How complete does the fossil record have to be?

Abstract

No science is based on complete knowledge, so the incompleteness of the fossil record is not necessarily a serious limitation in using palaeontological data to test geological or biological hypotheses. Completeness and adequacy of data are measured against sorne prior requirement which, in turn, should determine sampling strategy. Adequacy is considered for single samples, for detecting patterns by comparing samples, and for comparisons with neontology. Adequacy of single samples can be determined by sample size or by "growth of knowledge" curves. When comparing samples, standard sample sizes have serious disadvantages. Pattern detection requires appropriate density of samples. In stratigraphy using standard intervals has disadvantages. Bed by bed sampling is often better, but one sample per bed is usually inadequate. In comparing morphological data from the fossil record with more extensive neontologícai data, what matters is not that extra data are avaiiabie but whether or not they are necessary. Much neontological data is based on taxa identified morphologically anyway. Commonly problems with palaeontological data result as much from inadequate collection and recording as from inadequacies of the fossil record. We must be clear about our aims, and therefore what data are necessary, before data gathering begins.