At low temperature, graphite-potassium intercalation compounds of stage higher than one behave as molecular sieves as regards hydrogen [1,2]. At ambient and higher temperature, however, these compounds including the first stage one, chemically fix hydrogen by a completely different process. This reaction, which has been the object of previous papers [3,4], has been further studied so as to remove some of the previous uncertainties and imprecisions. The brown compound KC 8 after fixing hydrogen becomes blue, and saturates for an H K ratio of 2 3 . The reaction is reversible but does present noticeable hysteresis. Radiocrystallographic examination of the hydrogenated product shows that it involves a pure ternary phase of formula KC 8H 2 3 . The indexing of the 00 l reflections indicates a second stage compound: each intercalated layer is composed of two metallic planes in the presence of hydrogen (Fig. 1). This stage change can be interpreted by the pleating of the graphite layers [5]. Neutrocrystallographic studies confirm the preceding results and allow (Table 2) to conclude that the hydrogen forms a layer between the two potassium planes (Fig. 2). The belonging to the second stage of the phase KC 8H 2 3 is further confirmed by the fact that the free graphite spacings are susceptible of fixing new alkaline metal layers (K, Rb or Cs) to form new compounds referred to as “bi-insertion compounds” [6]. Attempts at hydrogen physisorption on the product during the course of chemical hydrogenation (Fig. 3), and the radiocrystallographic and kinetic measurements, clarify the finer points of the reaction. The hydrogen starts to be chemisorbed into the lattice of the compound KC 8 without any stage change (0A), then there appears progressively a second stage phase, unsaturated with hydrogen (AB). Saturation is only attained when the first stage phase has disappeared (BC). Low temperature hydrogen physisorption tests have been carried out on the second stage compound KC 24 during chemical hydrogenation. Figure 4 shows a linear decrease in the physisorption capacity, when the hydrogenation ratio increases. By extrapolation it is seen that this capacity should become zero for a chemisorption ratio H K in the neighbourhood of 2 3 . The results suggest the following hydrogenation model. Under the action of the gas, the intercalated metal forms double, hydrogenated layers, identical to those in the ternary KC 8H 2 3 . There also appear three types of graphitic spacings: free intervals, those occupied by lacunar single layers and those occupied by hydrogenated double layers. These spacings are distributed in such a manner that there appears a binary phase KC 12 n coexisting with a ternary phase KC 4NH 2 3 of equal or neighbouring stages n and N. Stoichiometry imposes a relationship between n and N. Figure 5 represents the variation of N as a function of n for various values of the hydrogenation ratio. At saturation the system would once again become single phase: a ternary, KC 24H 2 3 of sixth stage. In Table 3 are given the strongest 00 l reflections and the corresponding interplanar distances, characteristic of the compound KC 12 n and KC 4NH 2 3 . The experimental data taken from the X-ray diagrams, obtained for three KC 24 H y , compounds are compared to the values predicted by our model in Table 4. It is seen that the positions of the observed peaks are in good agreement with the 00 l reflections of the model, confirming the interpretation we propose. In conclusion, hydrogenation of the potassium layers of KC 8 or of KC 24 always leads to the formation of dense and hydrogenated metallic double layers of formula KH 2 3 . This process thus creates ternaries of higher stage than the starting binary.