Evaluation of Statistic Efficiency for some Characters of Artichoke Heads and Leaves.

A comparison test has been carried out for two years among 7 artichoke cultivars factorially combined with 3 plant densities, and for each of 42 obtained combinations, 45 artichoke plants, randomly chosen, were analysed for: weight, height and width of the main head, number of leaves per plant and their mean length, number and mean weight of secondary heads. For each of the examined characters and for each of the 42 samples, through a suitable statistical method, it was possible to evaluate the sample dimension, obtaining mean values characterized by half-widths of the tolerance interval, equivalent to 12.5; 10; 8; 6; 5; 4; 3; 2 % of arithmetical mean. This survey provided useful information. In particular it was evaluated a slight number (ni) of observations to carry out when the length and the width of the main head has been examined. Larger dimensions, but always measured, of the sample were considered in the evaluation of the number of leaves per plant, of their width and of the mean weight of the main head. Even larger samples are necessary for a careful evaluation of the mean weight of the secondary heads. It is more difficult to find out the number of secondary heads, because the values observed in each sample had frequency distributions that could not be considered, even appreciatively, normal distribution (of Gauss type). Consequently the high sample dimensions evaluated are found out to be hardly reliable.