Eighteen wether lambs, surgically equipped with abomasal and ileal cannulae, were used in two metabolism trials. Two levels of dietary Mg (.1 and .2%) and three levels of K (.6, 2.4 and 4.8%, dry basis) were fed with a 2 x 3 factorial arrangement. Each trial consisted of a 10-d preliminary period, a 7-d collection of feed, feces and urine for determining mineral balance and a 6-d sampling of feed, abomasal and ileal fluid and feces to determine flow and site of mineral absorption. Magnesium absorption occurred before the small intestine. Feeding 2.4 and 4.8% K decreased (P less than .05) Mg absorption by 24.4 and 61.2%, respectively. Preintestinal absorption of Mg was depressed linearly with increases in dietary potassium. Larger amounts of Mg flowed to the small intestine and were excreted in the feces when the high level of Mg was fed. A net secretion of Mg into the small intestine followed by a small amount of Mg absorption in the large intestine was also found. Serum Mg was depressed 7 and 10% when diets containing 2.4 and 4.8% potassium were fed, respectively. Magnesium level did not affect K absorption or balance. Potassium absorption, retention and urinary excretion increased (P less than .05) with increasing K level. Animals fed the low K diet had a negative absorption of K preintestinally with the small intestine being the main site of absorption. Increasing K level resulted in approximately equal amounts of K being absorbed in the stomach and small intestine.