ANIMAL NATURE OF DIATOMEiE. 419 



measurements in general a little exaggerated; not so 

 much, however, as those of Ehrenberg ; and the figures 

 (in Kiitzing's plates) are rather below the truth, which 

 we may explain on the supposition that for the sake of 

 economising space — a bad economy — he has not chosen 

 to figure the larger individuals. I may add, by way of 

 confirmation, the measurement of the striae which mark 

 the shields of every species in constant proportion. In 

 the iV". viridis, (the viridula of Ehrenberg, not of 

 Klitzing,) Ehrenberg states that 13 — 15 striae are com- 

 prised in T^ of a line, and Kiitzing 12 — 14. T find 

 seven of these striae constantly in a centimilliraetre, 70 

 in a decimillimetre, and then with a power of 686 I 

 always find 0*00098 met. between one stria and another, 

 with the utmost exactness. Having directly inquired of 

 Kiitzing himself, he politely told me, in reply, that he 

 made use of a micrometer by Plossl, marked upon glass, 

 in which the Paris line is divided into thirty parts, and 

 that he placed his object upon this micrometer whenever 

 he wished to measui-e it. At the same time he favoured 

 me with a copy of his micrometer magnified 100 

 diameters. Now in this copy every division, said to equal 

 ^ of a line, is 0'00498 met. ; hence such a line would be = 

 00149 met., or something more than half the measure of 

 the line of Vienna, which is 0-002638 met. Finally, I can- 

 not quit the subject of dimensions without making another 

 observation. When we examine many individuals of the 

 same species, we find both large and small in the process 

 of deduplication. And we find many of the same size col- 

 lected together whenever this is the case. Hence it seems 

 that deduplication may occur at any age. And it seems 

 to me that the greater the dimensions, the less frequent is 

 this process ; and I never saw the largest species doubled. 

 It is a question, when does this process cease ? Is it only 

 the larger individuals, in which no division takes place, 

 that propagate their species by true reproduction ? or are 

 some individuals destined from their earhest origin to mul- 

 tiply by deduplication and others by reproduction ? This 



