Feb. 1907. ] & A YA TA.— ON TA I WA NIA A ND ITS AFFINITY. 25 



ovules, and the shape of the seed, wing, albumen and embryo. 

 But it differs from that genus in the absence of the secondary 

 squama and in the number of the ovules (two in each scale). 

 These two points and the even more strikingly different habit 

 of the plant do not allow me to place it in Cunninghamia. 

 On this occasion, I thought whether I had not better regard 

 the plant as representing a new genus, Taiwania P This was 

 a difficult question for me to decide myself. I therefore sent 

 this specimen to Dr. M. T. Masters of the Linnean Society, 

 together with its description and figure, and asked him if I 

 might not be justified in making a new genus for this new 

 plant. In reply to my inquiry, that gentleman assured me 

 that he agreed with me on the matter. He also pointed out 

 that the foliage of this plant reminds one of that of Crypto- 

 meria but still more of Arthrotaxis, and the cone is very like 

 that of Tsuga. On his kind suggestion, I proceeded to examine 

 Arthrotaxis with the utmost care, and found that it resembles 

 Taiwania very closety, but differs from it in the form of the 

 cone, and still more in the general aspect. 



Arthrotaxis is Cupressus-like in its general form, as far as 

 I can learn from plates, 2) while Taiwania is not Cupressus-like, 

 but Cryptomeria-like. I am sorry that I have not ever seen 

 any specimen of Arthrotaxis; but I do not think that my 

 conception of Arthrotaxis acquired from plates can be widely 

 different from that obtained through studying specimen plants. 



As has been stated above, Cunninghamia, Arthrotaxis, 

 Cryptomeria. and Taiwania resemble in this point or that. 

 The question we are going to solve, is, which two of the 

 preceding three genera should Taiwania be inserted in? Sum- 

 marizing all the above accounts, we see that Taiwania comes 

 nearest to Cunninghamia in the form of its cones, while on 

 one side it resembles Arthrotaxis in its foliage, and on the 

 other it shows a close kinship to Cryptomeria in its habit. 



] ) My paper on this new genus was read before the Linnean Society, 5th, April, 

 1906. 



s ) Hooker:— Icones Plantarum, t. 559, and Engl, et Prantl. :— Nat. Pfl.-fam. 

 II.--1, p. 89, Fig. 59. 



