That is puzzling. 18% solvent is not common and you would expect very strong diffraction with such a low solvent content.
one possibility is that the NC symmetry is parallel to a crystal axis and is making monoclinic data appear to have an extra 2-fold axis. (There is a case I have heard of wherethe SG turned out to be P21 and not P2i2i2i - done by Michael Isupov. You could probably find a reference to it) I would look carefully at the pointless summary of the quality of the symmetry operators. If one 2-fold is better than the other two maybe process the data with that 2 fold only.. Of course another possibility is that the protein has been chewed up in crystallisation! Do you have an MR model? Eleanor