the nitrogen resonance positions of the methylpyridines are small except for the para-substituted compounds. This is rationalized qualitatively in terms of competitive inductive vs. hyperconjugative interactions between the substituent and the nitrogen, tert-Butyl substitution at the ortho position induces upfield shifts characteristic of -gauche interactions. Protonation or /V-oxide formation has little effect on the manner in which methyl groups influence the shielding, suggesting that the lone-pair electrons in the free base play only a minor role in the response of the nitrogen nucleus to alkyl substitution. The pyridinium ion resonance positions are markedly influenced by solvent and concentration. Picoline and lutidine chemical shifts do not correlate with electron densities calculated by the INDO method, but chemical shifts calculated by the Karplus-Pople method or the Witanowski modification thereof qualitatively reproduce trends. The validity of these types of calculations is assessed.

Nitrogen and carbon electron densities of the toluidines and xylidines have been recalculated by the INDO method; previously published errors have been corrected. Although the nitrogen-15 chemical shifts of these compounds still display the earlier suggested correlation with u and total electron densities, the calculated inverse correlation with proton electron densities has been shown to be incorrect. Methyl proton chemical shifts of these compounds display no meaningful correlation with the nitrogen shitts. The nitrogen chemical shifts of the toluidinium and xylidinium ions correlate moderately well with the "C chemical shifts of the analogous di-and tri-methylbenzenes.Recently, Adler and Lichter reported a correlation between the I5N chemical shifts of toluidines and xylidines, and the INDO-MO total and sigma electron densities at the nitrogen atoms."* A corollary of that calculation was the possible existence of a relationship between the nitrogen and methyl proton resonance positions. After publication of this work, two errors were discovered, and it is the purpose of this note to correct these errors and assess the implications of the corrected results. In addition, we report the nitrogen chemical shifts of the corresponding toluidinium and xylidinium ions (1) to show that when the lone-pair is absent, the nitrogen shifts correlate with methyl carbon shifts of the corresponding methylbenzenes (2) in the manner suggested3 for aliphatic amines. EXPERIMENTALSamples were commercially available and were purified by distillation just before use. Carbon tetrachloride and deuteriochloroform, used for the proton spectra, were also distilled just before use. The anilinium ions were prepared as 2~ solutions in trifluoroacetic acid, and under these conditions the anilines are completely protonated. Approximately 10% deuteriobenzene was added as an internal lock. Several samples, noted in Table 3, were run as the hydrochlorides in water, containing c. 10% D,O as internal lock.Proton spectra of the methylanilines were recorded on a Varian A-60A spectrometer using a sweep width of 50Hz and a sweep rate of 0.2Hzs-'. Chemical shifts were measured with audio sideband calibration * Author to whom correspondence should be addressed.t Jonas E. Salk Awardee, City University of New York, 1975. with respect to c. 2% internal TMS. The values reported are the averages of three measurements each in the forward and reverse sweep directions, and are accurate to k0.005 ppm. Nitrogen chemical shifts of the anilinium ions were determined at 10.09 MHz on a JEOL PS/PFT-lOO spectrometer equipped with the JEOL EC-100 data system. Chemical shifts were determined with respect to a 2 . 9~ solution of 95% enriched ammonium chloride in 1~ HC1 contained in a 2 mm capillary placed concentrically inside the 10 mm sample tube. Spectra were determined with complete proton decoupling, and were accumulated over a 4kHz range and 8 K words of memory, using 30" pulse widths and a 1.1 s repetition rate. The chemical shifts are assumed accurate to ItO.1 pp...