ANIMAL NATURE OF DIATO1OLE. 423 



36. AMPHORA. Individua libera, singularia, aperturis 

 mediis binis lateralibus, terminalibus nullis I. obsoletis. 



The Amphora are Cymlella with primary surfaces 

 equal, and secondary surfaces symmetrically convex, 

 instead of plane and inclined. AYe have yet to learn 

 whether the two median lateral perforations exist on one 

 part only or both. In the first case the analogy would 

 be complete; in the second, every Amphora might be 

 called a double Ci/mbella. And indeed the two indivi- 

 duals into which every Amphora divides itself, by de- 

 duplication greatly resemble two Cymbettte. Yet perhaps 

 this resemblance is only apparent. In the Cymbella there 

 is one primary surface, the dorsal, which forms the 

 convexity. Again, in the middle of the Amphorae the 

 convexity is itself referable to that one of the secondary 

 surfaces which remains. The Cymbetta, in subdividing, 

 give origin to two complete individuals. The two indi- 

 viduals proceeding from the division of the Amphora are 

 wanting in one of the two lateral convexities ; their lateral 

 surface, of new origin, ought to become convex, like the 

 other. In the Cymbella there occurs a simple division 

 or deduplication ; in the Amphora the division or dedu- 

 plication is succeeded by a species of reduplication. It 

 results, from this conformation, that in the Amphora the 

 navicular figure is only apparently similar to that of the 

 Navicula. The one is navicular in the secondary surfaces, 

 the other in the primary, because of the prominence of the 

 secondary. The division of Navicula is parallel to the 

 elliptical or rhomboidal surfaces ; in the Amphora it is 

 vertical to these (surfaces). I therefore absolutely exclude 

 the A. atomus from this genus, as it is a Navicula or a 

 Synedra, The doubt raised by Kiitzing as to the 

 A. elliptica appears a certainty to me ; no central aper- 

 ture in the primary surfaces being admissible. On the 

 same motive I assert that if the A. acutiuscula do truly 

 belong to this genus, the figure 1 (PI. V, fig. 32) is 

 incorrect, for it represents a median aperture; and 

 equally so is the third figure (PI. XX, fig. 18) of the 



