In a previous paper [1], I have commited an error of calculation which has been caused by a bad use of the values of some of the parameters which determine the paths of the spacelike and null geodesics.

I have made a first-order in M/r (corrected) calculation for small and large values of the distance, L, following the method (based on Fermat's principle) used in [2]. Also, I have explicitly used the equations of the spacelike geodesics of the Schwarzschild metric to relate the variables r₂, L and Ψ (see figure 1 of [1]).

For small values of L (≈1 m) the order of magnitude of this effect is very small (≈10¹⁷), so I have also made a second-order calculation which is not based on the first-order straight-line approximation. I have found a complete consistency between the corresponding results.

Results. Let be the relative proper time advance of light and let be the magnitude of the total angular variation of the anisotropy effect. In table 1 we give some values of .

Table 1.

L
1 m	5.44 x 10-17
10 m	5.44 x 10-16
1 km	5.44 x 10-14
1 km	1.27 x 10-10

Conclusion. We can no longer maintain our experimental proposal for interferometric measurement. It is most likely that this proposal is not feasible. It would require an interferometer sensitivity smaller than 10^-16 for an arm length of 10 m which, moreover, would be subjected to changes of slope.

Incidentally, we can see that the order of magnitude of this effect is 10^-10 (the one obtained in [1]) when L is of the order of 10⁴ km. So I can only set up the question (not proposal) of the possibility of measuring this effect by means of, for example, rockets or satellites in orbit around the Earth.

[1] Díaz-Miguel E 1993 Class. Quantum Grav. 10 2691

[2] D'Inverno R 1993 Introducing Einstein's Relativity (Oxford: Oxford University Press)