Although sponges are important components of benthic ecosystems of the Caribbean Sea, their diversity remained poorly investigated in the Lesser Antilles. By organizing a training course in Martinique, we wanted both to promote taxonomy and to provide a first inventory of the sponge diversity on this island. The course was like a naturalist expedition, with a field laboratory and a classroom nearby. Early-career scientists and environmental managers were trained in sponge taxonomy. We gathered unpublished data and conducted an inventory at 13 coastal sites. We explored only shallow water habitats (0–30 m), such as mangroves, reefs or rocky bottoms and underwater caves. According to this study, the sponge fauna of Martinique is currently represented by a minimum of 191 species, 134 of which we could assign species names. One third of the remaining non-identified sponge species we consider to be new to science. Martinique appears very remarkable because of its littoral marine fauna harboring sponge aggregations with high biomass and species diversity dominating over coral species. In mangroves, sponges cover about 10% of the surface of subtidal roots. Several submarine caves are true reservoirs of hidden and insufficiently described sponge diversity. Thanks to this new collaborative effort, the Eastern Caribbean has gained a significant increase of knowledge, with sponge diversity of this area potentially representing 40% of the total in the Caribbean Sea. We thus demonstrated the importance of developing exploratory and educational research in areas historically devoid of biodiversity inventories and systematics studies. Finally, we believe in the necessity to consider not only the number of species but their distribution in space to evaluate their putative contribution to ecosystem services and our willingness to preserve them.

The two main scientific tasks of taxonomy are species' delineation and classification. These two tasks are often treated differently, with classification accomplished by newly-developed phylogenetic methods, often based on molecular sequences, while delimitation of species is conducted by what is often considered to be an "old-fashioned" typological approach based on morphological description. A new "integrative taxonomy" has been proposed which maintains that species delimitation should be a multidisciplinary undertaking combining several independent datasets. Here we argue that the same principle is relevant to the classification of species. In the past 20 years, we assembled various datasets based on the external morphology, anatomy, cytology, spicule shapes, geography, reproduction, genetic sequences, and metabolomics of homoscleromorph sponges. We show how we used these datasets to describe new species of homoscleromorph sponges and to elucidate their phylogenetic relationships and their phylogenetic position within the phylum Porifera.