“…We can verify the scale of the effect of small-scale variations on the beam in two ways: first, with an analytic calculation, and second, by experimental data. For the former, ref (26) averages the product 〈P(I)∇ 2 E〉 over the beam cross-section, assuming a uniform beam intensity and a Heaviside-shaped P(I)=θ(I-I c ) that turns on at some critical intensity I c(26). This model, however, is too optimistic for our interferometer, as the influence of intensity noise will rise sharply when the beam intensity I(r) approaches the critical intensity at a certain radius from the center of the beam.…”