Comet Hyakutake was an ideal target for studying the near-nuclear region of a bright comet. On 1996 March 26.3–26.5 (UT) we observed the comet with an echelle spectrometer at the 4-m telescope atop Kitt Peak. High-resolution spectra of NH and CH are presented and compared with fluorescence models. At a spectral resolution of Δλ ≤ 0.2 Å two multiplets of ND are not blended by the overlying NH line forest. From these lines we derived a 3σ upper limit for the D/H ratio in NH, (D/H)NH≤ 0.006. We reevaluated theg-factors for the NH (0–0) band and presentg-factors for the strongest ND multiplets as a function of heliocentric velocity. Based on our measurements of the OH and NH (0–0) band we inferred a relative production rateQ(NH3)/Q(H2O) = (3.5 ± 1.0) × 10−3. We demonstrate that single-cycle fluorescence models can be used to describe the observed CH bands, if we introduce different rotational temperatures for the two spin states,T(F1) = 100 K andT(F2) = 130 K. These temperatures are believed to be remnants from a previous dissociative excitation. Obviously collisions or radiative processes cannot completely relax CH before the radical decays. The rotational temperature may thus provide clues to the identity of the CH parent(s). A theoretical analysis of CD suggests that several lines of the CDA–Xsystem are not masked by CH lines.