To the Editor.—— Chen and Magalhaes1recently showed an inverse correlation between serum albumin and paraprotein concentrations in patients with IgG multiple myeloma, but not in patients with IgA or light-chain myeloma. In exploring the implications of this relationship, several possibilities come to mind. First, the low albumin level might be due to cachexia. This is consistent with the correlations between albumin concentration and clinical stage of disease (see Fig 41 and estimated body myeloma tumor burden (see Fig 51). Second, albeit immunoglobulins have much less effect on oncotic pressure than albumin, unusually high concentrations of immunoglobulins might signal the hypothetical oncostat to reduce albumin synthesis. This is consistent with the facts that diminution of serum albumin in severe nephrotic syndrome and its absence in hereditary analbuminemia are associated with increases in other serum proteins, including lipoproteins. It is hard to think of a critical test of this hypothesis, ie,