Using conscious gastric fistula dogs, gastrin release stimulated by a 360-g meat-based solid meal was compared to that stimulated by two isocaloric (1 kcal/g), hypertonic (approximately 700 mosm/kg) nutrient liquids (Sustacal and Vivonex) and by two nonnutrient, hypotonic (60 mosm/kg) liquids (mannitol and coffee). Serum gastrin levels were measured at 15- to 30-min intervals over 120 min. Without a meal, serum gastrin levels remained stable. Effectiveness in stimulating gastrin release was coffee = mannitol less than Sustacal = Vivonex less than solid food; 2-hr integrated gastrin responses were 1.2, 2.6, 4.3, 5.6, and 12.2 ng/ml/min, respectively. The greater gastrin responses produced by nutrient liquids and meat meals could be explained by slower emptying and delayed acidification of gastric contents. We conclude that solid meals are preferable to liquid meals in studies of antral gastrin release.