Hi! I found a bug with the method multiplyBy in the class CairoMatrix. I think the bug come from the fact that you send to the primitive the self in the result field and also in the parameter field (I found the difference using the multiply method in VisualWorks and saw the difference on the graphics result of my work): self nbCall: #(void cairo_matrix_multiply (AthensCairoMatrix * self, AthensCairoMatrix * m , AthensCairoMatrix * self ) ) For example : a := AthensCairoMatrix new. c := AthensCairoMatrix new. c rotateByDegrees: 40. c translateBy: 50@50. a := a multiplyBy: c. a AthensCairoMatrix ( sx: 0.766044443118978 shx: 0.6427876096865393 shy: -0.6427876096865393 sy: 0.766044443118978 x: 6.1628416716219405 y: 70.44160264027587 ) If you just swap the 2nd field with the 3rd field the result change: self nbCall: #(void cairo_matrix_multiply (AthensCairoMatrix * self, AthensCairoMatrix * self , AthensCairoMatrix * m) ) a AthensCairoMatrix ( sx: 0.766044443118978 shx: 0.6427876096865393 shy: -0.6427876096865393 sy: 0.766044443118978 x: 50.0 y: 50.0 ) (the graphic result are good with the 2nd option and I have the same result with VisualWork) I think the problem come from this part of the doc of Cairo: "void cairo_matrix_multiply (cairo_matrix_t *result, const cairo_matrix_t *A, const cairo_matrix_t *B); Multiplies the affine transformations in a and b together and stores the result in result. The effect of the resulting transformation is to first apply the transformation in a to the coordinates and then apply the transformation in b to the coordinates." Because of Cairo change the result first with the new matrix A and store it in the result you change too the 2nd matrix after his application. I think to avoid some edge effect it will be more clean to send a new matrix for the result matrix and not "self". What do you think? Tristan.