IT has been found that prolonged administration of ‘luminal’ (and other barbiturates) to human beings (for example, epileptics) and animals greatly reduces the cholinesterase activity of the serum, so that it reaches 10–20 per cent of what it was before the treatment with the drug. No change in cholinesterase activity was found, however, when the drug was added to serum in vitro (2.5 mgm./c.c.). It appears that the barbiturate does not act directly on the enzyme asin the case of eserine and other anticholinesterase substances. The pronounced decrease in cholinesterase activity which was found in vivo might be interpreted as a counter-adaptation, the activity of the cholinergic system being reduced by the narcotic, reducing thereby the demand for cholinesterase, which, consequently, slowly decreases.