Reports of the disappearance and decline of amphibians have stimulated interest in developing more accurate methods for sampling amphibian populations. Specific sampling methods are often suggested because of logistical concerns or simple tradition, but the validity of underlying assumptions for particular methods are rarely confirmed. This study examined potential problems and biases with quantitative enclosure sampling of larval amphibians. Results indicate that movement within a pond by an investigator does not reduce the number of captured tadpoles any more than does approaching a sampling station from the shore. Disturbance created in taking a sample does not appear to reduce number of tadpoles captured in subsequent samples taken 5 m or even 1 m away. Net type can be critical in how efficiently animals are removed from the enclosure, but the effect of net type on capture efficiency may vary with the amount of vegetation present. Investigators should assess the validity of assumptions for enclosure sampling before using this technique to draw inferences about populations of larval amphibians.