The dissociation of nitric oxide from hemoglobin, from isolated subunits of hemoglobin, and from myoglobin has been studied using dithionite to remove free nitric oxide. The reduction of nitric oxide by dithionite has a rate of 1.4 X 10(3) M-1 S-1 at 20 degrees in 0.05 M phosphate, pH 7.0, which is small compared with the rate of recombination of hemoglobin with nitric oxide (25 X 10(6) M-1 S-1 (Cassoly, R., and Gibson, Q. H. (1975) J. Mol. Biol. 91, 301-313). The rate of NO combination with chains and myoglobin was found to be 24 X 10(6) M-1 S-1 and 17 X 10(6) M-1 S-1, respectively. Hence, the observed progress curve of the dissociation of nitric oxide is dependent upon the dithionite concentration and the total heme concentration. Addition of excess carbon monoxide to the dissociation mixture reduces the free heme yielding a single exponential process for chains and for myoglobin which is dithionite and heme concentration independent over a wide range of concentrations. The rates of dissociation of nitric oxide from alpha chains, from beta chains, and from myoglobin are 4.6 X 10(-5) S-1, 2.2 X 10(-5) S-1, and 1.2 X 10(4) S-1, respectively, both in the presence and in the absence of carbon monoxide at 20 degrees in 0.05 M phosphate, pH 7.0. Analogous heme and dithionite concentration dependence is found for the dissociation of nitric oxide from tetrameric hemoglobin. The reaction is cooperative, the intrinsic rate constants for the dissociation of the 1st and 4th molecules of NO differing about 100-fold. With hemoglobin, replacement of NO by CO at neutral pH is biphasic in phosphate buffers. The rate of the slow phase is 1 X 10(-5) S-1 and is independent of pH. The amplitude of the fast phase increases with lowering of pH. By analogy with the treatment of the HbCO + NO reaction given by Salhany et al. (Salhany, J.M., Ogawa, S., and Shulman, R.G. (1975) Biochemistry 14, 2180-2190), the fast phase is attributed to the dissociation of NO from T state molecules and the slow phase to dissociation from R state molecules. Analysis of the data gives a pH-independent value of 0.01 for the allosteric constant c (c = Kr/Kt where Kr and Kt are the dissociation constants for NO from the R and T states, respectively) and pH-dependent values of L (2.5 X 10(7) at pH 7 in 0.05 M phosphate buffer). The value of c is considerably greater than that for O2 and CO. Studies of the difference spectrum induced in the Soret region by inositol hexaphosphate are also reported. This spectrum does not arise directly from the change of conformation between R and T states. The results show that if the equilibrium binding curve for NO could be determined experimentally, it would show cooperativity with Hill's n at 50% saturation of about 1.6.