Arruda, Hermano vaz de2019-01-102019-01-101960-01ARRUDA, H. V. Aplicação da análise da covariância, num estudo sobre tamanho de canteiro para experiências com caféeiros. Bragantia, Campinas, v. 19, n. 1, p. 1-5,1960.1678-4499http://dx.doi.org/10.1590/S0006-87051960000100068http://www.sbicafe.ufv.br/handle/123456789/10866Plot sizes in coffee experiments carried out in the past varied considerably. The early ones usually had large numbers of hills per plot (50 or 30), but in later experiments they were reduced to 10, 9, or fewer hills per plot, usually each hill composed of 4 plants. The data available are not enough to permit a good judgement as to the best plot size for coffee experiments. The writer, by means of covariance analysis applied to a fertilizer experiment, has shown in this paper that 1 row of 5 hills gave comparable results to 3 rows with the same number of hills. It is suggested that the application of a similar analysis to other coffee experiments being carried out will give a fairly good estimate of an adequate plot size for coffee experiments.pdfpt-BROpen AccessCafeicultura::Implantação e manejo da lavouraArtigo