A partial integral is the complement of the partial derivative. The idea is that your integral has multiple variables but you integrate with respect to one at a time, while holding the other variables constant. The idea is not difficult if you have a good handle on partial derivatives. There is one additional thing you need to watch for, the form of the constant.

You remember way back in single variable calculus when you learned integration, you needed to remember to include the constant of integration. For example,
\( \int{2x~dx} = x^2 + C \)
In this example, the \(C\) is just a constant, i.e. just some real number. We use \(C\) as a placeholder, since it can have any value, \(3\), \(\pi\), \(17\), \(127\) and even \(0\).
The same idea holds for partial integrals, i.e. you need a 'constant' of integration. However, this constant is a bit different for multi-variable integration than it was for the single variable case. The 'constant' of integration here is actually a constant function of integration, i.e. it is a function of the other variables that are held constant when we integrate.

In general, it looks like this. For a function \(f(x,y)\), when we integrate this with respect to \(x\), we get

\( \int{f(x,y)~dx} = F(x,y) + g(y) \) where \(\partial F(x,y)/\partial x = f(x,y)\). Let's think about this.

If we were given a function that looks like \(F(x,y) + g(y)\) and we were told to take the partial derivative with respect to \(x\), we would get

\(\displaystyle{ \frac{\partial [F(x,y)]}{\partial x} + \frac{\partial [g(y)]}{\partial x} = f(x,y) + 0 }\)

\(\displaystyle{ \frac{\partial [g(y)]}{\partial x} = 0 }\) because \(g(y)\) has no \(x\)'s in it and, therefore, is considered a constant. And we all know that the derivative of a constant is zero.

So, other than the constant situation, integration should be pretty straightforward. Let's do an example.

Example 1 - - Evaluate \( \int{ 2xy~dx } \)
Since we are integrating with respect to \(x\), \(y\) is a constant. In fact, when integrating, we can pull the \(y\) outside the integral to make things easier to see. So we have
\( \int{ 2xy~dx } = y \int{2x~dx} = y(x^2) + g(y) = x^2y + g(y) \)
Notice that instead of \(C\), we have \(g(y)\), which is a function of \(y\). ( This function could also be just a constant. ) The point is, it can't have any \(x\)'s in it. It can have only constants and \(y\)'s ( or whatever variable(s) we hold constant during integration ).

Here is a great video clip with an example. He explains the details without getting too technical.

For definite integrals, we don't have worry about the constant, of course. However, we have an added tip that may help you avoid mistakes when you are tired or stressed, for example, during an exam. Let's look at an example.

Example 2 - - Evaluate \(\displaystyle{ \int_{a}^{b}{ 2xy ~ dx } }\).
\(\displaystyle{ \int_{a}^{b}{ 2xy ~ dx } = }\) \(\displaystyle{ 2y \int_{a}^{b}{ x~dx } = }\) \(\displaystyle{ \left[ x^2 y \right]_{x=a}^{x=b} = }\) \(\displaystyle{ (b^2-a^2)y }\)

Most of Example 2 is self-explanatory but we want to mention one thing we did that may look unusual. Notice that after integration but before substituting the limits of integration, we wrote \(\displaystyle{ \left[ x^2 y \right]_{x=a}^{x=b} }\). It would have been technically correct if we had written \(\displaystyle{ \left[ x^2 y \right]_{a}^{b} }\). However, writing \(x=a\) and \(x=b\), we are making sure that we remember to substitute for x, not y. It is very easy to get confused when we are tired or stressed. So writing it more clearly could help prevent lost points. So we highly recommend that you do it this way but, as always, check with your instructor to see what they expect.

Another twist with definite partial integrals is that the limits of integration do not have to be constants. They can be expressions like \(y+2\). This doesn't really change anything that you will do but it looks kind of strange when you first see it. Let's look at another example.

Example 3 - - Evaluate \(\displaystyle{ \int_{0}^{x}{2x+y ~ dy} }\).
\(\displaystyle{ \int_{0}^{x}{2x+y ~ dy} = }\) \(\displaystyle{ \left[ 2xy + y^2/2 \right]_{y=0}^{y=x} = }\) \(\displaystyle{ 2x^2 + x^2/2 = 5x^2/2 }\)

A few comments are in order.
1. Notice that we used the explicit notation for \(y=0\) and \(y=x\) for the limits. We do that a lot on this site to help you follow our examples.
2. If you have an expression in one or both limits of integration, the expressions will not contain the variable of integration under any circumstance. Here is something you will never see: \(\displaystyle{ \int_{0}^{y}{2x+y ~ dy} }\). Having an expression with the variable of integration in it does not make sense.
3. Notice that the result of the integration is a function of the variable that was considered a constant in the integration and the variable of integration is no where to be found. This will always be the case. In Example 3, y is not in the final answer and x was considered a constant of integration but in the final answer it is a variable.

We usually don't use partial integration in isolation. The beauty of integration is that we use this idea to solve double and triple integrals. You will get plenty of practice coming up, so, if you are comfortable describing areas in the plane, we recommend that you go straight to your next topic, double integrals.