Of the instruments for measuring the depth of colour of the blood, the double pipette of Hoppe-Seyler is quite the most delicate. A solution of carbonic oxide haemoglobin, accurately titrated, serves as the standard of comparison. The reliable preparation and conservation of the normal solution is however attended with such difficulties, that this method is not clinically available. In the last few years, Langemeister, a pupil of Kuehne's, has invented a method for colorimetric purposes, also applicable to haemoglobin estimations. The instrument depends on the principle, that from the thickness of the layer in which the solution to be tested has the same colour intensity as a normal solution, the amount of colour can be calculated. As a normal solution Langemeister uses a glycerine solution of methaemoglobin prepared from pig's blood. To our knowledge this method has not yet been applied clinically. Its introduction would be valuable, for in practice we must at present be content with methods that are less exact, in which coloured glass or a stable coloured solution serves as a measure for the depth of colour of the blood. There are a number of instruments of this kind, of which the "haemometer" of Fleischl, and amongst others, the "haemoglobinometer" of Gowers, distinguished by its low price, are specially used for clinical purposes. Both instruments give the percentage of the haemoglobin of normal blood which the blood examined contains, and are sufficiently exact in their results for practical purposes and for relative values; although errors up to 10% and over occur with unpractised observers. (Cp. K. H. Mayer.) Quite recently Biernacki has raised the objection to the colorimetric methods of the quantitative estimation of haemoglobin, that the depth of colour of the blood is dependent not only on the quantity of haemoglobin but also on the colour of the plasma, and the greater or less amount of proteid in the blood. These errors are quite inconsiderable for the above-mentioned instruments, since here the blood is so highly diluted with water that the possible original differences are thereby reduced to zero.