38 ARIL OF DILLENIADS AND SAMYDS. [BOOK i. 



If we take the anatropal ovule of Hibbertia volubilis, a 

 little before the expansion of the flower, we find around its 

 point of attachment a sort of circular thickening formed by 

 its great umbilical cord. The edge of this thickening, at 

 first at a tolerable distance from the exostome, continually 

 approaches, and in the expanded flower nearly reaches this 

 opening. I have not been able to trace the further evolution 

 of the ovules; but it is certain that the expansion does not 

 proceed much further, since, according to certain authors, the 

 ripe seed of this species has only a very short arillary mem- 

 brane at its base. 



In Tetracera, on the contrary, this organ is developed on 

 the seeds as a membranous, coloured sac, which is inserted 

 in a circle around the hiluni, and presents at its top a large 

 opening. The free edge of this orifice is more or less slashed 

 somewhat in the same way as a fimbriated petal. The enve- 

 lope adheres to the hilum of the seed, when the latter is 

 detached from the placenta, and completely covers the micro- 

 pyle which is beneath it, by the side of the point of attach- 

 ment, an essential character of a true aril. In a species 

 of Tetracera from Java, I have seen the aril cover the seed 

 completely : in other species it only partially covered the 

 surface of the seeds. 



Other genera of the same natural order, as Candollea, 

 Delima, Davilla, Pleuranda, &c., have, it is said, an aril very 

 like that of Tetracera. In the genera Pachynema and Hemi- 

 stemma, the aril is stated to be reduced to a cupule that 

 receives the base of the seed only. 



The genera Samyda and Casearia, although in their general 

 characters far removed from Dilleniads, have, however, an 

 aril, with a large opening and deeply laciniated border. The 

 seeds of these two genera are semi-anatropal, and only differ 

 from those that are completely anatropal by the position of 

 their point of attachment, which, instead of being quite close 

 to the micropyle, is as far from it as from the chalaza. The 

 raphe, very short in the first, only occupies the half of their 

 length. These seeds are completely hidden by the arillary 

 envelope, which inserted round the hilum, necessarily covers 

 the micropyle, a character sufficient to indicate an aril. 



