All who are concerned in the treatment of cancer must be interested in any evidence that is brought forward of a constitutional effect produced by the disease, particularly if it is of such a nature that it will lead to a laboratory test that will help in the diagnosis or guide treatment. Many such claims have been made and we publish today an article in which a modification of the Bendien test is described and in which a claim is made that, by this means, useful information can be obtained as to whether metastases are present or not.