After pointing out some of the aspects not yet well known of scarlet fever epidemiology in Brazil, the A. shows that during period 1940-44 the disease in the city of S. Paulo prevailed mostly in the austral autumn season (april-june); however no correlation, statistically significant, could be obtained comparing monthly waves of temperature and corresponding waves of the disease. Such a correlation has been found out for small-pox, in the same fiver-years period, both in Rio (-0.48 ± 0.10) and S. Paulo (-0.30 ± 0.12); and with absolute humidity variations (also in the previous month) in Belem (-0.38 ± 0.11) and Rio (-0.50 ± 0.10) - no data about A. H. being available from S. Paulo, Belem, Rio and S. Paulo showed a small-pox prevalence in the austral winter-spring seasons. Both in tropical cities (Belem, Recife, Salvador, Rio) and in temperate ones (Curitiba and Porto Alegre) - in S. Paulo the disease is not subjected to compulsory notification - measles distribution by four-months period --selected in accordance with the highest or lowest values of mean temperature and absolute humidity - induced to suppose that the disease was more uniformly associated with a low temperature. Several correlation coefficients statistically significant have been indeed obtained, between monthly morbidity rates and temperature variations in the same month (Recife - 0.26 ± 0.12, Salvador - 0.36 ± 0.11, Rio - 0.50 ± 0.10) and in the previous one (Salvador - 0.45 ± 0.10, Rio - 0.60 t = 5.72 and Porto Alegre - 0.38 ± 0.11); but also such coefficients have been found out between the same rates and absolute humidity variations in the same month (Recife - 0.27 ± 0.12, Salvador - 0.29 ± 0.12, Rio - 0.53 ± 0.09) and in the previous one (Salvador - 0.31 ± 0.12, Rio - 0.68 t = 7.08 and Porto Alegre - 0.35 ± 0.12).