A study was made of the renal pathology of 43 consecutive diabetic autopsies, and the pathology grouped under five headings, separating those cases who showed Kimmelstiel Wilson lesions (11) from the other groups, the Arteriosclerotic (4), Arteriolosclerotic (3), Pyelonephritis (4) and "other diabetic" group (21) . A comparison of the clinical syndrome revealed that the Kimmelstiel Wilson lesions occur principally in elderly females, who show albuminuria and often renal infection. The lesion was characterised by the presence of Hypertension with heart failure, Oedema at least of the dependent parts, Peripheral neuropathy, Diabetic retinopathy where the fundus is not obscured by cataract. This syndrome was not found togther in any of the other groups. The literature was discussed on these points. A further series of observations were then made on 9 patients who exhibited the syndrome outlined. They showed that while the syndrome could be discovered for the first time in its fully developed state (3 cases) that the march of progress shown by the other six cases showed that, while eye changes, oedema and peripheral neuropathy were early signs, hypertension developed later and progressed to failure if the disease was not terminated by one of the other hazards. Two cases coming to post mortem confirmed the clinical diagnosis. A study of ancillary methods of diagnosis showed that the presence of doubly refractile bodies in the urine was a useful point of differential diagnosis from all but the nepbr oti c stage of chronic glomerulonephritis, whilst that condition might be distinguished by repeated Addis counts. The massive degree of albuminaria was found to be a useful point in the fully developed syndromes, but not of clear differential diagnostic value, and liable to extra-renal variations. A study of renal clearances was made on normal diabetics (13) , advanced art eri o s cl er oti cs (5) , those with Kimmelstiel Wilson syndromes (7), and 1 case of chronic glomerulonephritis. The technique developed is described. These studies showed that the normal diabetic had no demonstrable alteration in renal function. The Kimmelstiel Wilson cases bad a considerably reduced function, the glomerular filtration rate being more reduced than the tubular maximum capacity for reabsorption of glucose, with the result that the renal thresholds tended to be high. The advanced arteriosclerotic cases had a more severe and more generalised disturbance of function with the result that their thresholds were low. The suggestion is made that the raised thresholds in Kimmelstiel Wilson syndromes may explain the delay in diagnosing the diabetes which is encountered, or even some of the cases which show the lesions at autopsy but were not known to be diabetic in life. The suggestion is also made that the high thresholds may lead to inadequate control if urine estimations only are used, and that this may be at least an aggravating factor in the production of the full degenerative picture of the Kimmelstiel Wilson Syndrome.