in the Seeds of Acantliacese. " 67 



Delile's plant, I have carefully compared the two, and although to the naked 

 eye the resemblance both of the seed and capsule is very striking, the result 

 of a careful microscopic examination has left some doubt of their identity. As 

 yet no seedlings have been raised of Mr. Holroyd's plant, and unfortunately he 

 possesses no specimen. 



The cells of the testa appear to me to be somewhat smaller in Mr. Holroyd's 

 seeds than in Delile's, but they certainly do not differ very materially in this 

 respect, the principal distinction between them being that in the former the 

 spiral fibre adheres closely to the membrane, whilst it is usually quite free in 

 the latter, the expansion of the fluid within the cell frequently detaching a 

 portion of the membrane and exposing the fibre, which is often singularly 

 confused and entangled : as, however, they agree in every other particular, it 

 may be a question whether the difference observable in Delile's seed be any- 

 thing more than the result of age destroying the elasticity of the fibre, and 

 thus preventing it from expanding simultaneously with the membrane. What 

 renders this the more probable is, that I find an entire conformity in the fibre 

 of another species of Acanthodlum gathered many years ago by Mr. A. P. Hove 

 in the Lymree desert of Guzerat, for which I am also indebted to Mr. Brown. 

 This seed bears a very close resemblance to the two already described, in the 

 structure both of the testa and the investing hairs, which here also are formed 

 of partially-cohering cylindrical tubes, containing one or two spiral fibres ; but 

 the hairs are more thinly scattered in the Indian seed, and consist of a smaller 

 number of tubes (about 5 — 8). 



The peculiar appearances of these seeds induced me to extend my inquiries 

 to other genera of the same natural family, with the view of ascertaining how 

 far the tendency to develope spiral hairs on the testa prevails among them, 

 and whether that peculiarity might afford any assistance in characterizing 

 genera. Although the presence of spiral cells is by no means universal in 

 Acanthacece, yet I have met with many examples of them, and with a con- 

 siderable diversity in the structure and arrangement of the hairs which clothe 

 the seeds. 



Besides Acanthodium, the only other genus in which I have met with fascicu- 

 late hairs is Blepharis, of which I have examined three species, two of them na- 

 tives of India {B. boerhaavicefolia and B. molluginifolia), the other B. ruhiifolia, 



k2 



