of these compounds at or above 150 °C gave the previously unknown azides 16 and 17, respectively, identified by their IR spectra at -196 °C and by the fact that they reverted to 14 and 15, respectively, when warmed to -10 to 0 °C [16: IR 2140 (s), 2120 (s), 1350 (s) cm"1. 17: IR 2130 (vs), 2110 (s), 1330 (s) cm"1]. The intensities of the azide absorptions increased with the pyrolysis temperature until ca. 380 °C, when a new and strong absorption at 2000 cm"1 appeared. The latter absorption increased in intensity till ca. 500 °C; the azide absorptions decreased over the same temperature interval. Above 500 °C, the 2000-cm"1 band started disappearing again, and new nitrile absorptions at 2225-2250 cm'1 appeared in its place. The latter absorptions remained unchanged at room temperature, and isolation and chromatographic separation of the material allowed their assignment to the two nitriles 22 and 23, which had been identified previously.4An optimal pyrolysis temperature for the observation of the 2000-cm"1 absorption was found at 490 °C. Under these conditions, only traces of the azides (16 or 17) remained, and only weak bands due to the end products 22 and 23 were present. The spectra recorded at -196 °C, following pyrolysis of either 14 or 15 at 490 °C, were identical, and we therefore assign them to a common intermediate, the carbodiimide 19. When the matrix was warmed to ca. -55 °C, the carbodiimide band at 2000 cm"1 disappeared, and the nitriles 22 and 23 did not appear. Instead, a new compound, C18H12N4, corresponding to a dimer of 19 was isolated. The two dimers formed from 14 and 15 were identical.15These observations are summarized and interpreted in Scheme II. The formation of the common intermediate 19 demonstrates that both 1-isoquinolylnitrene (18) and 2-quinolylnitrene (20) undergo ring expansion under rather mild conditions, i.e., the activation energies cannot be significantly higher than those required for thermolysis of the azides 16 and 17. It would be difficult to interpret the observed spectra in terms of the fused azirines 24 and 25 (Scheme II). These molecules would be expected