If we define the **differential**, **dx**, as an infinitessimal (vanishingly small) change in the coordinate x, then the derivative

can just be viewed as a **ratio of differentials** in **y** and **x**, the slope of a very short (infinitessimally short) line tangent to the function** y = f(x)** at a point of interest. From this point of view it's easy to rearrange our notation for the derivative of **f(x)**:

to a definition of the differential of **x**:

Although Gottfried Leibniz, who—independently but at the same time as Isaac Newton—founded calculus, didn't really intend this interpretation of his notation, it is correct and very convenient. It has been said that acceptance of Leibniz's notation on the European continent led to a decades-long leap ahead in calculus over the mathematicians of Newton's home, England, where the notation was not adopted until much later.

In order to get comfortable with the idea of the differential, let's use it in a linear approximation problem. We'll compare it side-by-side to the notation we developed before.

## Solution in both notations↓

We recall that the linear approximation of a function f(x) is

Our function is f(x) = x^{1/3} ; we'll center our approxmation at x_{o} = 27 and evaluate the result at x = 27.3:

for an error of about 0.7% from the calculated value, f(27.3) = 3.011.

In differential notation, that becomes:

So you see it's just a slight shift of notation to get to the same place. The linear approximation of a function near a point, x_{o}, is just the value of the function at that point plus the differential distance from it. The smaller **dx**, the smaller **dy** and the smaller the error in the approximation.

**dx** and **dy** are infinitessimally small changes in the coordinates x and y, respectively. The **derivative**, ^{dy}/_{dx}, may be considered to be a **ratio of differentials**.

Now let's let f(x) be the function that is the derivative of F(x):

We define the **antiderivative** of f(x) to be:

We've introduced the** integral symbol** here: ∫ . It's origin is Liebniz, who (and we'll learn why later) considered the integral to be a summation of infinitessimals (differentials), and wrote the word "summation" with an **f**: *fummation*. Thus the integral symbol looks a bit like an f, or the *f-hole* in a violin.

f-hole in a violin

We need a function which, when differentiated, gives **sin(x)**. We know that the derivative of **cos(x)** is **-sin(x)**, so the function **F(x) = -cos(x)** fits the bill:

There's just one thing . . . **G(x) = -cos(x) + 7** works, too, because the derivative of the constant, **7**, is zero. In fact, **F(x) = -cos(x) + C**, where **C** is an *arbitrary* constant (just pick one) works just fine. We say that the integral of **f(x) = sin(x)** is **-cos(x)** to within an additive constant. It's the same for all indefinite integrals: don't forget to add the constant:

An **indefinite integral** (antiderivative) can be found **to within an additive constant** (*See "An Important Question" below*).

We need a function which, when differentiated, gives **cos(x)**. We know that the derivative of **sin(x)** is **cos(x)**, so the function **F(x) = sin(x)** fits the bill:

This integral rounds out our basic trigonometric integrals:

Here we just need to work backward from the power rule of differentiation. We need a function that when differentiated gives **x ^{n}**, which has no constant coefficient. The solution is:

But of course there's a catch here ... did you spot it? **This indefinite integral fails if n = -1**, so we don't yet know how to find the antiderivative of f(x) = x^{-1}. Here it is:

First, it's entirely appropriate, and often done, to write this integral as

Now if we recall that the derivative of f(x) = ln(x) is 1/x, we have:

Because the derivative of f(x) = e^{x} is e^{x}, we have our easy integral:

We developed the derivatives of the inverse trigonometric functions in the implicit derivatives section. That leads to six very important indifinite integrals we will use often later:

Mathematically speaking, it would be a bummer if sin(x) + C wasn't (notwithstanding the C) the *only* indefinite integral of f(x) = cos(x). Mathematics doesn't deal that well with ambiguity. Uncertainty, yes, that's OK. Ambiguity is not. How do we know that sin(x) + C is the *only* indefinite integral of f(x) = cos(x)?

The answer lies with the Mean Value Theorem. Here's how it works:

Now we proved earlier, using the Mean Value Theorem , that the value of the derivative of a constant is zero (or if the derivative of a function is zero, then that function is constant) so we have

Thus we have proved that if G(x) and F(x) are both antiderivatives of f(x), then they can only differ by at most a constant.

This is a very important result in integral calculus because it proves the uniqueness (to within an additive constant) of an indefinite integral — there is no ambiguity!

Two important properties of indefinite integrals (presented without proof, for now) will help us to use the basic integrals developed above to solve more complicated ones. Using these with the substitution technique in the next section will take you a long way toward finding many integrals.

where k is a constant, and

Using just the integrals we've got so far, coupled with those two properties, it's possible to find the integrals of a whole class of more complicated functions. For example, this integral:

It's simple to try a substitution here:

Taking the derivative with respect to x of both sides, we get:

Now we substitute into the original integral, u for x^{5}- 3 and 5x^{4}dx for x^{4}dx, making sure to multiply by 1/5 outside of the integral to compensate for the extra factor of 5 we introduced. From there it's an easy matter to solve the integral (using the power rule of integration above) and then resubstitute for u:

It's not always easy to tell whether substitution (and for what?) is going to work. Sometimes you just have to try something based on your intuition and see if it works. In this case, notice that the degree of the denominator is one greater than that of the numerator, so when we determine *du*, the degrees of *du* and the numerator match. It doesn't always work out like it does here, but it's a good thing to try.

**1 + x ^{4}** is a good candidate for substitution:

Substitute, rearrange and solve the integral

Now back-substitute for u to get the final answer:

Make sure to differentiate your antiderivatives to see if they're correct. That's the nice thing about most integrals - you can easily check to make sure you've got it right.

Make the substitution **u = 1 + e ^{x}**, then

.

Easy peasy, lemon squeezy.

Find each of the following indefinite integrals, to within an additive constant, C, either directly from a known derivative, from the integral power rule or by substitution:

**xaktly.com** by Dr. Jeff Cruzan is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. © 2012, Jeff Cruzan. All text and images on this website not specifically attributed to another source were created by me and I reserve all rights as to their use. Any opinions expressed on this website are entirely mine, and do not necessarily reflect the views of any of my employers. Please feel free to send any questions or comments to jeff.cruzan@verizon.net.