Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

A parabola whose axis is oblique to the orthogonal coordinate axes is of the form $f(x,y)= 0$, for example

$$f(x,y) = 9x^2 + 24 xy + 16 y^2 + 22x + 46 y + 9=0.$$

Using algebra only it is airly straightforward to find its apex once you rearrange the equation above to

$$f= (3x+4y+5)^2 = 2(4x-3y+8).$$

The intersection of the tangent to the summit and axis gives $\left(-\frac{47}{25},\frac{4}{25}\right)$.

I would like to get the vertex result using calculus only. I suppose the way to do is to minimize the curvature radius (unless I'm answered a better way). How do you extend the formula I know for $\{x,f(x)\}$ curves —


to this $f(x,y)=0$ curve? Thanks.

share|cite|improve this question
yeah, item (17). – Will Jagy Sep 5 '12 at 4:43

What you want for a level set of a $C^2$ function $F$ is the gradient vector $\nabla F,$ let's take that as a row with $F_i = \frac{\partial F}{\partial x_i}$, then the Laplacian $\Delta F,$ the identity matrix $I,$ and the Hessian matrix with entries $F_{ij} = \frac{\partial^2 F}{\partial x_i \partial x_j}.$ Oh, for the identity matrix we use the Kronecker $\delta$ notation, $\delta_{ij}$ is $1$ if $i=j$ but $0$ if $i \neq j.$ For you $1 \leq i,j \leq 2.$ Also the Laplacian is the trace of the Hessian matrix. Go figure.

Next, define a square matrix $B = \Delta F \cdot I - \mbox{Hess} F,$ so $$ B_{ij} = \Delta F \cdot \delta_{ij} - F_{ij}. $$

We are in $\mathbb R^n,$ for you $n = 2.$ Then the mean curvature of a level set is $$ \frac{1}{(n-1) \, |\nabla F|^3} \; \; \left( \nabla \; F \cdot \; B \; \cdot \; \nabla F^T \right) $$

Finally, for a curve in the plane the mean curvature is the curvature, with the possible disagreement of a $\pm$ sign, about which nothing can be done anyway. If you prefer, instead of $B$ use $-B.$

This is a slightly revised excerpt from my first article, 1991. I do not know of any book where you would be likely to find this. Personal service, is what I'm saying.

share|cite|improve this answer
A great thanks for having taken soo much time for this answer. I am unable yey ttdrafting this anThanks for your long answer – sigismond kmiecik Sep 5 '12 at 20:32
A great thanks for having taken soo much time for this answer. A large part of what you wrote is not within my grasp so I cannot make much of it. However I'm amazed in the contrast between the ease of finding any tangents to this conic using partial derivatives on a given point and the sophistication – sigismond kmiecik Sep 5 '12 at 20:46
behind finding its vertex. – sigismond kmiecik Sep 5 '12 at 20:47

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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