STRUCTURE.] EXCRESCENCES ON BIXADS, 41 



second at the hilum, and the raphe between them would be 

 a sort of internode intimately adhering to the integument of 

 the ovule. This point of view I take to be more ingenious 

 than correct; for I cannot conceive that a portion of the 

 umbilical cord at first free, should be afterwards found im- 

 bedded in the tissue of the ovulary leaf, under the epidermis, 

 which is perfectly unbroken, and often under several crus- 

 taceous layers of the integument. Besides, the raphe 

 is generally a medial nervure which sends out along its 

 entire length, lateral nervures, and which branches at its 

 origin before reaching the chalaza. These facts prove that 

 this part is not more independent of the exterior ovulary leaf 

 than the midrib is of an ordinary leaf; and that it plays the 

 part of a midrib with respect to the integument, and of an 

 axis with respect to the more interior parts of the ovule ; but 

 they farther prove that there is a congenial adherence and 

 intimate blending between the axis and the nervure, just as 

 on the bracts of the Lime tree (Tilia) the midrib is joined 

 with the floral peduncle from its commencement. 



From the frequent examination of the passage from ortho- 

 tropy to anatropy in ovules, I am convinced that the umbi- 

 lical cord is never soldered to the ovulary leaf, and if I 

 rightly understand M. De Mirbel, these observations accord 

 with his. The axis of an orthotropal ovule continues in a 

 straight line that of the umbilical cord; the vessels of the 

 latter cross in a vertical direction the thickness of the exterior, 

 to reach the interior integument. But one side of the ovule 

 developing faster than the other, the interior integument is 

 forced to incline to the funicle on the opposite side ; and then 

 the vessels extending from the hilum to the chalaza traverse 

 the exterior integument obliquely, and these two points 

 coincide no longer. The hilum remaining fixed, the part of 

 the ovule between it and the chalaza increases rapidly, 

 separating the one from the other, thus forcing the vessels 

 joining them to lengthen into a raphe, in the thickness itself 

 of the testa. The point of attachment of the ovule, it will be 

 seen, is not changed during this transformation; and we 

 cannot admit the possibility of there being two opposite arils 

 on one seed, since there is never but one hilum, and the aril 



