The biovailability of theophylline from alcoholic and aqueous oral solutions was compared to that from an intravenous dose in 12 normal adults. The alcoholic elixir surprisingly gave rise to a significantly greater (114 ±14%, mean±sd amount absorbed than did the intravenous dose. The aqueous solution (99±8%) and intravenous dose were statistically indistinguishable in this respect, and, furthermore, the extent of absorption from a 300-mg dose of the aqueous solution was 99±10% of that from a 500-mg dose, and not statistically different. The aqueous solution was thus employed in three subsequent studies as a standard with which to compare 13 different types of theophylline tablets, all marketed in the United States. Of the 13 tablets, eight showed bioavailability statistically distinguishable from that of the standard. Nevertheless, for only two tablets could it be claimed with 95% confidence that the bioavailability was less than 95%. For none can it be stated at this confidence level that the bioavailability is less than 90%. Bioavailability studies should include criteria of clinical significance in addition to criteria of statistical significance. Contrary to the usual rationale behind choice of a bioavailability standard, nine of the 12 uncoated tablets appeared to allow more rapid absorption of theophylline than did the standard oral solution, an aqueous syrup. Increasing the dose of syrup decreased the rate of theophylline absorption. Orally administered drug solutions may have properties more absorption rate limiting than the disintegration of many brands of tablet.