Introduction

The species of the Cryptocoryne albida group are rather homogenous. Also other species have narrow, strap-shaped leaves, but they do not have a similar spathe, i.e. with a relatively long tube, and a spirally twisted limb of the spathe which has circular to elongate markings on the surface, lacks a collar, and has a more or less speckled inner surface of the kettle wall, the wall also having circular, alveolar depressions in the upper part.

The main taxonomic characters for the delimitation of related species in Cryptocoryne are the shape and colour of the limb of the spathe. Each species also has its own characteristic leaf-form, but in most cases this alone is not enough for a definite determination of a species. In the species dealt with here, the differences between the species lie in the exact form of the spirally twisted limb of the spathe and the differences in the brown, red, and purple colours of the surface, features that seem more quantitative than qualitative. These characters are difficult to distinguish in herbarium specimens, and only after study of living material has it been possible to clear up the limits of variation between the different species and in most cases possible also to distinguish dried specimens.

The shape of female flowers, stigmas, and olfactory bodies, the number and shape of male flowers, the colouration of the inner side of the kettle wall, and the colour of the valve closing the opening of the tube vary continuously within the group and have been found to be of little use in the delimitation of the species. The three species, as interpreted here, have adjacent but not (or only slightly) overlapping distribution areas (Fig. 6).

The Cryptocoryne albida group has been the subject of several different opinions (Engler 1920, De Wit 1971, 1983, Rataj 1975). Earlier (Jacobsen, 1976, 1977, and 1985), I put forward the view that the species of Cryptocoryne were narrow endemics. However, for two of the three present species this opinion is not tenable. Viewing the present three species, and their variation, in connection with the rest of the species of Cryptocoryne, it might be asked if they could just as well be considered one species ? With e.g. three subspecies ? If someone insists this is the case, I have no convincing arguments in my favor. The point of view does, to some extent, depend on the taxonomical characters used and the weight they are given.

Engler (1920), obviously aware of the difficulties connected with a sectional division of the genus, forwarded a "dispositio gregum" and placed C. unilocularis in Unitubulosae (because of the apparently lacking tube of the spathe), while C. retrospiralis and C. crispatula were placed in Bitubulosae.

De Wit (1971) maintained C. albida and C. balansae as separate species but included under C. retrospiralis the varieties costata, crispatula, and tonkinensis, but later (De Wit, 1983) accepted both C. albida (incl. C. costata), C. balansae, C. crispatula (incl. C. tonkinensis), and C. retrospiralis.

Rataj (1975) distinguished 7 species (plus C. usteriana Engl., from the Philippines), placed in two sections, viz. Sect. Retrospiralae Rataj and Sect. Costatae Rataj, a division for which I have found no basis (Jacobsen, 1977) considering that C. retrospiralis ssp. albida, which is conspecific with C. costata, and C. korthausae were placed in two different sections; the species recognized by Rataj were C. retrospiralis ssp. retrospiralis and ssp. albida, C. crispatula, C. bertelihansenii, C. tonkinensis, C. balansae, C. costata, and C. korthausae.

Recently another narrow-leaved species from India, C. consobrina Schott, has come into cultivation, and it has proven to have a chromosome number of 2n = 36 like the here mentioned species (Sivadasan, 1985), but it is not included in the present context as it has a rough surface of the limb of the spathe and a collar. C. cruddasiana Prain from Burma, which bears some resemblance to the three mentioned species, also has a rough surface of the limb of the spathe, and is likewise not included here. The well known C. spiralis (Retz.) Wydl. from India, also has narrow leaves, but the spathe is very different in lacking a tube and having a limb of the spathe which has transverse, irregular ridges.