This paper deals with a taxonomic study of two diatoms collected in Xizang (Tibet)
and, based upon the peculiarities of their rapheal system, a new taxonomic system of the Class Pennatae is proposed.
One of the diatoms was found in fifteen specimens collected in various localities (alt. 4100--5300 m) in eight counties in southern, western and northern Xizang, and has been identified as Amphiraphia xizangensis Chen et Zhu, a new genus and a new species, including a variety, major Chen et Zhu. This species is epiphytic in habit (sometimes becomes freefloating) and characterized by its two valves of a frustule dissimilar in structure: the epitheca has a dotted-line shaped shaphe with a more or less reduced central nodule and without the polar nodules, whereas the hypotheca has a typical raphe with both the well developed central nodule and polar nodules. These most striking structures have not been discovered in all the known species. Thus, based upon the pecularities of this genus, the authors have established not only a new family, Amphiraphiaceae Chen et Zhu, but also a new order, Amphiraphidales Chen et Zhu.
The other is a well known and widely distributed species, Rhoicosphenia curvata (Kütz.) Grun. This genus has always been placed in the Achnanthaceae. As We know that the alga does not have frustules with a pseudoraphe on one valve and a true raphe on the other as stated by Boyer (1927) and Smith (1950). In fact, it has only a rudimentary short raphe near each pole on the epitheca and a fully developed raphe on the hypotheca as mentioned by Cleve-Euler (1958), Hustedt (1962), and Patrick and Reimer (1966) and obse rved by the authors in the Xizang specimens. The alga is really similar to Amphiraphia xizangensis Chen et Zhu in having two valves dissimilar in rapheal type. Thus, based upon its generic characteristics the authors have established another new family, Rhoicospheniaceae Chen et Zhu, and placed it in the order Amphiraphidales Chen et Zhu too.
Regarding the systematic position of the Amphiraphidales, the following principles are considered:
1. It is evident that the free-floating or motile diatoms existing in the geological periods are really older than the epiphytic ones. This has been proved by some discoveries of the fossil diatoms.
2. Both epitheca and hypotheca of frustule of all the free-living diatoms, induding both centric and pennate species, are symmetrical in both valves and girdle-views, but, the epiphytic forms are usually asymmetrical, especially in a girdle-view.
3. The characteristics of rapheal system are often used by former algologists as a main taxonomic basis for considering category of pennate diatoms and affinity among them. The authors agree with this consideration.
4. In this paper Hustedt's classification of pennate diatoms is used as a basis for considering the phylogenetic problems of these diatoms, but the authors cite Schütt's Class Pennatae, instead of Hustedt's order Pennales; Sieminska's orders, Araphidales, Raphidioidales, Monoraphidales, and Biraphidales, instead of Hustedt's Suborders, Araphidineae, Raphidioidineae, Monoraphidineae, and Biraphidineae.
5. The rapheal system on the epitheca of the Amphiraphidales is really a reduced
structure derived from a typical form represented by the Biraphidales, and may be regarded as a transitional type from that of the Biraphidales to that of Monoraphidales.
6. Owing to the epitheca of the type genus Rhoicosphenia Grun. of the family Rhoi-
cospheniaceae lacking the central nodule and a great part of its axial area becoming a pseudoraphe, the authors consider that the family is more close to Monoraphidales than Amphiraphiaceae is in phylogeny.
According to the principles set forth above, a taxonomic system of the Pennatae is proposed as follows:
Class Pennatae Schütt
Order I. Araphidales Siem.
Order II. Raphidioidales Siem.
Order III. Biraphidales Siem.
Order IV. Amphiraphidales Chen et Zhu, ord.nov.
Family I. Amphiraphiaceae Chen et Zhu, fam. nov.
Family 2. Rhoicospheniaceae Chen et Zhu, fam. nov.
Order V. Monoraphidales Siem.
