Are higher taxa described earlier or later than expected by chance?

The record of species descriptions within a taxonomic group represents the product of a sampling process. How useful such a record is in inferences about biodiversity and evolutionary patterns can depend on the nature of this sampling process. Here we describe a test for two potential biases: the novelty bias, the preferential description of species from higher taxa with relatively few previously described species; and the familiarity bias, the preferential description of species from already described higher taxa. At the heart of the test is the determination of whether the description of the higher taxa proceeded at a rate faster (the novelty bias) or slower (the familiarity bias) than expected by chance given the total number of species described for each higher taxon. Ambiguity may arise if there is uncertainty in the exact order in which new species and higher taxa were described. We apply the test to description records for eight groups of orchids. A novelty bias is detected in two to three cases and familiarity bias may be present in one case. The results are discussed in relation to perception of morphological complexity and the potential for human vision-based bias in biodiversity assessments. © The Natural History Museum 2009.