A series of melt blended commercial linear and branched polyethylenes are used to explore the generality of blending laws. The measured relaxation modulus G(t), and zero shear viscosity eta(0), for each blend and blend fraction, have been compared with prediction for miscible blends, particularly using equations derived by Tsenoglou (1987). Plus or minus deviation between theory and measurement is dependent on the relative molecular weights of the blend components. We have found empirically that a generalised form of the blending law for G(t) and for eta(0), with a floating index C, provides an improved prediction of the blend fraction data. In particular the function defining C is non-symmetrical, from which we infer the significance of branching as well as molecular weight. The optimum value of the index differs for each of our blends, in the range 1.25 to 4, the variability being accounted for by the different degrees to which branched and linear polymers relax co-operatively in the melt. Blends of two near linear polymers do not fit the floating index prediction and conform more closely, though not precisely, to the original Tsenoglou rule.