MATH 241 -- Multivariable Calculus
Differentiability for October 20, 2010
We have now introduced directional derivatives of in all
directions at points
and seen the special cases
and
that give the partial derivatives:
We will next consider the following question: Is there some
single object that we can consider as ``the (total) derivative of f "
at We will see that the answer is yes, but the exact form
and the way this appears as a derivative is going to take some explanation.
This is perhaps the area where the ideas from 1-variable calculus need
to be augmented the most to deal with functions of several variables.
To prepare for this, let's go back and reconsider what it means for a
function of one variable, to be differentiable from a viewpoint
that might be new to you. The way you have probably seen this idea
defined in your previous calculus courses is this: is said to be
differentiable at if the limit
exists and if it does, then this limit computes the value of the derivative
The equation
can be rearranged to
If we rewrite then
and
Therefore this last equation can be rewritten as
What does this really mean? Well, note that the top is
where is the linear function whose
graph is the tangent line to at
Not only is
as
In fact, it is going to zero faster than
in the sense that the quotient is still going to zero after we divide
by Intuitively, we can factor
out of the top and
cancel with the bottom and what is left is still going to zero as In graphical terms, this means that when
exists, the graph
gets very close to the tangent line if we ``zoom in'' on the point
in a suitable way. Here is an example to show what this means.
We know is differentiable at
and
Hence Here is the graph
together with the tangent line plotted over smaller and smaller intervals.
Note how the graph quickly becomes indistinguishable from the tangent
line (up to the resolution of the screen):
> | ![]() |
![]() |
> | ![]() |
![]() |
> | ![]() |
![]() |
> | ![]() |
![]() |
Here is the plot of which indicates how the limit (*) above works in this case.
> | ![]() |
![]() |
This certainly seems to approach 0 as (This can also be shown, of course, via techniques including
L'Hopital's Rule for limits.)
On the other hand, here is a function and an
where
does not exist.
> | ![]() |
![]() |
Note what happens if we plot for any real
> | ![]() |
![]() |
We can see
while
no matter what the value of a is.
Intuitively, no straight line passing through (0,0) approximates the shape of the graph at all well near (0,0).
It is this idea of differentiability that extends well to functions of several variables:
Intuitively, we say is differentiable at
if there is some linear function
graph z =
approximates the graph
z = well near
More precisely, we say
is differentiable at
if there is some linear function
In fact there is only one "candidate" for the linear function: It is possible to see that
in is a plane in
So the question
we're asking is really -- does that plane approximate the shape of the graph
well near
Example 1. Let We claim that
is differentiable at
We have
and
First, we plot
and
together to see whether it looks like the plane is approximating the graph well.
We plot the graphs of f, ℓ together on a small rectangle with (1,2) at the
center:
> | ![]() |
![]() |
This looks pretty good, and it would look progressively better as we zoomed in towards
The next plot shows what happens if we graph the function
again on a small rectangle with (1,2) at the center:
> | ![]() ![]() |
![]() |
In fact, it is a theorem that if and its two first order partial derivatives
and
are continuous
on some open set U in containing
then
is differentiable at
That theorem applies
in this example, since
are all polynomial functions.
(Recall polynomials in x, y are continuous at all points in Next, we want to see what can "go wrong" if f is not differentiable at a point.
Example 2. Let if
and = 0 if
We will concentrate on the point Note that
and
So the "candidate" linear approximating function is
.
But the question is: does that linear function approximate the shape of well
near (0,0)?
> | ![]() |
![]() |
Next, the plot of
> | ![]() |
![]() |
What should our conclusion be in this example? How can we make it rigorous?
> |