The Zeeman effect in the photoexcitation spectrum of phosphorus in silicon has been studied up to 64 kG in the Faraday configuration with the field parallel to the three principal axes of the crystal. In this work, we have observed a slight decrease of the transverse effective mass with the field and we found a fair agreement between theory and experiment for the 2p± and 3p± levels. Better agreement is obtained, however, if the variation of the effective Bohr radii with the magnetic field is taken into account. For the 2p0 and 3p0 levels, the variation of the effective Bohr radius [Formula: see text] alone cannot be observed experimentally in the Faraday configuration, but our results indicate a much more important decrease than for the p± levels and a variational calculation is necessary in this case. The shift of the 2p0 and 3p0 lines obeys roughly the n2(n2 − 1) relation, which should not be observed in the case of strong interactions with upper levels.