1. There is a good correlation between dose of contrast medium injected and plasma concentration; this is initially largely independent of renal function. 2. The minimum concentration of contrast medium in the glomerular filtrate necessary to produce an appreciable nephrogram is probably in the region of 70 mg. iodine per cent. 3. If contrast medium is injected slowly over 8 to 10 minutes, the peak plasma concentration is not reached until 15 minutes and should be the period of maximum density of the nephrogram, and optimum period for nephrotomography. 4. Contrast medium acts as a simple osmotic diuretic; the concentration of contrast medium in the urine is accurately reflected by the diuresis, producing a "mirror image." 5. In patients with normal renal function and deprived of fluid for some 12-15 hours, the concentration of contrast medium in the urine continues to increase with increasing dose up to 140 cc. urografin 76 per cent, despite an increasing diuresis. 6. In patients with normal renal function the "optimal" dose of contrast medium, or one which will produce no further increase in concentration of contrast medium in the urine, will depend on the state of hydration of the patient at the start of urography. The better the dehydration, the higher will be the "optimal" dose. Maximum dehydration is not achieved with the usual period of 8 to 12 hours fluid restriction before urography. 7. Even in high dose urography, dehydration will improve the concentration of contrast medium in the urine in patients with normal renal function. 8. Drip infusion pyelography offers no advantage over high dose urography in which the equivalent dose of undiluted contrast medium is injected; a rapid injection will cause a higher peak plasma concentration and a denser nephrogram than a slow infusion.