A recent study into prescreening techniques to identify bones suitable for radiocarbon dating from sites known for poor or variable preservation (Brock et al. 2007, 2010a) found that the percent nitrogen (%N) content of whole bone powder was the most reliable indicator of collagen preservation. Measurement of %N is rapid, requires little preparation or material, and is relatively cheap. The technique reduces the risk of needlessly sampling valuable archaeological objects, as well as saving time and money on their unsuccessful pretreatment prior to dating. This method of prescreening is now regularly used at the Oxford Radiocarbon Accelerator Unit (ORAU). In the original study, linear regression analysis of data from 100 bones from 12 Holocene sites across southern England showed that when 0.76% N was chosen as a threshold, 84% of bones were successfully identified as containing sufficient (i.e. >1%) or insufficient (i.e. <1%) collagen for dating. However, it has been observed that for older, Pleistocene bones the failure rate may be higher, possibly due to the presence of more degraded, short-chain proteins that pass through the ultrafilters used in pretreatment, resulting in lower yields. Here, we present linear regression analysis of data from nearly 600 human and animal bones, antlers, and teeth, from a wide range of contexts and ages, to determine whether the 0.76% threshold identified in the previous study is still applicable. The potential of carbon:nitrogen atomic weight ratios (C:N) of whole bone to predict collagen preservation is also discussed.