The correct experiment design should be used to reduce the experimental error and, with that, maximize the precision of the information obtained. The objective of this work was to estimate the minimum number of families and of plants per plot for evaluation of root traits in a carrot population cultivated in agroecologic systems. The essays were conducted in the summer of 2006/2007 in two agroecologic properties in Brasília: "Associação Mokiti Okada", which follows model of Natural Agriculture in Brazlândia and "Núcleo Rural Taguatinga", following model of Organic Agriculture in Taguatinga, Brazil. We evaluated 100 half-sib families of carrot arranged in randomized block design with two replications and plots of 1 m². Twenty competitive plants per plot were harvested and individually evaluated for the traits: root length, root diameter, root mass, root xylem diameter, relationship of xylem diameter/root diameter, type of root tip, type of root shoulder, parameter a* of root xylem and phloem. Joint variance analysis of the group and of each system, were performed with information between and inside plots. Analyses of representativity minimum number of families and of plants to represent a carrot population were performed. A sample of 18 plants competitive/plot collected in essays with 2 replications, as well as 74 families, are enough to guarantee an appropriate evaluation of carrot half-sib families for the traits studied.
Daucus carota L.; minimun number of plants; minimun number of families; simulation
