Fossil specimens can be recovered easily only from exposed localities, so rock exposure area should represent a better proxy for rock availability than the frequently used outcrop (i.e., map) area. Data collected via remote sensing and GIS show that map area does not consistently correlate with exposure area in different regions. Proportional rock exposure is not geographically consistent and is influenced by a number of variables that are independent of outcrop area, including proximity to the coast, elevation, bedrock age, land use and lithology. These variables appear to be non-independent in their influence on rock exposure, and are not consistent in their effects across continents. The inconsistency in the correlation between outcrop and exposure area, and the variability in the influence of different factors on rock exposure, suggests that using outcrop area as a sampling proxy is poorly supported. The weaknesses in using outcrop area as a sampling proxy, highlighted by the lack of correlation with exposure area, suggest that a single accurate global sampling proxy may never be attained and it is premature to assume that paleodiversity curves can be corrected using such proxies. It is therefore preferable to work on a regional scale, comparing regional fossil collection data with a number of proxies representing all aspects of sampling. The lack of correlation between outcrop and exposure area suggests that the covariance detected between outcrop area and paleodiversity might be better explained by a common-cause model, and that geological megabiases may not have had as profound an effect on paleodiversity curves as previously thought.