P 



IX. 



ERRORS ABOUT GRAFTING. 



LANTS, to succeed when grafted upon each 

 other, must be in some way related, though as a 

 rule, they do not need to be as closely related as 

 they do for crossing. Usually, we may say that any 

 two plants of the same family may be grafted 

 together, while for crossing, the plants must gener- 

 ally be of the same genus. Thus the pear and 

 the quince may be grafted together, but will not 

 hybridize. The exact limits of grafting, however, 

 are not easily denned, and depend on causes which 

 are not understood. Why all varieties of pears, 

 which are so closely related, do not grow equally 

 well on foreign stocks, such as the quince and 

 thorn, we do not know. We know the general 

 limits of grafting only; the degree of success in 

 each case must be determined by trial. No expe- 

 rienced gardener, for example, would try to graft 

 a monocotyledon on a dicotyledon; in fact it is 

 almost impossible to graft monocotyledons at all. 

 The stories told therefore by the ancient Komans 

 of dates, olives, pomegranates and oranges all 

 growing from a single tree, may be regarded as 



(81) 



