A Microcolorimetric Method for Measuring the Oxygen Saturation of Blood

For purposes of measuring oxygen saturation in the arterial blood of newly born infants (1) it became necessary to develop a rapid and reasonably accurate method for small blood samples. It was further necessary to demonstrate that cutaneous blood samples could be considered identical with specimens drawn directly from an artery. The present communication describes the calorimetric method evolved for measuring blood oxygen content and capacity, and indicates its accuracy in oxygen determinations upon cutaneous blood, as compared with the results of Van Slyke-Neil1 manometric analyses (2) upon blood simultaneously drawn from the arteries themselves. Drabkin and Austin (3) showed in 1935 that Beer’s law of proportionality and absorption held good over an extensive range of concentrations for the pigments HbOz and Hb, a demonstration made possible by their specially devised chamber or cell. Solutions (such as blood) too concentrated in color for analysis by usual spectrophotometric methods could be spread in this apparatus to a film 0.07 mm. in thickness, thus allowing the transmission of light for registration by the spectrophotometer. Applications of the color differences between oxyhemoglobin and reduced hemoglobin to the study of oxygen saturation of blood in viva have been made by Kramer (4) and by Matthes and Gross (5). The principle is also utilized in the apparatus devised by G. A. Millikan. In such techniques a semimembranous portion of the body such as the pinna of the ear or the web between the thumb and fingers is placed between a source of light and the photoelectric cell. With suitable filters, or, as in Kramer’s method, without them, the changes in color range between that of HbOz and of Hb can thus be registered and converted to terms of oxygen saturation of the flowing blood. Brinkman and Wildschut (6) proposed a somewhat different method for the in vitro determination of oxygen saturation of a small sample of cutaneous blood. This required two galvanometer readings, the first from the original sample of blood as drawn, the second from a sample of the same