He seems to be saying (in the PDFs) that since Q(m.n), which is all > the factors including m and n after you apply his asymmetrical > difference quotient, must equal zero when the secant and tangent have > the same slope, that we can go ahead and just ignore them.