provide recommendations for isolating matrix from patch-quality effects on dispersal in both experimental and natural landscapes.
Keywords: connectivity, dispersal, landscape ecology, matrix, metapopulation, patch quality