We have devised a colorimeter for the clinical estimation of hemoglobin which possesses several advantages over instruments now in use. Those methods for the estimation of color density, which depend upon the comparison of a sample with an interrupted series of standards, have always been the most satisfactory. The eye is able to place a sample with more certainty between two members of a variable series than it is to compare a sample with a uniformly graded scale. The latter only becomes accurate when the error is absorbed into a great number of readings, and as yet there had been no instrument devised superior to the Fleischl-Miescher for such a procedure. For a single operation, comparison with a series offers much more certainty as to the correctness of the choice.
We have adopted the principle of the Sahli hemoglobinometer, modifying it to satisfy the above concepts. The Sahli instrument