A technics for prefreezing of blood plasma and serum is described in this paper. The method indicated by Strumia et al. (2), uses a rapid local freezing to obtain the shell-freezing, with refigerated alcohol bath, at temperatures around minus 35ºC. On our work, it has been found that normal horse blood plasma fulfils the instructions given by Strumia, although normal human blood plasma, very often, fails to give the expected results. This is very disadvantageous at the routine work. With the use of small amounts of solid carbon dioxide, spread over the flasks, in the refrigerated bath, it has been possible to start the chrystallization. The technics prescribes a rapid cooling, like the one used by Strumia, to bring the temperature down, to about plus 10ºC. and, with rotating device stopped, the solid carbon dioxide is applied for one minute simultaneously on each flask. Starting rotation again, it begins to form a very uniform shell around the walls of the flasks.