Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. It's 100% free.

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

In my textbook, the gravitational field is given by$$\mathbf{g}\left(\mathbf{r}\right)=-G\frac{M}{\left|\mathbf{r}\right|^{2}}e_{r}$$ which is a vector field. On the same page, it is also given as a three dimensional gradient$$\mathbf{g}=-\mathbf{\nabla\phi}=-\left(\frac{\partial\phi}{\partial x},\frac{\partial\phi}{\partial y},\frac{\partial\phi}{\partial z}\right)$$ As this second equation is also a vector field, why doesn't it contain a basis vector of some sort and why isn't $\mathbf{g}$ given as a function of something or other?

Also, how do you actually get from the gravitational potential field $$\phi=\frac{-Gm}{r}$$

to the second equation? I can see that you apply the operator $\nabla$ but how does that give you $\mathbf{g}$?

Thank you

share|cite|improve this question
I don't get what you mean by 'shouldn't g b a function of something or other'. It's a function of $\phi$(potential) isn't it? The second equation is more general, as it can handle any gravitational distribution; whereas the first only applies to single point mass. And the second equation does have basis vectors, it's written in cartesian shorthand, like normal 3D cartesian coordinates. – Manishearth Feb 23 '12 at 16:31
@Manishearth - Just showing my colossal ignorance! I'm a self learner and have only recently learned that $y=y\left(x\right)$ means $y$ is a function of $x$ so I was wondering why there's a $\mathbf{g}\left(\mathbf{r}\right)=$ for the first equation but only a $\mathbf{g}=$ for the second. – Peter4075 Feb 23 '12 at 19:08
Great! Even I'm a self-learner =D. The reason is that in the second equation, it's not $g(r)$ but actually $g(\left(x,y,z\right))$. Which becomes tedious. Actually when you write a quantity in bold its usually implied that its a vector field (or sometimes a plain vector). A vector field on its own need not have basis vectors, but while defining it, it does need to have them, as you pointed out. The rest is just a matter of the book switching bases without letting you know. Usually $(a,b,c)$ is cartesian unless specified. And if the equation has $x$,$y$,$z$, then its nearly always cartesian. – Manishearth Feb 24 '12 at 1:17
up vote 3 down vote accepted

We want to compute the gradient of

$$ \phi(r) = \phi(<x,y,z>) =\frac{-Gm}{|r|}=\frac{-Gm}{\sqrt{x^{2}+y^{2}+z^{2}}} $$

It is:

$$ \left<\frac{\partial}{\partial x} , \frac{\partial}{\partial y} , \frac{\partial}{\partial z}\right> \phi(<x, y, z>) = $$

$$ = Gm \left< \frac{x}{(x^2+y^2+z^2)^{3/2}} , \frac{y}{(x^2+y^2+z^2)^{3/2}} , \frac{z}{(x^2+y^2+z^2)^{3/2}} \right> = $$

Using $ (x^2+y^2+z^2)^{3/2} = |r|^3 $, we get

$$ = \frac{Gm}{|r|^3} \left< x, y, z \right> = $$

By definition $r=<x,y,z>=|r| e_r$

$$ = \frac{Gm}{|r|^2} e_r = $$

$$ = -g(r) $$

Note that $r$ is a vector, $e_r$ is a vector, $g(r)$ is a vectorfield (maps a vector r to a vector), $\phi(r)$ is a scalarfield (maps a vector r to a scalar).

share|cite|improve this answer
Thanks. Afraid I'm still a little confused with the maths notation. Can we do this in baby steps? When you say $\phi\left(x,y,z\right)$ I assume you mean $\phi$ is a function of $x,y,z$. But when you say $\nabla\phi=\frac{Gm}{|r|^{3}}\left(x,y,z\right)$ I think you mean $\left(x,y,z\right)$ is a vector, which is equal to the vector $|r|e_{r}$. So $\mathbf{g}=\mathbf{g}\left(\mathbf{r}\right)=-\mathbf{\nabla\phi}$. So why does the author say $\mathbf{g}=-\mathbf{\nabla\phi}$ and not $\mathbf{g}\left(\mathbf{r}\right)=-\mathbf{\nabla\phi}$? – Peter4075 Feb 23 '12 at 15:43
I made the notation easier to comprehend by using < > for vectors, eg. r = <x, y, z> – mtrencseni Feb 23 '12 at 16:00
@Peter4075 When we write a vector as $(a,b,c)$, the basis vectors are implied. No need to write I/j/k inside (technically, it's wrong to do that). Here they've implied cartesian coordinate by using derivatives with respect to x,y,z. I think they were assuming that you knew what $\nabla$ was. – Manishearth Feb 23 '12 at 16:37
@Peter4075 If you kept the I/j/k inside the brackets, it's wrong as then that would imply $\partial_x\hat{i}\text{ scalar multiplied by }\hat{i}$ etcetera. Notation with brackets is equivalent to scalar multiplication of bracket contents with basis vectors. You can't scalar multiply two vectors, so writing I/j/k inside is technically wrong. Though you need not be nitpicky about it. – Manishearth Feb 23 '12 at 16:43
By scalar multiply I mean scalar times vector, not the scalar (dot) product. Even though they behave similarly, we cannot interchange the two just as we cannot interchange dot and cross. – Manishearth Feb 23 '12 at 16:45

To add to @mtrecseni's derivation, the definition of scalar potential of a field $\mathbf{E}$ is a scalar field $\phi$ such that $\mathbf{E}=-\nabla{\phi}$.

We can easily derive this from the fact that the potential is a path integral (a path-independant path integral actually) $$\phi=-\int\limits_{(x_1,y_1,z_1)}^{(x_2,y_2,z_2)}\mathbf{E}\cdot \vec{dl}$$ We can write the path integral as three integrals, over x, y, z: $$\phi=-\int\limits_{x_1}^{x_2}E_x(x,y_1,z_1).dx-\int\limits_{y_1}^{y_2}E_y(x_2,y,z_1).dy-\int\limits_{z_1}^{z_2}E_z(x_2,y_2,z_2).dz$$ Basically we have first taken the particle from $x_1\to x_2$, keeing $y,z$ constant, and so on. $E_x$ is the x-component of $\mathbf{E}$.

Taking $\nabla$,since y and z are constant in the first integral and so on, the equation reduces to $\nabla\phi=-(E_x\hat{i}+E_y\hat{j}+E_z\hat{k})\implies \mathbf{E}=-\nabla\phi$

This derivation works for any conservative field.

share|cite|improve this answer

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.