22 EXPERIMENT STATION [Jan. 



forth the application of results there indicated to practical 

 methods of plant improvement. The methods used have served 

 to distinguish between those differences that are due to soil 

 and weather conditions and those peculiar to the plant which 

 are inherited and therefore of greatest interest to the practical 

 plant grower. The results which are in harmony with those 

 of other investigations indicate that our common varieties of 

 garden peas are composed of sub^'arieties or, more properly, 

 strains, which differ from each other in varying degrees, this 

 difference, however, being relatively fixed and permanent. 



'A planting of Excelsior peas proved to contain two strains, 

 one of which exceeded the other in productiveness by about 15 

 per ctmt. in an average of four years crops. While this differ- 

 ence is a pronounced one, it would have been totally hidden 

 by differences in yield due to environment had the two strains 

 not been planted separately and in large numbers. The separa- 

 tion of this more prolific strain should prove advantageous to 

 the gardener. 



According to the generally accepted belief, the selection of 

 superior individuals within this strain should result in further 

 improvement. This is doubtful. It is }irol)able that if any real 

 progress can be made by this method it will be small and of 

 doubtful permanency. We do not regard this as a proven fact, 

 but only as a strong probability. The efforts of the gardener 

 should then be directed to the isolation of their superior strains 

 where they exist in oiir common varieties. 



We may now outline a simple method of accomplishing this. 



The first step is to secure the best available stock of a variety 

 suited to the purpose of the grower. This should be planted 

 and cared for in the best possible manner. The most uniform 

 plot of ground available should be selected. The grower shoidd 

 study the plants carefully and select those which ap]K>ar the 

 most desirable, particularly as regards yield. The seed from 

 each plant should be gathered separately and given a number. 

 All shrunken or otherwise imperfect seeds should be thrown 

 out. 



The next spring the seeds from the several plants should be 

 planted, each in a row by itself, under the most uniform con- 



