Delimitation of species and identification of specimens to the species level continue to be difficult problems for practising entomologists, particularly those in tropical countries who often have no access to the holotype specimens or original literature of their local fauna. As a first step to the development of accurate Web-based species identification keys for Indian gryllids, we have examined the utility of morphological and song characters in correctly delineating species boundaries among 4 sympatric species of tree crickets of the Genus Oecanthus. Using a numerical taxonomic approach, phenetic clusters and ordinations were constructed on the basis of morphological and song characters. Quantitative and qualitative morphological characters were analysed independently and the results compared. The efficacy of the clustering and ordination techniques in species delimitation was examined by both internal and external allocation of individual specimens. Both the delimitation of species and the allocation of new specimens were 95 to 100% accurate using song or qualitative morphological characters. Quantitative morphological characters could also accurately delimit species, provided a large number of characters were used, irrespective of the specific characters chosen. For quantitative morphological characters, ordination was found to be more accurate than cluster analysis, both for delimiting species and in the allocation of new specimens.