well, I'll guess you just take the integral. So it would be:
∫(4,x)[te^(-t^2)dt]
=-1/2(e^(-t^2))|(4, x)
=-1/2(e^(-x^2))+1/2(e^(-16))
Or something like that.
__________________
Human history is the story of complaisance. While disaster is fresh in our memory, we take precautions. But as the memory of disaster recedes, we start to take risks.
Thinking about it with one variable, suppose you have
G(x) = ∫(4, x) [g(t) dt].
What is G'(x)?
The same principle will apply for f_x in your question, especially since f(x, y) is really just a single-variable function (there is no y in your expression).