The family is of special interest from the taxonomic stand point. On the basis of morphological differences, Schumann (1904) divided the Zingiberaceae into two subfamilies : the Zingiberoideae with many genera and Gostoidae with four genera namely, Costus (tropics), Monocostus (Peru), Dimerocostus (tropical America) and Tapeinochilus (Tabutu). The Costoideae differs from the rest of the family in showing vegetative and floral organization that is unique among the flowering plants. Therefore, morphologically, they form a natural group that deserves the status of a subfamily.
It is accepted that taxonomy should be based on evidences from as many disciplines as possible in the final evaluation of its units. But among the Costoideae, only Costus has been studied in detail. As the other genera of the Costoideae agree with Costus in anatomy, Tomlinson (1956) considers it reasonable to assume that the features that serve to distinguish Costus from the Zingiberaceae are equally diagnostic for the Costoidae as a whole. A list of diagnostic characters obtained from various disciplines are given in table I. These evidences also reveal that Costoideae is distinct and forms a natural group deserving the rank of a subfamily.
Certain-outstanding feature of the Costoideae are unique not only to the family, but also to the order Scitaminales as a whole. Perhaps, because of these Tomlinson (1956) has suggested that the Costoideae may possibly be given a family rank. However, although Tomlinson’s view is really encouraging, it is necessary to study the features of other genera of the Costoideae, with respect to various disciplines, in order to decide such an issue, instead of depending on the results obtained only on Costus.
The Zingiberaceae is further subdivided into three tribes Globbeae, Hedychieae and Zingibereae on the basis of the nature of lateral staminodes and the ovary. For the taxonomic consideration of tribes, there are no adequate data. However, it can be said that further research in the divergent fields on many more genera may only help in solving the taxonomic problems.
To a systematist the four families of the Scitaminales show an interesting floral organization, although they form a natural group because of certain morphological features common to all the families. A line of reduction in the number of stamens and ovules can be recognised in this order. The data available from various disciplines are too inadequate to verify the prevailing systematic arrangement of the families of the order. Nevertheless, the available embryological data consolidated in the table II unmistakably speak of a common ancestry of the families and their grouping in the order Scitaminales. At the same time certain distinguishing features of each family are probably suggestive of a family rank for each of them.