Review of the Scleractians as depth indicators (Cnidaria : Anthozoa). Before considering the use of fossil scleractinians as depth indicators, their living counterparts are presented in their morphological and ecological diversity. Jointly, the main environmental factors conditioning their geographical and depth distribution are pointed out. Intraspecific variation depending on environmental factors makes taxonomy of the group more complicate. Taxonomy is yet far from being satisfactory for many taxa, particularly in some areas. The terms hermatypic-ahermatypic as currently used by biologists and geologists, are not without ambiguities due to the original definition (WELLS, 1933) ; they are discussed together with complementary terms, which had been introduced in order to remedy to that ambiguity. In case of precise taxonomy and well recorded ecology, reference to living scleractinians permits extrapolation of paleoenviron-ments corresponding to fossil assemblages. It is important to distinguish between shallow water coral reefs of the tropical type and scleractinian constructions in deep water ; characters useful for this distinction are indicated. But the present contribution more specifically aims at demonstrating the usefulness of non-reefal scleractinians as indicators of depth (and temperature). According to WELLS (1967), the depth (and temperature) interval in common to the taxa present in a fossil assemblage can be estimated by reference to identical or related living forms. This provides an approximation for the paleoenvironment under analysis. It is essential to detect an eventual mixture of faunas issued from significantly different depth levels in order to avoid the pitfall of averaging incomparable data. At least for assemblages of a rather young geological past, extrapolations can be remarkably precise (examples of Pleistocene faunas from the Mediterranean). Howewer, the more remote in age the assemblages are, the less precise will be the results because more and more taxons in common with the Recent drop out. Analysis of morphofunctional adaptations, detached from any taxonomie basis, will provide, at best, only general information on the type of the'environment (deep or shallow) and of the substrate (soft or hard, stable or unstable).