Abstract Prolonged observations were made on the characteristics and the mechanism of production of the low potassium syndrome in a forty-two year old Chinese male with nephritis of unknown etiology. The syndrome consisted of intermittent bouts of muscular weakness, atony of the bladder, constipation and electrocardiographic changes of a broad flattened T wave, prolonged Q-T interval, prolonged P-R interval and dropped beats, all associated with low serum potassium levels. The plasma potassium level varied between 1.5 and 2.3 mEq./L. when the patient was on a regular diet (104 mEq. potassium daily), fell as low as 1 mEq./L. on a low potassium intake (30 mEq. potassium daily) and rarely exceeded 4 mEq./L. even when the potassium intake was six times greater than normal (625 mEq. potassium daily). The potassium balance was negative on an average potassium intake and became positive only when the daily potassium intake was increased to 425 mEq. or more. In general the serum sodium levels were low and sodium balance tended to be negative except when sodium intake was high. A reciprocal relationship between the potassium and sodium balances was apparent when the potassium intake was varied. Variations in sodium intake and balance, however, were not attended by reciprocal changes in the potassium balance. The plasma chloride level was usually low and chloride balance appeared to be related to the sum of the sodium and potassium balance in that chloride was retained when base was retained and vice versa. Potassium depletion appeared to affect adversely muscle strength and the vegetative nervous system. The severity of signs and symptoms referable to these systems varied in the presence of a constant plasma potassium level; nor was development of symptoms inevitable at any particular plasma potassium level. The height of the T wave of the electrocardiogram could be correlated with the plasma potassium level in acute experiments after potassium administration. Although electrocardiographic abnormalities occurred only during periods of potassium depletion, the day-to-day correlation between the height of the T wave and plasma potassium levels was poor. Glomerular filtration rate, renal plasma flow and maximum capacities of the renal tubules to excrete p-aminohippurate and to reabsorb glucose were approximately one-third of normal. The rate of glomerular filtration, but not the other functions, varied directly although not precisely with the level of daily potassium intake. During periods of low potassium intake and in spite of very low plasma potassium levels the kidneys continued to excrete urine containing appreciable amounts of potassium. Loss of potassium in the urine was not due to inability of the kidneys to defend by the usual mechanisms against either alkalosis or acidosis. Under conditions of low potassium intake as much as 50 per cent of the potassium filtered at the glomeruli appeared in the urine (normal = less than 20 per cent). Tubular excretion of potassium was easily demonstrable whenever the patient was on a high potassium intake. Potassium given during a period of potassium depletion and at a time when an abnormal amount was being lost in the urine demonstrated that the tubules were capable of reabsorbing additional potassium. A similar experiment performed during a period of high potassium intake resulted in a greater increase in the amount of potassium in the urine than in the glomerular filtrate. It is probable that both decreased absorption and increased excretion of potassium were a part of the renal tubular defect responsible for the excessive loss of potassium in the urine.