We have already studied the condition of glass-formation and the glass-formation range of borates, silicates and germanates. In these studies, however, we could not determine precisely the cooling condition which defines the glass-formation range, because the glassy stateis not a stable state, but a sub-stable one. These experiments were made under conditions which were determined for the sake of experimental convenience: namely, 1/80mols of specimen were melted and cooled naturally in a room. Therefore, it is necessary to examine to what extent the results of these experiments are effective in view of the glass structure. In this study experiments were carried out by changing the cooling rate, and the variation in the glass-formation range with various cooling rates was examined. These cooling processes included the followings: quick cooling by water, natural cooling in a room (cf. Curve I in Fig. 1), natural cooling in a furnace (cf. Curve II in Fig. 2) and slow cooling in a furnace controlled by a thermocontroller. These cooling rates are about 3×102, 10, 1.5×10-1 and 1.2×10-3°C/sec, respectively. The amount of molten glass is the same as that in the previous studies; crucibles employed are made of platinum or its alloy, which may have some effect especially in the case of the slow cooling in a furnace.Ternary borate systems have been chosen as the glass-forming system for the convenience of experiment, which have been divided into common systems and exceptional systems. The former include the B-type ternary system as the containing only the oxides of the a-group elements, the PbO-containing ternary system as the one containing both of the oxides of the a-group and the b-group elements, and the B2O3-Bi2O3-PbO system as the one containing only the oxides of the b-group elements. The results are shown in Fig. 1-19. These glass-formation ranges contain various critical lines of vitrification; the limit of the continuity of a network-structure (the AD-line in Fig. 2 and 3), the existing limit of necessary modifier ions for the network-formation (the B2O3-C line in Fig. 2 and 3), and the exchangeable limit of network ions represented by the number of b-group ions connecting B with B in the network-structure (the A1B2, A2B3, … lines in Fig. 8; cf. Table 1). The glass-formation range expressed by the above critical lines generally varies somewhat according to the variation in the cooling rate. Therefore the result of the glass-formation range under an arbitrary cooling condition has no absolute meaning. However, comparing Fig. 4 with Fig. 5, for example, we can see a similar variation in the glass-formation range in both cases. In the one case the modifier ions are not exchanged but the cooling conditions are changed, while in the other the modifier ions are exchanged but the cooling conditions are kept constant. This fact can be explained by assuming the 3-dimensional glass-formation range including the glass stability as shown in Fig. 7. When the modifier ion in the B2O3-PbO-RO system (Fig. 5) is smaller, so that its vitrified system is more unstable, the glass-formation occurs only in the high stability sections. The case is the same when the cooling rate is slower in a more stable vitrified system.We then studied the B2O3-MgO-BaO system (Fig. 9), the B2O3-TiO2-BaO system (Fig. 12), the B2O3-WO3-Li2O systems (Fig. 15) and the B2O3-K2O-Bi2O3 system (Fig. 17) as exceptional ternary systems and discussed the true feature of the anomaly of these systems. In the