A hypothesis that the ionic composition of drinking water might affect development of the ascites syndrome in broilers was investigated in two trials. The first trial comprised four groups of 650 male chicks. A control treatment was normal tap water and the other three treatments comprised the addition to the tap water of 1,000 mg/L sodium as NaCl, 5,000 mg/L NH4Cl, or 5,000 mg/L KHCO3, supplied from age 2 to 47 d. At Day 28, equally sized subsets of these groups were moved to individual cages, where they received a severe exposure to ambient cold. The development of the ascites syndrome was monitored by measurements of hematocrit and arterial blood oxygen saturation (PaO2) by oximetry, body weight, and examination of dead birds for cause of death. Mortality from ascites in cold-exposed birds from Days 28 to 47 was 28, 48, 40, and 16% in the tap water, NaCl, NH4Cl, and KHCO3 groups, respectively; only the NaCl mortality was significantly different from the tap water mortality. The KHCO3 treatment increased PaO2 (compared with tap water treatment) at Day 28 by 5.5% and at Day 35 by 10.5%, but not at Day 42. The KHCO3 caused a reduction in body weight, which was 13% less than the tap water group at Day 42, probably due to a chronic toxicity. The second trial specifically examined the same parameters with lower water levels of KHCO3 (3,000 and 1,000 mg/L), in comparison to a 10% feed restriction protocol, in order to clarify whether the increased PaO2 was due to a specific effect of the KHCO3 or was a metabolic manifestation of a reduced growth rate. The 3,000 mg/L KHCO3 treatment had no effect on PaO2, but the 1,000 mg/L treatment augmented PaO2 by 5.3% at Day 35 (but not at Days 28 or 42), without reducing the final body weight. The feed restriction group showed an elevated PaO2 of 5.4% at Day 35 (but not at Days 28 or 42), with no reduction in the final body weight. The inclusion of 1,000 mg/L of KHCO3 into the drinking water of broilers or a temporary 10% feed restriction may be means to augment PaO2.