Need to allow for slow nonlinearity in measurements of $n_2$

The components of the nonlinear susceptibility tensor $\chi_{1221}^{(3)}$ were determined by a method eliminating the contribution of the slow part of the nonlinearity. A comparison with an interferometric method made it possible to demonstrate a considerable contribution of the slow part of the nonlinearity in the measurements of $n_2$ under conditions of hard focusing of radiation in a sample. This comparison also showed that allowance for the slow part of the nonlinearity is essential in calorimetric measurements of $n_2$ in the case of pulse durations typical of $Q$-switching conditions.