changed due to decomposition. These thermal data, together with the equations of Schroeder-van Laar and PrigogineDefay,21 were thus used to construct the ideal binary phase diagram shown in Figure 3. The measured points for the pure phase materials and the eutectics have been included. The two calculated liquidus curves for the 1:1 and 2:1 cocrystals are indicated, and it can be seen that these are consistent with the measured eutectic temperatures. Since the calculated ideal liquidus curves of pure maleic and caffeine do not intersect those of the cocrystals in Figure 3, it appears from this ideal calculation that either the cocrystals are metastable states, which is not borne out by the DSC data or, more likely, the true liquidus curves, particularly for caffeine-rich liquids, deviate strongly from ideality due to liquid phase intermolecular interactions. Thus, Figure 4 shows a schematic binary phase diagram in which the liquidus curves of pure maleic acid and caffeine have been drawn to intersect those of 1:1 and 2:1 cocrystals, respectively. The two cocrystals’ liquidus curves also intersect giving, in total, three eutectic points. Comparison of Figures 3 and 4 now reveals the significant extent of nonideality which must exist on the caffeine-rich side of the binary system