in the Seeds of Acanthaceae. 75 



walled hexagonal cells. In H. fVallichii these tubercles are crowned with a 

 ring of very small deflexed hooks, which are wanting in H. purpurea. 



Gendarussa Neesiana, G. quadrifaria, G. tranquebariensis, G. orixen.sis, Ad- 

 hatoda Betonica, A.trinervia, A.argyrostachya, Rhinacanthus communis, Rungia 

 IVighHana, R. origanoides, R. parviflora, Perisfrophe pubigera, P. montana, 

 P. speciosa, P. lanceolaria. — Very little difference of structure is to be observed 

 in the seeds of the above species, which, until the subdivision of that genus by 

 Professor Nees von Esenbeck in his revision of the Acanthaceae, were all placed 

 together under Justicia. They are for the most part small, compressed, tri- 

 angular or cordate, with a loose spongy testa, consisting of nearly hexagonal 

 cells with thick side-walls, and so much crumpled as to form numerous closely 

 crowded hollow tubercles. The tubercles, which occupy the entire surface of 

 the seed, are usually blunt and rounded at their extremities ; but in Genda- 

 russa tranquebariensis and orixensis they are prolonged into decurved points. 

 These characters are less conspicuous in the genus Perisfrophe : the seeds of 

 P. pubigera and P. montana differ little, except in their larger size; but in 

 P. speciosa and P. lanceolaria the testa is thinner and more even ; the tuber- 

 cles also are very small and less numerous in P. speciosa, whilst in P. lanceo- 

 laria they are almost if not altogether wanting. 



These observations having been chiefly made on such seeds as could be 

 obtained from dried specimens, many of them in an unripe or imperfect 

 condition, it is scarcely possible that I should in every instance have avoided 

 mistakes : it would, however, be a great pleasure to me could I hope that this 

 very imperfect view of the subject might prove the means of leading to a more 

 accurate investigation of the seeds of this interesting family ; and I trust that 

 whoever may pursue the subject further, will make allowance for the difficulty 

 of attaining to perfect accuracy with such materials. 



EXPLANATION OF TAB. VI. 



Fig. 1. Fasciculate hair from the testa of a species oi Acanthodium [A.spicatum, Del.?), 

 collected in Upper Egypt by Mr. Holroyd. 

 2. Portions of single tubes of the same, more highly magnified. 



l2 



