The rate of secretion of the adrenal cortex of the anaesthetized dog as measured by the mouse eosinophil test is equivalent to about 11 \ g=m\ g cortisone/g fresh tissue/min.When the cortex of the dog's perfused adrenal is stimulated by ACTH, the increased rate of secretion is registered by the eosinophil test in the same way as by the 'cold test'.When the perfused adrenal was subjected to an increased blood concentration of potassium, only small fluctuations in hormone output were produced. These also were registered by both tests simultaneously.On the assumption that mineralocorticoids are not oxygenated in the C-11 position, these results may be interpreted to mean that, within the experimental errors of the assays, there is no variation in the production of (hypothetical) mineralocorticoids under the conditions of these perfusions.When Speirs & Meyer [1949] published their paper on the use of eosinophil counts in mice for the estimation of adrenocortical hormones, it appeared that their test would be much more sensitive than any known biological assay and permit the estimation of corticoids in arterial blood. This hope, however, has been disappointed, and our work had to be hmited to the estimation, in terms of cortisone, of the cortical hormone in extracts of dog adrenal venous blood under varying conditions. Furthermore, tests were carried out in parallel with a second biological assay, the so-called 'cold test'. The object of parallel assays is to get evidence for or against the identity of a biologically active substance. In this instance, it is known that eosinophils respond exclusively to corticoids with an oxygen atom in the C-l 1 position, whereas the response in the cold test is less specific and can be ehcited by substances without that grouping, for instance, by esters of deoxycorticosterone (DOC) [Vogt, 1951]. For a long time, it has been assumed that the adrenal gland secretes specific 'mineralocorticoids', and though it is no longer considered likely that DOC is a natural mineralocorticoid, the work of Spencer [1950] and of Tait, Simpson & Grundy [1952] and Simpson, Tait & Bush [1952] suggests the secretion of a mineralocorticoid of as yet unknown constitution. Provided this new compound or any other active secretory products of the adrenal cortex were not oxygenated in the C-ll position, they might influence survival time in the ' cold test ', while leaving the eosinophils un¬ affected. The secretion of such compounds might then find expression in discrepancies in parallel biological assays by the two methods. The test for discrepancy was made by comparing the results for each individual sample of plasma in terms of cortisone, further, by altering the rate of secretion and observing whether direction and magni¬ tude of change was reflected in both tests in an identical fashion. In order to alter the secretion rate in a way which might be expected to favour the formation of mineralo-