Multivariable calculus
Partial Derivatives of x sin(xy)
For f(x,y) = x sin(xy), the partial with respect to x is sin(xy) + xy cos(xy), and the partial with respect to y is x^2 cos(xy). The x-partial needs the product rule on x times sin(xy) plus the chain rule inside the sine. The y-partial needs only the chain rule.
Product rule outside, chain rule inside
Differentiating in means two rules stacked. The outer structure is a product, times , and the second factor is itself a composition whose inside is . Do the product rule first, then handle the sine.
With frozen, the inside has -derivative , so the chain rule gives . Substituting that back in and simplifying gives the answer.
In the picture is simpler. The leading is now a constant multiplier, so there is no product to differentiate; only the chain rule acts, and the inside has -derivative . That second is where the square comes from.
The mistake: losing the inner derivative
The most common wrong answer for the -partial is . That drops the inner derivative and leaves you one factor of short. Every time the inside of a trig function is a product of variables, the chain rule contributes the other variable.
- Differentiating in contributes , not 1.
- Differentiating in contributes , not 1.
- The outer in is differentiated only in the -partial, never in the -partial.
A second slip is writing the -partial as , which forgets the inner on the chain rule term. Check the shape of that term: it must carry both an from the product rule and a from the inside.
Reading the answer at a point
Take and . Then , so and . The cosine term switches off entirely.
The zero in the direction is honest information: at that point the sine has hit its peak, so nudging changes nothing to first order. Moving in still changes , because the leading factor of is growing.
Frequently asked questions
Do I need the product rule for the y-partial as well?
No. With held fixed, is a constant times , and a constant multiple does not trigger the product rule. Only the chain rule is needed, which brings down the inner derivative and gives .
Are the mixed second partials equal here?
Yes. Differentiating with respect to gives , and differentiating with respect to gives the same thing. Clairaut's theorem guarantees this, since is smooth everywhere.