In a recent article discussing the means of preventing the occurrence of chills following the transfusion of citrated blood, Lewisohn and Rosenthal1 in their conclusions said: "The use of solutions prepared with triple distilled water is indispensable." This advice has become traditional in our literature since the time when it was first introduced in an attempt to avoid the reactions that followed the intravenous administration of arsphenamine, apparently on the theory that, if one distillation is good, three distillations must of necessity be three times as good. The question, however, appears not to have been subjected to critical analysis.
In the ordinary practice of redistillation, distilled water is taken from a still of the Stokes, Barnstead or similar design and distilled again in a glass apparatus, usually made of pyrex glass but sometimes made of quartz, in which there is a tight connection, sometimes glass to glass, between the