The influence of sodium concentration and osmolality on net water and monovalent electrolyte absorption from or secretion into the intact human colon was studied in healthy volunteers. WHEN ISOTONIC SOLUTIONS CONTAINING NACL AND/OR MANNITOL WERE INFUSED INTO THE COLON: (a) a direct linear relationship between luminal sodium concentration (in the range of 23-150 mEq/liter) and rate of net water, sodium, and chloride absorption was found. No water absorption was found when sodium concentration in the luminal fluid was below 20 mEq/liter; (b) water and sodium absorption from the isotonic test solutions was not enhanced by addition of 80-250 mg/100 ml of glucose; and (c) the rate of water and sodium absorption was decreased markedly when chloride was replaced by bicarbonate in the test solution. WHEN THE COLON WAS PERFUSED WITH HYPERTONIC TEST SOLUTIONS CONTAINING NACL AND MANNITOL OR UREA: (a) water was absorbed from hypertonic NaCl solutions against a lumen-to-blood osmotic gradient of 50 mOsm/kg; (b) when the osmolality of the mannitol solution was increased, water entered the colonic lumen at a more rapid rate. The relationship between the rate of water entering the colon and the osmolality of the test solution was a parabolic one; (c) sodium and chloride entered the colonic lumen at a rate that was lineraly related to that of water entrance when the lumen-to-blood osmotic gradient exceeded 150 mOsm/kg; (d) water flow into the colonic lumen was identical when equimolar urea or mannitol solutions were infused; (e) neither urea nor mannitol was absorbed in significant amounts from the hypertonic solutions; and (f) our results suggest that the equivalent pore radius of the human colon is smaller than the molecular radius of urea (2.3 A).