262 BULLETIN 120, UNITED STATES NATIONAL MUSEUM. 



verf^ent races of separate species may more or less overlap in their 

 characters. Statistical studies and studies of the whole life-history 

 are probably essential to any adequate taxonomic treatment of a 

 goodly number of the " species " of Zellerlella^ Cepedea^ and Opalind. 

 I believe the present paper errs, if at all, on the side of too great 

 conservatism in the establishment of new species. 



But however uncertain may be some of the conclusions as to classi- 

 fication given in section 3, I hope the actual conditions found are 

 clearly presented. If the data are clear any one may exercise his 

 own judgment and preference as to the manner of expressing them 

 in a taxonomic system. It is too much a matter of the personal 

 judgment or even the individual taste of the student. I can claim 

 no more than this for some of the details of the taxonomy adopted 

 in this case. 



It may be well to consider what characters of the Opalinids are 

 the more trustworthy as a basis for taxonomic distinctions: 



Forvi of hody. — This is often quite various in a single infection. 

 Yet the character of the different shapes and their grouping about 

 a mean generally give the possibility of distinguishing infections of 

 one species from those of another. Yet these distinctions are very 

 difficult to put into words. A group of drawings is more effective 

 than many paragraphs of verbal description. One must be sure 

 to compare whole infections, and not merely individuals, or he may 

 be seriously misled. A daughter cell just come from longitudinal 

 fission will differ in shape from a daughter cell derived by trans- 

 verse fission, and each will differ from a full-grown individual. 

 When division is occurring rapidly, as in the presexual phases of 

 the life cycle, the daughter cells may not have time to restore the 

 characteristic shape between divisions, and extremes of divergent 

 form may result. We do not know much of the relative frequency 

 of cross and longitudinal fission, whether these are in constant pro- 

 portion for the family, or whether species differ in this regard, nor 

 do we knoAV whether there are racial or even individual differences 

 in this respect. General proportions, then, must be studied in whole 

 infections. Thickness of the body, and relative thickness of the 

 anterior and posterior ends of the body, often assist in demarcating 

 species. But there are many more special peculiarities of shape that 

 are safely and conveniently used in specific diagnosis, for example, 

 the presence or absence of posteriorly pointed individuals in the 

 infection, the character of such point when present, whether grad- 

 ually tapering or abrupt, sharp or rounded at the tip, straight or 

 bent. The character of the bend in the anterior end of the body and 

 the curvature of the body as a whole are also usable distinctions. In 

 general, while no single individual may be specifically distinctive 



