432 GENETICS IN RELATION TO AGRICULTURE 



(b) It is probable that the greater the number of strains to be compared the 

 more repetitions will be necessary, because of the greater area they will cover. 



(c) Small blocks, 5.5 feet square, give results similar to those of the row plots, 

 except that the reduction of experimental error is somewhat greater as a result 

 of repetition. Blocks repeated 8 or 10 times give results apparently about as 

 accurate as rows repeated 15 or 20 times. 



(rf) The rate of planting within certain wide limits, has little influence on 

 yield. 



(e) There is some competition between adjacent rows, especially when 

 varieties very different in habit of growth are planted side by side. The use of 

 blocks does away with this source of error. 



' ' ' ' ' L 1 ' ' ' ' ' ' ' ' M M ! i < I 



' ' M It I 



FlG. 179. Row-plot nursery (wheat) in which the rows are 16 feet in length with a 

 4-foot alley adjacent, thus making the beds 20 feet in width. (After Montgomery.) 



(/) Block plots and row plots at the usual rates of seeding will probably 

 correlate more closely with results in field plots than in plots where the plants are 

 spaced as in centgeners. 



(g) Where error is corrected by the system of repetition plots, check plots 

 would be used for the purpose of determining the experimental error. When 

 the variation in checks equals the variation in strains, no possible selection can 

 be made. 



Strain Tests. The chief purposes of strain tests are (1) separation of 

 types within commercial varieties with a view to standardization of 

 varieties and (2) selection of the most profitable strains within a variety. 

 The mixed condition of our varieties of vegetables calls for continual 

 attention on the part of the seedsmen and experiment stations in an 

 effort to bring existing varieties up to some definite standard. One 

 difficulty is found in the widespread use of synonyms, according to Work, 



