1957 AJCS 14(12):1957-1960 (2020) ISSN:1835-2707 doi: 10.21475/ajcs.20.14.12.2828 Optimum plot size for field experiments in sesame Janilson Pinheiro de Assis 1 , Roberto Pequeno de Sousa 1 , Paulo César Ferreira Linhares 1 , Eudes de Almeida Cardoso 1 , Walter Martins Rodrigues 2 , Joaquim Odilon Pereira 2 , Robson Pequeno de Sousa 3 , Aline Carla de Medeiros 1 , Neurivan Vicente da Silva 1 , Anderson Bruno Anacleto de Andrade 4 , Geovanna Alícia Dantas Gomes 1 , Mateus de Freitas Almeida dos Santos 1 , Lunara de Sousa Alves 1 1 Jitirana Research Group, Department of Agronomic and Forestry Sciences, Federal Rural Semi-Arid University, Mossoró, RN, Brazil 2 Center of Exact and Natural Sciences, Federal Rural Semi-Arid University, Mossoró-RN, 59625-900, Brazil 3 Computing Department, State university of Paraíba, Campina Grande-PB, 58429-500, Brazil 4 Plant Protection at the Federal University of Alagoas, Maceió, AL, Brazil *Corresponding author: janilson@ufersa.edu.br Abstract This work aimed to determine the appropriate plot size for field experiments in sesame. We performed a complete randomized block design experiment, using 14 sesame varieties and four replicates. The plots were composed of four rows of 0.8 m long, spaced 0.6 m apart, and 0.1 m between plants. The useful plot area (0.72 m 2 ), which was the two central rows, was divided into 12 basic units with one plant (0.06 m 2 ) each. The measures of sesame production were taken from the useful plot area. The appropriate size of the experimental plot was estimated using the intraclass correlation coefficient method and calculated the detectable difference between treatments. The optimum plot size for evaluation of sesame seed yield was 0.18 m 2 (useful area), taking into account the one-row border on the sides. Gains in experimental precision (12%) were occurred with increments in plot size up to eight basic units (0.48 m 2 ), using five replicates and four or more varieties. The increase in the number of replicates and plot size was more efficient than the increase in varieties number to increase the experimental precision. Key words: Sesamum indicum L, Intraclass correlation coefficient, Experimental precision, Experimental unit. Abbreviations: ρ _intraclass correlation coefficient method; d_detectable difference between treatments. Introduction Sesame (Sesamum indicum L.) is the ninth most cultivated oilseed in the world and its cultivation has great economic potential, due to the possibilities of exploitation, both in the national and international market (Mesquita et al., 2013). World production is estimated at 3.16 million tons, obtained on eight million hectares, with a productivity of 481.4 kg ha - 1 . Brazil is characterized as a small sesame producer with 15 thousand tons produced in an area of 25 thousand hectares and yields around 600.0 kg ha -1 , as it is planted in poor soils (Queiroga et al., 2007). After the fall in cotton production caused by the cotton boll weevil (Anthonomus grandis) cotton's breeding program is developing sesame studies to recommend varieties suitable for cultivation in the Northeast region (Queiroz and Beltrão, 2013). However, for a breeding program success, it is necessary to detect small variations among varieties during experiments, since the tendency is to decrease the difference among the new varieties. The challenge of breeders is to increase the experiment precision, allowing for genetic advances and, consequently, more productive and better quality materials (Silva, 2009). Thus, the execution of high precision experiments requires planning. Therefore, one of the fundamental questions is the appropriate size of the plot or experimental units. Plot sizes tend to increase with the progress of the breeding program, whereas the more advanced populations need larger plot size for experiments. With the advancement of generations, there is a reduction in the variation between the selected materials, requiring a higher number of plants to detect variation and make the selection. When the increase of plot size does not result in more precision, additional increases in accuracy will be obtained with the use of more replicates (Cargnelutti Filho et al., 2012). Several factors are involved in choosing the size and shape of experimental plots. Among them, soil heterogeneity is a crucial factor. Thus it is essential to have information about the area, in which the experiments will be carried out (Storck et al., 2016). Several methods have been used to estimate the optimal plot size, such as the modified maximum curvature method and the linear model segmented with plateau (Ferreira, 2007), either from uniformity assays or experiments that include treatment effects. The intraclass correlation coefficient method stands out among the estimation methods that take advantage of experimental data from