In this paper, preliminary to a series of investigations that the A. has the purpose to make about the influence of climatic factors particularly upon the prevalence of the most important acute infectious diseases in Brazil, he raises the question whether such factors do affect in this country the total death rates, as it is reasonable to suppose, according to what has been observed in temperate zones of northern and southern hemispheres. The inclusion of absolute humidity among other climatic factors to be dealt with seems justifiable according to Rogers and Stallybrass. Owing to scarcety of reliable data the A. was obliged to limit to a five-years period (1940-1944) the complete proposed investigation, which includes seven of the most important cities, scattered throughout the brazilian territory, from north to south - Belém, recife, Salvador, Rio, S. Paulo, Curitiba and Porto Alegre. Reference is made to their normal climatic conditions and monthly death-rates variations with their mean values and standard deviations. In a first part dealing with seasonal variations only for purposes of comparison, he points out that in there tropical cities of Brazil, without very clear seasonal differentiation, the curve of general mortality reached its highest point in austral autumn season and the remaining four (including Rio near the tropic) in the spring, with the exception of Curitiba, where the peak coincided with the summer season. He shows how such important causes of deaths, as diarrheas, common respiratory diseases and tuberculosis, whose seasonal distribution for each one of the seven cities is referred, may explain such seasonal variations. On a second part, a study is made of the general mortality distribution by four-months periods selected in accordance respectively with the highest or lowest values of rainfall and of mean temperature and humidity during period 1940-1944. Finally he compares the monthly waves of such climatic factors and the corresponding waves of total death - rates and finds through correlation coefficients 17 significant values with respect to their standard errors. Variations in the death - rates seemed to be perhaps more closely and uniformly associated with variations of mean humidity, as is indicated by coefficients ranging from + 0.3 to 0.6.