The Bulletin. 23 



lint and kernels, and conversely. (3) Varieties willi lari^e bolls gener- 

 ally have large seed and niediuni to small pcrfcntages of lint. (4) 

 Increasing the yi( Id of s(!cd cotton ])er acre, through proper fertili- 

 zation or favorable seasonal conditions, results generally in a slight 

 lowering of the percentage of lint to seed, or in other words in an 

 increased size and percentage of seed to lint, even of the same variety. 



(5) Varieties with large bolls and seed usually germinate well, grow 

 vigorously, and are less afi'cctcd by adverse seasonal and soil conditions 

 during their active growing period ; but are generally medium to late 

 in maturity, and hence may stand in danger of being cut off by frost 

 in the fall, if a comparatively long growing period is not afforded. 



(6) Good root and leaf development of a variety tend to increased 

 power of resistance to drought and insect and disease ravages. 



VARIETIES. 



. Six years ago 'the Department of Agriculture, by means of its 

 test farms, began comparative tests of varieties of cotton Avith the 

 purpose, primarily, of ascertaining, if possible, the varieties that 

 are most prolific of seed cotton per acre, when grown under our 

 conditions of soil and climate. During this time tests have been 

 made with thirty-seven varieties, the number ranging from seven 

 or eight in 1900 to twenty-five in 1904 in the different tests on 

 the different farms. It is felt from this accumulated data of five 

 years' tests that some very reliable and valuable information has 

 been derived, especially if taken and intelligently applied by the 

 individual farmers of the State in their farming operations. 



Valuation in Yield of Varieties. — In our variety tests we have 

 had some variety or varieties to yield 700 to 800 pounds of seed 

 cotton per acre more than other varieties in the same tests and grown 

 under identical conditions of soil, fertilization, and cultivation. This 

 variation in yield has been no uncommon occurrence in our experi- 

 ence. Take, for instance, the results at the Edgecombe farm during 

 the past five years. In 1900, in a test of eight varieties, the difference 

 between the variety yielding the largest amount of seed cotton per 

 acre and the one the smallest, was 565 pounds; in 1901 and 1903 in 

 tests of seven varieties each, the difference was 520 and 790 pounds 

 respectively; in 1903, 663 pounds, when nine varieties were incorpo- 

 rated, and 724 pounds in 1904 in a test of twenty-one varieties. The 

 average of these differences, during five years' tests, is almost equal to 

 seven-tenths of the average annual yield per acre of seed cotton in 

 JSTorth Carolina. To grow cotton cheaply per pound, more must be 

 produced per acre than is at present being done on an average. To 

 do this, better varieties must be planted, more thorough preparation 

 and cultivation be given to the land, and more intelligent fertiliza- 

 tion, either directly or indirectly, must be practised. It costs no 

 more to cultivate a prolific variety of cotton than one that has few 



