96 white males, 18-50 yrs., old were painstakingly selected with a view to excluding any factor likely to impair maximal urinary concentrating and diluting power; then they were subjected to a slightly modified Fishberg test. At first try, 76% concentrated to 1.022 and 58.3% diluted to 1.002, which are Fishberg''s standards. Repetition of the tests would sometimes, but not always, improve the results. Water excretion was the most inconstant and unreliable factor of all, extremes varying up to 300% with successive tests upon the same individual. If it were found feasible to modify Fishberg''s standards to 1.020 and 1.005, the percentages of subjects who attain these levels could be increased to 91.6 and 92.6, respectively.