FIELD CKOPS, 537 



experiment showed variations witliin tlie limit of probable soil error in all 

 cases tlie following year. An application of 1 cwt. of snlpbate of potash per 

 acre gave no certain increase but the use of 2 cwt. apparently increased the 

 yield 17.5 per cent. 



Six rice varieties? from Burma gave no yields of high promise. Six other 

 varieties from British Honduras are described as having blackish, light colored, 

 or ordinary grains, barley-shaped, and reddish grains, and gave yields ranging 

 from 1,344 to 4,272 lbs. per acre. Among other varieties tested Nos. 3, 75 

 (Suthra dhan), and Creole yielded 4,812, 4,728, and 3,948 lbs. per acre, 

 respectively. 



Attempts to obtain an artificial or controlled hybridization indicated a pro- 

 fusion of variants in carefully grown rice. The more promising varieties and 

 the Creole kind seem to be freer from variants than others. The so-called 

 barley rices were more variable than long-grain varieties. Notes are given on 

 single and multiple planting of early ripening varieties. 



Tests of Caravonica Silk, Mit-aflfi, Egyptian, Sea Island, and other cottons 

 have been abandoned because of failure to obtain sufliciently high yields to 

 cover the cost of cultivation. Notes are given on the quantity and quality of 

 the product of 6 varieties of cotton which are practically indigenous on the 

 borders of the colony contiguous to the Brazilian frontier. 



[Introductions of field crops], F. W. Strong (Rpt. Agi: Dept. East. Bengal 

 and Assam, 1909-10, pp. 7-11). — The author gives a progress report of experi- 

 mental growings of newly introduced varieties of cotton, jute, potatoes, sugar 

 cane, rice, oranges, wlieat, oats, and barley. 



Observations on the growth of nitrog'en-collecting' plants and nitrogen- 

 using' plants in mixed seedings, Tacke (Ztschr. Moorkultur u. Torfverwert., 

 S (1910), No. 5, pp. 233-236) .—The Moor Experiment Station has found in 

 experiments on highmoor soil that it is unnecessary to apply nitrogen if ade- 

 quately drained soil is carefully prepared and limed or marled and enough 

 clover be sown in the mixture. For a series of years it has been observed that 

 grass does better when sown with clover even during the first season. Both 

 sorrel (Rumex acetosella) and heather did well on highmoor soil when not 

 crowded by competing plants. They did equally well on plats to which 4,000 

 and 9,000 kg. of lime were added as marl. 



On lucerne; with notes on some other leguminous crops, B. Dyer (Neiv 

 Zeal. Dairyman, 14 {1909), No. 3, pp. 60-62; 1// (1910), No. ff, pp. 17-20).— 

 The author has collected and presented in tabular and descriptive form the 

 results of work on the efficiency of alfalfa and other legumes as nitrogen gath- 

 erers. He includes the work of Atwater, Hellriegel, Wilfarth, and others, and 

 that done at Rothamsted and at the Woburn Farm. 



On the measurement of correlation with special reference to some charac- 

 ters of Indian corn, H. L. Rietz and L. H. Smith (Illinois Sta. Bui. 1^8, pp. 

 291-316, figs. 3). — The writers briefly explain the nature and use of the correla- 

 tion table and the correlation coefficient, the methods of computing the co- 

 efficient, or r, the probable error and the regression coefficient. References are 

 given to other publications in which these methods are more fully explained 

 and used. They are also applied to some characters of Indian corn, grown 

 experimentally for other purposes, and previously reported (E. S. R., 20, p. 531). 



For the second year rotation corn, the correlation values of length to circum- 

 ference center about 0.43 and range from 0.203 to 0.623 for high and low yields, 

 illustrating a general tendency in this direction. The correlation between 

 length and number of rows was usually insignificant. " In circumference and 

 rows, . . . extremes presented arc 0.425 and 0.608 ... in length and weight of 



