The effects of parallax on geometric morphometric data

This study questions the use of parallactic images in the field of geometric morphometrics. Landmark data from both standard and distorted images of grids were compared to determine the differences between the data sets. The distorted images resulted from placing the camera too close to the specimen, thereby creating a parallax. Although a statistical significant difference existed between the standard and distorted grids, it was shown that because the parallax was consistent (due to a standardization of the camera set-up), the variation was small enough and the error was constant so that the data could still be used in subsequent geometric morphometric analyses. This could be important in geometric morphometric analyses, particularly when parallaxes in images are detected after the data collection has already begun and it is not plausible to reacquire the specimens. (C) 2002 Elsevier Science Ltd. All rights reserved.