The present research was accomplished in the Forest Mixed Ombrófila of São Francisco de Paula - FLONA, being this representative of great diversity in the State of Rio Grande do Sul. The objective was to esteem the accumulated burlap collected with four different collecting frames and the amostral intensity in each seasonal station. 15 random samples were collected for each frame by hectare in each seasonal station being separated in the stratus leaves, peels, branches and miscellany. It was used a completely random design in each station, for subsequent accomplishment for the united analysis of the experiments. The means, the variance, and the variation coefficient was calculated for all the strata, being accomplished the test of homogeneity of the variances. Through the simple aleatory sampling, being admitted a maximum sampling mistake of 10, 15, 20 and 25% of the esteem means, with 5% of error probability, the sampling intensity was calculated by hectare, for infinite population. The interaction between frame and station, in the joint analysis, was not significant for the total and the stratus, soon the behavior of the frames was the same in the different stations. The variances, verified among inside the frames of the stations, were homogeneous in relation with the accumulation of burlap. The effect of the frames for the total and for the stratus, branches, peels and miscellany they didn't differ to each other, only presenting significant difference in the leaves. There were not significant differences among the frames for the total, the most appropriate was like this the very small frame (0,25m²), and in the winter with 27 samples, in the spring, 25, in the autumn, 33 and in the summer, 15 units, are necessary to obtain 15% of difference between the means.
size of plot; sufficiency samples; homogeneity of variances
