Earnhardt et al. (1995) ask (1) if the census data of ISIS are accurate enough for North American regional zoo collection planning, and (2) can the "raw" zoo records presently available from ISIS substitute for a formal studbook and be used directly for genetic and demographic analysis. To check census, they compared ISIS Abstracts with summary SSPO reports in AZA's Annual Report on Conservation and Science (AZA-ARCS), assumed that they represented counts for the same population, and found only a moderate correlation (r = 0.835). We repeated their analysis and closely examined the 15 species with apparent census differences greater than 50. We find that most of the census differences arise because the two sources are not counting the same population. When nonzoo specimens are removed from census counts for just these 15 species, the overall correlation between ISIS and AZA-ARCS (for the full 68 species) rises to 0.974. We conclude that ISIS census data are reasonably accurate for the North American zoo population. Earnhardt et al. (1995) find "raw" ISIS data unsuitable for genetic and demographic species management. This conclusion came from comparing 10 ISIS species datasets to the respective North American regional studbooks. We re-examined the two taxa with the most reported discrepancies. For gorillas, studbook numbers on ISIS are indeed confused-in large part because multiple, conflicting studbook numbers are being assigned by different studbook keepers to individual specimens. For cinereous vultures, the large number of birth date discrepancies arise from records intentionally altered in the studbook, which substitutes amval date for estimated or unknown birth dates of wild-caught specimens. Whereas this studbook's minimuni age convention may well be useful for certain analyses, we find that the ISIS data accurately reflects the holding zoo's records. We generally agree with Earnhardt et al. (1995) that at present, "raw"

No abstract