Two related techniques handle relationships between x and y (or functions of x) that the ordinary explicit-function derivative table cannot differentiate directly.
Implicit differentiation. An equation F(x,y)=0 (e.g. the circle x2+y2=4) may define y as one or more functions of x implicitly — without being solved for y — and sometimes cannot even be solved for y in elementary terms (e.g. x4+x2y3−y5=2x+1). The method: differentiate both sides of the equation with respect to x, treating y throughout as a differentiable function of x (so any term in y picks up a factor dxdy via the chain rule — most simply as dxd(yn)=nyn−1dxdy), then solve algebraically for dxdy. Worked example: for x2+y2=1, 2x+2yy′=0⇒y′=−x/y.
Logarithmic differentiation. Handles power-exponential functions such as y=xx, where both the base and the exponent depend on x and neither the plain power rule nor the plain exponential rule alone applies. Method: (1) take log of both sides and simplify with the log laws; (2) differentiate implicitly; (3) solve for y′. For y=xx: logy=xlogx⇒yy′=logx+1⇒y′=xx(1+logx). The technique covers four general cases — constantconstant (derivative 0), variableconstant (ordinary power rule), constantvariable (ordinary exponential rule), and variablevariable (genuinely needs logs): dxd[f(x)g(x)]=f(x)g(x)[g′(x)logf(x)+g(x)f(x)f′(x)]. It is also useful more broadly — even for a function with no variable exponent — whenever it is built from several products/quotients/powers, since log turns those into sums/differences/multiples before differentiating, which is often far less work than a direct product-and-quotient-rule computation. …