The origin of the high reactivity and site selectivity of pyridine N-oxide substrates in O-pivaloyl hydroxamic acid-directed Rh(III)-catalyzed (4+2) annulation reactions with alkynes was investigated computationally. The reactions of the analogous pyridine derivatives were previously reported to be slower and to display poor site selectivity for functionalization of the C(2)-H vs the C(4)-H bonds of the pyridine ring. The N-oxide substrates are found to be more reactive overall because the directing group interacts more strongly with Rh. For N-oxide substrates, alkyne insertion is rate-limiting and selectivity-determining in the reaction with a dialkyl alkyne, but C-H activation can be selectivity-determining with other coupling partners such as terminal alkynes. The rates of reaction with a dialkyl alkyne at the two sites of a pyridine substrate are limited by two different steps: C-H activation is limiting for C(2)-functionalization, while alkyne insertion is limiting for C(4)-functionalization. Consistent with the observed poor site selectivity in the reaction of a pyridine substrate, the overall energy barriers for functionalization of the two positions are nearly identical. High C(2)-selectivity in the C-H activation step of the reaction of the N-oxide is due to a cooperative effect of the C-H Bronsted acidity, the strength of the forming C Rh bond, and intramolecular electrostatic interactions between the [Rh]Cp* and the heteroaryl moieties. On the other hand, some of these forces are in opposition in the case of the pyridine substrate, and C(4)-H activation is moderately favored overall. The alkyne insertion step is favored at C(2) over C(4) for both substrates, and this preference is largely influenced by electrostatic interactions between the alkyne and the heteroarene. Experimental results that support these calculations, including kinetic isotope effect studies, H/D exchange studies, and results using a substituted pyridine, are also described.