The rule, and why it works
To integrate , add one to the power, then divide by the new power: , for every n except −1.
Differentiating the answer proves it. The power n + 1 comes down in front and the power drops back to n, giving . The n + 1 on top cancels the n + 1 underneath, leaving . Nothing in that argument needs n to be a positive whole number, so the rule works for negative powers and fractional powers too. The one thing it needs is that n + 1 is not 0.
Negative powers
Take . Adding one to −3 gives −2, so the power goes up to −2, not down to −4. Divide by the new power, −2: . Check: differentiates to .
A fraction with a power of x underneath is a negative power, so write it that way first. , which integrates to . Check: differentiates to , which is .
The gold curve is and the plain curve is , for x > 0. The tangent to at (1, −1) is y = x − 2, with gradient 1, and is 1 at x = 1. At x = 2, has gradient , and is 0.25 there.
Fractional powers
A square root is the power ½: . Adding one gives the power , and dividing by is the same as multiplying by . So . Check: differentiating brings down , and , leaving , which is .
One over a root is a negative fractional power: . Adding one gives the power ½, and dividing by ½ doubles: . Check: differentiates to .
A constant is a power too. , which integrates to .
Rewrite as powers first
The rule handles one power of x at a time, so brackets and quotients are turned into sums of powers before integrating. , which integrates to .
A quotient with a single term underneath splits into separate terms. , which integrates to . Differentiating that gives back.
Signs need care when the power is negative. . The first term gives . The second gives , so the integral is . Check: differentiates to , the term that was there.
Why n = −1 is left out
Put n = −1 into the rule and it asks for , a division by zero. The formula breaks, and no repair of it can work, because no power of x differentiates to .
Here is why. A power differentiates to . To land on the power −1 the old power would have to be k = 0, and then the multiplier in front is that same 0: differentiates to 0, not to .
Yet does have an antiderivative. The region under from x = 1 to x = 6 has a definite size, and measured from 1 up to a moving right edge at x, that area grows at the rate , the height of the curve there. So the antiderivative exists; it is simply not a power of x. It is a new function, met in its own lesson. Powers close to −1 cause no trouble: integrates to .
The curve , with the region under it from x = 1 to x = 6 shaded. The region has an area, so has an antiderivative, but no power of x is that antiderivative.
The usual mistakes
Moving a negative power the wrong way. Adding one to −3 gives −2. Writing for the integral of subtracts one instead.
Dividing by a fraction the wrong way up. Dividing by means multiplying by , so integrates to , not .
Integrating the top and the bottom of a quotient separately. must be split into first; integrating the numerator and the denominator on their own and dividing gives a function that does not differentiate back.
Using the rule on . It gives , which has no meaning. needs the new function from its own lesson.