A new volumetric method for estimating fecundity in Decapoda is presented. Its precision and amount of time spent in the estimation were determined and compared to other two methods available in the literature (Hopkins, 1962, as modified by Diaz, 1974;and Bogorov and Zenkevich, 1947, as modified by Ros and Menocal, 1978). The new volumetric method was found to b e more precise and faster than the other two methods referred.Methods for estimating the number of eggs in Decapoda have been developed in an attempt to determine fecundity in relation to population dynamics. In most such methods a regression or functional relation has been established between volume or weight of the sample and the number of eggs in the berry. Then, the total number of eggs in the egg mass can be estimated. However, chitinous connections among eggs introduce some error in the estimate. The error originates in estimating the volun~e occupied or the weight of the connections. A variable amount of materials such as sand grains and the degree of humidity in the subsample magnify the error. In order to prevent such errors, most methods described in the literature include disaggregation of the egg mass; however, this is tedious and time-consuming (e.g.

Methods and designs have been validated for sampling coral reefs and reef-associated communities, but the accuracy and equivalence of results obtained from studies of mangrove-root communities can not be assured. Thus, it is convenient to establish the accuracy and reliability of sampling methods on a practical basis. Line interception, point interception, and quadrat sampling were used to estimate several basic features of the community associated with mangrove roots. Species-number estimates, cover estimates, sampling efforts, and some statistical properties of the three methods were compared. A sampling design should be chosen if it yields estimates with the greatest accuracy at the lowest cost. Results showed that sampling of mangrove roots could start by using 4 × 4-cm quadrats with a mean sample size of 50 for detecting the number of species, and for establishing species cover it appears that the point-interception method with a sample size of 50 would be appropriate. The effort involved in such a sequence would be approximately 10 h for 20 roots. This combined strategy is clearly laborious and time-consuming in comparison with the use of just a 200-point interception, which would require only 3.25 h for the same number of roots. Although it would cost more than the above methods, a fast and accurate technique based upon existing computer hardware to estimate cover from photographic or video material would result in a census of the whole community. This might prove to be a reliable way to approach the problem.