A low-magnesium diet was fed to four sheep, each of which had been surgically prepared with a rumen fistula, a tube into the cranial one-third of the omasum, a tube to the cranial one-third of the abomasum and a re-entrant duodenal cannula. Mg, as gluconate or acate or acetate, was continuously infused for 12-14 d in turn into (1) the caudal duodenal cannula, (2) the abomasum, (3) the omasum, (4) the rumen. A continuous infusion of the chromium-ethylenediaminetetraacetic acid complex (CrEDTA) was maintained to the rumen. The abomasal effluent which flowed through the cranial duodenal cannula was continually sampled and the flow of Mg calculated from the concentrations of Mg and CrEDTA. Blood and rumen fluid samples were taken and urine and faeces collected during each period of Mg infusion. 2. The Mg infused to either the abomasum or omasum was completely recovered at the duodenum, indicating a lack of net absorption of Mg from these stomach compartments. In contrast, 13-7-18-7 mmol (36-61%) of the Mg infused to the rumen was not recovered at the duodenum which suggested that a substantial net absorption of the infused Mg occurred from the reticulo-rumen. Absorption of Mg caudal to the pylorus was not related to the site of Mg infusion and averaged 3-28 +/- 0-56 (SEM) mmol/d. 3. Compared with the intraruminal infusion, the post-ruminal infusion of Mg was associated with decreased plasma and rumen fluid Mg concentrations, decreased urinary Mg exretion, decreased Mg balance and increased faecal Mg excretion. 4. It is concluded that no significant absorption of Mg occurs from either the omasum or abomasum in sheep and that the reticulo-rumen is the principal site of Mg absorption before the pylorus. Absorption of Mg post-ruminally is insufficient to maintain normal Mg status in the animal.
