ANIMAL NATURE OF DIATOMEjE. 493 



view, of the Licmophoreae (excluding the genus Licmo- 

 pJiora) with the Gomphonemese ; the necessary fusion 

 into one only, of the two families Striatelleae and Tabel- 

 lariese, the character of median aperture given as dis- 

 tinctive of the second being absolutely false, with the 

 separation of the genus Terpsinoe as a distinct type ; the 

 removal of the Coscinodiscese, and probably of the 

 Tripodiscese, among the Melosirese; the grouping together 

 of the three families, Anguliferse, Biddulphieae, and 

 Angulatse, excluding Lithodesmium as uncertain; and 

 leaving there the genus Odontella, which demonstrates 

 the affinity of the entire group with the Melosireae ; 

 finally, the decided separation of the Actiniscese from all 

 the other Diatomese. 



These changes being admitted, and the principle being 

 conceded, to its full extent, not to allow absolute value to 

 any character taken separately, the Diatomeae would be 

 divided into two sections ; the Actiniscese, and all others 

 that could be comprised under the name loricatte, unless 

 we should refuse to acknowledge that the first (section) 

 have really any bond of alliance with the second, except 

 the siliceous nature of the dermo-skeleton. In the 

 loricatcB it would be hasty to separate the genus Terpsinoe 

 from all the others, which might be reduced to eight 

 families only: — 1, Eunotiese; 2, Fragillariese, (uniting 

 with them the Meridiese, the Striatelleae, and the Tabel- 

 lariese ;) 3, Melosirese, (comprising the Coscinodiscese, 

 Tripodiscese, Anguliferge, Biddulphieae, and Angulatse ;) 

 4, Cocconoideae ; 5, Achnantheae; 6, Cymbelleae; 7, 

 Naviculeae (with aU the Surirelleae ;) 8, Gomphonemeae 

 (with all the Licmophoreae, except the genus Licmophora.) 

 These families, too, might be separated into two groups, 

 according as the temnogenesis takes place by simple 

 deduplication, (Eunotiese, Fragillarieae, Naviculeae, Gom- 

 phonemeae, Cymbelleae, and Cocconoideae,) or by redu- 

 plication, (Melosireae, Achnantheae.) I do not believe,. 

 however, that this distinction is so essential as it may 

 seem ; for even in this case it appears to me that the 



