NEW FORMS OF MARINE DIATOMACE^. 



519 



sidered as only a single frustule, while the Cymbella or Cocconema would be called 

 a double one. 



I confess I feel disposed to consider both in the same light, as double forms. 

 If that be correct, then what is usually called the valve of an Amphora will be 

 an entire frustule, and what is usually called the entire frustule (as in figs. 63 to 

 72, for example) will be considered as two frustules in the act of self-division, but 

 still united by the connecting membrane. 



One reason why I incline to this view is, that what, on this principle, are to 

 be called single frustules, which are now regarded as halves or single valves, are 

 often much more frequent than the entire, or, as I should say, double frustules. 

 We never see the halves of Navicular separated in this way. Secondly, Each single 

 valve, or as I should say frustule, has three nodules ; but the entire or double 

 frustules have six. Thirdly, In Amphorae, the two halves of the double or entire 

 frustule are almost always separated by a rectangular space, apparently homo- 

 logous with the connecting membrane, which is seen in other genera when in a 

 state of self-division. When this is absent, as in fig. 81 e, or in fig. 76, the form 

 has exactly the appearance of a twin frustule of Cymbella. Lastly, If we regard 

 the so-called valves as entire frustules, they become perfectly analogous, in gene- 

 ral form and structure, to Cymbellse or Cocconemata ; and we shall find, in the 

 remarkable group of Complex Amphorae, next to be described, that the segments 

 in many species, indeed in most, have a still greater resemblance to Cymbellse or 

 Cocconemata. 



Let us now consider that group. 



B. Complex Amphorce. 

 Of this sub-group, as I have already stated, there are, in these materials, as many 

 as of the simple group just described. These species are as follows : — 



82. Amphora lyrata, n. sp. 



83. 

 84. 

 85. 

 86. 

 87. 

 88. 

 89. 

 90. 

 91. 



Milesiana, n. sp. 

 elongata, n. sp. 

 quadrata, n. sp. 

 excisa, n. sp. 

 nobilis, n. sp., G. 

 Arcus, Greg. 

 Grevilliana, Greg, 

 complexa, n. sp. 

 fasciata, n. sp. 



92. Amphora sulcata, Breb. 



93. 



acuta, n. sp. 



94. 

 95. 

 96. 



crassa, Greg, 

 pusilla, n. sp. 

 granulata, n. sp. 



97. 

 98. 

 99. 



cymbifera, n. sp. 

 proboscidea, n. sp. 

 costata, Sm, 



100. 



bacillaris, n. sp. 



The first species of this group which was figured was A. costata, Sm. But the 

 remarkable peculiarity of its structure was not specially noticed. I subsequently 

 noticed several species in the Glenshira Sand, in which the complex structure had 

 attracted my attention, such as A. Arcus, A. Grevilliana, and A. crassa, which 

 I described and figured. I also pointed out that the same structure occurs in 

 A. costata, Sm. In the new materials, besides acquiring a more accurate know- 

 ledge of the three new species just named, all of which I here figure a second 



