It doesn't rotate the diaphram, but the presence of the grid modifies the shape of the housing so that the diaphram is obscured when viewed from the side and sound enters directly from the open, axial end of the housing. The grid 'shadows' the direct line of sight to the face of the diapharam.
These things are so small as to be perfectly omnidirectional through nearly the entire frequency range so that doesn't make much of a difference except for the very highest frequencies. Even then the response difference 90 degress off axis is minimal as apparent on the no-grid graph above and on the polar response graph made with the grids (not shown here), yet I can hear a touch of additional 'air' on-axis (end-address) with the short grid installed, and usually make somewhat of an effort to orient the mic with that in mind. In any case they are far more omnidirectional at the highest frequencies than larger mics are, so orientation with respect to frequency response is considerably less important.
Mostly just trying to clarify differences between the way the two measurments were made, even if it's pretty much inconsequential.