560 Journal of Agricultural Research voi. xix. no. h 



Of these 21 lines, 7 have curlydwarf tendencies. Twelve appear among 

 the best 33 lines chosen in 191 8 on vine characteristics alone. This 

 leaves two, 443 and 452. in the doubtful list. 



IINE 

 NO. REMARKS. 



416. Third unit a curlydwarf; others vigorous though lacking in uniformity. 



44.3. A very good type; first and second unit not especially vigorous. 



442. A good vigorous type. 



434. A good vigorous type. 



510. A good vigorous type. 



451. A good vigorous type. 

 450. Not a very good type. 



412. Uniform and fairly vigorous. 

 433. A good vigorous type. 



424. Vines not a good type. 



415. A good type and fairly vigorous. 



448. All have more or less curlydwarf tendency. 



422. Lacking in vigor but type good. 



446. First four units have curlydwarf tendencies. 



414. First unit with curlydwarf tendencies. 



425. Very good and vigorous type. 



452. Not very vigorous but type good. 

 430. First unit very vigorous; others good. 

 480. Fairly vigorous. 



492. A very good type. 



419. Some curlydwarf tendencies. 



In Table XIII data are presented showing the 3-year average numeri- 

 cal production of tubers per single-stemmed hill. In Tables XIV and 

 XV a portion of this data is used to show the lack of correlation between 

 yields in numbers and in weight, as well as the relation between numeri- 

 cal tuber production and good vine characteristics. There is in Table 

 XIII a maximum variation of 1.29 tubers per hill. Between the average 

 of the 20 highest and 20 lowest (Tables XIV and XV) there is a variation 

 of 0.79 tubers per hill. Considering that these are single-stemmed hills, 

 this variation is rather pronounced. But the data are apparently no 

 more reliable than weight records in pointing out promising lines. Low 

 numerical tuber production does not always mean low production by 

 weight, and such records are apparently of little value as a check on 

 degenerate tendencies. Fluctuation^ in tuber production in different 

 seasons has been even greater than the variation between lines in any 

 one season. These 108 lines averaged 2.05 marketable tubers per hill in 

 1916, 3.78 in 1917, and 2.44 in 1918. 



In Table XVI the 33 lines with promising vine characteristics have 

 been assembled with their 19 18 yields of marketable tubers. The average 

 yield of these 33 lines is 21,142 pounds per acre. In Table XVII the 

 33 heaviest-yielding lines of 1918 are assembled for comparison. Their 

 average yield is 24,109 pounds, while the 191 8 average of the 108 Rural 

 New Yorker lines was 19,787 pounds per acre. The lines chosen upon vine 



