The number of replicated small plat tests required in regional variety trials
1939
Harrington, J.B.
1. A study was made of data from several hundred replicated small plat tests of barley and wheat to determine the number of tests required for accurate comparative results in a given area. The tests studied were distributed throughout the farming area of Saskatchewan and totalled 27,573 plats, each plat consisting of four 10-foot rows. 2. For the Province as a whole random lots of 40 tests gave close approaches to the results obtained with 261 tests (1935 results). 3. For separate soil-climatic and cereal variety zones a quarter to a half as many tests as run in 1935 and 1936 gave results similar to those from the larger numbers, except that differences were less frequently significant. The drought in 1937 reduced the number of satisfactory tests to 138 and in only three zones could the number that year be considered larger than necessary. 4. For sampling of local areas within zones results obtained from distinctly fewer tests than those run did not reveal or bring out differences that otherwise were found. Similarly, the differential reactions of the varieties to particular conditions were not shown clearly where small numbers of tests were used. 5. It was concluded that the number of cooperative tests run in Saskatchewan in 1935 to 1937 was not excessive and that materially reducing the number of such tests would have been detrimental to the purpose for which they were planned.
显示更多 [+] 显示较少 [-]