Table 1. Substituent (ring B) -ortho -meta -paraF + 0-17 (a) + 0-31 (a) + 0-15 (a) + 0-337 + 0 -062Cl + 0-22 (a) + 0 -38 (a) + 0-20 (a) + 0-17 (6) + 0-373 + 0-227Br + 0-25 (a) + 0-24 (6)+ 0-232 CH3 - 0-14 (a) - 0-07 (a) - 0-14 (a)- 0 -069 0-170 NO2 + 0-75 (a)+ 0-710 + 0-778 OH+ 0-05 (a) - 0 -32 (a) + 0-121 - 0-37OCH3 - 0-34 (a) + 0-04 (a) - 0 -28 (a) + 0-11 - 0-27 (6)- 0 -268 N(CH3)2 - 0-80 (a)- 0-91 (b) - 0-83Values (a) obtained with ring A unsubstituted, with carbon tetra-chloride as solvent (K = 0.033). Values (b) obtained with a 4-N02-group in ring A, with methylchloride as solvent (K = 0.033). Unmarked values are from ref. 3 for comparison.Unobtainable due to the preferred hydrogen bonding from the ring B pherolic hydroxy group. A method has been devised which permits estimation of the Hammett a-constant in a manner circumventing many of these objections. In the system : the group being investigated (R) is physically isolated from the O H N hydrogen bond regardless of its position on ring B. Its effect on the basicity of the nitrogen may be measured, none the less, by the strength of the hydrogen bond on the phenol group of ring A as determined by the exact frequency of the O H stretching vibration. Table 1 shows various a-constants obtained from the expression:= voH . K where vOH is the frequency shift of the substituted homologue (in wave-numbers) compared with the unsubstituted parent compound, and K is a factor employed to bring these wave-number shifts into the range of the classic Hammett -values.The figures above show that when steric effects are eliminated, the ortho and para -constants are nearly the same. A more elaborate discussion of the principles involved, and a description of experimental detail, will be published elsewhere.