Have worked through to get some sort of answer, but find the evaluation (simplest bit) impossible - wood for trees, I think. given d2(e^(xy^2)/dy^2 I need to evaluate for x=2, y=0 now using chain rule d(e^(xy^2))/dy = d(e^u)/du du/dy where u=xy^2 & d(e^u)/du = e^u => x(2y) e^(xy^2) fine for 1st order moving on d(x(2y) e^(xy^2))/dy removing constants 2x(d(ye^(xy^2))/dy use product rule d(uv)/dy = v du/dy + u dv/dy u =e^(xy^2) & v=y => 2x(e^(xy^2)(d(y)/dy)+y(d(e^xy^2))) chain rule again where u=xy^2 & d(e^u)/du = e^u 2x(ye^(xy^2)(x(d(y^2)/dy)))+e^(xy^2) which shuld be d2(e^xy^2)/dy2 => 2x(xy(2y)e^(xy^2)+e^(xy^2) Now evaluate for x=2, y=0 everything goes to 0 ?????? I'v gone throught he difficult bits in the book, now the simple stumps me!!! Please help before I bang my head against the wall again!!!