IMPROVEMENT OF PLANTS. 261 



should be borne in mind that " the possibilities 

 are by no means exhausted, but it is quite pos- 

 sible that the descendants of these hybrids will 

 yield valuable sorts." 



In many cases the cross, or its descendants, 

 may possess the desired characters of one 

 parent, while those desired from the other parent 

 may be entirely lacking. In that case "it would 

 be advisable to cross the offspring with that 

 parent * whose characteristics did not appear; 

 for, by so doing the tendency to transmit those 

 particular characters will be increased, for this 

 tendency is itself variable. "~|~ 



At the same time, the individual plants of the 

 original cross should not be discarded for sev- 

 eral generations, for there is in the offspring a 

 slight atavistic^ tendency, or a tendency to re- 

 vert to the character of some remote ancestor; 

 hence, at any time an individual plant may 

 appear which presents the very characters 

 desired. 



In no instance can the plant-improver afford 

 to neglect any condition or advantage which 

 will tend to induce the desired variation. 



(c] Process of Cross-pollination. This con- 

 sists in the transference of pollen from a flower 

 of one plant selected to be crossed to the stigma 

 of a flower from the other plant selected. In 



* Year-book, l8gg, p. 484. 

 ^Year-book, 1898, pp. 355-357- 



