Estimate of the number of experiments and importance of morphoagronomic characters in lettuce

Abstract

In scientific research, the repetition of the experiment and the selection of variables-response improve the precision of the results and make the conclusions more assertive, allowing generalizations from the joint analysis. However, there are no official rules about how many times experiment should be repeated, as well as which variables best discriminate the evaluated treatments. For this reason, this research aimed to estimate the number of repetitions of experiments and indicate the importance of morphoagronomic characters of lettuce cv. ‘Veneranda’. Data from four experiments were considered, in which four soil covers were evaluated (SoilTain DW^® geosynthetic blanket, white and black polypropylene mulching and uncovered soil) in the completely randomized design, with five replications. The number of experiments was estimated by the repeatability coefficient and the importance of the characters, estimated by Singh criteria. In general, the repeatability coefficients were of medium magnitude, in which the number of measurements ranged from 1 to 9 experiments and, for yield, four experiments are enough to accurately indicate the best treatments. The most discriminant variables were the number of leaves (39.47%) and their length (20.79%) and plant height (17.06%); leaf width and root mass variables may be discarded in future studies (0% contribution). The results indicate that the execution of a single experiment, common in lettuce researches, does not allow accurate conclusions, and should be a criterion to be considered in the experimental planning.