- 89 



After eliminating the first and sixth rows from 

 the north ends of plats 5 and 9, the north ends only were 

 taken, since the data from the south ends were not depend- 

 able due to injuries sustained:. The variation was as fol- 

 lows: For plat 5, Minimum 11.6JS, maximum 25^, and extreme 

 variation Z6,6f=; for plat 9, minimum SO.S^S, maximum 30^, 

 and extreme variation 50.3J5. 



The table shows an extreme variation in green 

 weight of stalks between the rows of the same plat from 

 5.9 per cent in plat 24, to 111,9 per cent in plat 5. E- 

 liminating plats 5 and 9, the maximum extreme is, 65.9^, oc- 

 curring in plat 8, and the average variation is about 28 

 per cent . The number of stalks harvested in the row seems 

 to have little influrnce in the variation. 



Rows 1 and 6 of plat 5 being on the edges of the 

 plats, where conditions for growth were better, should 

 clearly n6t be coxinted in final yields . To secure com- 

 parable results rows 1 and 6 of plat 9 were also eliminatdd. 

 Row 6, plat 8, being next to plat 9, is clearly not repre- 

 sentative. Rows 5 and 6, plat 16, and row 1 plat 17 were 

 markedly poorj but row 4 plat 16 and row 2 plat 17 were 

 markedly good. The yields^ indicate that the south series 

 beginning with plat 14 improved in productiveness to about 

 the middle of plat 15, from whence a gradual reduction is 

 shown till the border between plats 16 and 17 is reached. 

 A few rows beyond this show an increase, after which there 



