This study was designed to evaluate a new drug delivery system. The authors undertook to determine if microdroplets prepared by encapsulating volatile anesthetics with a membrane of lecithin could be used for local anesthesia. Local anesthesia was determined by monitoring the response of the rat to tail clamping and electrical stimulation of the skin following the intradermal injection of the microdroplets. Microdroplets were prepared from isoflurane, enflurane, halothane, methoxyflurane, diethyl ether, chloroform, and heptane. Although all microdroplet preparations produced local anesthesia, only methoxyflurane microdroplets produced an ultra-long duration of local anesthesia (approximately 24 h). Further characterization of the methoxyflurane microdroplets revealed two important differences from conventional local anesthetics. First, the local anesthetic effect of methoxyflurane reached a plateau that did not change significantly for 20 h while the injection of lidocaine and bupivacaine resulted in a peak effect that returned to baseline within 1 and 3 h, respectively. Second, the anesthetic effect of methoxyflurane remained essentially localized to the site of injection, while the anesthetic effect of lidocaine and bupivacaine migrated 15 cm in less than 1 h. The toxicity and safety of methoxyflurane were evaluated. When administered over the dosage range 1–16% (v/v) intradermally, or by injections into muscle, or by repeat injections every 4 days for 16 days, all animals regained their pre-treatment response to painful stimulations, and there was no evidence of gross injury to tissue. Deliberate intravenous injection of 0.8 ml of 6.7% (v/v) methoxyflurane microdroplets had no apparent anesthetic or toxic effect. The present study demonstrates that methoxyflurane microdroplets produce an anesthetic effect that is highly localized, stable in intensity, ultra-long in duration, and reversible.