4 Station Bulletin 368 



Howard 17 is used as a measuring stick in evaluating new varieties 

 and recording their characteristics in the University of New Hampshire 

 variety tests. 



METHODS 



Beginning in 1934, straw l)crry variety tests were conducted so that 

 the results could be subjected to statistical analysis. Previous to 1941, 

 randomized complete block designs were employed, usualh' with four 

 replications. As a large number of varieties were to be treated in 1941- 

 1942 the use of a randomized complete block design was questioned both 

 Ijecause of the relativel)' small amount of land available and the lack of 

 know ledge concerning soil uniformit)' in this area. Arrangement of the 

 plots in a Latin square was ruled out principally because the number of 

 replicates would require a much larger area of land than was available. 

 The writer, therefore, adopted a lattice design similar to that used b)' 

 Cox, Eckhardt, and Cochran i for testing varieties of corn in Iowa. This 

 design was selected on the basis of their statement that "as the nuiul)er 

 of varieties increases, these designs {i.e., Latin square and randomized 

 complete blocks) may become less efficient through failure to eliminate 

 soil heterogeniety. Furthermore, the Latin square design becomes cuiu- 

 bersome because it requires replicates equal in numi)er to the varieties." 



In 1941, two lattice designs were set up, one consisting of 25 varie- 

 ties dug from the field and transplanted the same da\' and the other of 

 16 varieties, 13 of which were obtained from distant commercial nui'- 

 series. It was deemed advisable to plant the "shipped-in" plants in a 

 separate lattice because previous experience had show n that such plants 

 sometimes fail to become established as quickl\' as field-dug plants. Lresh- 

 dug Howard 17, Fairfax, and Pathfinder plants were included in this 

 smaller lattice so that direct comparisons could be made with the varie- 

 ties in the larger lattice. All transplants were selected for uniformitx of 

 size and vigor. 



In all years individual plots ccjnsisted of short lengths of single rows, 

 and the data collected were subjected to analysis of variance. 



Plants were transplanted into the plots in Alay of each year begin- 

 ning with 1934. They were grown by the matted row system, and the 

 fruit harvested the summer of the year next following planting. Onl)' one 

 crop of fruit was harvested from a given planting. 



RESULTS 



Yield. X'arietal yield from 1935 to 1942 is presented in Tahle 7, and 

 yield compared to Howard 17 for the same period on a percentage basis 

 is presented in Table 2. In 1935, Aberdeen, Clermont, Dorsett, Dunlap, 

 Fairfax, and Stevens Late out\ ielded Howard 17, largely because the 

 Ho\vard 17 plants were attacked by cyclamen mite and suffered more or 



iCox, Gertrude M., Eckliardt, Robert C, and Cochran, W. G. The analysis of hittice and 

 tripple lattice experiments in corn varietal tests. Iowa Agr. Exp. Sta. Res. Bui. 281: 1939. 



