1919] HAVATA—PROTOMARATTIA 85 



in the elongated linear sori, but differs from it in the absence of 

 ridges between the sori and in the dorsiventral rhizomes. Gwynne- 

 Vaughan (loc. cit.) says that in the specimens examined by him 

 there was no suggestion of dorsiventrality in the rhizome of Arch- 

 angiopteris, and the leaf arrangement and the vascular structure 

 indicated a radial symmetry. So far as I can judge from the 

 figure given by the author/ however, there are some indications of 

 dorsiventrality in the rhizome, as can be seen in the upward bending 

 of the stipes of the leaves and in the fact that the rhizome ascends 

 somewhat obliquely toward the apex; and it also may be inferred 

 that the remaining portions of the rhizome not given in the figure 

 very probably run horizontally. In Archangiopteris Sotnai, 

 recently discovered in Formosa, and two other new species from 

 Tonkin, which will be described later, the rhizomes are prostrate 

 and show very clear signs of dorsiventrality. 



In the summer of 191 6 I was sent to Tonkin for collecting. I 

 found there two new species of Archangiopteris and a t}pe of a new 

 genus closely related to the latter. All these plants present an 

 appearance very similar to other ferns, such as Coniogramme 

 japonica, C. fraxinea, Diplaziopsis javanica, or Diplazium hanta- 

 mense, and occur in very small numbers amidst a multitude of the 

 previously mentioned ferns. As it is very rarely that one has the 

 opportunity to meet with these plants of Archangiopteris and 

 the allied new genus, it may not be entirely out of place if I should 

 tell how I was led to discover these very rare and interesting 

 ferns. 



Some 10 years ago I was greatly interested in learning of 

 Henry's discovery of Archangiopteris, representing as it did an 

 ancient type of the Marattiaceae, and I wondered whether there 

 might not exist another species of the genus in Formosa, the flora 

 of which I have since then been studying. In 191 5, when examin- 

 ing collections sent by the late T. Soma from Formosa, I found 

 among them a curious looking fern labeled Gymnogramme japonica. 

 A glance at the specimen showed me that it was another type of 

 Archangiopteris, which was then named and published as A. SomaiJ 



* Gwynne-Vaughan, D. T., loc. cit. pi. 10. figs, i and 2. 



^Hayata, B., Icones Plantarum Formosanarum. 5:256; 6:154. />/. 79. 



