[en] The experimental technique for the direct determination of the second refractivity virial coefficient is described. The absolute measurement of the refractive index n combined with an expansion technique for obtaining the higher-order coefficients of the Lorentz-Lorenz expansion LL=((n²-1)/(n²+2))p^-1=A /SUB n/ +B /SUB n/ p+C /SUB n/ p²+... leads to precise values of density p.A /SUB n/ is the ideal molar refractivity, which is readily determined from the absolute measurements of n in terms of pressure, whereas B /SUB n/ , C /SUB n/ ,... are the higher-order molar refractivity virial coefficients, which are obtained from expansion experiments. The expansion method consists in measuring the sum of optical path lengths of two similar cells: one of them is filled with the gas at density p, and the other is evacuated. After the expansion the density is nearly halved and one measures again the optical path lengths. In order to cancel the small differences in volume and path lengths between the two cells, the process is reversed. Because the linear term in density remains the same before and after the expansion and only the quadratic and higher-order terms change, we can determine the refractivity virial coefficients B /SUB n/ , C /SUB n/ ,... from the change in the optical path lengths. The measurements for the determination of B /SUB n/ and C /SUB n/ for methane, nitrogen, and five mixtures were carried out at 323.15K and pressures up to 450 bar. The mixed-interaction constant for methane and nitrogen derived from the experimental second refractivity virial coefficient is compared with those obtained from the geometric and linear mixing rule as well as Lorentz combination