This study was accomplished on a more comprehensive basis to evaluate previous questions that were raised from a preliminary article about the genetic structure of Cryptocarya moschata populations. Thus, through the analysis of 40 polymorphic allozyme loci, allele frequencies were estimated from 335 individuals of 11 natural populations of C. moschata from six hydrographic basins of São Paulo state and Serra da Estrela, Rio de Janeiro, Brazil. Estimates of Wright's F statistics were done through the analysis of variance, presenting average values of <img border=0 width=32 height=32 id="_x0000_i1026" src="../../../../../../img/revistas/bn/v4n1/img/a04car(f).jpg" align=absmiddle > or = 0.352, <img border=0 width=32 height=32 id="_x0000_i1027" src="../../../../../../img/revistas/bn/v4n1/img/a04car(0p).jpg" align=absmiddle > or = 0.285 and <img border=0 width=32 height=32 id="_x0000_i1028" src="../../../../../../img/revistas/bn/v4n1/img/a04car(f2).jpg" align=absmiddle > or = 0.097. These results indicated that individuals within populations must be panmitic, and that the diversity among populations is fairly high, being superior to what would be expected for groups of plants having a full-sib family structure. From estimates of <img border=0 width=32 height=32 id="_x0000_i1029" src="../../../../../../img/revistas/bn/v4n1/img/a04car(0p).jpg" align=absmiddle>obtained for populations taken two at a time, the model of isolation by distance was tested; data did not fit the model, showing that <img border=0 width=32 height=32 id="_x0000_i1030" src="../../../../../../img/revistas/bn/v4n1/img/a04car(0p).jpg" align=absmiddle>did not increase by the respective increasing of the geographic distance. The estimated gene flow of 0.55 migrants per generation corroborated the pronounced populational differentiation, indicating that drift effects should be more important than the selection ones. The effective population sizes found from the sampled populations showed that there was an adequate genetic representativeness of the samples for those with relatively low values of <img border=0 width=32 height=32 id="_x0000_i1031" src="../../../../../../img/revistas/bn/v4n1/img/a04car(f2).jpg" align=absmiddle>. Though, under a metapopulation context, the effective population size was 17.07 individuals, indicating that sampling performed for the species corresponded to 88.44% of the maximum effective size obtained from 11 populations with a <img border=0 width=32 height=32 id="_x0000_i1032" src="../../../../../../img/revistas/bn/v4n1/img/a04car(0p).jpg" align=absmiddle>of 0.285, equivalent to only 5.09% individuals for the total sampled. Management and conservation strategies aimed at preserving high intrapopulation genetic variation in C. moschata would imply in the maintenance of populations with great number of individuals. Moreover, for the preservation of the species as a whole, the maintenance of many such populations would be mandatorily recommended, which denotes that the conservation of large areas of Atlantic rain forest should be necessary to hold its evolutionary dynamics.