The thermal behavior of salts of the type Co(NH.sub.3).sub.6X.sub.3 X = Cl, Br, I has been the object of considerable interest since early in the twentieth century. As noted, some of the early measurements on the trichloride complex recorded what appeared to be an "anomaly" occurring in its heat capacity. Our studies, however, indicate a far more complex behavior in which (a) DSC measurements between 100 and 300 K show a glass transition at ca. 200 K corresponding to the previously-mentioned "anomaly"; (b) the glass transition is observed in both the heating and cooling cycles, but the exact details are dependent on the rate at which the process is carried out in either direction; (c) parallel crystallographic studies using samples from the same batches as those in the DSC studies document that the origin of the thermal changes are due to torsional motions of the -NH.sub.3 hydrogens about the Co-N vector, evidence of the formation of a complex structure (condis crystalline mesophase) with a variety of order-disorder (entropically-diverse) material of differing stability and mobility; (d) inasmuch as there are four independent Co(NH.sub.3).sub.6.sup.+ cations in the asymmetric unit (Z' = 3), an additional reason for the complexity of the DSC observations is due to the sequential degree of torsional disorder induced thermally during the heating cycle on each of the four individual specimens of the asymmetric unit; (e) similar behavior is observed in the cooling cycle, but, the "re-orientation" of the thermally excited -NH.sub.3 ligands to a "cooler" more-orderly condition is considerably more sluggish (hysteresis)-thus, the observed differences in the recorded rates mentioned above; (f) finally, and not surprisingly, observations carried out with different crystals of (1) the same batch and (2) different batches show the same general behavior, but the specific details vary considerably, indicating that the history (provenance) of the crystal preparation is crucial to the exact results observed in each case.