Nerve-growth factor, a substance specifically promoting, in vitro, the growth of nerve fibres from explants of embryonic dorsal root and sympathetic ganglia, has been purified from the venom of Vipera russelli. The purification proved remarkably difficult and presented some unusual features. The final product, which was homogeneous to all the tests available to us, was found to be a glycoprotein of molecular weight 37000 and containing about 20% of carbohydrate. From the values of the sedimentation and diffusion coefficients it may be shown that the axial ratio is approximately 1:7; the molecule has, therefore, an unusual shape. The substance is highly basic (isoelectric point approximately pH 10.5). It is also relatively stable: neither trypsin nor chymotrypsin affect its biological activity. However, activity is completely destroyed on treatment with pepsin. Comparison with nerve-growth factor from other sources does not, with the data at present available, reveal any striking similarities. The difficulties involved in the isolation and characterization of this factor from snake venoms are briefly discussed.