BURBANK'S HORTICULTURAL NOVELTIES 167 



From this discussion it may easily be seen that my evi- 

 dence relies, for a large part, on experiments which are not 

 yet finished and the ultimate result of which cannot yet be 

 estimated. For the description of the methods used, this 

 is of no importance, and in many cases the older experiments 

 with their practical results will have to be alluded to. 



Burbank's first catalogue was published in 1893. It is 

 now 13 years old. The varieties described therein are, of 

 course, older, but they are only a small number in comparison 

 with his present stock. The larger part of his experiments are 

 younger, and only a few of his pedigrees cover more than ten 

 years, as, for instance, those of the plums. 



A special feature of Burbank's work is the large scale 

 on which his selections are made. It is evident that in a 

 variety gf mixed condition, or in the offspring of a hybrid, and 

 even in ordinary fluctuating variability the chance of finding 

 some widely divergent individual increases with the number 

 of the plants. In some hundred specimens a valuable sport 

 can hardly be expected, but among many thousands it may 

 well occur. The result depends largely upon these great num- 

 bers. In one year he burned up sixty-five thousand two and 

 three year old hybrid seedling berry bushes in one great bon- 

 fire, and had fourteen others of similar size. He grafts his 

 hybrid plums by the hundreds on the same old tree, and has 

 hundreds of such trees, each covered with the most astonish- 

 ing variety of foliage and fruit. Smaller species he sows in 

 seed-boxes and selects them before they are planted out, 

 saving, perhaps, only one in thousands or ten thousands of 

 seedlings. Thornless brambles, spineless cacti, improved 

 sweet grasses (Anthoxanthum odoratum), and many others 

 I saw in their wooden seed-boxes being selected in this way. 



The same principle prevails in the selection of the species 

 which are submitted to his treatment. Here, also, the result 

 depends chiefly upon the numbers. He tries all kinds of 



