408 Annals New York Academy of Sciences 



Future Developments 



At present the data required to construct a classiiication by this method is 

 not available. Progress in diatom taxonomy depends upon its being obtained. 

 So far as ultrastructure is concerned, there are techniques for collecting the 

 data (cf., p. 399). The more difficult problem is to make it available. As can 

 be seen when Helmcke and Krieger's (1953, 1954, Helmcke et al., 1961) work is 

 compared with other published electron micrographs of diatoms, the only 

 method of reproduction that is really adequate is the making of photographic 

 prints. The cost of pubUcation of sufficient of these to cover most species of 

 diatoms would be prohibitive. The most feasible method of building up files 

 of micrographs will be by the exchange of duplicate prints between workers, or 

 their institutions, in much the same way as herbarium specimens are now ex- 

 changed. It is to be hoped that diatomists who have the facilities for electron 

 microscopy will enter into such a scheme. The desirability of stereomicro- 

 graphs has already been stressed, and also the necessity for adequate resolution. 

 A low magnification micrograph permitting verification of identity should 

 accompany those showing the detail of the ultrastructure, and adequate docu- 

 mentation of the origin of the specimen is essential. 



Collection of information about cell contents, on the other hand, depends 

 upon the development of a technique of preparation that will enable details of 

 both this and the valve structure to be seen in the same specimen. Now that 

 the phase-contrast microscope is available, this should be possible. I plan to 

 attempt it in the immediate future, but, in the words of the old proverb, two 

 heads are better than one, and there is more likelihood of success if others also 

 try to find a method. When such a technique is available, the same problem 

 as with ultrastructure will arise: the examination of large numbers of species 

 and the dissemination of the resulting information so that, as far as recent 

 diatoms are concerned, a volume of knowledge about cell contents comparable 

 to that about valve structure is available. Here again, the quantity involved 

 is Ukely to make publication impossible and the most satisfactory alternative 

 will probably be exchange of preparations. 



Not until we know the ultrastructure and the cell contents of most of the 

 species in a group will it be possible to consider whether, and if so in what way, 

 the taxonomy of the group can be remodeled on sounder lines. At present all 

 that is pertinent is to suggest that the methods of numerical taxonomy (Sneath 

 and Sokal, 1962) are likely to be of great use at that stage. As Sneath (1962) 

 has pointed out, at least 40 or 50 independent characters of each operational 

 taxonomic unit {e.g., individuals being classified into species or species being 

 classified into higher groups) need to be taken into consideration when using 

 the method of numerical taxonomy to construct a natural classification. If, as 

 has been normal practice, we rely on intuition rather than calculation to eval- 

 uate overall resemblance, our judgments are likely to be sound only if we take 

 note of a comparable number of characters. It is this which makes it essential 

 that diatom taxonomists should no longer confine themselves to studying 

 cleaned frustules under the light microscope, but should observe the cell con- 

 tents and the ultrastructure and make use of the information these provide in 

 their classifications. 



