The effect of missing plants on yield in field experiments with cotton was investigated through a previous study about the relation between mean production by plant and mean available area by plant of a given plot. This relation was described by a 3rd degree polinomial curve but linearity may be assumed between the limits of 0,100m² and 0,300m²/plant, that correspond to current spacings used by farmers. Based on the assumption of linearity, the following adjustment of data is proposed: N Y1a = k1Y1 - b(k1A1 - Sk1A1/N) i = l with b = S (k1A1) (k1Y1) - (Sk1A1) (Sk1Y1)/N S (k1A1)² - (Sk1A1)²/N and A1 = ne(c + 2d) - e(L1 - 2m1d) Y is the actual yield of a plot, N the number of plots in the experiment, k the quotient of perfect stand by actual stand and A is the area considered as really used by the plants obtained by deducting exceeding areas from large skips. Adjustments, made for several experiments, showed that the equation was efficient in most of the cases. However, its validity is questionable when skips are totally or partially linked to treatment effect, and there are cases, when no compensations of the gaps occur, in which the regression of yield on final stand permits better adjustment.