A number of carbamoyl- and N(1)-substituted analogs of physostigmine were synthesized and their in vitro potencies (IC50 values) vs. human erythrocyte and brain (cerebral cortex and caudate nucleus) acetylcholinesterase (AChE) and electric eel AChE and against human brain and plasma butyrylcholinesterase (BChE) were compared to the potencies of physostigmine and other traditional anticholinesterases. In general, increasingly hydrophobic, simple nonbranching carbamoyl groups (as in octyl-, butyl- and benzylcarbamoyl eseroline) did not greatly alter potency vs. AChE whereas increasingly hydrophobic N(1)-substitutions [i.e., N(1)-allyl-, -phenethyl and -benzylphysostigmine] decreased potency vs. AChE. In contrast, increasing the hydrophobicity of both the carbamoyl and N(1) groups increased the potency of the compound against BChE. Furthermore, quaternarization at the N(1) position (physostigmine methosulfate) increased potency vs. AChE but reduced potency vs. BChE. Bulky, branched carbamoyl groups (e.g., N-benzyl-N-benzyl-allophanyl eseroline) were all poor anticholinesterases. N-phenylcarbamoyl eseroline was as potent as benzylcarbamoyl eseroline against AChE yet was 50 to 100 times less potent than the benzyl analog vs. BChE. Therefore, the phenyl substitution appears to increase greatly the selectivity of the compound for AChE. Although it is not possible to determine whether physostigmine analogs that are potent in vitro might be of interest in vivo, these results do show that the structure of physostigmine can be changed significantly while retaining biological activity.