In this publication, we begin the composition of a catalogue of arboreal pollens of the south regions of Brazil, which will serve for a posterior study of the humus formations of Brazil. The sequence of the families corresponds on the phylogeny of the families of plants; for this reason we begin with the polycarpicae, especially with the families magnoliaceae s. lat., Lauraceae, Annonaceae and Myristicaceae. Though the phylogenetic sequence of the different families is yet a very hypothetical one (compare WETTSTEIN, 1944, and HUTCHINSON, 1946 and 1959), we verified similarity between the pollens of the studied families. Comparations between the different genus did not give us a safe phylogenetic result. The pollens of the different genus of Lauraceae, excepting Cryptocarya, look so similar that, by this, we see no possibility for the determination of genus and species. In the contrary, we find in the genus of Annonaceae, and also of the Magnoliaceae s. lat., typical forms and structures. We call attention for the summary of pollen-systematics on the end of this publication. The preparations were manufactured by the method of acetolysis (cit. by ERDTMAN, 1952 and 1954). Because of the very delicate exine of all the pollens, excepting Drimys and Virola, it was necessary to work with great caution to avoid destruction. At the same time, we treated, in all the cases, an object by the method of Wodehouse for comparation (WODEHOUSE, 1935).