High-resolution resonance-enhanced multiphoton ionization (REMPI) spectroscopy of a pulsed nitrogen beam is applied to determine the (1, 0) to (4, 0) line positions (J″ ≤ 2) of the a1Πg ← X1Σ+g transition. A strong a.c. Stark effect is observed even next to the REMPI threshold pulse energies and is tentatively correlated with certain energy levels near the three-photon energy. At the same time, the REMPI signal, which is not exactly proportional to the a ← X Franck-Condon factors, is dominated by the near-resonant enhancement due to these levels. This is particularly pronounced in the case of the (3, 0) band. In the three other cases, precise zero pulse energy values of the line positionscan be determined by linear extrapolation. The signs of the slopes for these extrapolations are opposite to those expected from the theoretical expression for the dynamical Stark shift. The obtained band origins are used to recalibrate the literature data which, so far, have exhibited major mutual disagreement. A Dunham-type least-squares fit of more than 3100 available tabulated spectral line positions yields a greatly improved mathematical description of this transition for v′ ≤ 15 and v″ ≤ 27, as well as ofa few A3Σ+u and b′1Σ+u levels. The X-state vibrational levels cannot be adequately described by a simple v″ + 12 polynomial. It is, however, found that the application of two separate polynomials for v″ ≤ 5 and v″ > 5 can remove all discrepancies. There are two conclusions from these calculations. The first is that there is an obvious onset of enhanced anharmonicity of the N2 ground state for v″ = 5 which is not observed for the isoelectronic CO molecule. This effect has already been found in earlier, shorter, and less accurate Dunham expansions. These earlier polynomials, however, do not take into account the sudden change of the vibrational frequency at v″ = 5 which must be concluded from two independent experiments by R. E. Miller. This small but distinct change suggests the mathematical treatment of the vibrational levels in the two ranges specified above. An unambiguous interpretation of the effect is not possible at present. No nearby 1Σ+g state is known which might act as a perturber, and, alternatively, a change of the electronic configuration needs to be quantified by a detailed theoretical study. Accurate measurements of the X-state vibrational levels are still rather sparse. Our work strongly suggests further experiments need to be done at a ≤0.001 cm−1 accuracy level that cover the full level range from v″ = 0 to at least v″ = 15 in order to obtain an even more unambiguous representation of this state and to reveal more details of the rotational structure.