Xl] HOMOLOGIES OF POLLEN-TUBE, ETC. 95 



protoplasm of the growing pollen-tube; the generative 

 nucleus is either passed on to the end of the tube and 

 handed over to the oosphere direct, or it first undergoes 

 division into two, and one of these two is thus handed 

 on, the other becoming disorganised. Or, in some cases, 

 both are handed over to the embryo-sac. 



It is clear from the foregoing that the pollen-grain 

 is not a mere cell which elongates into a tube, but that 

 it germinates and forms in its interior a delicate cell- 

 tissue, consisting of two to five nucleated cells, in some 

 cases (e.g. most Angiosperms) not even completely 

 separated off by a cell-wall, but in others (e.g. especially 

 Cycadese) very distinctly clothed with cellulose coats. 



The fact that the generative nucleus with its surround- 

 ing protoplasm i.e. the nucleated generative cell always 

 occurs, and is passed into the pollen-tube to be handed 

 over to the oosphere, shows that it is the one cell of this 

 tissue which is of prime importance ; while the fact that 

 the other cells vary in number and completeness, and 

 soon undergo disorganisation, or are even not formed at 

 all, indicates that they are of less importance. 



All the evidence, therefore, goes to show that we have 

 in these rudimentary tissue formations remnants of some 

 processes which in the ancestors of the pollen-grain had 

 more significance than they have at present, and that the 

 reason they are formed at all as transient and incomplete 

 structures, is due to that conservative principle of heredity 

 which alone explains why rudimentary and noAV useless 

 structures are developed at all. It is evident that our 

 only hope of explaining what these rudimentary structures 

 represented in the ancestors, is to piece together all the 

 stages of evolution shown by the pollen-grain in its origin 

 in the anther, and we find that the formation of the 

 sporogenous mass and its surrounding degenerated layers 



