An (in)ordinate fondness for diversity: the fossil record of the Hexapoda

Since the first Phanerozoic global diversity analyses based on large datasets were published three decades ago, the meaning and biases of the fossil record have been strongly questioned. One key argument in favour of the robustness of compilations of the fossil record based on counts of taxa for defined time units is that large amounts of additional data do not change significatively the diversity patterns obtained from earlier compilations. No clade has such a profound impact on our perception of the evolutionary, taxonomic and ecological complexity of Phanerozoic life on Earth as the Arthropoda, and among them the astoundingly diverse insects. In spite of the vagaries, limitations and biases of the fossil record, the Devonian to Holocene distribution of fossil Hexapoda provides an unequalled model for macroevolutionary analysis. Following a conservative protocol widely employed in global Phanerozoic diversity studies, we focus on the distribution of fossil families of the Hexapoda, since this rank may represent an acceptable compromise between taxonomic stability and recognizability and paleobiological representativeness in respect to subordinate taxa (genera and species). Our inventory of the fossil hexapod families started as an updated version of Ross & Jarzembowski’s (1993) compilation in The Fossil Record 2 (Benton, 1993), based on a survey of the primary literature and searches of the Zoological Record and GeoRef (last accessed December 2005). For practical reasons, the stratigraphic framework follows that of Benton (1993). These two datasets were analyzed separately and contrasted with that of the recent monograph by a consortium of mainly Russian specialists (Rasnitsyn & Quicke, 2002). The aims of our study are threefold. Firstly, a comparison of the three datasets provides an opportunity to evaluate the stability of the patterns of the hexapod fossil record in terms of taxonomy and expanding knowledge. Secondly, we computed a number of diversity metrics based on orders and families for the three datasets, starting with the most intuitive component of diversity, taxonomic numerosity (unweighted taxonomic richness). Thirdly, we attempt to evaluate the fidelity of the fossil record of insects by assessing which taxa fail to become incorporated in it and by using stratigraphic congruence methods.