Osyris, Loranthus an</ Viseum. 205 



be made to r<;fer even to the apex of the embryonary sac : and in Loranthus, 

 if my explanation be correct, it is at once obvious that the radicle can have 

 no primary relations to the ovulum, and, indeed, no secondary ones to those 

 parts of the seed from which have been derived the expressions regarding its 

 direction. To include all these anomalies, the wording of the law must per- 

 haps be made to refer to the pollen tube. 



1 now pass to a recapitulation or summary of my ideas of the structure of 

 Santalum, Osyris, Loranthus and f^iscum. 



In Santalum the ovulum consists of a nucleus and an embryonary sac, 

 pi'olonged beyond both the apex and base of the nucleus : the albumen and 

 embryo are developed in the part above the septum, the part below and the 

 nucleus remaining unchanged. The embryo is developed from the pollinic 

 vesicle. The seed has no actual proper covering, and no other theoretical 

 covering than the incorporated upper separable part of the embryo-sac. 



In Osyris the ovulum is reduced to a nucleus and an embryonary sac, which 

 is prolonged in the same directions as Santalum, but not to such a degree 

 beyond the apex of the nucleus. The seed is formed outside the embryo-sac, 

 and is absolutely without proper tegument, or whatever covering it may have 

 did not enter into the composition of the ovulum. The embryo appears to be 

 developed at some distance from the anterior end of the pollen tube. 



In Fiscum, the modifications appear to me to be two : in the one, an evident 

 cavity exists in the ovarium; and the ovulum appears to be reduced to an 

 embryonary sac hanging from one side of the base of a nipple-shaped or coni- 

 cal placenta ; in the other, the ovulum is reduced to an embryonary sac, but 

 this is erect, and has no such obviously distinct point of origin as in the first. 

 In both, the albumen has no other proper covering than the incorporated 

 embryonary sac ; and, at least in the last, the embryo appears to be a direct 

 transformation of the pollinic vesicle. 



In Loranthus, each ovulum appears to be reduced to an embryonary sac ; 

 the albumen is developed either partly within the sac, or entirely, or almost 

 entirely, without it. The embryo is a growth from the ends of the continua- 

 tions of the pollen tubes outside the anterior ends of the embryo-sacs ; and is, 

 in one modification, exemplified by L.globosus, up to a certain period exterior 

 even to the albumen. In L. bicolor the albumen has no proper tegument ; 



2 E 2 



