From p. 79 in Brown's and Churchill's "Complex Variable and Application":

Let the function $f(z) = u(x, y)+iv(x, y)$ be analytic in a domain $D$, and consider the family of level curves $u(x, y) = c_1$ and $v(x, y) = c_2$. Prove that these families are orthogonal.
Specifically, show that if $z_0 = (x_0, y_0)$ is a point in $D$ which is common to two particular curves $u(x, y) = c_1$ and $v(x, y) = c_2$, and if $f'(z_0) \ne 0$, then the lines tangent to the curves at $(x_0, y_0)$ are perpendicular to each other. Then, the question gave a clue that:

$\frac{\partial u}{\partial x}+\frac{\partial u}{\partial y}\frac{dy}{dx} = 0$ and $\frac{\partial v}{\partial x}+\frac{\partial v}{\partial y}\frac{dy}{dx} = 0$ (*)

At first, I thought that the authors meant $\frac{\partial u}{\partial x} = \frac{\partial u}{\partial x}\frac{dx}{dx}+\frac{\partial u}{\partial y}\frac{dy}{dx} = 0$, $\frac{\partial v}{\partial x} = \frac{\partial v}{\partial x}\frac{dx}{dx}+\frac{\partial v}{\partial y}\frac{dy}{dx} = 0$. However, that cannot be the case, since we were told that $f'(z_0) = u_x(x_0, y_0)+iv_x(x_0, y_0) \ne 0$. Furthermore, $\frac{\partial u}{\partial x} = 0$ implies that $u(x, y)$ is constant in a direction parallel to the $x$-axis, but that certainly is not the case in general. So, how do you get the two equalities in (*)?

Furthermore, what is the significance in $f'(z_0) \ne 0$? In the next question, we are asked to show that with $f(z) = z^2$, the level curves $u(x, y) = x^2-y^2=0$ and $v(x,y)= 2xy = 0$ are not orthogonal. But a straight-forward computation $\nabla u · \nabla v = u_xv_x + u_yv_y = 4xy - 4xy = 0$, a constant zero. What did I gloss over?


Hint. To do this problem efficiently it is probably best if you take the point of view of two-variable real calculus. The level curves $u=c_1$ and $u=c_2$ are orthogonal at a certain point if their normal vectors are orthogonal at that point. These normals are $$\nabla u =\Bigl(\frac{\partial u}{\partial x},\frac{\partial u}{\partial y}\Bigr) \quad\hbox{and}\quad \nabla v =\Bigl(\frac{\partial v}{\partial x},\frac{\partial v}{\partial y}\Bigr)\ ,$$ provided they are not zero vectors. Can you finish the problem by explaining why these vectors are perpendicular in the context of your question?

  • $\begingroup$ Actually, I figured out why the condition that $f'(z) \ne 0$ is necessary. If not, then $\nabla u, \nabla v$ are zero vectors, which have no directions. Nonetheless, what are the two equalities in (*)? $\nabla u$ is supposed to be normal to the line tangent to $u$ at $(x_0, y_0)$, but I am not seeing the expression in (*) as a directional vector. It looks more than $u_x$ to me. $\endgroup$ – Andy Tam Jun 30 '14 at 14:22
  • $\begingroup$ If you regard $u(x,y)=c_1$ as defining $y$ implicitly in terms of $x$, then the first expression in $(*)$ is $du/dx$ (NB not $\partial u/\partial x$). $\endgroup$ – David Jun 30 '14 at 14:42
  • $\begingroup$ I see it now. TY .... $\endgroup$ – Andy Tam Jun 30 '14 at 17:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.