Three experiments were done to determine the effects of administering procaine penicillin G to broiler chickens. In exps. 1 and 2, groups of 100 male, day-old chicks received one of the following treatments: (1) control, no penicillin; (2) 27.5 mg penicillin kg−1 added to the diet; (3) an amount of penicillin added to the drinking water equivalent to the amount consumed the previous day by the birds in group 2; and (4) same route of administration as group 3, except the dose was one half of the daily intake by group 3. In exp. 3, both female and male (1200 each) broiler chickens were given diets that contained either 0, 1.1, 2.2 or 4.4 mg kg−1 of penicillin. Each experiment was terminated when the birds were 42 d of age. Route or level of penicillin administration had no effect (P > 0.05) on mortality or feed conversion in any of the experiments. When the penicillin was added to the diets of the birds in exps. 1 and 2, their body weights at 21 and 42 d were 3.5 and 2.9%, respectively, less than those of the control group, but when it was administered via the drinking water the weights were 1–3% higher at these ages than those of the control group. Lower monetary returns (> 0.05) were obtained for the chickens in the penicillin groups than in the control group; these differences (6.6–15.6%) were due mainly to the cost of the antibiotic. In exp. 3, higher body weights (P < 0.05) at 21 d (3.6–5.4%) but not (P > 0.05) at 42 d were obtained for the birds given penicillin-supplemented diets than for those fed the control diets. No measurable levels of residue (limit of detection 100 μg kg−1) were found in samples of kidney, liver and muscle (breast and thigh) from birds in exp. 2, which had received 27.5 mg penicillin kg−1 for the 42-d experimental period. Key words: Broiler chickens, growth promoter, procaine penicillin, route of administration, residue levels