How Domestic Varieties Originate 211 



the seeds from one-half of the progeny being carefully 

 selected year by year, and all those from untypical plants 

 discarded, whilst in the other half all the seeds from all 

 the plants, whether good or bad, are saved and sown. In 

 the one case, it will be seen, we are fixing the tendency 

 to "come true," for this is all that constitutes a horticul- 

 tural variety, a brood very much like all its parents. 

 In the other case, we are constantly eliminating the 

 tendency to "come true" by allowing every modifying 

 agency full chance. So the very act of taking seeds only 

 from plants that have "come true," tends still more 

 strongly to fix the hereditary force within narrow limits. 

 Working against this restrictive force, however, are all 

 the agencies of environment and atavism, so that, fortu- 

 nately, now and then a seed gives a "rogue," or a plant 

 widely unlike its parents, and this may be the start for a 

 new variety. 



With bud-multiplied varieties, however, the case is 

 very different. Here every seed may be sown, as in the 

 illustrative case above, because the seedlings are not 

 wanted for themselves, but only as stocks on which 

 to bud or graft the desired varieties. So there is no seed 

 selection in the ordinary propagation of apples, pears, 

 peaches, and the usual orchard fruits. The seeds are 

 taken indiscriminately from pomace or the refuse of can- 

 ning or evaporating factories. Moreover, many such 

 varieties are hybrid, and when propagated by seed, split 

 up into many forms. But every annual garden vegetable 

 is always grown from seeds more or less carefully saved 

 from plants that possess some desired attribute. There is 

 no reason why the tree fruits should not reproduce them- 



