ANIMAL NATURE OF DIATOME^. 395 



more. Now, this superficies merits the more considera- 

 tion because, referring it, as analogy requires, to the 

 primary surfaces, it offers an exception to the general 

 law that the transverse striae are wanting. For these are 

 very evidently continuous, and therefore may be com- 

 pared to those upon the lateral or secondary surfaces of 

 Denticula, of Odontidium, and Biatoma. We noticed 

 the same condition before, treating of Odontidium, in the 

 Syrinx annulata of Corda. But there remains .that other 

 ciaaracter of the primary surfaces which seems to have 

 the more organic importance, — the longitudinal division. 

 I have not seen the intermediate gradations, but I 

 have certainly seen two individuals, one superimposed 

 upon the other, and adhering closely, which might fairly 

 be regarded as proceeding from the deduplication {sdop- 

 piamento) of one individual. This brings new support to 

 the affinity of this genus with the Melosirce suggested by 

 Klltzing. That microscopical observations are to be 

 interpreted with the severe scrupulosity of critical logic 

 is well known to those who habitually use that valuable 

 instrument ; and the prudent caution will be more re- 

 quisite when we have before us a fact at variance with 

 many others. I ought not, therefore, to suppress a 

 doubt that occurred to me relative to the continuity of 

 these transverse striae over the primary surfaces. In 

 hundreds of individuals I have succeeded in obtaining a 

 front view of the margin, and in seeing it crossed by 

 those thick striae which correspond to the radii of the 

 lateral surfaces. These are among the microscopic ob- 

 jects which may be regarded as gigantic, and in respect 

 to which we may banish all suspicion of illusion. But 

 it may be supposed that in aU such cases I have had 

 before me only one of the lateral valves, and that the in- 

 terstitial ring was wanting. Individuals geminate through 

 an antecedent deduplication might have removed that 

 doubt. ■ But these are rare, and owing to the complica- 

 tion produced by the superposition of four similar trans- 

 parent valves, I could not safely decide whether the striae 



