The specific retention volumes, V°g, at 60°C are reported for 10 solutes with OV-101, OV-3, OV-7, OV-11, OV-17, OV-22, and OV-25 stationary phases. Baseline resolution of all components of the mixture is obtained with a glass capillary column coated with a predicted (window-diagram) optimum intimate blend of the two end-point phases, 45% OV-101 + 55% OV-25. Separation is not obtained, however, with the neat phase of closest phenyl content, OV-17. The same mixture of phases coated onto previously deposited and in situ polymerized polydimethylsiloxane gives retentions that correspond precisely to those calculated on the assumption that the underlying polymer layer mimics neat OV-101. The method employed to deduce the column weight of polymer is detailed, and the same result, although less precise, is found when pure OV-25 is used as probe solvent. However, peak asymmetry is clearly evident in the latter instance and OV-25 does not appear to wet the in situ prepared polydimethylsiloxane. Inaccuracies in the method are therefore attributed to interfacial effects. A graphical method of representation of the selectivity of the pure polymethylphenylsiloxanes in terms of weight percent compositions of OV-25 + OV-101 is given, and it is shown that the latter correspond closely in phenyl content to the former. OV-7 (20% phenyl content) or its equivalent appears to offer sufficient efficiency as well as selectivity for inclusion in a set of standard stationary phases. Plots of the absolute retentions of the test solutes regress linearly against weight percent composition of this phase in intimate admixture with OV-101.