Information theory deals with the relative frequencies of nominal classes by treatment of average uncertainty,

[EQUATION]

where p_i is the relative frequency of the i^th class, connected with observation of the system. Applied directly to proportions of the taxa in a collection, the equation yields a diversity measure. One may then generate an equitability measure E=s^prime/s, where s^prime is the number of taxa necessary to yield the observed diversity if the proportions of taxa were random and s is the observed number of taxa. Applied to foraminifer data from Sabine Lake, La.-Tex., diversity/equitability parameters define salinity gradients more clearly than the presence of particular taxa. Similarly, where applied to invertebrate fossils from the Mississippian of Scotland, these parameters make it possible to subdivide a transgressive sequence in finer detail than an analysis of ta onomic composition. Interpretations in terms of community structure are not justified, but empirical treatment of contemporaneous and successional patterns appears to be a useful paleoecologic tool.

End_of_Article - Last_Page 706------------