The rearrangement of singlet phenylnitrene (1a) to 1-azacyclohepta-1,2,4,6-tetraene (3a) has been studied computationally, using the CASSCF and CASPT2N methods in conjunction with the 6-31G*, cc-pVDZ, and 6-311G-(2d,p) basis sets. Ring expansion from the (1)A(2) state of 1a is predicted to occur in two steps via 7-azabicyclo[4.1.0]-hepta-2,4,6-triene (2a) as an intermediate. The rearrangement of 1a to 2a is estimated to have a barrier of ca. 6 kcal/mol and to be rate-determining. Azirine 2a is unlikely to be detected, because of the small calculated barrier (ca. 3 kcal/mol) to its rearrangement to 3a. At the CASPT2N/6-311G(2d,p)//CASSCF(8,8)/6-31G* + ZPE level of theory, the reaction 1a --> 3a on the lowest singlet potential energy surface is calculated to be exothermic by 1.6 kcal/mol. This reaction is predicted to be ca. 19 kcal/mol less exothermic, but to have a barrier ca. 9 kcal/mol lower than the analogous ring expansion of (1)A’ phenylcarbene (1b) to cyclohepta-1,2,4,6-tetraene (3b). Factors which contribute to these and other energetic differences between the ring expansion reactions of la and Ib are discussed. The lowest singlet state of planar 1-azacyclohepta-1,3,5-trien-7-ylidene (4a) is an open-shell singlet (1(1)A "), which is calculated to be ca. 20 kcal/mol above 3a and to be the transition state for enantiomerization of 3a. Unlike cycloheptatrienylidene (4b), 4a is predicted to have a triplet ground state and an energy difference between 1(1)A " and the lowest triplet (1(3)A ") of ca. 1 kcal/mol. The similar geometries of and small adiabatic energy difference between these two states of 4a is probably the reason why triplet 4a has not been detected by EPR.