A CHARACTERISTIC finding in malignant disease is an increase in the circulating glycoproteins, particularly the serum al-globulin. Such a change has frequently been reported in man (Winzler, 1960; Tombs, James and Maclagan, 1961) and in animals (Darcy, 1960). There has been some speculation as to the source and mechanism of production of the excess ax-glycoprotein. Suggestions have ranged from an increased production by the normal mechanism, or by injured or necrotic tissue, to a preferential utilisation of non-glycoproteins, leaving an apparent preponderance of glycoprotein in the serum (Shetlar, 1961).Good evidence that the liver is the normal source of serum acx-glycoproteins has only recently become available (Spiro, 1959;Hochwalde, Thorbecke and Asofsky, 1961). If, in cancer, the liver is stimulated to produce more ax-glycoprotein, it might be expected that the organ itself would contain more than it normally does. Very little evidence is available at present on the distribution of acidic glycoproteins within normal tissue or tumours. In a previous investigation (Tombs, Burston and Maclagan, 1962) chromatographic analysis of the cytoplasmic proteins of a number of organs showed that a glycoprotein very similar to if not identical with the serum a,-globulin was widely distributed; in particular it was present in the liver. Quantitative analysis of several human livers showed that in cancer the more basic and exclusively cytoplasmic proteins were depressed, and the acidic components, including the oc,-globulin were increased. This was so whether tumours were present in the liver or not. These results suggested that the tumour in some way caused the liver to produce more a,-globulin than it normally does. However, it was felt desirable to obtain more precise estimates of individual proteins than was possible from the somewhat heterogeneous chromatographic fractions. Thus in contrast with the previous experiment, which was a general analysis of all the cytoplasmic proteins, we sought a method of obtaining a,-globulin component in a pure state, and its relatively precise estimation. Perchloric acid was introduced (Winzler, Devor, Mehl and Smyth, 1948) as a reagent for the selective denaturation and precipitation of all the serum proteins except a glycoprotein fraction which remains soluble. While this fraction is by no means homogeneous (Biserte, Havez, Laturaze and Hayem-Levy, 1961) the major component is the a1-glycoprotein. It seemed reasonable to suppose that the use of perchloric acid on tissue extracts would give good preparations of the alglobulin. This was found to be so, but unexpectedly a number of basic cytoplasmic proteins were also found to be soluble in perchloric acid: these substances, one of which exhibits variation in cancer, have not previously been described.